Volume 28

Editor-in-chief: Bruce McCormick Guest Editors: Fiona Martin Nigel Hollister

December 2012
ISSN 1353-4882

SPECIAL EDITION Intensive Care Medicine

IMPORTANT NOTICE
Periodically we need to check that Update in Anaesthesia is being posted to the correct recipients. If you wish to continue receiving the journal please send the following details to us: Name Institution Postal address

Email address (if you have one) Number of recipients wishing to receive Update at this address Please send this information: By email to: [email protected] or by mail to: Carol Wilson Pound Cottage Christow Exeter EX6 7LX United Kingdom

page 2

Update in Anaesthesia | www.anaesthesiologists.org

Anaesthesia
Contents
4 Editorial The Global sepsis Alliance - fighting a global disease Sebastian N Stehr and Konrad Reinhart 6 Editors notes Bruce McCormick GENErAL PrINCIPLEs 7 Intensive care medicine in resource-limited settings: a general overview Martin W Dnser 11 18 22 27 Systematic assessment of an ICU patient Sebastian Brown, Sophia Bratanow and Rebecca Appelboam Intensive care medicine in rural sub-Saharan Africa - who to admit? RM Towey and John Bosco Anyai Identifying critically ill patients - Triage, Early Warning Scores and Rapid Response Teams Tim baker, Jamie Rylance and David Konrad Critical care where there is no ICU: Basic management of critically ill patients in a low income country Tim Baker and Jamie Rylance 133 141 SEPSIS 145 156 Management of sepsis with limited resources Kate Stephens Abdominal compartment syndrome William English Management of burns Nigel Hollister Management of drowning Sarah Heikal and Colin Berry

Update in

mICrOBIOLOgY 160 Bugs and drugs in the Intensive Care Unit Simantini Jog and Marina Morgan CArDIOVAsCULAr 169 Inotropes and vasopressors in critical care Hannah Dodwell and Bruce McCormick 177 Management of cardiac arrest - review of the 2012 European Resuscitation Guidelines Paul Margetts

mONITOrINg 32 Monitoring in ICU - ECG, pulse oximetry and capnography Ben Gupta 37 43 51 Invasive blood pressure monitoring Ben Gupta Central venous cannulation Will Key, Mike Duffy and Graham Hocking Cardiac output monitoring Thomas Lawson and Andrew Hutton

rEsPIrATOrY 183 Acute respiratory distress syndrome (ARDS) David Lacquiere 188 192 Hospital-acquired pneumonia Yvonne Louise Bramma and radha Sundaram An introduction to mechanical ventilation Fran OHiggins and Adrian Clarke

199 Tracheostomy Rakesh Bhandary and Niraj Niranjan rENAL 207 Acute kidney injury - diagnosis, management and prevention Clare Attwood and Brett Cullis 215 223 Renal replacement therapy in critical care Andrew Baker and Richard Green Peritoneal dialysis in acute kidney injury Brett Cullis

GENERAL CARE 59 Acid-base disorders in critical care Alex Grice 67 74 79 88 Delirium in critical care David Connor and William English Sedation in intensive care patients Gavin Werrett Nutrition in the critically ill Sophia Bratanow and Sebastian Brown Evidence-based medicine in critical care Mark Davidson

NEUrOmUsCULAr DIsEAsE 228 Neurological causes of muscle weakness in the Intensive Care Unit Todd Guest 233 Tetanus Raymond Towey 240 243 247 253 261 Brainstem death Niraj Niranjan and Mike Duffy Cultural issues in end-of-life care Sara-Catrin Cook and Carol Peden Diabetic ketoacidosis Claire Preedy and William English Emergency management of poisoning Sarah Heikal, Andrew Appelboam and Rebecca Appelboam Management of snake envenomation Shashi Kiran and T A Senthilnathan

TrAUmA 95 Management of major trauma Lara Herbert and Ruth Barker 107 112 119 Management of head injuries Bilal Ali and Stephen Drage Acute cervical spine injures in adults: initial management Pete Ford and Abrie Theron Thoracic trauma Anil Hormis and Joanne Stone

MISCELLANEOUS

125 Guidelines for management of massive blood loss in trauma Srikantha L Rao and Fiona Martin 130 Rhabdomyolysis Michelle Barnard

Update in Anaesthesia | www.anaesthesiologists.org

page 3

Contents

Anaesthesia
Guest editorial
The Global Sepsis Alliance fighting a global disease
Only in the past thirty years has sepsis been recognized as a very common disease of global proportions and impact. Initially underdiagnosed and unrecognized, it is now accepted that sepsis, a clinical syndrome defined by the presence of both infection and a systemic inflammatory response,1 is most probably one of the leading causes of death in the world.2 727,000 patients were hospitalized with a primary diagnosis of septicaemia or sepsis in the United States in 2008, more than double the number of patients documented in 2000. 3 In-hospital deaths were more than eight times more likely in patients with a diagnosis of septicaemia or sepsis compared to other diagnoses.3 These estimates concern an environment of a developed, modern intensive care setting. There is very little data available for the developing world, where the majority of worldwide deaths related to sepsis are to be expected due to the prevalence of HIV/AIDS, malaria and maternal sepsis. It has been proven that the introduction of evidence-based guidelines focussing on early recognition, emergent antibiotic treatment and application of fluids and vasopressors can reduce sepsis-related mortality.4 It is unclear to what extent these interventions can be translated to a developing world setting.5 A multitude of local, national and international organisations and societies dedicated to sepsis have developed over the past years. The Global Sepsis Alliance (GSA) was launched in September 2010 as part of a Merinoff Symposium of the Feinstein Institute for Medical Research on Long Island, to take on sepsis as a global problem. The GSA was founded by the World Federation of Societies of Intensive and Critical Care Medicine (WFSICCM), the World Federation of Pediatric Intensive and Critical Care Societies (WFPICCS), the International Sepsis Forum (ISF), the Sepsis Alliance USA (SA) and the World Federation of Critical Care Nurses (WFCCM) to coordinate global efforts against sepsis and to speak with one voice. In the meantime, the member organisations of the GSA represent over 600,000 health care professionals from more than 70 countries (Table 1). The GSA has set out to Speak in One Voice offering consistent, easily understood messaging to governments, philanthropies and the public. The GSA has set goals to provide opportunities supportive of global interaction and defined output. As a first step, the GSA has developed a definition of sepsis that facilitates communication with the lay public: Sepsis is a life threatening condition that arises when the bodys response to an infection injures its own tissues and organs. Sepsis may lead to shock, multiple organ failure, and death, especially if not recognized early and treated promptly. Sepsis remains the primary cause of death from infection despite advances in modern medicine, including vaccines, antibiotics, and acute care. Millions of people die of sepsis every year worldwide. Large scale studies are necessary to find out more about possible interventions to reduce sepsis-related morbidity and mortality. A major goal of the GSA is to assist societies and initiatives in the process of developing proposals for experiments, trials, projects and programs in support of researchers, caregivers and the public, especially in securing funding to implement such efforts. The GSA is to be empowered to easily identify and access resources and people of common purpose and intent within and without the scientific community. The 2005 World Health Organisation Health global report on global child death considers that 80% of global child deaths are related to severe infections associated with pneumonia, malaria, measles, neonatal sepsis, and diarrhoea.6 One exemplary project supported by the GSA is the development and implementation of sepsis demonstration projects in the poor districts of Ugandan, both urban and rural, in collaboration with the Ministry of Health, Makerere University College of Health Sciences, Mbarara University of Science and Technology and the Centre for International Child Health, University of British Columbia. The GSA will employ its contacts to regionally and globally disseminate the initiatives experiences, findings and lessons learned. The GSA will focus on addressing with equal commitment and vigour the needs of both adults and children in the developed and developing world. The GSA urges the medical community to recognize sepsis as a medical emergency, requiring the administration of fluids, antibiotics and other appropriate treatments of infection within one hour of first suspecting a case of sepsis. This is also possible in regions without modern intensive care units, using a less sophisticated approach.7 In conclusion, the global burden of sepsis is high and is increasing, especially in the developing world. The

Update in

Editorial
Sebastian N Stehr Konrad Reinhart Chairman of the Global Sepsis Alliance Department of Anesthesiology and Intensive Care Medicine Friedrich-Schiller-University Jena Germany
page 4

Update in Anaesthesia | www.anaesthesiologists.org

Anaesthesia
Table 1. Membership of the Global Sepsis Alliance.

Update in

Membership of the Global Sepsis Alliance Founding organizations International Sepsis Forum (ISF) Sepsis Alliance (SA) World Federation of Pediatric Intensive and Critical Care Societies (WFPICCS) World Federation of Societies of Intensive and CriticalCare Medicine (WFSICCM) World Federation of Critical Care Nurses (WFCCN) Committed organizations American Thoracic Society (ATS) Australia and New Zealand Intensive Care Society (ANZICS) Belize Medical and Dental Association Centre for International Child Health Chilean Society of Critical Care Chinese Society of Critical Care Medicine Dutch Meningitis Initiative Emirates Intensive Care Society German Sepsis Society and German Sepsis Aid Gruppo italiano per la Valutazione degli interventi in Terapia Intensiva (GiViTI) Hellenic Sepsis Study Group International Forum for Acute Care Trialists (InFACT) International Pan Arab Critical Care Medicine Society Latin American Sepsis Institute Maventy Health International Society of Critical Care Medicine Spanish Edusepsis Network Surgical Infection Society (SIS) Survive Sepsis United Kingdom Sepsis Trust

use of current evidence-based knowledge must be applied to reduce the worldwide high sepsis mortality rate. Healthcare professionals and laypersons must be taught that sepsis is an emergency requiring urgent treatment. The GSA will focus on programs to better understand that sepsis is an emergency and to foster a greater understanding of the medical burden of sepsis among the public and is planning a World Sepsis Day for 2012. The GSA encourages all concerned groups and societies to learn from each other and to join forces in the fight against sepsis at a global level and to become a member of the GSA. More information is available on the GSA website at www. globalsepsisalliance.com. REFERENCES
1. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R, et al. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock: 2008. In: Critical Care Medicine 2008. p 296327.

2. Marshall JC, Reinhart K. The Global Sepsis Alliance: building new collaborations to confront an under-recognized threat. Surg Infect (Larchmt) 2011; 12: 12. 3. Hall MJ, Wiliams SN, DeFrances CJ, Golosinsky A. Inpatient Care for Septicemia or Sepsis: A Challenge for Patients and Hospitals [Internet]. NCHS Data Brief. [cited 2011 Nov. 9]; 62(June 2011). Available from: http://www.cdc.gov/nchs/data/databriefs/db62.pdf. 4. Levy MM, Dellinger RP, Townsend SR, Linde-Zwirble WT, Marshall JC, Bion J, et al. The Surviving Sepsis Campaign: results of an international guideline-based performance improvement program targeting severe sepsis. Intensive Care Medicine 2010; 36: 22231.

5. Cheng AC, West TE, Limmathurotsakul D, Peacock SJ. Strategies to Reduce Mortality from Bacterial Sepsis in Adults in Developing Countries. PLoS Med 2008; 5: e175. 6. Bryce J, Boschi-Pinto C, Shibuya K, Black RE, WHO Child Health Epidemiology Reference Group. WHO estimates of the causes of death in children. Lancet 2005; 365: 114752. 7. Kissoon N, Carcillo JA, Espinosa V, Argent A, Devictor D, Madden M, et al. World Federation of Pediatric Intensive Care and Critical Care Societies: Global Sepsis Initiative*. Pediatric Critical Care Medicine 2011; 12: 494503.

Update in Anaesthesia | www.anaesthesiologists.org

page 5

Anaesthesia
Editors notes
Dear Readers, Welcome to this Special Edition of Update in Anaesthesia, which focuses on Intensive Care Medicine. This specialty has developed greatly over the last 30 years, however development of dedicated intensive care units (ICUs) in more poorly resourced countries has only come about in the last few years. We think of an ICU as a location in the hospital where the sickest patients are admitted for more invasive monitoring and more aggressive organ support and therapy. Inherently these monitors and treatments incur far higher costs than standard ward care, making them unachievable in many settings. However, equipment is not the major factor that sets the ICU or high dependency unit (HDU) apart from the other wards of a hospital; it is the expertise and numbers of the ICU staff that confers the most dramatic advantage in providing effective care for the critically ill. Nursing staff numbers, and therefore the nurse to patient ratio, vary starkly between the general wards (around one to sixty in the description of a Ugandan ICU by Towey and Anyai, on page 16 of this edition of Update) and the ICU (ideally 1:1, but commonly 1:4 or 1:6). In addition it is the quality of training and experience of these nursing staff that has a major impact on patient care, particularly where staff morale allows good retention of staff and longevity of careers in the ICU. In addition to good nursing care, close attention to the detail of basic good medical care by trained and experienced clinical officers and doctors, probably has a far greater impact on patient outcome than use of expensive, invasive equipment. In fact there are few interventions in ICU for which the evidence remains relatively unequivocal, examples being nursing patients in the semi-recumbent position (30 degrees head up) to decrease the incidence of ventilator associated pneumonia and administration of antibiotics to patients with sepsis within one hour or presentation. Therapies such as steroids and activated protein C for septic shock, despite encouraging early randomised control studies, have now been proven to be ineffective or harmful. Many of the more technical strategies for providing advanced respiratory support to patients with intractable hypoxia, such as extra-corporeal membrane oxygenation and high frequency oscillation ventilation, have very little supporting evidence. So we are left in a situation where timely basic interventions are likely to bring about the greatest improvements in mortality and morbidity of critically ill patients, manoeuvres such as effective airway management and haemodynamic resuscitation in trauma, early antibiotics and surgical source control in sepsis. These strategies are available in most healthcare settings around the world. This edition of Update in Anaesthesia attempts to provide an overview of the essential aspects of care of the critically ill and critically injured, with particular focus on practices that are most relevant and achievable in poor resource settings. For most topics in our speciality we have tried to achieve a balance between making the text relevant to workers where high-tech equipment is not available and achieving appropriate coverage of the topic for areas where some level of more advanced equipment may be available. In many parts of the world, health centres that are geographically close to each other may vary greatly in their resources, due largely to the influence of alternative funding streams from non-government organisations. I hope that this edition is useful. I would appreciate your feedback at [email protected]. The articles do not cover this subject fully and suggestions for further ICM topics would be welcomed. This edition is available, along with the full back catalogue of Update in Anaesthesia at http://update.anaesthesiologists.org

Update in

Bruce McCormick Editor-in-chief Consultant in Anaesthesia and Intensive Care Medicine Exeter, UK

Update Team
Editor-in-chief Bruce McCormick (UK) Guest Editors Nigel Hollister (UK) Fiona Martin (UK) Editorial Board Douglas Bacon (USA) Aboudoul-Fataou Ouro Bangna (Togo) Martin Chobli (Benin) Gustavo Elena (Argentina) SS Harsoor (India) Kazuyoshi Hirota (Japan) David Pescod (Australia) Jeanette Thirlwell (Australia) Isabeau Walker (UK) Zhanggang Xue (China) Jing Zhao (China) Chief Illustrator Dave Wilkinson (UK) Typesetting Angie Jones, Sumographics (UK) Printing COS Printers Pte Ltd (Singapore)

page 6

Update in Anaesthesia | www.anaesthesiologists.org

Intensive care medicine in resource-limited settings: a general overview

Martin W Dnser Correspondence Email: [email protected]
WHAT IS INTENSIVE CARE MEDICINE? Intensive or critical care medicine refers to the medical specialty which focuses on the management of critically ill patients. Critical illness describes a general state which may arise from various medical pathologies (e.g. trauma, infection, acute coronary syndrome, stroke etc.) and leads to the impairment of vital (consciousness, circulation or respiration) or single organ functions (e.g. kidney or liver function). Furthermore, intensive care includes the care of patients after major surgery or the observation of patients in whom critical illness may rapidly occur. INTENSIVE CARE MEDICINE IN RESOURCEPOOR SETTINGS In Western countries, the first intensive care units (dedicated hospital wards where intensive care medicine is practiced and critically ill patients are cared for) were established in the 1950s and 1960s following the last European polio epidemics.1 Since then the number of intensive care units (ICUs) has grown steadily and intensive care medicine has gained importance as a medical specialty in its own right. The majority of acute care hospitals in high-income countries now run one or more ICUs. The most frequent pathologies leading to ICU admission in Western countries are cardiovascular diseases, major surgery, sepsis and respiratory failure. Although the first ICUs were introduced to select resource-poor settings shortly after intensive care medicine started to develop, the majority of critically ill patients in less developed countries, harboring around two thirds of the world population, still do not have access to intensive care.2 Few data exist on the current state of intensive care medicine in less developed countries, but there seems to be wide variability in the availability of ICUs in these countries, ranging from non-existent to sophisticated centres in selected private hospitals catering for a few privileged patients. Recent data from the Republic of Zambia revealed that only 29 ICU beds exist for the entire country of 12.9 million people and only 7% of hospitals providing surgical services run an ICU. Even in those hospitals with ICUs, basic equipment is lacking and an oxygen supply is only inconsistently available.3 Similar data were reported from other African or Asian regions. In countries like Bangladesh, India and Nepal, there has recently been an important increase in the availability of intensive care units, although shortage in staffing, lack of basic equipment, poor maintenance of equipment and interrupted supplies often pose major challenges. In addition, the medical profession in less developed countries is, in general, not set up to provide formal training in intensive care medicine. Knowledge about important recent progress in the field is frequently absent. These factors inevitably result in a lack of recognition of intensive care medicine as a medical specialty in resource-poor settings. As a consequence, disproportionately high mortality rates have been reported for selected critical illnesses in developing countries.4 DIFFERENCES IN INTENSIVE CARE MEDICINE BETWEEN HIGH INCOME AND RESOURCEPOOR SETTINGS Intensive care medicine between Western and less developed countries not only differs in equipment and material availability, but also in the patient populations treated in the ICU.5 In less developed countries critically ill patients admitted to the ICU are characteristically younger, suffering from less premorbid conditions. The underlying diseases leading to ICU admission in resource-poor areas differ geographically from those seen in high income countries. While in Northern developing countries (e.g. central Asian countries) ICU admission diagnoses are similar to those reported from high income countries, tropical and infectious diseases are among the leading causes of critical illness in developing countries in South Asia, South America and Africa. Trauma and sepsis are far more common in ICUs of developing than Western countries. Disease severity at ICU admission is typically higher in resource-poor settings, while the number of interventions and procedures performed is smaller compared to critically ill patients admitted to ICUs in high income countries. Irrespective of the ICU admission diagnosis, mortality rates of critically ill patients are consistently higher in less developed than in high income countries.5

Summary
This introductory article gives an overview of intensive care medicine in developing countries and contrasts its development with high-income countries. The second part of this manuscript aims to give the reader a general overview of the basic aspects and requirements of ICUs and intensive care medicine in resource-poor settings.

ICU STAFFING
An ICU needs the presence of well trained and

Martin W Dnser MD DESA Global Intensive Care Working Group of the European Society of Intensive Care Medicine. Department of Anesthesiology, Perioperative Medicine and General Intensive Care Medicine, Salzburg General Hospital and Paracelsus Private Medical University MllnerHauptstrasse 48 5020 Salzburg Austria

Update in Anaesthesia | www.anaesthesiologists.org

page 7

General Principles

Anaesthesia

Update in

experienced ICU workers 24-hours-a-day, 7-days-a-week. An ideal ICU team consists of nurses, specially trained in intensive care medicine, one or more intensivists (physicians specialized in providing intensive care medicine) and a variable number of nurse assistants, technicians and cleaners. In many resource-poor settings, the role of the intensivist is taken over by a nurse anaesthetist or an anaesthetic clinical officer. This is a practicable and legitimate policy since maintenance and restoration of vital functions is one of the key fields of anaesthesia. If the intensivist is not a medical doctor, it is advisable that a physician is available to assist in the care of the critically ill patients underlying disease. Ideally, the intensivist in charge should be a physician specially trained in intensive care medicine. In some Western countries (e.g. the United Kingdom), specialized postgraduate training programs for intensive care medicine exist. In addition, diplomas in intensive care medicine can be taken from international intensive care societies (e.g. the European Society of Intensive Care Medicine). Due to the wide-ranging lack of health care personnel and qualified staff in many resource-poor settings, the anaesthetist/physician caring for the ICU often has to fulfill additional medical duties in the operation theatre or hospital, particularly at night and during weekends. This frequently leaves the ICU unattended by an intensivist and places more responsibility on the ICU nurses, making them the key players of the ICU team. Trust and good communication with the intensivist in charge, as well as continuous education, adequate training and a strong team spirit, are of outstanding importance for ICU nurses in resource-poor settings. ORGANIZATIONAL ASPECTS OF AN ICU An ICU can be organized in different ways. Larger hospitals in particular often run specialized ICUs caring for critically ill patients with selected diseases; for example surgical, pediatric, neurosurgical, cardiac, medical or burns ICUs. Although this may have some benefits for certain patient populations, recent data indicate that multidisciplinary ICUs caring for patients with different pathologies may result in better care. In any case, it is important to understand that caring for a critically ill patient, irrespective of the underlying disease, must include an interdisciplinary approach, involving integration of physicians from other medical specialties such as neurologists, surgeons or pediatricians. Mutual respect is a prerequisite for fruitful interdisciplinary communication. In a closed ICU one or more intensivist is principally responsible for the care of all patients admitted to the ICU. This organizational structure is in contrast to the open ICU where different physicians, who are not continuously present in the ICU, care for single critically ill patients. Organization of ICUs as closed units, including the presence of a an intensivist, has been shown to result in lower mortality, less complications, a reduced length of ICU stay and lower costs, when compared to open ICUs.6 If hospitals are too small to implement a 24hour intensivist service, telemedical assistance by external intensivists may be used to support decision making and patient care.7 Although most reports on intensive care telemedicine originate from highincome countries (the United States and Australia), personal experience of the author suggests that regular (e.g. weekly) telemedical counseling by experienced intensivists can be a valuable tool to improve patient care in ICUs in resource-limited areas.

CONSTRUCTIONAL ASPECTS OF AN ICU Even though intensive care medicine can be supplied under several circumstances and at various locations, an ICU in a resource-poor setting has certain constructional requirements. Non-leaking roofs, closable windows/doors, solid walls and, whenever necessary, a functional heating system must be available to protect patients and staff from adverse climate influences. Floors and walls should be easily washable to allow effective cleaning. Light and a stable electricity supply are further indispensable prerequisites to run an ICU. Stable electricity supply, on the one hand, includes the availability of a power generator (e.g. driven by gasoline or diesel), providing electricity in case of power cuts. On the other hand, in many resource-poor settings, voltage stabilizers need to be placed in the main electrical line supplying the ICU, in order to prevent voltage peaks that may damage delicate medical apparatus such as mechanical ventilators or patient monitors.

5 1 4 3 2
Figure 1. Intensive care unit in a rural African hospital - 1, patient monitor; 2, suction machine; 3, oxygen concentrator; 4, mechanical ventilator; 5, mosquito net.

Running water with a constant supply of soap is essential to reduce cross-infection between critically ill patients. In areas where malaria and other insect-transmitted infectious diseases are endemic, mosquito nets should be available for each ICU bed to protect patients from insect bites during evening and night times (Figure 1). Air filtering and room climatization are not essential, but can greatly help to maintain clean air and adjust room temperatures and air pressure to patient needs. Although no scientific data have so far proven that isolation of patients with resistant bacteria, such as methicillin-resistant Staphylococcus aureus, can reduce transmission of these bacteria to other patients, an ICU should include a room to isolate patients. For certain infectious diseases, such as open pulmonary tuberculosis or certain viral haemorrhagic fevers, isolation is obligatory. When spatial isolation is required, the patient should not be in isolation from medical and nursing care. The nurse base, an integral part of the ICU, should be placed centrally and allow full sight on as many ICU beds as possible (Figure 2). OXYGEN, PRESSURIZED AIR AND SUCTION One of the most important drugs required in the ICU is oxygen. Oxygen can be stored and supplied in various ways. Oxygen

page 8

Update in Anaesthesia | www.anaesthesiologists.org

Figure 2. View from the nurse base of a Mongolian ICU.

Figure 3. Air compressor supplying the pressurized air system of a Mongolian ICU, with the pressure regulator indicated by the white arrow.

concentrators provide 90-100% oxygen but rely on a constant electricity supply and usually do not provide oxygen flows higher than 4-6L.min-1. While this is sufficient to treat neonates and infants with respiratory insufficiency, in many cases it is inadequate to oxygenate larger children or adults with respiratory failure. In contrast, oxygen cylinders can provide pure oxygen at high flow rates and are independent of electricity supply, but need to be replenished at regular intervals. This must be addressed in advance before the last cylinder has emptied, leaving the patient with respiratory distress without oxygen. Central oxygen systems are the most efficient and convenient way to store and supply ICUs with oxygen. The source of oxygen of a central oxygen system can either be a special oxygen tank storing oxygen at low temperatures, or a bank of oxygen cylinders. Both of these require regular maintenance and replenishment. The tubing of the pressurized oxygen system must consist of a non-oxidizing material, typically copper. In countries where no professional companies offer installation of medical air systems, refrigeration engineers usually have sufficient experience in installing copper/pressurized gas lines. Pressurized air, used to run mechanical ventilators, can similarly be administered either by direct connection of a compressor to the mechanical ventilator or preferably by connecting a compressor to a central air system, providing pressurized air through single outlets at each ICU bed. Although specific medical air compressors exist, oil-free industrial compressors, with a pressure regulator as well as additional air filters, provide comparable air qualities. These are more easily affordable in resource-poor settings (Figure 3). Where oil-free compressors are available air filters need to be placed in the air lines and before air enters the ventilator. Although oil spilling into the patients respiratory system is the by far most relevant danger, more frequent complications are acute blockade of line or air filters in the ventilators. Central suction units may be connected to the pressurized air system, but usually depend on special suction generators, which can be cumbersome to find and install in resource-limited areas. BASIC RESOURCE REQUIREMENTS OF AN ICU Although intensive care medicine, above most other medical specialties, relies on technical devices and material resources, it is

crucial to consider that no apparatus can replace an alert ICU worker at the bedside. Nonetheless, certain technical devices are required to support the work of the ICU staff. These typically include patient monitors, suction machines and mechanical ventilators. While patient monitors measuring ECG, respiratory rate, arterial blood pressure and oxygen saturation should be available at each bed, suction machines and mechanical ventilators can be used specifically for patients in need of these devices. The technical aspects of mechanical ventilators must be considered, because the majority of available ventilators depend on a dual supply of pressurized oxygen and air. In ICUs where neither pressurized air nor adequate stores of pressurized oxygen are available, only ventilators with internal air compressors together with an external oxygen source (e.g. from an oxygen concentrator or an oxygen cylinder) can be used. Infusion and syringe pumps allow drugs and fluids to be administered at exact rates and dosages, but, in the clinical practice of resourcepoor settings, may well be replaced by mechanical drop regulators or close clinical surveillance by a nurse. Any device not depending on electricity increases patient safety during power cuts, particularly when vital drugs (e.g. catecholamines) are infused. Despite being a lifesaving intervention, renal replacement therapy in patients with acute kidney failure is usually unavailable in resource-poor settings. Given that neither intermittent hemodialysis nor continuous hemofiltration is superior in terms of patient survival, and that hemofiltration is more time and resource-consuming, intermittent hemodialysis is the technique of choice to treat patients with acute kidney failure in resource-poor settings. Although data on the use of peritoneal dialysis in critically ill patients with acute kidney failure are conflicting, peritoneal dialysis may be an option if local experience is available. Similarly, a basic set of essential disposable materials, drugs and laboratory tests need to be available to adequately and safely care for critically ill patients. These sets usually do not need to include highend materials or a large variety of drugs or tests, but should focus on the basic needs of critically ill patients treated in the respective ICU. Furthermore, small numbers of essential materials, drugs and tests warrants expert use by the ICU staff and facilitates stock maintenance.

Update in Anaesthesia | www.anaesthesiologists.org

page 9

THE ICUS PLACE IN A RESOURCE-POOR HOSPITAL Intensive care medicine is an integrative medical specialty, requiring close cooperation with several other medical disciplines and technical services (e.g. laboratory services, blood bank etc.) in the hospital. Therefore, to assure adequate and efficient care of critically ill patients, other medical departments and hospital services need to be prepared and trained to manage the needs of critically ill patients.2,4 Since ICUs in resource-poor settings are either non-existent or have only recently been established, acceptance of ICU services among colleagues from other medical specialties (who have so far cared for critically ill patients on the hospital ward) is a frequent problem. After establishing an ICU in a resource-poor hospital, referral and admission rates are often low. If patients are admitted this typically occurs at a pre-terminal stage, where ICU interventions may fail to safe the patients life. This can lead to a perception amongst ward staff and relatives that patients are transferred to the ICU to die. Integration of ward physicians into ICU care (e.g. during daily rounds or regular discussions at the bedside), together with education of the hospital staff about when to admit patients to the ICU are ways to increase acceptance of newly established ICU services in resourcepoor hospitals. When ICU services are well-established and accepted, unavailability of ICU beds is a far greater problem. ICU bed capacities need to be coordinated with the emergency department and the operation theatre at regular intervals each day. From a practical standpoint, ICUs should always have the capacity to admit unplanned critically ill patients. This can be organized by leaving one ICU bed in the hospital unoccupied or having the facility to discharge one patient rapidly to an appropriate hospital ward. INTENSIVE CARE MEDICINE WITHOUT WALLS Provision of intensive care medicine is not only restricted to the ICU. In order to prevent patients being admitted too late, after they have developed irreversible shock or organ failure, the intensivist can play a valuable role in assessing patients before ICU admission (e.g. in the operation theatre or the emergency department) or after ICU discharge (post-ICU review). In several hospitals, intensivists play a key role in resuscitation teams or medical emergency teams. The function of these teams within a hospital is described in a later article. Implementation of medical emergency teams in hospitals of high-income countries reduced the rates of unexpected cardiac arrests on non-ICU wards.8 In addition to providing resuscitation and emergency care, intensivists may further assist physicians from other medical specialties with certain clinical problems (e.g. prescription of parenteral/enteral nutrition, provision of palliative care, cannulation of central vessels or assessment of surgical and anaesthetic risks). CONCLUSION Intensive care medicine is a comparatively young medical specialty which has grown rapidly to become an essential component of modern hospitals. Many hospitals in resource-poor settings do not run ICUs and critically ill patients frequently receive suboptimal care with unacceptable levels of mortality. When implementing intensive care medicine in resource-poor settings several staff, constructional, organizational and resource aspects need to be considered.

Table 1. Ten basic principles of intensive care medicine.

1 No medical apparatus can replace the presence of an ICU worker at the bedside. 2 No diagnostic test can replace a thorough patient history, chart review or systematic clinical examination. 3 Supportive therapy is life-saving, challenging and may distract the intensivists attention from searching for the underlying cause of critical illness. Always try to identify why a critically ill patient is sick and do everything to treat this condition.

4 Always ask why a patient is deteriorating or fails to improve. Never accept or explain treatment failures simply by disease severity. 5 Do not over-sedate. Only sedate agitated patients or those with certain diseases (intracranial hypertension, acute lung or circulatory failure). 6 Do not overhydrate patients. Although fluid resuscitation can safe lives in the acute phase, indiscriminate infusion of fluids at later stages leads to complications (e.g. sepsis), prolongs ICU stay and increases mortality.

7 Do no harm! Be aware that every intervention and drug applied in the ICU carries the potential to harm the patient. 8 As soon as the patient has stabilized do everything to reduce invasive support. 9 Always consider the therapeutic consequence before performing diagnostic tests (e.g. imaging studies). 10 Do not indiscriminately order laboratory tests but only measure these values where relevant and pathologic information can be expected.

1. Berthelsen PG, Cronqvist M. The first intensive care unit in the world: Copenhagen 1953. Acta Anaesthesiol Scand 2003; 47: 1190-5. 2. Baker T. Critical care in low-income countries. Trop Med Int Health 2009; 14: 143-8. 3. Jochberger S, Ismailova F, Lederer W et al. Anesthesia and its allied disciplines in the developing world: a nationwide survey of the Republic of Zambia. Anesth Analg 2008; 106: 942-8. 4. Dnser MW, Bataar O, Tsenddorj G et al. Differences in critical care practice between an industrialized and a developing country. Wien Klin Wochenschr 2008; 120: 600-7. 5. Dnser MW, Baelani I, Ganbold L. A review and analysis of intensive care medicine in the least developed countries. Crit Care Med 2006; 34: 1234-42. 6. Topeli A, Laghi F, Tobin MJ. Effect of closed unit policy and appointing an intensivist in a developing country. Crit Care Med 2005; 33: 299 306. 7. Lilly CM, Cody S, Zhao H et al. Hospital mortality, length of stay, and preventable complications among critically ill patients before and after tele-ICU reengineering of critical care processes. JAMA 2011; 305: 2175-83.

REFERENCES

8. Jones DA, DeVita MA, Bellomo R. Rapid Response Teams. N Engl J Med 2011; 365: 139-46.

page 10

Update in Anaesthesia | www.anaesthesiologists.org

Systematic assessment of an ICU patient

Sebastian Brown*, Sophia Bratanow, Rebecca Appelboam *Correspondence Email: [email protected]
INTRODUCTION The Surviving Sepsis Campaign1 and World Health Organisation surgical checklist2 have demonstrated that use of a systematic review and checklist approach, to optimise patient management and safety, improves outcome. Reducing surgical mortality is dependent upon the ability to recognise and rescue patients who develop complications.3,4 Improved survival of patients treated in critical care has been attributed to improvements in the processes of care, rather than the introduction of individual therapies or diagnostic modalities.5 Furthermore, the strict implementation of dedicated processes of care, often called care bundles, improves ICU and hospital mortality.6 In this article, we describe a head-to-toe assessment and treatment strategy to guide the daily or night review of intensive care patients. This systematic assessment incorporates the current evidence and care bundles that contribute to improve outcome. HISTORY If the patient is awake, introduce yourself and explain who you are and what you intend to do. Whether they are awake or asleep, try to avoid focusing intently on the monitors and charts, thereby ignoring the patient. Although we have more monitoring and tests available to us, the focus of our attention should always be the patient, their symptoms and clinical signs. The patients presenting complaint (e.g. pneumonia) will usually be the primary focus of your assessment, but after the first few days of admission the emphasis may shift to other priorities; a patient may recover from intra-abdominal sepsis but be left with respiratory failure due to acute respiratory distress syndrome (ARDS) or underlying airways disease. Details of the past medical history and the presentation of the primary pathology may be difficult to obtain, but information should be sought from relatives, the ambulance crew, referring hospitals and general practitioners or hospital specialists caring for the patients chronic medical conditions. When you encounter a patient for the first time, it is worth sitting down to read the current and old hospital notes in full (including specialists letters and old investigations), in order to form a complete picture of the patients medical history. The physicians traditional wisdom that 90% of the diagnosis is in the history is equally applicable to patients on the intensive care unit. Some patients will only be able to answer your questions for a short period before clinical deterioration or sedation prevents this. The information that you obtain from them may be vital, for example sudden onset of chest and abdominal pain whilst vomiting in a septic patient, suggests a perforated oesophagus, a diagnosis that can easily be missed without a suggestive history. If a patients response to treatment is not as predicted, review the presentation and consider whether the working diagnosis is correct. Decisions made regarding admission to ICU, require some knowledge of patients physiological reserve, their quality of life and their own attitude to such treatments. These decisions should be made and documented preemptively rather than when a catastrophic deterioration occurs. Patients may remain in the ICU for some weeks. Experienced intensivists are able to plot the next few days of a patients ICU stay, allowing goals to be set for certain aspects of the patients illness. In spite of this, unexpected events occur relatively frequently and it is important have flexibility to focus on whatever issues arise. EXAMINATION Physical examination of the patient and their observations can often occur together. A systematic approach must be used and a head-to-toe system is appropriate. Each section focuses on history, clinical examination and observations. Even though this approach is labour-intensive it is this type of attention to detail that may make a difference in a patients progress in ICU. For example, identifying and removing a cannula that has been in for 5 days, is not being used and shows erythema around it, may prevent an episode of Staphylococcal bacteraemia. Try to avoid making assumptions about other what other medical staff have done; if a trauma patient has been moved rapidly from the emergency department to theatre for abdominal bleeding, when they arrive

Summary
A structured approach to assessment and management improves outcome. Management should incorporate best evidence and current care bundles. Checklists, such as FASTHUG, can aid the complete assessment of the ICU patients needs. Good documentation and communication between health professionals, the patient and family are vital parts of a daily review. Many of the disease processes and therapies described are discussed in more detail in later articles in this edition.

Sebastian Brown Specialist Trainee Sophia Bratanow Specialist Trainee Royal Devon and Exeter NHS Foundation Trust Exeter Rebecca Appelboam Consultant in Intensive Care Derriford Hospital Plymouth UK

Update in Anaesthesia | www.anaesthesiologists.org

page 11

General Principles

Anaesthesia

Update in

in the ICU and there is no documentation that a secondary trauma survey was completed, then the ICU team must take responsibility to perform it. Most would choose to assess the primary organ failure first, so in a head-injured patient, start with the central nervous system. Head/central nervous system General considerations If the patients primary pathology is a head injury, cranial surgery or a cerebral event then your assessment should be adjusted accordingly. The patients Glasgow Coma Score (GCS) should be recorded - for head-injured patients this is most usefully done when sedation has been stopped. If a painful stimulus is applied to assess the motor response, avoid repeating this procedure by different clinicians more than once a day. A full cranial and peripheral nerve examination should be performed daily where indicated - for example in those with fluctuating neurology due to a cerebral abscess or Guillain-Barr syndrome. Note the pupil size and reaction. Over-sedation is undesirable for a number of reasons and performing daily sedation breaks reduces length of stay on ICU.7 A sedation score such as the Richmond Agitation-Sedation Scale (RASS) may be used to monitor and titrate sedation appropriately.8 Delirium occurs in 1580% of critical care patients. It increases mortality and causes cognitive decline in the long-term.9 Delerium should be regularly sought and quantified using the CAM-ICU score and management steps, such as treatment with haloperidol, applied if appropriate.10,11 Despite the availability of adequate methods of analgesia and appropriate monitoring, pain control can be poor in ICU. Pain scores should be recorded and analgesia reviewed daily, particularly in postoperative patients. Most of the techniques that are applicable for postoperative patients on the surgical ward can be used in ICU and it is useful for intensivists to learn regional techniques such as rectus sheath and epidural insertion. Simple analgesics such as paracetamol should be prescribed routinely, although non-steroidal anti-inflammatory drugs are usually avoided in the critically ill. Patients with intracranial pathology Patients at risk of raised intracranial pressure should ideally be treated at centres with specialist input and, where available, intracranial pressure (ICP) monitoring should be considered for those requiring sedation and at risk of high ICP. Local policies targeting cerebral perfusion pressure (CPP, usually >60mmHg) should be followed when the ICP is greater than 20-25mmHg or when there is clinical or radiological evidence of a raised ICP. Standard neuro-protection includes treating patients head-up 30 degrees with the endotracheal tube taped rather than tied (to minimise obstruction to cerebral venous drainage), ventilation to a PaCO2 of 4.5-5kPa and the maintenance of a PaO2 greater than 8kPa. Glucose should be in the normal range and steps should be taken to avoid hyperthermia. Disorders of sodium metabolism are common in brain injury. Serum sodium should be maintained at the upper normal range. Ensure adequate sedation, analgesia and muscle relaxation. Seizures need prompt treatment and phenytoin is the preventative antiepileptic of choice. The administration of mannitol and hypertonic

saline is controversial, but they are often reserved for use in patients with high ICP or suggestive physical signs, for example a blown (fixed, dilated) pupil.16 Hyperventilation is a short-term measure to reduce critically high ICP before surgical intervention, but should be considered a rescue therapy only. In many centres around the world, ICP monitoring is not available, so patients with severe head injury are sedated and managed as above for 48 to 72 hours. After this time, daily sedation breaks allow assessment of their underlying condition. Respiratory and ventilation General considerations A past medical history of respiratory disease, including lung function tests, and current respiratory issues should be noted. The patients airway and respiratory system should be examined. If an endotracheal tube is in place, note that the length at the teeth is as documented at insertion and check its position is correct on the most recent chest Xray. Often it is only possible to auscultate the chest anteriorly and in the axillae. The ventilator settings should be inspected and the measured tidal volume, minute volume, peak and plateau pressures noted. Note whether the patient appears comfortable on these ventilator settings, in particular whether they are fighting (co-ordinating poorly with) the ventilator or display an increased work of breathing. The patients saturations and, where available, arterial blood gases should be inspected and trends noted. Regular arterial gas measurements of PaO2 and PaCO2, assessment of the PaO2:FiO2 ratio and pH are useful in guiding your ventilation strategy. If the clinical appearance, oxygenation or blood gases are not satifactory, then you must address this by altering the ventilator mode, settings or level of sedation to improve the situation. Set targets for gas exchange; these should be specific to each patient, so that a patient with severe COPD may have a target SaO2 of 88% or above. Acute respiratory distress syndrome (ARDS) ARDS occurs in up to 14% of ventilated patients, and carries a mortality of 40-60%.12 It arises as a complication in both pulmonary and non-pulmonary conditions and is diagnosed according to specific criteria (see Box 1). Low tidal volume ventilation of 6ml.kg-1 and a conservative fluid management strategy should be used in patients with ARDS.13,14 Aim for plateau pressures below 30cmH2O, allowing hypercapnia if necessary.15 High PEEP has been shown to be beneficial for patients with confirmed ARDS (PaO2/FiO2 < 200mmHg),16 and in severe left ventricular failure. Early paralysis with neuromuscular blocking agents may improve outcome in patients with ARDS with a PaO2/ FiO2 ratio < 150mmHg.17 Weaning The ICU clinicain should implement a strategy for gradual weaning of ventilation, from mandatory positive pressure ventilation to a progressive reduction in pressure support, to levels that simply compensate for the resistance of the endotracheal tube and the circuit. Tracheostomy is often used in the ICU to aid weaning from ventilation, and most are now placed using a percutaneous dilational

page 12

Update in Anaesthesia | www.anaesthesiologists.org

Box 1. Proposed new definition of ARDS (European working group and awaiting formal publication).

Mild ARDS
Timing Hypoxaemia Origin of oedema Radiological abnormalities Additional Physiological Derangement

Moderate ARDS

Severe ARDS

Acute onset within 1 week of a known clinical insult or new/worsening respiratory symptoms PaO2/FIO2 201-300mmHg with PEEP/CPAP 5cmH2O PaO2/FIO2 200mmHg with PEEP 5cmH2O PaO2/FIO2 100mmHg with PEEP 10cmH2O

Respiratory failure not fully explained by cardiac failure or fluid overload Bilateral opacities N/A Bilateral opacities N/A Opacities involving at least 3 quadrants Minute volume >10L.min-1 or compliance < 40ml.cmH2O-1

technique. The strength of a patients cough, their secretion load and swallow function should be assessed. In patients with a tracheostomy, the ability of the patient to tolerate deflation of the cuff and use of the speaking valve are important indicators of weaning progression. Where available, extubation to non-invasive ventilation may reduce the risk of reintubation in patients with COPD.18 Circulation A comprehensive examination of the cardiovascular system should be performed daily. This should include auscultation of the heart sounds and lung bases. Peripheral perfusion, pulses and the presence of peripheral oedema should be noted. Oedema will be present in the lower back and sacrum of a patient that has been supine for a prolonged period and this is a common finding in those who have been critically ill. Spontaneous clearance of oedema, with an accompanying diuresis, is usually a sign that an acute episode of sepsis is resolving. It is useful to chart observations of heart rate, blood pressure, capillary refill and interventions, such as fluid and inotrope administration, graphically, in order to identify trends. Baseline and serial ECGs are important in patients with ischaemic heart disease, to assess for ischemic changes associated with acute deterioration of the patient. Where available, transthoracic (TTE) or transoesophageal (TOE) echocardiography are useful in evaluation of the structure and function of the right and left ventricles and heart valves. The use of goal-directed fluid therapy, guided by cardiac output monitoring is controversial but may be of benefit in early sepsis,19 however many units lack the required equipment for this. The clinical response to fluid administration and, where available, central venous oxygen saturations (ScvO2) may be employed to guide the use of fluids, inotropes and vasopressors (See Box 3).19 The use of steroids should be reserved for refractory shock.5 Abdomen and nutrition The abdomen should be fully examined at least daily, as it is a concealed source of infection and subsequent driver of inflammation in critical illness. The presence of any surgical drains should be noted and the trends of collection volumes noted to see if further surgery is required, or whether the drain can be removed. Abdominal

pressure measurements may be required if abdominal compartment syndrome is suspected on examination. Where available the serum lactate provides a non-specific indicator of pathologies such as bowel ischaemia, that are difficult to detect clinically. Nasogastric (NG) tube placement should be confirmed on a daily basis by pH testing or chest Xray if being used for feeding. The NG tube should be removed as soon as it is no longer needed. The patients daily weights should be recorded as a basic nutritional assessment. The typical critical care patients energy needs are approximately 25kcal per kg per day.20 This may be doubled in severe sepsis, trauma and burns. Oral intake, NG feeding and any gastric residual volume should be used to calculate energy intake. If available, dietician support and the use of feeding guidelines,21 will aid adequate calorie, protein, fat, essential amino-acid and mineral input. If NG feeding fails, consider the use of post-pyloric feeding via a tube inserted through the stomach into the proximal small bowel. The potential for re-feeding syndrome should be considered in patients with poor dietary input prior to their ICU admission. Bowel output should be recorded, and diarrhoea noted and tested for infectious organisms such as Clostridium difficile that causes pseudomembranous colitis. Other causes of diarrhoea such as overflow, drugs, high-osmolar feed and intestinal ischaemia should be considered. Delayed gastric emptying is indicated by large aspirates from the NG tube. This is relatively common in critically ill patients and early administration of prokinetics, such as metoclopramide or low-dose erythromycin, is often required. Aperients may be required for constipation. Early enteral nutrition, is recommended to prevent stress ulceration of the stomach22 and to preserve mucosal integrity. Ranitidine or a proton pump inhibitor, such as omeprazole, should be given to ventilated patients who are not yet established on enteral feeding.22,23 Parenteral nutrition (PN) should be reserved for those patients in whom enteral feeding is contraindicated or failing.24 Renal, fluids and electrolytes The urine output should be charted every hour where appropriate. Most urinary catheters are colonised with bacteria, but these are usually not clinically significant. However, catheters should be removed if not

Update in Anaesthesia | www.anaesthesiologists.org

page 13

Box 2. Abbreviated surviving sepsis care bundle.

SURVIVING SEPSIS CARE BUNDLE (ABBREVIATED)1 Initial resuscitation (first 6hours) 1. Begin resuscitation immediately if hypotensive or lactate > 4mmol.L-1. Targets are: a. CVP 8-12mmHg b. MAP 65mmHg - norepinephrine or dopamine are first-line vasopressors. Use epinephrine as second-line in norepinephrine/dopamine refractory shock. If possible use an arterial catheter to guide vasopressor infusions. c. d. e. Urine output 0.5ml.kg-1.h-1. Do not use low-dose dopamine infusions for renal protection. Central venous O2 saturations 70% or mixed venous 65%. If venous saturation target missed: i. ii. iii. Consider further fluid, Transfuse packed red cells to a haematocrit of 30% and/or, Start dobutamine infusion 5-20mcg.kg-1.h-1. Do not increase cardiac index to supranormal levels.

2. Ventilation a. 6ml.kg-1 tidal volumes. Aim for plateau pressure 30cmH2O. b. Permissive hypercapnia may be required to minimize plateau pressures, except in patients with intracranial hypertension. Obtain appropriate cultures as long as this does not significantly delay antibiotic administration. Two or more blood cultures (one percutaneous culture and cultures from each vascular access device in place > 48hours). Perform imaging studies promptly to confirm and sample any source if safe to do so. Begin broad-spectrum antibiotics with good penetration to presumed source and active against likely pathogens as soon as possible, but at least within 1 hour of recognizing sepsis or septic shock. 3. Diagnosis a. b.

4. Antibiotic Therapy a.

b. Combination therapy should be considered for Pseudomonas infection or in neutropaenic patients, until culture susceptibilities are available. c. a. Stop antibiotic therapy if the cause is found to be non-infectious. Hydrocortisone < 300mg per day in divided doses can be considered for fluid and vasopressor-refractory shock. haemodialysis (IHD) or continuous veno-venous haemodiafiltration (CVVHD) in outcome, but CVVHD may be better tolerated in patients who are cardiovascularly unstable.28 Thrombocytopaenia is a common complication of renal replacement therapy and is usually due to consumption by the extracorporeal circuit, but other causes such as heparin induced thrombocytopaenia (HIT) should be considered. BLOOD TEsTs All of the patients blood tests should be reviewed and trends noted this is most easily viewed when plotted on a chart. Where available, ICU patients require daily measurement of renal function, electrolytes and haematology indices. Magnesium and calcium levels, clotting function and blood grouping for transfusion are frequently required. Low levels of magnesium (<0.7mmol.L-1) and phosphate should be treated by intravenous supplementation. 5. Steroids

required or in patients who are anuric due to renal failure. The trends in renal function and electrolytes should be examined frequently and correlated with the patients progress as a whole. The patients fluid administration should be reviewed and the daily and cumulative fluid balances noted. The use of crystalloid versus colloid fluid is still debated. The use of starch-based colloids does not improve survival and may cause renal impairment.25,26 The SAFE study showed no benefit of albumin over saline in all ICU patients and subgroup analysis suggested albumin may reduce mortality in sepsis, but increase it in traumatic brain injury.27 Dialysis or renal replacement therapy (RRT) may be required in hyperkalaemia, fluid overload, uraemia, acidosis, or poisoning due to a filterable toxin. There is no difference between intermittent

page 14

Update in Anaesthesia | www.anaesthesiologists.org

A conservative transfusion strategy is usually recommended in stable critically ill patients; aim for a haemoglobin level above 7g.dL-1,29,30 although a higher haemoglobin concentration may be targeted in patients with ischaemic heart disease and septic shock.19 Transfusion practice is greatly affected by local availability of donor blood and by the prevalence of diseases such as malaria within the population. Where available, platelet transfusion is usually guided by consultation with a haematologist, but should always be considered when: the platelet count is <5 x109.L-1 regardless of bleeding, 5-30 x109.L-1 in active bleeding, or <50 x109.L-1 when surgery or invasive procedures are planned. Some neurosurgery centres may aim for a platelet count of >100 x109.L-1 in cases of intracranial haemorrhage. Blood glucose control has been controversial, with a major recent study demonstrating that tight glucose control worsens outcome.31 Use an intravenous infusion of short-acting insulin or regular injections of subcutaneous insulin to keep blood sugar levels between 5 and 8mmol.L-1. MICrOBIOLOgY Sepsis can begin insidiously and may be difficult to recognise, but should be suspected if the patient is not progressing as expected. Many patients are relatively immunocompromised in response to their primary illness, and tend to develop secondary episodes of sepsis several days after admission. A thorough CNS, respiratory, cardiac and abdominal examination, looking for stigmata of infection, should be completed to identify the likely sources of infection. Management should follow the surviving sepsis bundle (see Box 2). Blood cultures and other microbiology samples should be taken and appropriate antibiotics administered within 1 hour.1 Each hour that appropriate antibiotic administration is delayed increases mortality by 8%.32 Microbiology input should be sought and antibiotics tailored to local pathogens and their known sensitivities. Antibiotics should be reviewed on a daily basis and stopped after an appropriate response and duration. RADIOLOgY Current and past imaging should be reviewed as required. A competent person should check every diagnostic test and document the results, to ensure that relevant information is not missed and that patients do not undergo unnecessary harmful procedures involving exposure to Xrays. There is no evidence that routine daily chest radiography is superior to clinically indicated studies. MEDICATIONs Scrutinise the patients medication chart on every ward round, stopping any unnecessary drugs and antibiotics. If the patient has impaired renal or hepatic function, special consideration should be made for the risks of each medication administered and the remaining medication should be dose-adjusted. Levels may be required for certain medications, such as digoxin and phenytoin. Each medication should be reviewed in light of the current diagnosis and issues affecting the

patient, for example the presence of ACE inhibitors or non-steroidal anti-inflammatory drugs in acute kidney injury. Ensure that, when appropriate, the patients usual drugs are restarted (e.g. antihypertensive drugs after an episode of sepsis). VAsCULAr ACCEss Routinely check any vascular access catheters for each patient. Your unit should have robust system for documenting the insertion date of each of these. If sepsis develops and no other source is evident, replace all venous and arterial catheters. There is no evidence to support routine replacement of venous catheters after a certain number of days, but suspicion of infection should increase the longer a cannula is in situ, particularly if there are local signs of infection (erythema, pus). FASTHUG The application of a final series of checks helps to ensure that all elements of good supportive care are in place. The FASTHUG assessment is one such system in common use (see Box 3).33 This simple assessment covers many aspects of important ICU care, that are often neglected, but will reduce the incidence of ventilator associated pneumonia (VAP), deep vein thrombosis (DVT), stress ulcers and malnutrition. Box 3. FASTHUG mnemonic33 FASTHUG F Feeding - Ensure nutrition has been assessed and that the patients nutrition needs are being met. A Analgesia - Pain should be assessed and pain relief given for the patients disease process and medical interventions (including as part of sedation strategy). S Sedation - Sedation should be assessed and patients not over-sedated. Daily sedation breaks should be considered. T Thromboprophylaxis - All patients should receive prophylactic dose subcutaneous low molecular weight heparin unless contraindicated. TED stockings or calf/foot pumps should be applied.

H Head-up - The head of the bed should be elevated to 30 - 45 degrees to reduce gastro-oesophageal reflux and nosocomial pneumonia in ventilated patients. U Ulcer prophylaxis - Ranitidine should be prescribed for ventilated patients, not established on enteral feeding. Once enteral feeding is established, it should be discontinued. G Glucose control - Aim to keep glucose levels 150mg.dL-1 (8mmol.L-1) using a validated protocol.

Update in Anaesthesia | www.anaesthesiologists.org

page 15

DOCUmENTATION Document all of your findings in a systematic way. Always clearly date and time your assessment of the patient. Make a clear problem list, followed by a plan for the day that relates to the problem list. It is useful to tick off the items on the plan as they are completed. FAmILY/NExT Of KIN Ask who the immediate family are and whether they have had any discussion with members of the nursing or medical staff. Should someone speak to them today to keep them up-to-date with changes in the patients condition? Document any discussions that you do have. OThEr POINTs Discuss the resuscitation status of the patient and check that any decisions about the levels of care offered in the case of clinical deterioration have been documented. Ask the nurse looking after that patient whether they have any other issues that have not been resolved in your assessment. Ask any other members of the team whether they have anything else to add. Explain your main findings and plans to the patient, in as much detail as appropriate. SUMMARY This system will guide you to perform a comprehensive assessment of your patient. In ICU rigorous attention to detail can make the difference between survival and death. Combine clinical skills with knowledge of current evidence to reach a diagnosis and guide your management of each patient you encounter. REFERENCES
1. Dellinger RP, Levy MM, Carlet JM, Bion J, Parker MM, Jaeschke R et al. Surviving Sepsis Campaign: international guidelines for management of severe sepsis and septic shock: 2008. Critical Care Medicine 2008; 36: 296-327.

8. Sessler CN, Gosnell MS, Grap MJ, Brophy GM, ONeal PV, Keane KA et al. The Richmond Agitation-Sedation Scale: validity and reliability in adult intensive care unit patients. Am J Resp Crit Care Med 2002; 166: 1338-44. 9. Ely EW, Shintani A, Truman B, Speroff T, Gordon SM, Harrell FE, Jr. et al. Delirium as a predictor of mortality in mechanically ventilated patients in the intensive care unit. JAMA 2004; 291: 1753-62. 10. 11. Ely EW, Inouye SK, Bernard GR, Gordon S, Francis J, May L et al. Delirium in mechanically ventilated patients: validity and reliability of the confusion assessment method for the intensive care unit (CAM-ICU). JAMA 2001; 286: 2703-10. Ely EW, Margolin R, Francis J, May L, Truman B, Dittus R et al. Evaluation of delirium in critically ill patients: validation of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU). Critical Care Medicine 2001; 29: 1370-9.

12. Rubenfeld GD, Caldwell E, Peabody E, Weaver J, Martin DP, Neff M et al. Incidence and outcomes of acute lung injury. NEJM 2005; 353: 1685 93. 13. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. NEJM 2000; 342: 1301-8. 14. Wiedemann HP, Wheeler AP, Bernard GR, Thompson BT, Hayden D, deBoisblanc B et al. Comparison of two fluid-management strategies in acute lung injury. NEJM 2006; 354: 2564-75. 15. Hickling KG, Walsh J, Henderson S, Jackson R. Low mortality rate in adult respiratory distress syndrome using low-volume, pressure limited ventilation with permissive hypercapnia: a prospective study. Critical Care Medicine 1994; 22: 1568-78. 16. Briel M, Meade M, Mercat A, Brower RG, Talmor D, Walter SD et al. Higher vs lower positive end-expiratory pressure in patients with acute lung injury and acute respiratory distress syndrome: systematic review and meta-analysis. JAMA 2010; 303: 865-73. 17. Papazian L, Forel JM, Gacouin A, Penot-Ragon C, Perrin G, Loundou A, et al. Neuromuscular blockers in early acute respiratory distress syndrome. NEJM 2010; 363: 1107-16. 18. Burns KE, Adhikari NK, Keenan SP, Meade M. Use of non-invasive ventilation to wean critically ill adults off invasive ventilation: meta analysis and systematic review. BMJ 2009; 338: 1574. 19. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, et al. Early goal-directed therapy in the treatment of severe sepsis and septic shock. NEJM 2001; 345: 1368-77. 20. Reid C. Nutritional requirements of surgical and critically-ill patients: Do we really know what they need? Proc Nutr Soc 2004; 63; 467-72. 21. Heyland DK, Dhaliwal R, Drover JW, Gramlich L, Dodek P. Canadian clinical practice guidelines for nutrition support in mechanically ventilated, critically ill adult patients. J Parenter Enteral Nutr 2003; 27: 355-73. 22. Cook D, Heyland D, Griffith L, Cook R, Marshall J, Pagliarello J. Risk factors for clinically important upper gastrointestinal bleeding in

2. Haynes AB, Weiser TG, Berry WR, Lipsitz SR, Breizat AH, Dellinger EP et al. A surgical safety checklist to reduce morbidity and mortality in a global population. NEJM 2009; 360: 491-9. 3. Ghaferi AA, Birkmeyer JD, Dimick JB. Variation in hospital mortality associated with inpatient surgery. NEJM 2009; 361: 1368-75. 4. Khuri SF, Henderson WG, DePalma RG, Mosca C, Healey NA, Kumbhani DJ. Determinants of long-term survival after major surgery and the adverse effect of postoperative complications. Annals of Surgery 2005; 242: 326-41; discussion 41-3. 5. Vincent JL, Singer M. Critical care: advances and future perspectives. Lancet 2010; 376: 1354-61. 6. Robb E, Jarman B, Suntharalingam G, Higgens C, Tennant R, Elcock K. Using care bundles to reduce in-hospital mortality: quantitative survey. BMJ 2010; 340: 1234. 7. Kress JP, Pohlman AS, OConnor MF, Hall JB. Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. NEJM 2000; 342: 1471-7.

page 16

Update in Anaesthesia | www.anaesthesiologists.org

 patients requiring mechanical ventilation. Canadian Critical Care Trials Group. Critical Care Medicine 1999; 27: 2812-7. 23. Marik PE, Vasu T, Hirani A, Pachinburavan M. Stress ulcer prophylaxis in the new millennium: a systematic review and meta-analysis. Critical Care Medicine 2010; 38: 2222-8. 24. Casaer MP, Mesotten D, Hermans G, Wouters PJ, Schetz M, Meyfroidt G et al. Early versus Late Parenteral Nutrition in Critically Ill Adults. NEJM 2011; 365: 506-17. 25. Schortgen F, Lacherade JC, Bruneel F, Cattaneo I, Hemery F, Lemaire F, et al. Effects of hydroxyethylstarch and gelatin on renal function in severe sepsis: a multicentre randomised study. Lancet 2001; 357: 911-6. 26. Brunkhorst FM, Engel C, Bloos F, Meier-Hellmann A, Ragaller M, Weiler N et al. Intensive insulin therapy and pentastarch resuscitation in severe sepsis. NEJM 2008; 358: 125-39. 27. Finfer S, Bellomo R, Boyce N, French J, Myburgh J, Norton R. A comparison of albumin and saline for fluid resuscitation in the intensive care unit. NEJM 2004; 350: 2247-56. 28. Vinsonneau C, Camus C, Combes A, Costa de Beauregard MA, Klouche K, Boulain T et al. Continuous venovenous haemodiafiltration versus

intermittent haemodialysis for acute renal failure in patients with multiple-organ dysfunction syndrome: a multicentre randomised trial. Lancet 2006; 368: 379-85. 29. Reiles E, Van der Linden P. Transfusion trigger in critically ill patients: has the puzzle been completed? Crit Care 2007; 11: 142. 30. Hebert PC, Wells G, Blajchman MA, Marshall J, Martin C, Pagliarello G et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. Transfusion Requirements in Critical Care Investigators, Canadian Critical Care Trials Group. NEJM 1999; 340: 409-17. 31. Finfer S, Chittock DR, Su SY, Blair D, Foster D, Dhingra V et al. Intensive versus conventional glucose control in critically ill patients. NEJM 2009; 360: 1283-97. 32. Kumar A, Roberts D, Wood KE, Light B, Parrillo JE, Sharma S, et al. Duration of hypotension before initiation of effective antimicrobial therapy is the critical determinant of survival in human septic shock. Critical Care Medicine 2006; 34: 1589-96. 33. Vincent JL. Give your patient a fast hug (at least) once a day. Critical Care Medicine 2005; 33: 1225-9.

Update in Anaesthesia | www.anaesthesiologists.org

page 17

Anaesthesia
Genral Principles
Intensive care medicine in rural sub-Saharan Africa who to admit?
Ray Towey* and John Bosco Anyai *Correspondence Email: [email protected]
INTRODUCTION St.Marys Hospital Lacor is a not-for-profit, church hospital situated in northern Uganda. It is in a rural area which, until recently, has suffered considerable insecurity and is in one of the poorest areas of rural sub-Saharan Africa. 1 The hospital has significant overseas support and patient care is subsidised in order to fulfil its mission of serving the poorest patients to the highest standards possible. There are approximately 500 hospital beds and 5000 operations are performed in the theatre block per year. Since July 2005 we have prospectively collected data on outcomes of all patients admitted to the ICU. The data of over 2000 patients is stored on an Access database. ICU STAFFING AND INFRASTRUCTURE The eight-bed ICU has 8 trained nurses and 4 assistant nurses. There is one anaesthetic officer assigned to the ICU who also covers the emergency theatres in the night, with one nurse anaesthetist. One overseas anaesthesiologist, has been attached to the ICU for the last 9 years and he is the only physician with a clinical responsibility totally to the ICU. All patients are admitted under the care of the admitting physicians, who also have duties in the main wards, labour ward, outpatients and theatre. There are no other dedicated ICU medical staff. The majority of the nurses on the ICU are not rotated around the main wards, as is often the custom in other institutions, so that a core of locally trained specialised ICU nurses has been retained. The ICU has no capacity for peritoneal dialysis or haemodialysis. There are no infusion pumps or blood gas analysis and it is only occasionally possible to estimate serum electrolytes. There are currently three Glostavent ventilators (Diamedica, UK) in ICU with a fourth in theatre and an adequate number of pulse oximeters and non-invasive blood pressure machines. The ratio of trained nurses to patients varies from 1 to 4 to 1 to 8. The ICU is a large open-planned area with two cubicle spaces, situated close to theatre (Figure 1).2 With a physical capacity of eight beds, and with the added possibility of admitting more patients on trolleys if required, the ICU is rarely physically short of beds to accept referrals from the hospital clinicians. However the nursing staff number is fixed so in busy times the ratio of nurses to patients suffers.

Update in

Summary
This article describes some of the factors to be weighed up when considering which patients are appropriate for admission to an intensive care unit in a country with limited resources. The authors describe their experience running an ICU in a rural part of Uganda, and use the audited outcomes of a cohort of 2,202 patients admitted over a six-year period.

RM Towey Consultant Anaesthesiologist John Bosco Anyai Anaesthetic Officer St.Marys Hospital Lacor Gulu Uganda

Figure 1. The intensive care unit at St Marys Hospital Lacor, Uganda.

page 18

Update in Anaesthesia | www.anaesthesiologists.org

ADMISSIONS POLICY Admission of patients to the ICU is open to any clinician, with no strict policy to guide this. Clinicians have discovered by a process of clinical experience how best the ICU could serve their patients and inappropriate admissions have been identified by ward round feedback on a daily basis. The difficulty of deciding who to admit to the ICU has both ethical and clinical factors. The concept of futility remains an issue no matter what resources are available and it remains an issue of continuous discussion in many guises. If ICU admission is refused then the patient will receive the level of care that is offered on the general wards, and so this must also be evaluated in order to compare that offered in the ICU. General ward care The patient ratio on the wards may range from one trained nurse to 30 patients to one trained nurse to 60 patients, with the night shifts often the most stretched. Any critically ill and unstable patient who is denied access to the ICU will then be admitted to the ward, where both the nurse to patient ratio, and the experience of the individual nurses to deal with these patients, is far less favourable than in ICU. However, we do not have data from ward patients for direct comparison to the ICU patient population. The issue of who to admit to the ICU and also who to discharge back to the ward is, in practice, an ongoing discussion between the ICU and ward-based clinicians. ICU ward rounds are conducted three times per day and the suitability of each patient for discharge back to the wards is discussed, in light of new referrals and the need to maintain a good nurse to patient ratio. ANALYSIS OF ADMISSIONS TO ICU The annual rate of admissions to ICU has grown over the last 5 years, from 264 patients in 2006 to 449 patients in 2010. The ICU mortality has remained at a steady level, ranging between 26 to 36% (Figure 2).

Thirteen arbitrary diagnostic groups were used. The development of our ICU has largely been driven by a need to meet the demands of the surgical, obstetric and gynaecological services. Medical admissions comprise 7% of admissions with a mortality of 47%. We have encouraged the development of a high dependency area within the general medical ward, containing oxygen concentrators and pulse oximeters, so that rational oxygen therapy can be administered without the need for ICU admission for this reason alone.

Figure 3. Outcomes in 2202 patients over the last 6 years, shown according to their admission specialty. Referred means admitted to the ICU but later referred to another hospital, usually the teaching hospital in Kampala.

Head injured patients, judged clinically to be unsuitable for ward management are admitted to the ICU, but our policy has always been not to undertake advanced respiratory support, with intubation and ventilation, in these patients. The limited number of nurses and ventilators, along with the expectation of poor outcomes, even with prolonged ventilation, has ensured that this policy persists today. On rare occasions intubation and ventilation has been commenced, when sputum retention is considered to be a major factor in deteriorating coma or when early referral to an urban area is an option. Of 282 head-injured patients admitted, 108 died, 154 were discharged to the ward and 20 were referred on to Mulago Hospital, Kampala for further care. Overall mortality for head injured patients admitted to ICU was 37%. All patients undergoing thyroidectomy are admitted to the ICU postoperatively for at least one night, as our experience is that a small number of patients develop airway problems postoperatively on the main ward. These are not reliably recognised and effectively managed on the general ward. Figure 3 shows that postoperative general surgical patients form the largest diagnostic group in our ICU patient population. The mortality for this group is 22%, perhaps reflecting that our non-physician anaesthetists recognise the importance of adequate preoperative resuscitation. During the intraoperative period active resuscitation, cardiorespiratory monitoring and respiratory support is continuous.

Figure 2. Analysis of admission to ICU at St Marys Hospital Lacor, with mortality data. Upper section of bar = number referred; second section from top = number discharged to ward; third section from top = number discharged home; bottom section = number died

The work of Fenton and colleagues, assessing the mortality of Caesarean sections in Malawi, demonstrated that 80% of deaths occur in the postoperative period.3 It is likely that general surgical deaths have a similar postoperative emphasis in Africa. Figure 3 shows our outcomes in 2202 patients over the last 6 years, shown according to their admission specialty.

Update in Anaesthesia | www.anaesthesiologists.org

page 19

We now view the postoperative period as the time of greatest risk to the patients - the general wards have a poor nurse to patient ratio, limited monitoring and limited senior staff available ward, meaning that monitoring of hypoxia and adequate blood and fluid replacement is difficult to establish. The non-physician anaesthetist assigned to the ICU is ideally situated, equipped and trained to identify high-risk postoperative patients needing blood, fluid and/or oxygen therapy, as well as pain relief. Ventilation in ICU Among the general surgical patients, 181 (23%) were also given intermittent positive pressure ventilation (IPPV) and the overall mortality of this sub-group was 52% (Figure 4).

they represent a small subgroup, patients suffering respiratory failure following snakebite and poisoning have a good outcome in our practice. Similarly, patients who developed acute pulmonary oedema occurring under general anaesthesia formed a small group with very good outcomes when treated with postoperatively ventilation. The majority of these patients had longstanding anaemia exacerbated by acute haemorrhage, predisposing to development of pulmonary oedema.

Figure 5. Outcomes of all ICU patients treated with IPPV.

Figure 4. Outcomes of ventilated patients by diagnostic group.

Postoperative IPPV is the main invasive ICU intervention that we can offer. Our theatre anaesthetist is trained to identify patients with cardiorespiratory instability during surgery or in the immediate postoperative period, that may benefit from postoperative IPPV. These patients are transferred to ICU for ventilation. We use the same type of ventilator/anaesthesia machine (the Glostavent) in theatre and in ICU, and this has made this process of postoperative IPPV much easier to manage and teach. Among the postoperative surgical patients who received IPPV, 83 (47%) survived. It is our view that the majority would have died if managed on the general ward without the facility for IPPV. A similar experience of postoperative support is seen with the 162 obstetric and gynaecological patients, in whom the overall mortality was 26%. Within this group 51 patients were given IPPV with a mortality of 53%, with 24 survivors. Again we believe that the majority would have died if they had received only ward care. Figure 5 shows that in the 378 patients from all diagnostic groups, who received IPPV, the mortality was 56%, with 150 survivors. Eight patients suffered snakebite with a neurotoxic venom, where muscle fasciculation and respiratory distress required IPPV. Six of these patients were among the survivors. Six patients were admitted with some form of poisoning, often a pesticide (organophosphate) compound with five, all of whom received IPPV, surviving. Although

Tetanus Tetanus is a disease category which requires special consideration as its treatment lends itself particularly to the skills of anaesthetists and intensivists. We have classified tetanus into neonatal, child and adult. In our experience neonatal tetanus outcomes have been very poor, with a mortality of 77% in 30 neonates admitted. Four patients were given IPPV and all died. We no longer offer IPPV for neonatal tetanus. The introduction of magnesium sulphate therapy for adult and child tetanus has, in our view, contributed to a major improvement in our outcomes, in comparison with the care previously offered in the general ward. Our protocols have been published in a previous edition of Update in Anaesthesia4 and follow the advice of anaesthetists in Sri Lanka.5 The mortality in 65 adults with tetanus receiving ICU management was 48% and in local conditions we consider this a remarkably good outcome, giving 34 survivors. In 23 of these patients tetanus was so severe that, despite large doses of magnesium sulphate, the spasms could not be controlled and so they were sedated, paralysed, ventilated and subsequently tracheostomy performed. There were 7 survivors in this group (mortality 80%). In our experience tetanus in children has a much better prognosis with an overall mortality of 15% from 33 children. In 11 children the spasms were so severe that they were sedated, paralysed, ventilated and received tracheostomy, yet only one of these children died (mortality 9%). The management of severe tetanus in children and adults is very demanding on the ICU nurses and anaesthetists, as it may require IPPV for up to 4 weeks, but is one of the most rewarding conditions to treat. Venous access is a challenging problem in these long stay patients and femoral and internal jugular lines are usually required. Severe burn patients are managed in the ICU and our mortality is 50%.

page 20

Update in Anaesthesia | www.anaesthesiologists.org

CONCLUSION The majority of patients admitted to our ICU in rural sub-Saharan Africa were postoperative surgical patients. The nurse to patient ratio, close supervision and assessment by anaesthetists and basic ABC interventions is superior to that available on the general wards. We believe that this has produced better outcomes for many diagnostic groups and consequently admissions rates have rapidly increased over that last six years, although we recognise that comparative data for patients receiving ward-based care is not available. The main sustainable and inexpensive invasive intervention offered in our ICU is postoperative IPPV. This is indicated in patients with reversible respiratory insufficiency and/or haemodynamic instability - conditions that would likely lead to death on the general wards. The most dramatic effective intervention for medical patients was IPPV for poisoning by pesticides. General medical patients remain a very small percentage of our admissions over the 6 years of data collection. Tetanus patients had good outcomes compared with ward care and very early in our experience all tetanus patients were treated in the ICU. Snake bite patients with neurotoxic paralysis also did well with

IPPV. Rational use of oxygen with oximetry monitoring and oxygen therapy using oxygen concentrators has proved to be a sustainable inexpensive and effective treatment for hypoxia from whatever cause. We continue to train the general ward clinicians and nurses on basic principles of rational oxygen therapy so that ICU admissions for this sole reason are reduced. REFERENCES
1. Available at: www.lacorhospital.org 2. Towey RM, Ojara S. Intensive Care in the Developing World. Anaesthesia 2007; 62 (Suppl): 32-7. 3. Fenton PM, Whitty CJM, Reynolds F. Caesarean Section in Malawi: prospective study of early maternal and perinatal mortality. British Medical Journal 2003; 327: 587-90. 4. Towey R. Tetanus: a review. Update in Anaesthesia 2005; 19: 43-8. 5. Attygalle D, Rodrigo N. Magnesium as a first line therapy in the management of tetanus: a prospective study of 40 patients. Anaesthesia 2002; 57: 778-817.

Update in Anaesthesia | www.anaesthesiologists.org

page 21

General Principles

Anaesthesia
Identifying critically ill patients: Triage, Early Warning Scores and Rapid Response Teams
Tim Baker*, Jamie Rylance, David Konrad *Correspondence Email: [email protected]
INTRODUCTION One of the most important tasks in a hospital is prioritising which patients to treat first. This is known as triage and should involve the quick and accurate detection of life threatening or serious illness. Triage facilitates timely clinical care and prioritises the use of the hospitals resources according to clinical need. The patients at highest risk can be cared for in emergency rooms or intensive care units, so that equipment and human resources can be concentrated at the point of greatest need. In emergency departments in high income countries, formal triage systems are ubiquitous. In low income countries, triage is often absent or of insufficient quality and may be one of the weakest parts of the health system. Queue-based systems are common, without effective mechanisms to prioritise the critically ill. Further identification of critical illness also takes place after admission to hospital. Such ward-based triage involves the regular assessment of clinical needs in order to detect the deteriorating in-patient.1 Rapid Response Teams, that may include staff from the intensive care unit, are called to provide emergency and critical care for ward patients identified by triggers in ward-based triage. Ward-based triage and Rapid Response Teams are relatively new and are becoming increasingly utilised in high income countries, but remain in their infancy in low income countries. triage and make recommendations for triage-on-arrival, Patient report The patients themselves provide the first source of triage information. However, assessing a patients illness severity through self-report is confounded by problems of both under reporting due to a lack of recognition and by over reporting in order to gain quicker medical attention. Nurse or clinician intuition Utilising the intuition of experienced clinical staff is quick and simple and many critically ill patients will be correctly identified in this way. However, such a subjective assessment is prone to considerable bias if used alone. Intuition has been adopted into most triage systems in high-income countries as an additional criterion for increasing sensitivity. Ability to walk Patients who are unable to walk by themselves are often critically ill. Noting the inability of a patient to walk is almost instantaneous and in a number of studies inability to walk has correlated with outcome.2 It is likely that this is a proxy measurement for a number of abnormalities, including reduced conscious level. Vital signs Abnormal vital signs (heart rate, respiratory rate, systolic blood pressure, conscious level, body temperature, oxygen saturation) have been shown to predict mortality in a low-income setting. In a Single Parameter Score (SPS) system, triggers are defined for each vital sign and any patient with one or more observation falling beyond these triggers is categorised as an emergency (see Table 2). The sensitivity and specificity of SPS depends on the defined triggers, and one limitation is that some seriously ill patients will be missed. SPS is well suited to inpatient use where time or resources are limited because of its simplicity, particularly when combined with a nurses intuition. Heart rate Heart rate at either extreme is associated with increased mortality; bradycardia is often a pre-terminal event and tachycardia a common finding in critical illness. Co-interpretation of heart rate and blood pressure abnormalities can be particularly useful.

Update in

Summary
This article describes the principles of triage of newly admitted patients and those who deteriorate on the ward. The existing systems are explained and the authors suggest those that are best suited to practice in a lowresource setting.

Dr Tim Baker MB ChB Department of Physiology and Pharmacology Karolinska Institute Section for Anaesthesia and Intensive Care Karolinska University Hospital Stockholm Sweden In this article we describe the available methods for

Dr Jamie Rylance BMBS ward-based triage and Rapid Response Teams that are Malawi-Liverpoolrealistic and feasible for use in hospitals in low-resource Wellcome Trust settings. Blantyre Malawi Dr David Konrad MD PhD Department of Physiology and Pharmacology Karolinska Institute Section for Anaesthesia and Intensive Care Karolinska University Hospital Stockholm Sweden

FORMS OF TRIAGE The methods for prioritising patients, including their advantages and disadvantages, are summarised in Table 1. Some of the methods such as vital signs and danger signs are suitable for prioritising according to clinical need and will be described in detail below. Others, such as prioritising by ability to pay, or unselective queue systems are in common use but do not help in identifying the critically ill.

page 22

Update in Anaesthesia | www.anaesthesiologists.org

Table 1. Methods for prioritising patients.

Method

Advantages

Disadvantages

Patient report Incorporated into most triage systems (patient history)

Quick and simple

Liable to distortion, or difficult to interpret

Intuition Quick and simple. May incorporate The intuition of an experienced health staff considerable clinical information Ability to walk Quick and simple Proxy measurement for a number of abnormalities Quick and simple. Uses routinely collected medical data. Well validated in highincome countries Scoring systems are well validated in highincome settings

Medical professionals may not accurately identify patients at risk If used alone will miss some critically ill patients and include some stable patients

Vital signs Single Parameter Score (SPS)

May miss some critically ill patients

Compound physiological abnormalities Early warning Scores

Time-consuming Minimum information available on the performance of the scores in resource limited settings

Danger signs ETAT3 World Health Organisation quick check

Incorporates early emergency May be time-consuming management with triage. System shown Emergency treatments may delay the to be beneficial when non-skilled staff triage of other patients. provide healthcare (for example in children using IMCI) Sophisticated, likely to have good sensitivity and specificity SATS validated in South Africa Rapid directed care by specialist Complex, requires training. Not validated in low-income setting

Complex triage systems SATS4 Manchester Triage System Specialty e.g. Pregnant women redirected to the obstetrician. Trauma teams represent a hybrid system with multiple specialties involved Queuing First-come-first-served

Minimal use of severity markers; may delay the point at which action is taken

Simple. Requires minimal oversight therefore personnel concentrate on healthcare delivery

Doesnt prioritise according to clinical need

Means to pay Direct admission to a fee-paying ward

Ethically difficult to defend Does not promote equity of access to healthcare.

ETAT=Emergency Triage and Treatment; IMCI=Integrated Management of Childhood Illness; SATS= South African Triage Score

Update in Anaesthesia | www.anaesthesiologists.org

page 23

Table 2. Single parameter score.

Adult normal Critical illness range triggers

Heart rate 50-90 <40min-1 >130min-1 Respiratory rate 10-20 <8min-1 >30min-1 Systolic BP 100-140 <90mmHg

Conscious level Reduced conscious level is a common finding in critically ill patients. The simple AVPU scale of conscious level (A = Awake; V = responds to Voice; P = responds to Pain; U = Unresponsive) allows conscious level to be objectively measured and documented, and deterioration identified. The Glasgow Coma Scale is good for predicting prognosis, but may be overly complex for triage. Body temperature Abnormal temperature can indicate severe illness. However, temperature has relatively less value where febrile illness may be trivial but common, for example adult malaria in endemic regions. Oxygen saturation The World Health Organization recommends the use of oxygen for patients with raised respiratory rate or oxygen saturations less than 90%. Where pulse oximeters and oxygen concentrators or other delivery devices are available, using hypoxaemia as a marker of critical illness is reasonable. Research is required in low-income country settings to identify which patients would benefit most from oxygen therapy. Compound scores of physiological abnormality - Early Warning Scores (EWS) Combining several vital signs may improve the accuracy of triage decisions. Compound scores encompass multiple measurements, each of which may be graded according to the degree of derangement. An aggregated score summarises all of this data into a single number (see Table 3). With increasing score the risk of mortality rises.2 A threshold can be chosen above which the patient is labelled as critically ill or an action is taken. For example a healthy person with all measurements within the normal range has an EWS of 0, whereas a critically ill patient has been defined as anyone with a score of 5 or more.

Conscious level Awake, alert Any sudden deterioration Responds only to Pain or Unresponsive Temperature Oxygen saturation 36C - 37.5C 95% >39C <90%

Respiratory rate Respiratory rate is one of the most sensitive single physiological parameters in prediction of mortality. However, medical staff are often reluctant to count respiratory rate due to perceived time pressure and it is unfortunately often omitted. Systolic blood pressure Hypotension is a good marker of severe disease and can reflect diverse pathologies such as depleted intravascular volume, loss of vascular tone in septic shock, poisonings or cardiac failure. Although both systolic and diastolic values are relevant, for simplicity in triage, the systolic alone can be used.
Table 3. Early Warning Scores.

3 Respiratory Rate Heart rate Systolic BP Temperature Conscious level (AVPU) Trauma Mobility <71

2 <9 <41 71-80 <35

0 9-14

1 15-20 101-110

2 21-29 111-129 >199 38.5

3 >29 >129

Commonly used EWS

41-50 81-100

51-100 101-199 35-38.4 Alert No Waking

Reacts to voice Yes with help

Reacts to Pain

Unresponsive

The South African Triage Score (SATS) adds the following stretcher/ immobile

Note: EWS= Early Warning Score; BP= Blood Pressure; the threshold for critical illness is 5 for EWS or 7 for SATS Example: A patient with a respiratory rate of 32 (3 points), heart rate 120 (2 points), systolic BP 110 (0 points), temperature 37 (0 points) and conscious level V (1 point) has an EWS of 6 points and is therefore critically ill.

page 24

Update in Anaesthesia | www.anaesthesiologists.org

Early Warning Scores give a fuller picture of physiological derangement than single parameter scores and are generally more sensitive, but less specific, depending on the chosen threshold. In other words most patients who are critically ill will be identified using an EWS, however in addition some stable patients will be mistakenly labelled critically ill. Scoring systems have been further refined by the addition of other factors, for example taking into account age as part of a modified EWS, age specific values to improve paediatric use, or the presence of trauma (South African Triage Score, see Table 3). A recently proposed score from the UK includes the electrocardiogram (ECG) as a parameter, further improving sensitivity, but increasing complexity.5 Early Warning Scores have been devised from data from well resourced settings with relatively low rates of HIV. In other areas, such as in sub-Saharan Africa, the limited evidence available suggests that EWS perform less well. Differences in study endpoints, however, have made comparisons difficult and the impact of the higher prevalence of HIV is not well known. Such limitations require that EWS systems are properly validated in the setting in which they will be used. Danger signs Danger signs are physiological findings or conditions that indicate that the patient is critically ill. This is the basis for the Paediatric Emergency Triage and Treatment (ETAT) guidelines that have been developed by the World Health Organization.3 The danger signs form a checklist that is simple to follow, standardizes the triage process and allows coupling of triage to early life saving interventions. A patient with any danger sign is classified as an emergency and the checklist indicates which investigations or treatments should be initiated. For example a patient showing the danger sign reduced conscious level is an emergency, hypoglycaemia should be suspected and the airway should be kept clear. The introduction of ETAT in resource limited settings has been shown to reduce mortality, but adult guidelines have been harder to draft. Complex triage systems In high-income countries, triage systems based on algorithms involving diagnosis and physiological parameters are most commonly used. Examples are the Manchester Triage System and the Australian Triage Scale. The validity of these systems in low-income countries has not been determined, they are time consuming and their implementation requires extensive training. The triage system with most relevance to low-income countries is the South African Triage Scale (SATS).4 SATS (previously called the Cape Triage Score) is a nationwide triage system in South Africa established in 2006.4 It is based on a modified EWS using physiological signs (see Table 3) and adds in several diagnoses and clinical signs termed discriminators, such as pain, hypoglycaemia, seizures and nurses intuition, that can modify the triage category. Each patient is given one of four categories: red (immediate), orange (very urgent see within 10 minutes), yellow (urgent see within 60 minutes) and green (delayed priority). SATS assessment takes 2-4 minutes. It has been validated in South Africa, and is increasingly promoted in similar settings in other countries. Combining triage and treatment Many triage systems categorise patients into three levels of urgency:

Emergency; Priority; Non-urgent. Clear documentation of the triage findings and the patients category of urgency can be done with a stamp on the patients notes, or colour coding stickers, such as red for emergency, yellow for priority and green for routine. Integrating the triage system with the provision of clinical care allows rapid action to be taken.6 Emergency patients can be transferred to a resuscitation room, a senior clinician can be called and emergency treatments can be given. Where danger signs are used, emergency treatments can be recommended for each finding. However, with increasing complexity of intervention, the system becomes less of a triage tool and more a management guide. Provision of clinical care may delay the triage of subsequent patients. Ward-based triage and Rapid Response Teams Physiological derangement amongst in-patients is a precursor to adverse events such as cardiac arrest and death. This observation has led to the development of systems for early recognition and treatment of such individuals, and the advent of Rapid Response Teams (RRT), also known as Medical Emergency Teams (MET). Regular and systematic ward-based triage can be used for all patients based on single parameter or compound scores. Where ward staff recognise abnormalities beyond a predefined trigger level, they call for medical support. The RRT may consist of physicians, specialist nurses, anaesthetists or others with specific skills in acute medical care. Some are combined with ICU outreach services, whilst others are independent. Such heterogeneity of structure and service makes generalisation of their impact difficult. Before-and-after studies in single centres have shown reduced cardiac arrest rates and even reduced hospital mortality by approximately 20% in high-income countries.7 These findings have been confirmed by a recent meta-analysis.8 RRTs are uncommon and have not been formally studied in lowincome countries. While the model is attractive, significant alterations may be necessary depending on the available facilities. Research should be directed at identifying areas of maximum impact. RECOMMENDATIONS The following are the authors recommendations for triage in a hospital with limited resources. Suggested audit criteria for triage-on-arrival are presented in Table 4. Triage on arrival to hospital Every hospital should have a formal triage system for new patients. It is important to consider when and where triage should be done, who should do it and how it should be done. When and where Triage should precede registration processes and payment for services, where these are demanded. Emergency patients should be triaged and treated irrespective of ability-to-pay. The triage area should be close to the hospital entrance and should provide privacy. A resuscitation or emergency room should be situated nearby for the immediate treatment of the critically ill. Who As triage has the potential to save lives and reduce costs, it should

Update in Anaesthesia | www.anaesthesiologists.org

page 25

be a prioritized activity and senior staff should be appointed where possible. This is often a skilled nurse with specific training. However, where resources are limited, triage can be performed by any staff member doctor, nurse, auxiliary nurse, porter, gateman, record clerk, cleaner as long as they understand the principles of triage and have been trained in recognising the emergency patient.
Table 4. Suggested audit criteria for triage-on-arrival.

ward-based triage system. Vital signs should be checked regularly and we recommend the use of the Single Parameter Score system with the triggers in Table 2. An EWS could also be used but it is more complex and time-consuming and may be difficult to implement. The intuition of the ward nurse is a valid additional criterion. Rapid Response Teams Where patient deterioration has been identified, an immediate response should follow. Interventions can include emergency ABC treatments such as freeing a blocked airway, providing oxygen, or intravenous fluids, or calling senior staff and ensuring that the patient is seen first on the ward round. Staff should be trained in the acute management of the critically ill patient and there should be clear guidelines for the most common emergencies. Rapid response teams should be considered where resources allow, but may only be useful where intensive care facilities are present. CONCLUSION Although seen as a vital part of hospital systems in much of the world, triage in low-income countries has been neglected. Introducing simple and realistic triage for patients on arrival and on the wards prioritises resources, based on clinical need, and has the potential to reduce mortality.6 Research is required into the optimal methods of triage when resources are limited; which triggers should be used for instituting medical care and which treatments are appropriate; the effects of HIV status on triage scores; and the cost-effectiveness of Early Warning Scores and Rapid Response Teams. REFERENCES
1. NICE. Acutely ill patients in hospital: recognition of and response to acute illness in adults in hospital. National Institute for Health and Clinical Excellence Clinical Guideline 50. 2007.

Triage initiated

Patients arriving to hospital should be given a triage category There should be sufficient documented evidence to support this decision

Triage action appropriate

There should be documentation of all vital signs: walking status, airway patency, HR, RR, SBP, conscious level, temperature and, where available, oxygen saturation A patient with reduced conscious level should have their blood glucose checked or sugar given, within 5 minutes A patient with any overt bleeding should have immediate measures to prevent further blood loss A patient with respiratory distress (RR >30min-1 or oxygen saturation <90%) should be given oxygen within 5 minutes

Triage decision implemented

Emergency patients should be seen by a clinician within 10 minutes Urgent patients should be seen by a clinician within 60 minutes

2. Rylance J, Baker T, Mushi E, Mashaga D. Use of an early warning score and ability to walk predicts mortality in medical patients admitted to hospitals in Tanzania. Trans R Soc Trop Med Hyg 2009; 103: 790-4. 3. WHO. Emergency Triage Assessment and Treatment (ETAT). Geneva: WHO Press; 2005. 4. Wallis PA, Gottschalk SB, Wood D, Bruijns S, de Vries S, Balfour C. The Cape Triage Score -- a triage system for South Africa. S Afr Med J 2006; 96: 53-6. 5. Kellett J, Deane B, Gleeson M. Derivation and validation of a score based on Hypotension, Oxygen saturation, low Temperature, ECG changes and Loss of independence (HOTEL) that predicts early mortality between 15 min and 24 h after admission to an acute medical unit. Resuscitation 2008; 78: 52-8.

Triage categories reasonable

Patients who die within 48 hours of admission should have been classified as emergency

HR= Heart Rate; RR= Respiratory Rate; SBP= Systolic Blood Pressure How Triage must be quick and simple. After a triage decision has been made, patients should be moved quickly to appropriate areas in order to decongest the assessment area. The choice of triage system should depend on the available human and physical resources. Whichever is used, it should be fairly and consistently applied. The most useful methods of triage are likely to be the vital signs, ability to walk, the triage nurses intuition and the danger signs described above. Ward based triage Staff shortages often preclude regular full reassessment in resource-poor settings. Identifying the deteriorating inpatient is however necessary for reducing mortality and all hospitals should have at least a basic

6. Molyneux E, Ahmad S, Robertson A. Improved triage and emergency care for children reduces inpatient mortality in a resource-constrained setting. Bull World Health Organ 2006; 84: 314-9. 7. Konrad D, Jaderling G, Bell M, Granath F, Ekbom A, Martling CR. Reducing in-hospital cardiac arrests and hospital mortality by introducing a medical emergency team. Intensive Care Med 2010; 36: 100-6. 8. Chan PS, Jain R, Nallmothu BK, Berg RA, Sasson C. Rapid Response Teams: A Systematic Review and Meta-analysis. Arch Intern Med 2010; 170: 18-26.

page 26

Update in Anaesthesia | www.anaesthesiologists.org

Critical care where there is no ICU: Basic management of critically ill patients in a low-income country
Tim Baker* and Jamie Rylance *Correspondence Email: [email protected]
INTRODUCTION Critical care can be defined as all care given in hospital to patients with serious and reversible disease. The burden of critical illness is especially high in developing countries.1 Over 90% of maternal deaths, child deaths, deaths from sepsis and deaths from trauma occur in developing countries.2 50% of child deaths in hospitals occur within 24 hours of arriving at the hospital. One survey from South Africa found that as many as one in four of medical admissions is critically ill.3 High-income countries can afford resource-intensive and sophisticated services for managing critical illness. In countries with much lower healthcare spending, there is a need for inexpensive critical care, but there are many barriers to its provision. Processes for prioritising and caring for critically unwell patients are not routinely implemented. Life saving drugs and equipment are not immediately available. Medical guidelines often lack relevance and treatments may not be evidence-based for resource-poor settings. Staff training in the management of critical illness is uncommon, and intensive care units (ICUs) are rare. Critical care has not been promoted as it cuts across traditional disciplines and lacks advocates. Critical care need not be expensive. Cheap treatments such as adequate fluid resuscitation to children with diarrhoea and intravenous dextrose for hypoglycaemia can be life saving. Emergency triage and treatment for children in a hospital in Malawi costs only US$1.75 per patient and has reduced hospital mortality by 50%.4 Oxygen therapy can cost between one and six US dollars per day (WHO figures). In this article we describe critical care services which are feasible in a district hospital in a low-income country. We focus on the hospital structure, routines and basic clinical management and do not discuss advanced interventions such as mechanical ventilation and dialysis, or care within specialist intensive care units. THE HOSPITAL STRUCTURE Critically ill patients arriving at hospital require early identification and treatment. An appropriate physical environment can facilitate this. Formal triage systems at the entrance to hospital should divide the patients into urgent and routine cases, and direct the urgent cases to a resuscitation room or emergency department. Dividing the patients in this way can be both clinically and cost effective, as resources can be focused on those who have the most pressing clinical needs. The emergency department should be adjacent to the hospital entrance and have facilities designed for managing emergency patients. There should be resuscitation bays or rooms for immediate treatments, with emergency drugs and equipment always at hand. Medical staff should be present or on-call 24 hours-a-day and have senior staff who can be called quickly for complicated or serious cases. Treatment rooms should be spacious to allow a team of several health professionals to work efficiently together and communication between practitioners must be prioritised; a quiet place with good access to radiology, laboratory and surgical provision is ideal.

Summary
Many of the life saving therapies delivered in an ICU can be delivered in the most resource-poor settings. This article describes how infrastructure, personnel and clinical care may be designed to deliver effective care for all critically ill patients.

Within the hospital, at least 1-2% of beds should be assigned for the critically ill.5 This means at least 4-8 beds in a 400-bed hospital. An ICU can concentrate expertise and resources and provide good critical care. Staff can receive directed training in managing the critically ill, effective routines can be set up and emergency drugs and equipment can be kept near the patients who need them most. However, there is a risk that an ICU could divert already scarce resources from the rest of the hospital: it should provide treatments and facilities consistent with the rest of the healthcare system. Where a separate ICU is not possible, Tim Baker MB ChB designating beds on a general ward as critical care or Department of Physiology and Pharmacology high dependency beds improves medical oversight. Karolinska Institute Where resources allow, hospitals can introduce a Rapid Section for Anaesthesia and Response Team. This is a team of hospital staff trained Intensive Care in critical care who may be summoned to support the Karolinska University care of seriously ill patients on a general ward. Rapid Hospital Response Teams can improve communication between Stockholm the wards and the ICU. They can provide critical care Sweden treatments outside the designated ICU and provide on the job critical care training to general staff. Jamie Rylance BMBS Malawi-Liverpool-Wellcome IDENTIFYING THE RIGHT PATIENTS Trust Triage is the quick and accurate detection of patients Blantyre with life threatening illness. Formal triage systems are Malawi

Update in Anaesthesia | www.anaesthesiologists.org

page 27

General Principles

Anaesthesia

Update in

ubiquitous in hospitals in many parts of the world, but in low-income countries triage is often absent or of poor quality. Queue-based systems are common and can result in delays for the critically ill patients and less rational prioritisation of the hospitals resources. Every hospital should have a formal triage system for new patients. Triage should precede registration processes and payment for services. The triage area should be close to the hospital entrance and be near to (or part of ) the emergency department. As triage has the potential to save lives and reduce costs, it should be a prioritised activity, with senior staff appointed where possible. Triage must be quick and simple. The choice of triage system depends on the available human and physical resources. Most triage systems involve vital signs, early warning scores or danger signs. Vital signs Abnormal vital signs (heart rate, respiratory rate, systolic blood pressure, conscious level, body temperature, oxygen saturation) have been shown to predict mortality in a low-income setting. Limits or triggers can be defined for each vital sign and any patient with one or more observations falling beyond these triggers is categorised as critically ill (see Table 2 in previous article). Early warning scores Combining several vital signs may improve the accuracy of triage decisions. Compound scores or Early warning scores encompass multiple physiological measurements, each of which may be graded according to the degree of derangement.6 An aggregated score summarises all of this data into a single number. With increasing score the risk of mortality rises. A threshold can be chosen above which the patient is labelled as critically ill or an action is taken. Danger signs Danger signs are physiological findings or conditions that indicate that the patient is critically ill. The danger signs form a checklist that is simple to follow, standardises the triage and allows for the coupling of triage to early life saving interventions. A patient with any danger sign (for example reduced conscious level) is classified as an emergency, and the checklist indicates which investigations or treatments should be initiated. Ward-based triage Further identification of critical illness also takes place after admission to hospital. Such ward-based triage involves the regular assessment of clinical status in order to detect the deteriorating inpatient. Vital signs should be checked regularly and the triggers used for defining critical illness. The intuition of the ward nurse is a valid additional criterion. Postoperative critical care Postoperative patients can leave theatre in a critical state, due to the effects of the surgery and anaesthesia. Many of these patients have a good prognosis if they receive adequate critical care for a limited period of time. Indeed, many ICUs have begun as postoperative units. Hospitals with ICUs should have the capacity to manage the critically ill postoperative patient. Hospitals without an ICU should have a recovery room, where the patients can be cared for directly after the

operation, and critical care or observation beds on the general ward. An initiative for predicting postoperative risk, the Surgical Apgar Score, has recently been developed.7 Based on three intra-operative parameters: estimated blood loss, lowest heart rate and lowest mean arterial pressure, it has been shown to provide an objective indication of risk and could be used for post-operative triage in hospitals in lowincome countries (see Appendix 1). Criteria for admission to ICU An ICU should have well defined admission criteria. These criteria depend on the facilities and expertise available but should be based on the hospitals triage systems. The goal is to admit the patients to the ICU who could most benefit from the critical care, i.e. those who have life threatening conditions and have a reasonable chance of recovery. Equally important are discharge criteria. Those patients who have sufficiently improved and no longer require critical care, or those who are judged to be too severely ill to benefit from the available care should be discharged from the ICU to free up beds for other critically ill patients. SIMPLE ROUTINES Although hard evidence of effective critical care interventions is lacking, it is clear that earlier treatment, more intensive monitoring and more goal-based systems have been beneficial. Increasing staff to patient ratios improves all of these and may be the single most important factor for successful critical care. Regular physiological observations can identify deterioration early and monitor the success of interventions. Frequent assessment by medical staff is similarly important twice daily ward-rounds of critically unwell patients and 24 hour access to a clinician should be routine. The most effective interventions for the critically unwell patient are simple, but need to be carried out quickly. Emergency drugs and equipment such as diazepam, oropharyngeal airways, oxygen delivery equipment, intravenous fluids and giving sets should be kept on the ward and always be available. A full list of emergency drugs and equipment is in Appendix 2. Keeping these well stocked is challenging: supplies may be erratic, used items may not be replaced and equipment may be borrowed for use elsewhere. For an efficient emergency service, these disruptions must be minimised. A list should be kept on the ward and daily stock-taking and equipment testing by designated in charge clinical staff should be carried out. Critically ill patients should not be required to pay before they have access to life-saving therapies and relatives and staff should not need to leave the ward to find or purchase the treatments. The patients observations, received treatments and fluid balances should be regularly documented. This enables early recognition of the deteriorating patient, monitors the success of the care and reduces errors in drugs prescription and dispensing. Documentation can also be useful for quality control and audit. Basic hygiene routines including hand washing before and after patient contact and use of disposable gloves should be rigorously followed to reduce nosocomial infections.

CLINICAL MANAGEMENT
Effective clinical management of the critically ill patient involves concentrating on the common and easily preventable causes of

page 28

Update in Anaesthesia | www.anaesthesiologists.org

mortality. These are often described with the ABCDE acronym. The clinician should begin by assessing the Airway, treat any abnormality found, then successively assess and treat the Breathing, Circulation and Disability (neurological dysfunction) before moving on to Everything Else. This approach is used in emergency and critical care training all over the world and has been found to be effective and easy to remember, even in stressful circumstances. The details of clinical management are covered in detail in several other articles in this edition of Update. IMPROVING QUALITY OF CARE Staff should be adequately trained in caring for critically ill patients. Training includes both pre-service in colleges and universities, and in-service through courses and seminars. Evaluation and feedback from external senior critical care specialists can be valuable. National and local guidelines and standards for managing the critically ill are rare and should be developed to encourage improved care. All hospitals should have a system of audit for evaluating the care they are providing. Additionally, specific case discussion as part of mortality and morbidity meetings (M and M) is useful. This evaluates the strengths and weaknesses in the medical care of fatal cases. It is fundamental to their success that blame attribution is not pursued, but sensitive discussions may identify specific areas for improvement. ETHICAL ISSUES Critical care brings with it several specific ethical issues. These issues are extremely important and health staff should have an understanding of them. Critically ill patients and their relatives should be treated sensitively and with respect. Decisions are frequently made that have a huge impact on the patients lives and sometimes the patient is incompetent to make decisions themselves due to their illness.
Table 1. Requirements for a simple critical care service.

End-of-life care has great social, cultural and religious, as well as medical, importance. Continued active treatment of a patient, who will not benefit from it, should be avoided. Palliative care services and adequate pain-relief can improve support for patients and their carers at the end of life. The point at which treatment becomes palliative rather than active depends on the wishes of the patient, informed by senior medical opinion and may often be affected by cultural norms. CONCLUSION The requirements for a simple critical care service are summarised in Table 4. In order to ensure such services are routinely available, critical care needs to be moved up the policy agenda. Strong advocacy is required, even if hampered by the lack of critical care specialists. Training in critical care is crucial and should be aimed at both newly trained and professionally established healthcare workers. While there remain so few critical care physicians, other clinicians must be trained in providing effective care to critically unwell patients. Increasingly, hospitals in less developed settings are developing links with those from high-income countries. This has the potential to improve training, and act as a catalyst for improvement in care. However, care should be taken that such initiatives are truly collaborative and well grounded within the existing hospital systems. Research is required to give a better understanding of critical care in low-income settings, to evaluate the clinical effectiveness of critical care interventions and to establish their cost effectiveness. The WHO recently began promoting surgical services as a way to reduce mortality and morbidity. Critical care should be next in line.
1. Dunser MW, Baelani I, Ganbold L. A review and analysis of intensive care medicine in the least developed countries. Crit Care Med 2006; 34:1234-42.

REFERENCES

Hospital structure Identifying the right patients Routines for Critical Care

Emergency department ICU or critical care beds on a ward Triage on admission Triage on wards Increased nurse:patient ratio Regular observations Regular ward rounds Senior medical review Emergency drugs and equipment to hand, restocked, no need to pay Documentation

Clinical management Improving the quality of care

ABCDE approach Supportive care Training and supervision Guidelines Audit and clinical governance

Update in Anaesthesia | www.anaesthesiologists.org

page 29

2. Baker T. Critical care in low-income countries. Trop Med Int Health 2009; 14:143-8. 3. van Zyl-Smit R, Burch V, Willcox P. The need for appropriate critical care service provision at non-tertiary hospitals in South Africa. S Afr Med J 2007; 97: 268-72. 4. Molyneux E, Ahmad S, Robertson A. Improved triage and emergency care for children reduces inpatient mortality in a resource-constrained setting. Bull World Health Organ 2006; 84: 314-9.

5. 6. 7.

Watters D, Wilson I, Leaver R, Bagshawe A. Care of the Critically Ill Patient in the Tropics. Oxford: Macmillan Publishers; 2004. Rylance J, Baker T, Mushi E, Mashaga D. Use of an early warning score and ability to walk predicts mortality in medical patients admitted to hospitals in Tanzania. Trans R Soc Trop Med Hyg 2009; 103: 790-4. Haynes AB, Regenbogen SE, Weiser TG, Lipsitz SR, Dziekan G, Berry WR, et al. Surgical outcome measurement for a global patient population: validation of the Surgical Apgar Score in 8 countries. Surgery 2011; 149: 519-24.

APPENDIX 1. The 10-point Surgical Apgar Score7

0 point Estimated blood loss (ml) Lowest mean arterial pressure (mmHg) Lowest heart rate (beats per min) >1000 <40 85*

1 point 601-1000 40-54 76-85

2 points 101-600 55-69 66-75

3 points 100 70 56-65

4 points

55*

*Occurrence of pathologic bradyarrhythmia, including sinus arrest, atrioventricular block or dissociation, junctional or ventricular escape rhythms, and asystole also receive 0 points for lowest heart rate. The Surgical Apgar Score is calculated at the end of any operation, from the estimated blood loss, lowest mean arterial pressure and lowest heart rate entered in the anesthesia record during the operation. The score is the sum of the points from each category.

page 30

Update in Anaesthesia | www.anaesthesiologists.org

APPENDIX 2. Emergency equipment and drugs for use in Critical Care. This is a generic list and should be modified according to local and national resources, pathologies and formularies. Modified from: WHO Essential Trauma Care Guidelines; WHO Generic Essential Emergency Equipment List; Baker et al Standards for Good Quality Emergency and Critical Care in Low Income Countries (unpublished).

Equipment Drugs Clock with second hand Oral rehydration solution Gloves - clean IV glucose 5% Gloves - sterile IV glucose 50% (or other concentration 10%) Sharps disposal IV crystalloid (Normal saline or Ringers lactate) Running water Diazepam Soap Paracetamol Oropharyngeal airway (adult and paediatric sizes) Parenteral penicillin (or equivalent) Suction machine (foot powered or electric) Parenteral gentamycin (or equivalent) Suction catheters - size 16FG Parenteral quinine (or equivalent) Laryngoscope (working and spare batteries) Ketamine Endotracheal tubes adult and paediatric sizes Lignocaine for local anaesthesia Rigid neck collar Epinephrine (adrenaline) Sandbags/towel rolls and head restraints Atropine Chest tube & underwater seal (or equivalent) Frusemide Sterilised surgical set for small procedures Nifedipine or other anti-hypertensive Oxygen concentrator/cylinder with face masks or nasal prongs and tubing Aminophylline Pulse oximeter Salbutamol (for inhaler or nebuliser) Resuscitator bag and mask (Ambu bag) Hydrocortisone Stethoscope Insulin Foetal stethoscope IV/IM opioids e.g. morphine Blood pressure monitoring equipment Naloxone IV cannulae adult size (e.g. 18G) Thiopentone IV cannulae paediatric size (e.g. 22G, 24G) Succinylcholine IV giving sets Non-depolarising muscle relaxant Needles Oxytocin/ergotamine Syringes at least 2ml, 5ml Magnesium sulphate Lumbar puncture needles Phenobarbital/phenytoin Urine catheters & bags Gauze and bandages Skin disinfectant Torch (and spare batteries) Electricity 24hours/day Telephone or other emergency communication system Light suitable for clinical examination Bedside blood glucose testing device and strips Thermometer Refrigerator Weighing scales adult and paediatric Nasogastric tubes Spacer device for inhaled salbutamol

Update in Anaesthesia | www.anaesthesiologists.org

page 31

Monitoring in the ICU

Anaesthesia
Monitoring in ICU - ECG, pulse oximetry and capnography
Ben Gupta Correspondence Email: [email protected]
INTRODUCTION In the intensive care unit (ICU), monitoring a patients physiological parameters is an important part of their overall care package. Monitoring alerts you to any deterioration in a patients condition and also helps you to asses their response to treatment. In this article we will consider three of the most commonly used electronic monitoring systems - electrocardiogram (ECG), pulse oximetry (SaO 2) and capnography. While these monitoring systems are important and useful, it should be remembered that they are always an addition to, rather than a replacement for, good clinical monitoring of heart rate, blood pressure, capillary refill time, respiratory rate, neurological status and urine output. WHAT ARE THE BENEFITS OF MONITORING? ECG, SaO2 and CO2 monitoring systems require a source of power and usually additional consumables such as the pulse oximetry probe or the D-fend water trap of the capnograph. They also require technical expertise for maintenance and repair. These aspects all make them potentially problematic in a low resource environment. However the benefits include: Additional clinical information. ECG, SaO2 and CO2 give very useful information about your patients cardiorespiratory function. This information is continuous and in real time and so is especially useful in critically ill patients. Non-invasive. These monitors are non-invasive, and so are well tolerated. Early warning system . The monitors alarm systems can be adjusted to detect deviation of parameters from acceptable levels, thus providing a prompt warning of any change in the patients condition. Careful attention to the trends of these deviations will alert you to early signs of clinical Ben Gupta deterioration. Specialist Trainee Bristol School of The following section considers each monitoring system Anaesthesia in turn, assessing what it can, and what it cant, tell UK you. The physics behind each type of monitor is not described, but can be found in previous editions of Update in Anaesthesia.1-4 PULSE OXIMETRY (SaO2) If allowed only one form of monitoring, many anaesthetists would choose a pulse oximeter, reflecting how useful and informative this equipment can be. Most pulse oximeters are stand-alone units, usually battery-powered, but it may also be incorporated as part of a larger multipurpose monitor. It consists of a sensor probe, usually placed on the patients finger, and a screen to display measured values. What can it tell you? The most important information available from this monitor is the arterial blood oxygen saturation (SaO2), which is given as a percentage. In a normal person breathing air, a value of 96-100% is normal. Note that in smokers and those with chronic lung disease the value is likely to be slightly lower, around 92-95%. Critically ill patients, particularly with a primary (e.g. pneumonia) or secondary (e.g. acute respiratory distress syndrome) lung disease will have impaired gas exchange and low oxygen saturations. The target level of saturation, achieved by administration of oxygen and ventilation, should be set with reference to their usual respiratory status. For example it is reasonable to aim for an SaO2 of 88% in a patient with underlying chronic lung disease with an intercurrent infection. Most SaO2 monitors display a value for the heart rate and emit an audible tone in time with the heart beat. The pitch of the pulse tone varies as the SaO2 level changes; it is difficult to guess the SaO2 when you hear the tone in isolation, but a change in the pitch of the tone alerts you to look at the monitor. A pulse waveform is also included on the screen of some monitors - this gives information about the quality of the signal and indicates whether a low recording is likely to be genuine. In general, if a good signal is received, this indicates that perfusion to the patients extremities is good. This also has specific role where the perfusion of a limb is at risk, for example following trauma or vascular surgery. A weak or absent signal, should alert you to assess the patients perfusion and blood pressure. Be aware that the signal will transiently disappear if the blood pressure cuff inflates on that arm.

Update in

Summary
Monitoring serves an important role in ICU as an adjunct to clinical evaluation. It is important that all staff are familiar with their equipment in order to use it safely and to the maximum benefit of the patient. In particular, the settings of the alarms must be tailored to each patient, with consideration to their age, weight and pathology.

page 32

Update in Anaesthesia | www.anaesthesiologists.org

What cant it tell you? The SaO2 only tells you part of the picture regarding oxygen delivery to the tissues, as this is also dictated by the haemoglobin level and the cardiac output. A patient with a hemoglobin level of 3.5g.dL-1 may have a SaO2 of 100%, but have poor oxygen content in their blood and therefore low oxygen delivery to the tissues. The SaO2 value is determined by the effectiveness of both ventilation (i.e. the pump function of the lungs) and gas exchange across the alveolar-capillary membrane. However ineffective ventilation (for example, due to airway obstruction, opioid excess or neuromuscular weakness), causes type 2 respiratory failure, in which CO2 accumulates. Pulse oximetry gives no indication of the arterial CO2 level - a drowsy patient may have a reassuringly normal SaO2 level if oxygen is being administered, but may have severe respiratory acidosis with a PaCO2 over 10kPa and be close to cardiovascular collapse. Tips for successful use Bright or fluorescent ambient light causes interference with signal detection, as can bright sunshine. The effects of external light can be minimized by covering the hand and probe with a dark material. Patient movement can lead to artifact on the trace and inaccurate measurements - this is a particular problem with combative or agitated patients. If transferring a patient on an uneven road, taping the probe to the finger so that it moves with the limb may help. Different types of probe are available and if you have a choice, use one that is appropriate for your circumstances. All probes work in the same way; they are just designed to fit different sized patients and different parts of the body. Smaller ear probes are available as well as special probes for children. If no paediatric probe is available, an adult finger probe can be placed around a small childs hand or foot. Paediatric probes often come as single use stickers designed to go around a babys hand or foot - when supplies are short these can be re-used after wiping gently with a cleaning swab, as long as they are not soiled. These probes will also fit well onto an adults finger and do not fall off during transfers. In a cold or shocked patient try placing the probe on a more central site. A finger probe can be clipped inside the patients mouth to detect through their cheek. Alternatives are the nose or earlobe. A small ear probe can also be placed onto the cheek, the lip or onto a nostril. ELECTROCARDIOGRAM (ECG) MONITORING ECG monitoring in ICU usually involves display of one lead - lead 2 - and measures the electrical activity of the heart along its long axis from right to left. Three electrodes are required for this - one on the right shoulder (usually red), one on the left shoulder (usually yellow) and one placed on the left side of the chest (usually green). Lead 2 is felt to detect most arrhythmias, which is the main role of ECG monitoring in the ICU setting. What can it tell you? The heart rate is calculated by the monitor by averaging the number of complexes over a set period of time. If the patient has an irregular rhythm such as atrial fibrillation the rate is calculated most accurately if the calculation period is set at the longest available.

Arrhythmias are usually diagnosed by setting the alarm limits at a high and low limit, to detect tachy- and bradyarrhthmias. The default alarm settings are usually appropriate for a healthy adult undergoing anaesthesia, but may be inappropriate for critically ill adults or for young children. An adult with sepsis may have a heart rate of 120 per minute, which will be continuously above the default high heart rate setting. The alarms can be adjusted manually to levels that would represent a clinically significant deviation from their current reading. Some monitors allow you to set the upper and lower alarm limits at 10% above and below the current measured value. Some more advanced modules are able to recognize patterns and diagnose arrhythmias, however it is often down to the clinician to identify the nature of the arrhythmia (artifact due to movement or shivering is commonly interpreted as ventricular fibrillation). It is useful to use ECG and pulse oximetry in conjunction; onset of a broad complex tachycardia with loss of the pulse oximetry waveform indicates pulseless ventricular tachycardia (VT), a medical emergency. It is often useful to print a rhythm strip on a piece of paper in order to study the rhythm more closely (e.g. looking for P-waves). It is also usually possible to pause the screen to allow further analysis. Be aware that some ECG signals are misread by the monitor, for example large T-waves may be counted as separate QRS complexes, doubling the measured rate. Again this can be resolved by comparing to the waveform and heart rate of the pulse oximeter. Multi-channel monitors that combine ECG, pulse oximetry (and invasive arterial blood pressure) usually default to show the heart rate from the ECG reading, but this can be changed to read from a different channel. What cant it tell you? When a patient develops myocardial ischaemia, a single lead ECG may show morphology changes if the ischaemia happens to be in the area of the heart that corresponds to the single lead position. Otherwise it will be missed. You should request a full 12-lead ECG to assess all areas of the myocardium if you suspect myocardial ischaemia. A normal ECG trace does not always indicate a well patient; in the case of a pulseless electrical activity (PEA, formerly electro-mechanical dissociation) cardiac arrest, the patient has no cardiac output despite the fact that the ECG may be displaying normal sinus rhythm. Always check that what the monitor tells you corresponds with your patients clinical appearance. Tips for successful use Poor quality ECG monitoring can be due to poor contact between the electrodes and sweaty or dirty skin. Clean the skin thoroughly and allow it to dry completely before applying electrodes. If a patient is shivering or moving around then interference on the screen may give the appearance of an arrhythmia. If you do not have any ECG sticker type electrodes, you can improvise by using a small piece of saline soaked gauze to couple the ECG lead to the skin. CAPNOGRAPHY (CO2) MONITORING There are various types of CO2 monitoring available but most systems consist of a connector, placed in series with the patients breathing

Update in Anaesthesia | www.anaesthesiologists.org

page 33

More basic systems do not give a waveform but are nonetheless very useful. Is CO2 present in the patients expired gas? This information alerts you to a serious problem with the patient. If there is no CO2 detected, either: the patient is not being ventilated e.g. displaced endotracheal tube, circuit disconnection, or no CO2 is being delivered to the lungs because circulation has ceased (cardiac arrest). Both of these situations obviously need urgent attention. Where available it is mandatory to have capnography present during induction and intubation of the critically ill, in order to rapidly confirm tracheal intubation, or to identify incorrect placement. How much CO2 is present? The CO2 reading at end-expiration (end-tidal, ET-CO2) most accurately represents the PaCO2, but note that the ET-CO2 level is generally 0.5kPa lower than the PaCO2. However this difference is not predictable in all patients, particularly those with major mismatches between perfusion and ventilation of their lungs. For some conditions, such as acute head injury, the PaCO2 level is critical and so, where available, arterial sampling is useful. Even if performed very irregularly, it can be used to quantify the end-tidal:arterial CO2 difference, so that capnography can be used more effectively to alter the patients ventilator settings. CO2 output from the lungs is dictated by: 1. The rate at which CO2 is produced and transported to the lungs (i.e. the patients metabolic rate and cardiac output). 2. The patients minute ventilation (MV) - this is the volume of gas moving into the lungs per minute and is the product of respiratory rate and tidal volume. The relationship is inverse, i.e. if the MV rises the PaCO2 falls. The monitor will give you a number in mmHg, kPa or percentage, with the percentage very close numerically to the kPa value, since atmospheric pressure is 101kPa. As a guide, 4-6kPa or 35-45mmHg are normal values in healthy non-smokers. Often the trend in CO2 and the rate of rise or fall is more important than the actual value. For example, a rising CO2 in a ventilated patient could indicate that their lungs are becoming less compliant, as they develop ARDS, or that their metabolic rate has increased, as they develop sepsis. The capnography waveform In order to recognize abnormal waveform patterns a normal waveform for a circle system is shown in Figure 2. What cant it tell you? All of the scenarios described above are guides to be assessed in line with clinical review of the patient. Sudden loss of the capnograph trace may represent equipment failure or blockage of the sample tubing with water vapour. For this reason the sample tubing should always be on the ventilator side of a heat and moisture exchanger.

b
Figure 1. Monitoring screens for a patient who has sinus beats followed by couplets and triplets of pulseless ectopic beats. In (a) the heart rate source (circled) is the pulse oximeter, correctly reading the effective pulse as 40bpm. In (b) the ECG has been selected as the heart rate source, recording a rate of 188 per minute (square), that corresponds to the number of ECG complexes, whether they generate an effective cardiac output or not. Either of these selections can be misleading if the ECG or pulse oximetry monitoring is viewed in isolation.

system that is attached, via a sampling line, to a monitor. The monitor analyses the gas and the values are displayed on a screen. To get the most accurate reading, the connector in the breathing system should be placed as close as possible to the patients mouth. If it is placed distant from the patients mouth, falsely low readings result, as alveolar gas is diluted with fresh gas from the circuit tubing. What can it tell you? CO2 monitoring systems can tell you three important things: 1. Whether CO2 is being detected in the patients expired gas or not, 2. The partial pressure of CO2 (capnometry), 3. It provides a continuous CO2 waveform plotted against time (capnography).

page 34

Update in Anaesthesia | www.anaesthesiologists.org

Figure 5. Up-sloping phase 3. An upslope to the phase 3 plateau (arrow) is often in seen in patients with a prolonged expiratory time e.g. in lung disease such as chronic obstructive pulmonary disease (COPD). A longer time for expiration may be required if the patient is being mechanically ventilated. Figure 2. Normal capnography trace; Phase 4 corresponds to the onset of inspiration, Phase 1 corresponds to inspiration. During this phase the waveform should return to zero. If it doesnt, this indicates an element of rebreathing; Phase 2 corresponds to the onset of expiration. As alveolar gas containing CO2 mixes with dead space gas the level of CO2 in the breathing circuit rises. Phase 3 is the plateau phase and corresponds to expiration of pure alveolar gas. The CO2 value at the end this phase is the end tidal CO2 (ET-CO2) and is the value displayed by the monitor. It is normal to have a very slight upslope during the plateau phase. The level of CO2 detected in the circuit drops as fresh gas is inspired.

The following are examples of abnormal traces that you might see on the ICU:

B
Figure 3. Rebreathing - the baseline does not return to zero (arrow) and may increase over time. This is commonly caused by exhausted soda lime or in inadequate gas flows for the breathing circuit in use. Figure 6. Superimposed waveform. Sometimes you might see smaller waves or oscillations on the normal CO2 waveform. This might represent cardiac oscillations whereby a pressure wave from the heart is transmitted to the airway (A). Alternatively it might represent regular attempts by a patient to breathe over the top of mechanical ventilation (B) - as a cleft in the CO2 waveform.

SETTING ALARMS A major factor in the effective use of monitors to alert nurses and clinicians to a change in the status of the patient, is the sensible use of the machines alarm systems. Most monitors come with preprogrammed alarm settings, set by the manufacturer for an average adult patient. Typical settings are as follows (but may differ between manufacturers):
Figure 4. Sudden decrease in CO2 waveform. This may represent an interruption in ventilation e.g. breathing circuit disconnection or a sudden decrease in cardiac output e.g. cardiac arrest.

SaO2: Will alarm if less than or equal to 94% ET-CO2: Will alarm if less than 4kPa or greater than 6kPa

Update in Anaesthesia | www.anaesthesiologists.org

page 35

 ECG: Will alarm if heart rate less than 60 or greater than 100 beat per minute. Some monitors will detect arrhythmias. As you become more familiar with the patient you are treating you may decide to alter the alarm settings. For example, a fit young patient may have a normal heart rate of 40-50 and you will want to alter your alarm settings accordingly. If you do not do this, then the alarm will sound continuously, be repeatedly silenced and will loose its power to alert you to clinically important changes. As evidence of this, you may notice that the silence or suspend buttons on your monitors wear out well before the other buttons! Beware of relying completely on the monitor alarms - someone may have re-set them to limits that are not appropriate to your patient, or even turned the alarm system off altogether. CONCLUSION Using capnography, pulse oximetry and ECG monitoring can be an invaluable addition to treating a patient in the ICU setting, increasing safety and optimising treatment. Remember that all monitors are only

as good as the person using them - think about what you are measuring, set your alarms appropriately and always use them in conjunction with clinical examination. FURTHER READING
1. Lee J. ECG monitoring in theatre. Update in Anaesthesia 2000; 11:16 32. Available at: http://update.anaesthesiologists.org/2000/06/01/ ecg-monitoring-in-theatre/ 2. Brown Z, Gupta B. Electrical signals and their measurement. Update in Anaesthesia 2008; 24,2: 164-9. Available at: http:// update.anaesthesiologists.org/2008/12/01/biological-signals-and their-measurement/ 3. E Hill, MD Stoneham, S-J Fearnley. Pulse oximetry - practical applications. Update in Anaesthesia 2008; 24,2: 156-9. Available at: http://update.anaesthesiologists.org/2008/12/01/practical applications-of-pulse-oximetry/ 4. Grant McFadyen. Respiratory gas analysis. Update in Anaesthesia 2008; 24,2: 170-3. Available at: http://update.anaesthesiologists. org/2008/12/01/respiratory-gas-analysis/

page 36

Update in Anaesthesia | www.anaesthesiologists.org

Invasive blood pressure monitoring

Ben Gupta Correspondence Email: [email protected]
INTRODUCTION Invasive (intra-arterial) blood pressure (IBP) monitoring is a commonly used technique in the Intensive Care Unit (ICU) and is also often used in the operating theatre. The technique involves the insertion of a catheter into a suitable artery and then displaying the measured pressure wave on a monitor. The most common reason for using intra-arterial blood pressure monitoring is to gain a beat-to-beat record of a patients blood pressure. ADVANTAGES OF IBP MONITORING Continuous beat-to-beat blood pressure monitoring is useful in patients who are likely to display sudden changes in blood pressure (e.g. vascular surgery), in whom close control of blood pressure is required (e.g. head injured patients), or in patients receiving drugs to maintain the blood pressure (e.g. patients receiving inotropes such as adrenaline). DISADVANTAGES OF IBP MONITORING The arterial cannula is a potential focus of infection, although arterial lines become infected far less frequently than venous lines, especially central venous lines. The arterial catheter can lead to local thrombosis which may result in emboli travelling down the limb or occasionally arterial occlusion this is rare if the catheter is kept flushed with saline and an appropriate vessel is chosen. The radial, femoral and axillary arteries may be used, as may the arteries of the foot, the posterior tibial and dorsalis pedis arteries. Where possible, the brachial artery should be avoided as this is an end artery and has no collateral supply occlusion of the brachial artery will result in loss of blood supply to the lower arm. Any drug inadvertently administered into the arterial line may form crystals and cause catastrophic ischaemia of the limb. Examples of drugs with which this has been reported are thiopentone and antibiotics. All arterial lines should be clearly labelled and the tubing colour coded (usually with a red stripe) to avoid confusion. Drugs should never be administered via the arterial line. The insertion of an intra-arterial blood pressure monitoring system can be difficult and time consuming, especially in shocked patients. This can potentially distract from other problems that need more urgent attention.

Summary
Where the facilities exist, invasive blood pressure monitoring provides continuous monitoring of changes in blood pressure in unstable patients with critical illness. It allows also frequent assessment of arterial blood gases and lactate, in order to gauge progression of disease and the response to therapies. This article describes the theory of the technique in order that inaccuracy in measurement can be minimised. We also provide practical tips for insertion of appropriate catheters. Arterial pulse contour analysis for estimation of cardiac output is described in more detail in a later article in this section.

The technique allows accurate blood pressure readings at low pressures, for example in shocked patients.  The trauma of repeated cuff inflations is avoided in patients who are likely to need close blood pressure monitoring for a long period of time e.g. ICU patients. Intravascular volume status can be estimated from the shape of the arterial pressure trace, either by eye or by waveform analysis by a specific device e.g. a pulse contour analysis system. IBP measurement allows accurate assessment of blood pressure in certain patients not suitable for non-invasive blood pressure monitoring, e.g. patients with gross peripheral oedema in ICU or morbidly obese patients.

The monitoring equipment, spare parts and cannulae are expensive when compared to non invasive methods of blood pressure monitoring. The arterial monitor requires an electrical supply which will limit its usefulness in some settings. COMPONENTS AND PRINCIPLES OF IBP MONITORING The components of an intra-arterial monitoring system can be considered in three main parts (see Figure 1): 1. the measuring apparatus, 2. the transducer, 3. the monitor. Ben Gupta Specialist Trainee Bristol School of Anaesthesia UK

The indwelling arterial cannula is convenient for repeated arterial blood sampling, for instance for arterial blood gases. This is not usually the sole reason for insertion of an indwelling arterial catheter.

Update in Anaesthesia | www.anaesthesiologists.org

page 37

Monitoring in the ICU

Anaesthesia

Update in

safety features such as high and low mean blood pressure alarms and tachycardia and bradycardia alerts.

Figure 1. Components of an arterial monitoring system.

The measuring apparatus The measuring apparatus consists of an arterial cannula (20G in adults and 22G in children) connected to tubing containing a continuous column of saline which conducts the pressure wave to the transducer. The arterial line is also connected to a flushing system consisting of a 500ml bag of saline pressurised to 300mmHg via a flushing device. Formerly 500IU heparin was added to this fluid, but many centres now consider this to be unnecessary. The flush system provides a slow but continual flushing of the system at a rate of approximately 4-5ml per hour. A rapid flush can be delivered by manually opening the flush valve. There is also usually a 3-way tap to allow for arterial blood sampling and the ejection of air from the system if necessary. The three-way tap must also be clearly labelled as arterial, to minimise the risk of inadvertent intra-arterial injection of drugs. For small children a smaller volume of flush is administered via a syringe driver, so that it is not possible to over-administer fluid by repeated flushing of the arterial cannula. The transducer A transducer is any device that converts one form of energy to another for example, the larynx is a type of physiological transducer (air flow converted to sound). The output of transducers is usually in the form of electrical energy. In the case of intra-arterial monitoring the transducer usually consists of a flexible diaphragm with an electric current applied across it. As pressure is applied to the diaphragm it stretches and its resistance changes, altering the electrical output from the system. The transducers used are differential pressure transducers and so must be calibrated relative to atmospheric pressure before use. The monitor It is not necessary for the anaesthetist to have an in-depth understanding of the internal workings of the monitor. Modern monitors amplify the input signal; amplification makes the signal stronger. They also filter the noise from the signal unwanted background signal is removed with an electronic filter - and display the arterial waveform in real time on a screen. They also give a digital display of systolic, diastolic and mean blood pressure. Most monitors incorporate various

Figure 2. Invasive blood pressure monitoring (boxed). The waveforms are usually colour coded (red for the arterial trace) and the monitor displays the systolic/diastolic BP, with the mean arterial BP in brackets below.

ACCURACY OF IBP MONITORING The accuracy of intra-arterial monitoring is affected by several important physical principles - the oscillation, natural frequency, damping and resonance of the system Oscillation A swinging pendulum is an example of a system that oscillates. When a pendulum is pushed (energy is put into the system), it moves away from its resting position, then returns to it. The resting position for a pendulum is at the bottom of its arc of swing and is dictated by gravity. However, the pendulum doesnt usually just return to the resting position, but tends to overshoot, swinging past the resting point in the opposite direction to the original push. This cycle continues until all the energy put into the system has been dissipated. The tendency of a system to move either side of set point is referred to as its tendency to oscillate. Damping Imagine you have two identical pendulums. One has recently been well greased at its point of rotation (fulcrum) and the other is stiff from rust. When an equal sized force is applied to each, the well greased one will oscillate freely around the set point but the old rusty pendulum may barely move. This is because much of the energy put into the system will be used up or damped in overcoming the frictional force of the rusty axis. The rusty pendulum will tend to oscillate at smaller amplitude (i.e. smaller swings) and for a shorter period of time than the well greased one. How freely a system oscillates following an input of energy is dependant on the degree of damping in the system. A well damped system tends not to oscillate freely whereas a poorly damped system may oscillate wildly. The amount of damping inherent in a system can be described by the damping coefficient (D) which usually lies between 0 and 1 (but can be greater than 1). A system with a D value greater than 1 describes a system that is over-damped, will

page 38

Update in Anaesthesia | www.anaesthesiologists.org

not oscillate freely and takes a long time to initially move away from and to return to its resting point (a high friction pendulum). A D value less than 1 and approaching 0 describes a system that is underdamped, that oscillates freely, moving rapidly away from its resting point and back again, but tends to overshoot and then oscillate around the resting point (a low friction pendulum). A D value of exactly 1 is known as critical damping.

If the input of energy into a system is occurring at the same frequency (or close to) the natural frequency, a phenomenon called resonance occurs and the output amplitude of the oscillations is greatly magnified. In the case of intra-arterial blood pressure monitoring this could lead to over-reading of the systolic blood pressure. Arterial pulsation is a complex sine wave and is composed of many individual sine waves. It is therefore important that the natural frequency of the measuring equipment (the catheter and column of saline etc) does not correspond to any of the component frequencies of the arterial pulsation input. This is achieved by making sure that the natural frequency of the measuring system is raised above any of the component frequencies of the arterial sine waveform. The characteristics of the measuring equipment that will ensure that the natural frequency of the system is higher than that of the arterial pulsation are: Arterial catheter must be short and with the maximum gauge possible, Column of saline must be as short as possible, The catheter and tubing must be stiff walled, The transducer diaphragm must be a rigid as possible. SETTING UP THE ARTERIAL LINE AND TROUBLE SHOOTING The usual location for insertion of the arterial catheter is the radial artery. The advantage of the radial artery is that it is superficial, easily accessible, and there is a collateral blood supply to the hand from the ulnar artery. It is advisable to perform Allens test to detect adequacy of collateral supply to the hand via the ulnar artery, although the test is not infallible and can only be performed in conscious patients (see Figure 4). The brachial artery should be avoided if at all possible (no collateral supply); the femoral artery, the ulnar artery, arteries of the foot and ankle, and even the axillary artery should be used in preference if necessary. Whichever location of artery is used, the distal limb should be monitored regularly for signs of emboli or distal ischaemia. Insertion of a radial arterial line This should be performed as an aseptic technique. The wrist should be cleaned with alcoholic chlorhexidine solution prior to cannulation and, in conscious patients, the skin should be infiltrated with 1% plain lidocaine. The arm should be abducted in the anatomical position and the wrist should be hyper-extended to aid cannulation (the radial artery is brought closer to the skin surface and the hand moved out of the way). This is most conveniently done by an assistant. If an assistant is not available use tape to secure the patients hand fingers extended over a bag of fluid (see Figure 5). There are various types of stiff, short arterial catheter available. Some feature a simple cannula over needle design, similar to an intravenous cannula and some incorporate a guide-wire as part of a Seldinger technique the needle is inserted, a wire passed through the centre of the needle, the cannula threaded over the wire (Figure 6). The correct cannula to use is the type with which you are most comfortable. Ideally a cannula with an injection port should not be used, as this

Figure 3. Graph showing the effect of different levels of damping on the oscillation of a measuring system.

Oscillations are undesirable in physiological measuring systems. These systems require accurate measurement of a maximum amplitude (for instance, that caused by the arterial pulsation), with a rapid response time and rapid return to the set point, ready for the next measurement. The ideal level of damping applied to a measuring system is a compromise between achieving a rapid response time and accurate reflection of maximum amplitude i.e. a system with D close to 0, and needing a system that returns to the resting point without excess oscillation (D around 1). In the case of an IBP monitoring system this would represent the difference between using very compliant measuring apparatus (compliant catheters, tubing) i.e. D approaches 0, and very stiff or non-compliant equipment i.e. D is closer to 1. The value of D chosen for physiological measuring systems such as IBP monitoring equipment lies between 0.6 and 0.7 it is known as optimal damping. Natural frequency and resonance A pendulum of set length and with a set weight at the end will always oscillate at exactly the same frequency, no matter what the initial starting point of the oscillation. In other words, whether you give the pendulum a small push or a really hard shove it will make the same number of oscillations per unit time (although the amplitude of the oscillations will alter). This is why pendulums can be used to keep time. Any system such as this will have a frequency at which it naturally oscillates. This frequency is known as the natural frequency.

Update in Anaesthesia | www.anaesthesiologists.org

page 39

a
Figures 4a and 4b. Allens test. Ask the patient to make a fist, use your thumbs to occlude the patients radial and ulnar arteries. Ask the patient to unclench their fist the palm will remain pale (a), whilst the blood supply is still occluded. When you remove your thumb that is occluding the ulnar artery, the palm will flush red if the ulnar artery is functional (b).

Figure 5. Technique for securing the patients wrist extended, using adhesive tape and a fluid bag. Figure 6. Two 20G arterial cannulae. The lower cannula has a guidewire that can be slid into the artery through the needle to allow smooth placement of the cannula (inset).

may be confused with an intravenous cannula - if such a cannula is used, the injection port should be taped over and the cannula clearly labelled as arterial. Make sure that you tape the cannula securely in position, and take care not to kink the cannula as you do so. Sometimes it is advisable to suture the arterial line in place.

The arterial catheter should be connected to the tubing, the transducer secured in a position approximately level with the heart and the

page 40

Update in Anaesthesia | www.anaesthesiologists.org

a
Figure 8. If location of the artery is difficult an alternative method involves positioning your thumb so that the radial pulse is running directly under the centre of your thumb. Then advance the cannula at 30 degrees under the centre point of your thumb.

If you hit the artery but fail to cannulate it a couple of times, it is often wise to move to the other wrist; the artery will go into spasm following repeated trauma making cannulation progressively more difficult.

Inserting an arterial catheter in shocked patients is very difficult. Do not waste time making repeated attempts to do so; resuscitation of the patient is more important!

b
Figure 7. Cannula insertion. The usual insertion technique is to palpate the artery with the fingers of one hand and locate the artery with the cannula at an angle of about 30 degrees (a).Once a flashback has been obtained the cannula should be brought level with the skin and then advanced 2-3mm further (b). This should ensure that the entire tip of the cannula, rather than just the needle, is within the arterial lumen. At this stage either the cannula can be advanced over the needle or the guide-wire introduced.

After attaching the catheter to the saline column take great care to ensure there are no air bubbles in the system before flushing it. If you suddenly obtain a very high blood pressure reading, check the position of the transducer; it may have fallen on the floor! If you lose the waveform on the monitor or it decreases in amplitude, the catheter may be kinked or blocked with a blood clot, or there may be an air bubble damping the trace. After checking that your patient has a pulse, you can try making sure the wrist is extended, aspirate any air bubbles and then flush the catheter, or withdraw the catheter slightly to check it is not kinked. Note that over or under-damped traces will give false blood pressure values. An under-damped trace will overestimate systolic pressure and underestimate diastolic pressure as the system over oscillates. A low amplitude, over-damped trace will underestimate the systolic blood pressure and overestimate the diastolic blood pressure. Fortunately, the value for the mean arterial blood pressure is little affected and can usually be taken as accurate.

transducer zeroed - that is, closed to the patient and opened to atmosphere to obtain a reading of atmospheric pressure. It is often convenient to tape the transducer to the patients upper arm to ensure it is level with the heart. Practical tips and trouble shooting Figures 7 and 8 show common techniques for arterial cannula insertion he radial artery is very superficial at the wrist. Often when you think you cant find it, you have in fact transfixed it (a technique some people use preferentially). Remove the needle and then slowly withdraw the cannula, aspirating using a 5ml syringe attached to the hub all the time. As the tip of the cannula re-enters the artery, blood will flow into the syringe briskly. From this point slowly advance the cannula whilst rotating the cannula in a twisting motion about its long axis. This technique will salvage the cannulation more often than not.

PULSE CONTOUR ANALYSIS Useful clinical information can be obtained by looking at the pattern of the arterial waveform on the monitor. A large swing or variation in peak amplitude of the systolic pressure that coincides with the ventilatory cycle often indicates that the patient is hypovolaemic.

Update in Anaesthesia | www.anaesthesiologists.org

page 41

 Conscious patients who are in respiratory distress may also have a large swing on the arterial pressure trace, due to large changes in intrathoracic pressure. A narrow width, high amplitude pulse combined with tachycardia tends to indicate hypovolaemia. The angle of the upstroke of the arterial waveform may give an estimate of myocardial contractility; a steeper upstroke indicates greater change in pressure per unit time and higher myocardial contractility. In practice, this only provides a rough assessment of myocardial contractility.

Analysis of the arterial waveform has been developed mathematically to calculate cardiac output. The term pulse contour analysis is usually used to refer to the cardiac output monitoring systems employed in the PiCCOTM (Pulsation Medical Systems, Germany) and LiDCOTM Plus (LiDCO Ltd, UK) monitors. The PiCCOTM and LiDCOTM systems both measure cardiac output using both the shape and the area under the arterial pulsation curve. For both techniques a haemodilution method is used to calculate the cardiac output and calibrate the pulse contour analyser. Note that this means that both systems require central venous access. By knowing the exact shape and area under the arterial pulsation curve at the time of calibration, future arterial pulsation curves can be compared and the cardiac output at that point in time extrapolated. The way in which these two systems calculate the initial cardiac output differs in that the PiCCOTM uses haemodilution of cold saline and the LiDCOTM uses haemodilution of lithium. The LiDCOTM cannot be

used in patients on lithium therapy or for up to two hours following the administration of non-depolarizing muscle relaxants. Both systems need regular recalibration by re-measuring the cardiac output using haemodilution. All the factors previously mentioned that alter the accuracy of the arterial waveform (air bubbles, kinking etc) will affect the cardiac output value that the system gives. The two systems also alter in terms of the mathematical modelling they use to perform the pulse contour analysis. Recently these systems have been adapted so that they no longer need to be calibrated, but use population data to generate measurements (e.g. the LiDCOrapide TM). Further clinical evaluation of these systems is needed. Further description of these techniques can be found in the article on page 51. SUMMARY Invasive arterial monitoring is a highly useful tool, which allows close blood pressure monitoring for patients undergoing major surgery and the critically ill. It is also useful for repeated arterial blood gas analysis and as an access point for obtaining other blood samples. It is important to understand the principles of biological measurement systems in order to optimise their performance and allow troubleshooting when performance is poor. FURTHER READING
J S Gravenstein and David A Paulus. Clinical Monitoring in Practice (2nd edition). Published by J B Lippincott, Philadelphia, 1987. M K Sykes, M D Vickers, C J Hull. Principles of measurement and monitoring in Anaesthesia (3rd edition). Published by Blackwell Science Publications, 1991.

page 42

Update in Anaesthesia | www.anaesthesiologists.org

Central venous cannulation

Will Key*, Mike Duffy, Graham Hocking *Correspondence Email: [email protected]
INTRODUCTION Central venous cannulation is a relatively common procedure in many branches of medicine, particularly in anaesthesia and intensive care medicine. An estimated 200 000 central venous access procedures are carried out each year in the United Kingdoms National Health Service and over 5 million in the United States. Historically central venous access was gained by a surgical cut-down procedure, but central venous catheters (CVCs) are now predominantly inserted percutaneously, using a technique first described by Seldinger in 1953. There are many different types of catheter and a number of different sites suitable for central venous access. Site selection depends on numerous factors including the indication and duration of access, anatomy of the patient, local resources and operator skill and experience. INDICATIONS The main indications for CVC insertion are: Administration of drugs - Irritant drugs - Long-term treatment (chemotherapy, antibiotics) - Long stay patients (e.g. tetanus) - Parenteral nutrition Haemodynamic monitoring - Central venous pressure - Mixed/central venous oxygen saturations Difficult peripheral access Haemofiltration / haemodialysis Insertion of pacing wires or pulmonary artery catheters. CATHETER SELECTION There is a large range of catheters available and selection should be based on site, reason for insertion and length of use. In anaesthesia and intensive care medicine the main considerations are catheter length and the number of lumens. Three to five lumens are ideal for critically ill patients, allowing multiple drug infusions, but the lumens are usually narrow with a high resistance to flow and so less effective for rapid infusion of fluid during resuscitation. Larger, shorter catheters such as an 8.5F introducer sheath are better suited to this purpose. Types of catheter Single/multiple lumens Peripherally inserted central catheters (PICC) Tunnelled (the catheter travels a few centimetres under the skin before entering the vessel, in order to decrease the incidence of line infections). Specialized - dialysis catheters - continuous central venous saturation monitoring.

Summary
Central venous catheters are extensively used in ICUs in high-income countries, but they remain beyond the facilities available in many developing world ICUs. In developing countries, their main use is as access for delivery of irritant drugs such as catecholamine infusions and for intravenous access in patients requiring prolonged organ support due to illnesses such as tetanus. The different sites of insertion are described and the common techniques for insertion are outlined.

Lumen size Larger catheters allow greater fluid flow (e.g. for resuscitation and haemofiltration) but have a greater risk of significant haemorrhage or air embolism during insertion or inadvertent disconnection. They also have a significant dead space to consider during administration of potent drugs such as vasopressors - the narrow gauge lumens of multiple channel lines Will Key are better for this purpose. Larger catheters are more Specialist Registrar in likely to cause thrombosis or late stenosis of the vessel. Anaesthesia Royal Devon and Exeter Impregnated catheters NHS Foundation Trust A number of manufacturers make catheters impregnated UK with antimicrobial agents, such as chlorhexidine and silver sulfadiazine, in an attempt to reduce catheter Mike Duffy related infections. Many ICUs use these lines routinely Consultant in Intensive despite the higher cost, the potential for development Care of drug resistance and the inconclusive evidence for Derriford Hospital reduced morbidity and mortality. Plymouth PRINCIPLES OF INSERTION In well resourced settings, ultrasound guided insertion has become standard practice. As this is unavailable in most low-income countries, landmark techniques are emphasized here. The basic preparation and equipment required for CVC insertion is the same regardless of site or technique. A UK

Graham Hocking Consultant Anaesthetist Sir Charles Gairdner Hospital Perth Western Australia

Update in Anaesthesia | www.anaesthesiologists.org

page 43

Monitoring in the ICU

Anaesthesia

Update in

suitable clinical area should be chosen where full aseptic technique can be observed. A trained assistant is useful and the patient should be monitored with continuous ECG, oxygen saturations and blood pressure measurement. The suggested essential equipment is listed in Table 1. There is evidence that the use of a dedicated lines trolley increases compliance with best practice. Confirm that the CVC is still needed and select the most appropriate route (see below). Explain the procedure to the patient. Strict asepsis at the time of insertion is a major factor in reducing line related infections - wear sterile gloves, a gown, mask and theatre hat. Drape the surrounding areas of the patient and bed as thoroughly as possible. Good positioning and identification of anatomical landmarks will minimise the risk of failure and complications. In conscious patients local anaesthetic should be used.
Table 1. Suggested essential equipment for CVC insertion.

It is well worth having all of your equipment laid out in a logical order before you proceed. Make sure that you are familiar with the set provided. Patient on a tilting bed, trolley or operating table Hat, mask and sterile gown and gloves Large sterile drapes and gauze swabs Antiseptic solution (chlorhexidine in alcohol) Local anaesthetic agent with needle and syringe Saline flush Appropriate central venous catheter set Three-way taps Scalpel blade Sutures
Figure 1. The Seldinger technique.

precipitate arrhythmias and intravascular loss of the guidewire has been reported. You can usually get some idea of the length required prior to insertion by laying the catheter on the patients chest prior to insertion. Map out its course and note the distance between the tip at the medial right second intercostal space and the site of skin puncture. Remove the guidewire and aspirate and flush all the lumens with saline to check for free flow. Finally secure the catheter in place with sutures and a sterile non-occlusive dressing. SITE SELECTION (Figure 2) There are a number of approaches to the central venous system and these veins may be deep structures, running close to arteries, nerves and other structures (e.g. the pleura in the case of the subclavian vein). You must know the deep and surface anatomy of the area to undertake a landmark technique safely. 2-D ultrasound is increasingly used, where resources allow, and is well suited to the internal jugular, femoral and peripheral approaches. Ultrasound allows visualisation of the vessel, confirmation of placement of the wire within it and identification of anatomical variation. The main entry sites are: Internal jugular vein Subclavian vein Femoral vein External jugular vein Veins of the arm or antecubital fossa (basilic or cephalic veins).

General technique (Seldinger) The most common method of insertion is the catheter-over-guidewire (Seldinger) technique (see Figure 1). The vein is punctured with a small gauge needle (18 or 20G) attached to an empty syringe and blood is aspirated easily. If the blood appears bright red, is at high pressure or pulsatile consider the possibility of an arterial puncture. The guidewire commonly has a J-shaped tip to reduce risk of damage to the vessel wall and help negotiate tortuous vessels. It should advance and withdraw easily at all times. After removing the needle, a dilator is then passed over the guidewire and a small incision made in the skin to allow the dilator to advance through the skin, subcutaneous tissues and a short distance into the vein (further passage along the vein may cause damage to the vessel or distal structures). Gentle skin traction and a twisting motion aids passage of the dilator and prevents kinking of the guidewire. Remove the dilator and insert the catheter over the guidewire. The guidewire is held whilst the catheter is advanced to the desired length. Care should be taken not to advance the guidewire with the catheter, as this may

page 44

Update in Anaesthesia | www.anaesthesiologists.org

 Ease of learning Puncture of visible/ palpable vein versus blind, landmark technique. Applicability to patients of different ages. Equipment available Availability of suitable apparatus (including ultrasound) Cost. SPECIFIC CVC SITES
Figure 2. Common sites for intrathoracic CVC placement.

Internal jugular vein The internal jugular vein (IJV) is most frequently chosen site for CVC insertion. It is a potentially large vein with a lower risk of pneumothorax compared with the subclavian approach. Inadvertent arterial puncture can be controlled easily with manual compression. Many approaches have been described depending on the level of the neck at which the vein is punctured. A high approach reduces the risk of pneumothorax but increases the risk of arterial puncture. For lower approaches the converse is true. With experience this route has a low incidence of complications. Anatomy (Figure 3) The IJV arises from the jugular foramen at the base of the skull and is a continuation of the sigmoid sinus (within the skull). It descends in the neck in the carotid sheath, with the carotid artery and the vagus nerve. It lies initially posterior to internal carotid artery before becoming lateral then anterolateral to the artery. Behind the medial end of the clavicle it joins the subclavian vein to form the brachiocephalic vein. The vein has dilatations at both ends, the superior and inferior jugular venous bulbs. Cannulation can be difficult in the morbidly obese, as landmarks are often obscured and those patients with very short necks or limited range of movement can also be a challenge. The IJV is unilaterally absent in 2.5% of patients and is outside the predicted path in 5.5% of patients. The right IJV offers some advantages in that it tends to be larger and straighter than that on the left, it is more convenient for the right-handed practitioner and avoids the possibility of thoracic duct injury. Positioning The patient is supine, arms by their sides, with a head down tilt to distend the veins and reduce the risk of air embolism. The head should be slightly turned away from the side of cannulation for better access (excessive turning should be avoided as it changes the relationship of the vein and artery and can collapse the vein). The patients neck can be extended by removing the pillow and putting a small towel under the shoulders.

Factors determining choice Duration of use Consider a tunnelled line for prolonged administration of antibiotics or where intravenous access has become difficult. Femoral lines are only appropriate for use for up to 48 hours due to the higher infection risk. Suitability of site for planned CVC use e.g. for CVP measurement, catheter tip must be in thorax. Operator Knowledge and practical experience of the technique it is better to have a few clinicians in each area who perform all the central venous cannulations and gain experience (a central venous access team). Technique characteristics Success rate for cannulation and central placement Complication rate

Figure 3. The anatomy of the right internal jugular and subclavian veins

Update in Anaesthesia | www.anaesthesiologists.org

page 45

Technique Stand at the head of the patient and palpate the mastoid process and the sternal notch. The entry level is half way along a line joining these two landmarks. Palpate the carotid artery at this level and check your entry point is lateral to this. It is sometimes possible to ballot the vein which can aid accurate needle placement. Keeping your finger gently over this point (even small amounts of pressure can collapse the vein) insert the needle at 30-40 to the skin directed caudally towards the nipple on the same side (in females guess where it would be if it were a male) aspirating as you go. The vein is usually very superficial and only 0.5-2cm under the skin. Practical problems If the vein is not found recheck your landmarks, ensure the patient is adequately head down and consider rehydration if the patient is hypovolaemic. After a failed attempt to locate the vein, continue to aspirate as you slowly withdraw the needle; the vein may have collapsed on the way in and be transfixed as the needle has gone through the posterior wall. Resist the urge to advance the needle deeper into the tissues, as you are most likely to be in the wrong place rather than too superficial. If ultrasound is available use it to check that the anatomy is normal. Subclavian vein The subclavian vein (SCV) has a calibre of 1-2cm in adults and is thought to be held open by its surrounding tissues, even in severe circulatory collapse. It is often preferred for long-term central access as it is generally more comfortable for patients, can be easily tunnelled and has a lower risk of infection and other long-term complications. This route may also be preferred in trauma patients with suspected cervical spine injury. This route is best avoided in patients requiring long-term renal replacement, as there is a significant risk of venous stenosis, causing problems for existing or future arteriovenous fistulae. The subclavian route is best avoided in patients with abnormal clotting or bleeding diatheses, as the vessels are inaccessible to direct pressure after inadvertent arterial puncture. Serious immediate complications are uncommon but occur more frequently than other routes. Pneumothorax is one of the most common major complications with an overall incidence of 1-2%. This figure increases to 10% if multiple attempts are made. Although possible in some patients, visualisation of the subclavian vein with ultrasound is difficult in most. Anatomy The SCV is a continuation of the axillary vein as it reaches the lateral border of the first rib (Figure 3). It ends at scalenus anterior where it joins the internal jugular vein, to form the brachiocephalic (inominate) vein, behind the medial end of the clavicle. Its only tributary is the external jugular vein and it lies anterior and parallel to the subclavian artery throughout its course. The cervical pleura lies behind the artery. Initially the vein arches upwards and across the first rib and then inclines medially, downwards and slightly anteriorly across the insertion of scalenus anterior.

Right

Left

Right clavicle

a
Left

Right Right clavicle

Posterior border of right sternocleidomastoid

Figure 4. Right IJV cannulation, (a) view from caudal aspect; (b) view from cranial aspect.

page 46

Update in Anaesthesia | www.anaesthesiologists.org

Try an alternative route on the same side unless Xray is available to confirm there is no pneumothorax. External jugular vein As the external jugular vein (EJV) lies superficially in the neck it is often visible or palpable, which negates many of the complications of the deep vein approaches. It is a useful when expertise is lacking, for emergency fluid administration and in cardiac arrests where no carotid pulse is palpable. A long catheter will not reliably thread into the SCV (due to the presence of valves and other anatomical abnormalities), so it is usual to use a short peripheral cannula.

Right

Sternocleidomastoid

Anatomy The EJV drains blood from the superficial facial structures and scalp and passes down in the neck from the angle of the mandible, crosses the sternocleidomastoid muscle obliquely and terminates behind the middle of the clavicle where it joins the SCV. The vein is variable in size and contains valves which may prevent the passage of the guidewire and catheter. There is a wide range in EJV size and prominence due to natural variation and disease states. Positioning As for IJV. Technique Standing at the head of the patient identify the EJV as it crosses the sternocleidomastoid. Insert the needle into the vein where it is most easily seen or palpated.

Right clavicle Jugular notch Left

Figure 5. Right SCV cannulation (mannequin).

Positioning The patient should be positioned as for the internal jugular approach with the head down to fill the veins and reduce the risk of air embolism. Technique The right SVC is usually preferred as this approach avoids damage to the thoracic duct. However in the presence of unilateral lung pathology, cannulation should be performed on the same side so that a pneumothorax will not affect the healthy lung. The infraclavicular approach is most commonly used, where the needle is inserted into the skin slightly below the lower border of the clavicle, at the junction of the middle and medial thirds of the clavicle. The needle is kept in the horizontal plane advancing medially, posterior to the clavicle aiming for the sternal notch. The needle should not pass beyond the sternal head of the clavicle. Practical problems If you are unable to get beneath the clavicle consider starting more laterally and bending your needle upward slightly. Some axial traction on the arms by your assistant and a pillow or rolled up towel between the shoulder blades may also improve success. If you still cannot find the vein, direct the needle a little more cephalad - place your finger fully into the sternal notch and aim for the middle of it. Do not persist after repeated attempts as the complication rate increases dramatically.

Figure 6. Cannulation of the external jugular vein.

Practical problems If the vein is not visible or palpable, press on the skin above the middle of the clavicle and reduce drainage into the SCV, thereby distending the vein (Figure 6). Alternatively ask the patient to do a valsalva manoeuvre, tilt the patient more head down or hold in inspiration if ventilated. If there is difficulty threading the guidewire or catheter, try twisting whilst advancing or flushing saline through the catheter as you insert it. Slowly moving the head from one side to the other may also help. Caution should be used when manipulating the wire

Update in Anaesthesia | www.anaesthesiologists.org

page 47

with the needle attached as there is a risk of the needle shearing of the end of the wire (a plastic cannula is safer) Femoral vein The femoral vein (FV) may be cannulated with low risk of serious short-term complications and, for this reason, is preferred by less experienced operators. This route is also useful in urgent situations when the patient is coagulopathic and is perhaps the safest central vein in children requiring resuscitation, where central access is needed for vasopressor therapy. The large diameter of the FV allows large fluid volumes to be removed and infused and is commonly used in the ICU for placement of short-term haemofiltration catheters. Femoral catheters are better suited to ventilated, sedated patients as excessive movement can cause kinking of the catheter and mechanical complications. The CVP measurement from a femoral catheter can be affected by intra-abdominal pressure, although in ventilated patients values correlate well with those from intra-thoracic catheters. Arterial puncture or femoral nerve damage are both possible if insertion is too lateral. The risk of infection in the medium and long-term is higher with femoral catheters compared with most other routes because of the greater degree of bacterial colonisation found in the groin compared to other sites. There is also an increased risk of thromboembolic complications compared with internal jugular and subclavian approaches. For these reasons femoral catheters should be removed within 48-72 hours of insertion. Anatomy The FV starts at the saphenous opening in the thigh and runs alongside the femoral artery to the inguinal ligament where it becomes the

external iliac vein. In the femoral triangle the FV lies medial to the artery in the femoral sheath. Positioning The patient should be supine with a pillow under the buttocks to elevate the groin. The thigh should be abducted and externally rotated. Technique Palpate the femoral artery 2cm below the inguinal ligament and insert the needle 1cm medial to the pulsation and aim cephalad and slightly medially at an angle of 20-30 to the skin. In adults the vein is usually 2-4cm below the skin. In children the FV is more superficial so the angle should be 10-15. Cannulation can be difficult because of the lack of landmarks especially in obese patients. Practical problems It can be difficult to feel the arterial pulsation especially in obese patients. Get an assistant to retract the abdomen if this is a problem and recheck the landmarks. As with the internal jugular approach 2-D ultrasound, if available can be very useful to assess anatomy and guide the needle. As with other routes, ensure no digital pressure is collapsing the vein. The antecubital veins The superficial, palpable veins of the antecubital fossa provide a very safe route for central access. Risk of infection is lower than other routes and lines can be used for longer periods (e.g. TPN, prolonged antibiotic courses or chemotherapy). A long catheter is required (around 60cm) to thread the tip into the central veins and for this reason flow rates are low, with large dead space making them less useful for resuscitation and inotropes. Tip position is important as migration can occur with movement of the arm (up to 7cm in cadaveric studies but femoral artery around 2cm in vivo).
femoral vein pubic tubercle greater saphenous vein

anterior superior iliac spine inguinal ligament

femoral nerve

sartorius muscle

adductor longus muscle

insertion

Anatomy Two main veins are available but the more medial basilic vein has a smoother, more direct route to the SCV. The more lateral cephalic vein turns sharply to pass through the clavipectoral fascia and also has valves at its termination. These factors frequently cause difficulty in advancing the catheter. The basilic vein ascends along the medial side of the forearm before moving anterior to the medial epicondyle, where it is joined by the median cubital

Figure 7. Anatomy of the right femoral vein.

page 48

Update in Anaesthesia | www.anaesthesiologists.org

comfort as the line does not cross the joint and will be less prone to kinking and other mechanical complications. If difficulty in threading the catheter is encountered first check the tourniquet has been released and check you are definitely in the vein. Flushing with saline as the catheter is advanced may facilitate passage through valves. Further abduction of the arm may also help. ChECKs BEfOrE UsINg ThE CAThETEr It is important to ensure that the catheter is within the vein prior to use. This is best done by transducing the pressure waveform or comparing synchronous arterial and venous blood gases. Dark blood at low pressure is not always a reliable sign especially in a hypoxaemic, poorly perfused patient. The position of catheters that enter the chest (i.e. jugular or subclavian approach) should be confirmed on chest Xray (Figure 9). The tip of the catheter should lie in the SVC, just above its junction with the right atrium. On chest Xray it should be above or overlying the right main bronchus. Check there is no pneumothorax.

Figure 8. Anatomy of the peripheral veins of the right arm.

vein. It then runs along the medial edge of the biceps muscle to the middle of the upper arm, where it pierces the deep fascia and runs alongside the brachial artery, becoming the axillary vein. The cephalic vein ascends on the front of the lateral side of the forearm to the front of the antecubital fossa, where it communicates with the basilic vein via the median cubital vein. It ascends along the lateral edge of the biceps muscle until it reaches pectoralis major, where it pierces the clavipectoral fascia to pass beneath the clavicle, where it usually terminates in the axillary vein (occasionally it may join the EJV). Positioning Apply a torniquet to the upper arm and select the best vein. The medial side of the arm is best for the reasons mentioned above. Lie the patient supine with the arm abducted at 45 to the patient and the head turned towards the ipsilateral arm (this may help prevent the catheter passing into the IJV). Technique Estimate the length of catheter required to reach the SCV. Insert the cannula supplied in the set and remove the needle. Thread the catheter through the cannula and advance it 2-4 cm before releasing the torniquet. Continue to advance the catheter until the desired length is inserted. The cannula is often designed to tear apart to remove it from the catheter. Other sets contain a guidewire and dilator for a Seldinger technique which is useful for smaller vessels. Practical problems In critically ill patients numerous attempts at venepuncture and cannulation have usually occurred, leaving vessels thrombosed and unusable. Looking more proximally may reveal untouched veins, especially on the inner aspect of the upper arm. 2-D ultrasound can be very useful for locating and checking patency of veins as well as guiding the needle. A more proximal approach can improve patient
Figure 9. Chest Xray showing optimal CVC tip position (arrow).

COMPLICATIONS Complications occur in up to 10% of CVCs and can be divided into mechanical, infectious and thromboembolic aetiologies, the most common of which are listed below. The complication rate is dependant on a number of factors including site, patient factors (concurrent illness and variations in anatomy) and operator skill and experience. Interventions recommended to prevent complications are also listed below. There are no absolute contraindications to central venous cannulation, as it can be a lifesaving procedure, but serious complications, including death, may occur during insertion or ongoing use of a CVC. Operator training and experience are important factors in reducing complication rates and experienced help should be sought after repeated attempts. The frequency of mechanical complications is six times greater than after a single attempt. Caution should be used to try and avoid complications in high risk patients and may influence site of access. There is a higher risk of pneumothorax with the subclavian approach and as the vessels are not amenable to direct compression. This site is least appropriate in patients with severe respiratory disease or bleeding diatheses. Penetrating abdominal trauma or known inferior vena caval disruption would make the femoral approach less desirable.

Update in Anaesthesia | www.anaesthesiologists.org

page 49

Mechanical complications Arterial puncture Haematoma Pneumothorax Haemothorax Haemorrhage Arrhythmias during procedure Cardiac tamponade Respiratory obstruction Thoracic duct injury Brachial plexus injury Infectious complications Local infection Bacteraemia, sepsis Thromboembolic complications Thrombosis of vessel Thrombus formation Venous air embolism Catheter/guidewire embolism Interventions to prevent complications Use antimicrobialimpregnated catheters Insert in the subclavian vein Strict asepsis at insertion Avoid antibiotic ointment Remove catheter when promptly when no longer required Recognize risk factors for difficult catheterization and seek experienced assistance Avoid femoral route Use ultrasound during internal jugular insertion ULTRASOUND GUIDANCE In 2002 the National Institute for Clinical Excellence (NICE) in the United Kingdom recommended the use of ultrasound for the elective placement of CVCs into the IJV. Since this time the use of ultrasound use has increased dramatically. More clinicians are becoming experienced in its use and there is now increasing evidence showing a reduction in number of passes, failure rates, arterial puncture and time to placement and infectious complications using this technique. Ultrasound guidance is particularly suited to the IJV, FV as well as peripheral veins. It is not possible to visualise the subclavian vein easily with ultrasound, due to the shadow cast by the clavicle, however the axillary vein can be visualised more laterally on the chest. The ultrasound image provides information about the patency and location of the vessel and can be used to guide the needle in real time (Figure 10).
Figure 10. Ultrasound images of internal jugular vein. Application of pressure using the probe identifies the vein as the compressible structure.

SUMMARY Central venous access is a commonly performed procedure that can be lifesaving, but is associated with significant complication rates. Operator experience, familiarity with the range of sites available, along with sound knowledge of anatomy and use of ultrasound can help to minimise some of the mechanical complications. Strict asepsis at the time of insertion, use of impregnated catheters, proper maintenance and timely removal can minimise infective complications.

FURTHER READING

The National Institute for Clinical Excellence. Guidance on the use of ultrasound locating devices for placing central venous catheters (NICE technology appraisal No. 49). London: NICE, 2002. Duffy M, Sair M. Cannulation of central veins. Anaesthesia & Intensive Care Medicine 2007; 8: 17-20.

Taylor RW, Palagiri AV. Central Venous Catheterization. Crit Care Med 2007; 35: 1390-96. McGee D C, Gould M K. Preventing Complications of Central Venous Catheterization. New Engl J Med 2003; 348: 1123-33. Craven J. Large Veins of the Neck. Anaesthesia & Intensive Care Medicine 2004; 5: 4-5. Pronovost P, Needham D, Berenholtz S et al. An Intervention to Decrease Catheter-Related Bloodstream Infections in the ICU. New Engl J Med 2006; 355: 2725-32. Schuster M, Nave H, Piepenbrock S, Pabst R, Panning B. The carina as a landmark in central venous catheter placement. Br J Anaesth 2000; 85: 192-4.

page 50

Update in Anaesthesia | www.anaesthesiologists.org

Cardiac output monitoring

Thomas Lawson and Andrew Hutton Correspondence Email: [email protected]
DEFINITIONS Before describing the different types of cardiac monitoring, it is essential to understand some basic definitions: The cardiac output is determined by, and therefore can be manipulated by, alterations to the heart rate or rhythm, the preload, the contractility and the afterload. Cardiac output informs us of global blood flow and therefore oxygen delivery (the product of cardiac output and blood oxygen content), but does not describe delivery of oxygen to each organ, whose function must be assessed individually. An appreciation of how information is measured and derived using a cardiac output monitor is essential in order to use the information accurately and appropriately. Thorough assessment of a patients clinical status strongly influences interpretation of the measurements made. For example, cardiogenic shock and obstructive shock due to tamponade will both give a low cardiac output, but can be differentiated by the patients clinical signs. There is a wide variability between practitioners in the measurement and interpretation of cardiac output data, using the various techniques that are currently available. For this reason, clear evidence that they benefit patient outcome is difficult to obtain.

Summary
Estimation of cardiac output has an important role in patient management during anaesthesia and critical care. Cardiac output can be measured in a number of ways, from simple clinical assessment to invasive haemodynamic monitoring. Advanced monitoring techniques are often used when clinical signs are difficult to interpret. While there has been much research looking at the accuracy of these monitors there is currently very little information in the literature regarding the effectiveness of these monitors in improving clinical outcomes.

Cardiac output is the volume of blood ejected from each of the ventricles of the heart per minute, and is therefore the product of stroke volume and heart rate. The unit of cardiac output is L.min-1. Cardiac index is the cardiac output of a patient referenced to their body surface area and has units of L.min-1.m-2. Stroke volume is the volume of blood ejected by each contraction of the ventricle and is determined by the preload, afterload and contractility. The stroke volume is usually 60-80ml for an average sized adult. Preload describes the tension developed in the ventricular wall at end-diastole (i.e. at maximal filling just prior to contraction). This tension is difficult to measure and end-diastolic pressure is taken as a surrogate (or estimate) measurement. It is mainly determined by venous return and gives an indication of the filling pressure of the ventricle. Contractility refers to the amount of work the heart can generate, at given levels of preload and afterload, and is estimated by the maximum rate at which the ventricle can generate a change of pressure over time. Inotropy is used to explain an increase in the work done by the heart that is independent of heart rate, preload and afterload. Afterload is the tension that needs to be generated in the ventricular wall in order to eject blood into the arterial system during systole. This is largely determined by the resistance of the arterial system the systemic vascular resistance (SVR). It is calculated by: SVR = Mean arterial pressure (mmHg) Central venous pressure (mmHg) x 80 Cardiac output (L.min-1) (Recall that Ohms Law describing electrical resistance is analogous to this: V = IR) Mean arterial pressure (MAP) is the average arterial blood pressure throughout the cardiac cycle. As 2/3 of the cardiac cycle is spent in diastole, and 1/3 in systole, MAP may be calculated using the formula: MAP = Diastolic BP + 1/3(Systolic BP - Diastolic BP) Ejection fraction is the fraction of total blood in a ventricle that is ejected per beat. It applies to both the left and right ventricles. It gives an index of contractility. Normal value is in the region of 55-65%.

Thomas Lawson Registrar Andrew Hutton Consultant Department of Anaesthesia Derriford Hospital Plymouth UK

Update in Anaesthesia | www.anaesthesiologists.org

page 51

Monitoring in the ICU

Anaesthesia

Update in

CLINICAL INDICATORS OF CARDIAC OUTPUT The interpretation of data from invasive haemodynamic monitoring is made in light of the clinical examination. No single clinical sign can be used to make an accurate assessment of cardiac output. Heart rate, blood pressure, pulse strength at various sites, patient colour, respiratory rate and core to peripheral temperature gradient all give an indication to a patients haemodynamic status. Note that although blood pressure is often used as an indicator of cardiac output, it is frequently unhelpful. Blood pressure may be maintained by intense peripheral vasoconstriction in the face of a perilously low cardiac output. A patients ability to compensate for a haemodynamic insult is highly variable, depending on age, premorbid status and other comorbidities. An example is the rise in the diastolic pressure in early hypovolaemic shock, associated with peripheral vasoconstriction that is usually only seen in young, fit individuals. In addition, clinical parameters such urine output, capillary refill time and cognitive function give a guide to end organ perfusion. Change in heart rate, blood pressure and central venous pressure in response to a straight leg raise is useful to predict a patients response to a fluid bolus. Measurement of lactate and base deficit in arterial blood and, in particular, the trend of these variables over time gives non-specific information about a patients organ perfusion. Lactate is produced by anaerobic metabolism, and is an indicator of tissue hypoperfusion. It is measured on most modern blood gas machines. It can be used to monitor therapy, as it will fall as oxygen delivery improves, and as liver perfusion (which enables lactate metabolism) increases. The oxygen saturation in central venous blood (ScvO2) also gives a global indication of haemodynamic status, is useful in directing fluid therapy3 and is a reliable surrogate of mixed venous oxygen saturation (see under pulmonary artery flotation catheters, below). Learning point blood pressure is a poor indicator of cardiac output. OVERVIEW OF THE ROLE OF CARDIAC OUTPUT MONITORING Mortality in sepsis increases by 15-20% for each organ failure that a patient develops.1 Organ failure results when delivery of oxygen is inadequate for the organs requirements. Since the 1980s, research has suggested that optimisation of oxygen delivery (a product of cardiac output and blood oxygen content) in high risk surgical patients prevents organ failure and improves mortality.2 This has been investigated early in critical illness, and prior to, during and after surgery.2 Although no single study provides categorical evidence, the weight of evidence suggests that therapies directed at enhancement of oxygen delivery (goal-directed therapy) should be our aim. There is also increasing evidence that, while hypovolaemic septic patients need fluid to optimise their cardiovascular delivery of oxygen, excessive liberal fluid therapy may be harmful. The major factor limiting this field of clinical practice has been development of a monitoring device that will reliably and accurately guide our use of fluid therapy - to recognise where fluid is needed and give enough, but not too much. Measurement of filling is difficult. We aim to apply Starlings Law, where cardiac performance

improves with stretching of the ventricular muscle fibres, to a certain optimal point beyond which further stretching impairs performance (see Figure 1). To apply this strategy we would like to know the left ventricular end-diastolic volume (LVEDV) and monitor changes in the LVEDV as we give fluid boluses. The best surrogate estimate of LVEDV we have is to use a pulmonary artery catheter (PAC) to measure pulmonary artery occlusion (wedge) pressure, which gives us an estimate of left atrial pressure, which is in turn and estimate of LVEDP, which is a surrogate of LVEDV (and makes assumptions about normal compliance of the left ventricle). This is not a reliable measure of filling, particularly given the effects of ventilation, applied PEEP and the anatomical location of the catheter tip in different lobar pulmonary artery branches. Thermodilution using the PAC does provide an accurate measurement of cardiac output, which can be measured continuously given the correct equipment, however use is diminishing in many parts of the world, due to concerns over safety and lack of robust evidence to support their use.

Figure 1. Simplified explanation of Starlings Law.

Currently, the main focus of research and development is towards less invasive monitors with inherently lower risks of use. Broadly these are monitors that use Doppler analysis of the aortic blood velocity (viewed from the oesophagus) or monitors that analyse the shape of the arterial waveform (pulse contour analysis). Some of the cardiac output monitors that rely on arterial waveform analysis, use thermo- or indicator dilution to obtain an accurate estimate of cardiac output, which can then be used to calibrate continuous analysis of the waveform, transduced from a modified arterial catheter. In order to make these easier to set up and use, more recent models calibrate their pulse contour analysis using population data, based on age, weight and height. The disadvantage is that the population data is derived from healthy volunteers, and so is not validated for patients with abnormal vascular resistance, which undoubtedly has a major effect on derived indices such as stroke volume. The oesophageal Doppler also uses population data to estimate aortic diameter. However, even if we are sceptical about the absolute numbers generated, these monitors can be reliably used to observe trends in stroke volume, and the effect of interventions such as fluid administration. The key feature is to determine whether the patient is fluid responsive; meaning

page 52

Update in Anaesthesia | www.anaesthesiologists.org

that a bolus of fluid augments their cardiovascular performance (for example their stroke volume), thereby improving oxygen delivery. Fluid responsiveness implies that we have moved the patient up the Starling curve. A current and future area of development is the use of stroke volume variation (SVV) or pulse pressure variation (PPV) that is measured from the transduced arterial waveform. We have long observed that hypovolaemia causes an exaggerated swing in systolic pressure during the respiratory cycle; SVV and PPV quantify this swing or variation as a single number. Again, it is a change in the number, rather than the absolute value that is useful in assessing the fluid responsivenesss of your patient. From a pragmatic perspective, these monitors are most useful when observing the effect of a single intervention (such as fluid administration) in isolation - this is often difficult during the changing stimuli of surgery, or when the physiological response to sepsis is changing rapidly. Measurements are most plausible when interventions and pre- and post- stroke volume, SVV or PPV measurements are performed during a lull in other stimulating activity. DOPPLER ULTRASOUND AND ECHOCARDIOGRAPHY Ultrasound Ultrasound is any high-frequency sound wave. Ultrasound is used medically to create a 2 dimensional image by using a probe to transmit high-frequency sound waves (1-5MHz) into the body, and to detect the waves as they are reflected off the boundaries between tissues interfaces. By using a mathematical model involving the speed of sound and the intensity and timing of each echos reflection, the distance from the probe to the tissue boundaries is calculated, and used to create a two- dimensional image. Doppler ultrasound When sound waves are reflected from a moving object, their frequency is altered. This is the Doppler effect. By using an ultrasound probe to visualise directional blood flow, the phase shift (i.e. the change in frequency before and after reflection off moving red blood cells) can be determined. This, together with the cross-sectional area of the blood vessel being observed (measured or estimated) can be used to determine flow, where: Flow = area x velocity Oesophageal Doppler Theory of technique A Doppler probe is inserted into the distal oesophagus (Figure 2) and is directed to measure the blood flow in the descending aorta at about 35 to 40cm from the incisors. The monitor calculates cardiac output using descending aorta diameter, which is either obtained from an age-related nomogram or measured directly in newer machines. The ventricular ejection time, corrected for heart rate (the corrected flow time, FTc), gives an indication of preload and the peak flow velocity (PV) estimates the contractility of the ventricle. Newer probes incorporating M-mode Doppler measurement may improve accuracy and reliability.

Practical application The technique is straight-forward, easily learned and relatively noninvasive. The disposable probes are easy to insert, however some expertise must be gained in recognition of intracardiac and pulmonary artery signals. Continuous measurement is possible, although frequent positional adjustments are needed. Some user variability is inevitable. The cardiac output data is best used as a trend to guide the effectiveness of interventions such as fluid challenges. Wave form interpretation A full description of the use of oesophageal Doppler is beyond the scope of this article but guidance can be obtained from the NHS Technology Adoption Centre at http://www.ntac.nhs.uk/searchresent. aspx?search=cardioQ Advantages Minimally invasive Minimal interference form bone, lung and soft tissue Quickly inserted and analysed Little training required The system is small and relatively portable Paediatric probes are available. Disadvantages May require sedation User dependent Interference from surgical instruments (e.g. diathermy) Depends on accurate probe positioning Probe may detect other vessels e.g. intracardiac/intrapulmonary Assumes a constant percentage of cardiac output (approx 70%) enters the descending aorta. May therefore be inaccurate in a hypovolaemic patient where flow may be redirected to the cerebral circulation. Contraindicated in the presence of oesophageal varices. Transthoracic echocardiography Echocardiography is cardiac ultrasound and can be used to estimate cardiac output by direct visualisation of the contracting heart in real time. Echocardiography is becoming widely accepted as one of the safest and most reliable cardiac output monitors in the critically ill. A focused echocardiogram can be performed in a matter of minutes and assist in determining the cause of haemodynamic instability. Using transthoracic echocardiography four views are obtained (parasternal long axis, parasternal short axis, apical, and subcostal), and it is possible to make an assessment of ventricular function and size of cardiac chambers with these.

Update in Anaesthesia | www.anaesthesiologists.org

page 53

Figure 2. Image of descending aortic waveform obtained using an oesophageal Doppler probe, CardioQ (Courtesy of Deltex Medical).

Table 1. Summary of variables obtained from oesophageal Doppler Variable Description Interpretation Height Peak velocity The highest detectable aortic flow can be used as a measure of afterload, vascular resistance and contractility Slope of upstroke Mean acceleration Measure of contractility Width of base Flow time Left ventricular ejection time, i.e. duration of aortic blood flow. When corrected for heart rate gives an index of preload (e.g. if base is narrow suggests hypovolaemia) Area under waveform Stroke distance Distance a column of blood travels along the aorta during each curve ventricular systole Stroke distance Afterload Stroke volume SVR Aortic cross-sectional area Shown by a reduction in waveform height and base Practical application A multiplane transducer is inserted into the oesophagus and stomach, where various standardized views are gained. Advantages A large amount of haemodynamic information is available beyond just cardiac output.

Transoesophageal echocardiography Theory of technique A specialized probe is inserted into the oesophagus, providing real-time, high resolution ultrasound images. Both qualitative and quantitative values for cardiac output are available, using a two dimensional crosssectional area measurement, a Doppler flow measurement at that point and the heart rate.

page 54

Update in Anaesthesia | www.anaesthesiologists.org

Disadvantages The probes are still expensive and the machinery is large and bulky. Various levels of examination skill are required and these take time and resources to learn. A full study can take over twenty minutes. Some form of local pharyngeal anaesthesia or sedation is required to tolerate the probe. There is a risk of trauma from the probe, although the risks are low in patients with no oesophageal disease. The probes generate a degree of heat and are therefore not suited to continuous measurement. As the technology advances and costs decrease, TOE may find more applications in theatre and the ICU. DILUTION METHODS These techniques require: a marker substance that can completely mix with blood, remain within the circulatory system, and is minimally metabolised. a central vein (into which the marker substance is injected) and a peripheral artery (from which the arterial content of the substance can be measured) must be cannulated. As long as blood flow between the injection and measuring sites is constant, flow (i.e. cardiac output) can be calculated from the area under a concentration versus time graph, using a modified StewartHamilton equation. Advantages of dilution methods Less invasive than PAFC (see below). Disadvantages of dilution methods Can only be used to calculate cardiac output in ventilated patients in sinus rhythm. Specific heart-lung interaction is required for the calculation of stroke volume variation (SVV) and pulse pressure variation (PPV) Invasive and associated morbidity / mortality. User dependent. Can underestimate cardiac output in low output states.

Lithium Dilution Monitoring Lidco and PulseCO and Lidcoplus Theory of technique This technique combines the techniques of lithium dilution (Lidco and Lidcoplus) and pulse contour analysis (PulseCO). A small dose of lithium is injected into a peripheral vein and an ion selective electrode is attached to a peripheral arterial line. The area under the curve of a plot of lithium concentration against time allows calculation of the cardiac output. This information is then used to calibrate the PulseCO which provides beat-to-beat cardiac output measurement, using pulse contour analysis of the arterial waveform. Practical application The convenience of this system is that it uses catheters which are likely to be in place or are likely to be needed in a critically ill patient. The system requires some familiarity to set up, but is relatively quick. The total dose of lithium is small and is clinically insignificant. Calibration is recommended every 8 hours, or after any significant change in the patients clinical condition. Advantages A figure for stroke volume variation is produced and provides an indicator of volume responsiveness to fluid therapy. Disadvantages The system cannot be used for patients taking lithium and those who have recently received vecuronium or atacurium. The monitor performs poorly in the presence of atrial fibrillation and other tachyarrhythmias. The system is prone to technical difficulties related to damping and resonance within the measurement system (see page 38). Thermodilution pulse contour monitoring PiCCOplus Theory of technique This technique utilises thermodilution in combination with Pulse Contour Analysis (PulseCO) to measure cardiac output, and correlates well with the PAFC (below). Stroke volume variation (the mean difference between the highest and lowest arterial pressure wave peaks over 30 seconds) gives an indication of the blood volume status of the patient. The system is calibrated using intermittent cold transpulmonary thermodilution, where cold fluid is injected through a central venous catheter and traverses the pulmonary circulation. A curve of blood thermodilution is measured in a systemic artery and, in addition to cardiac output, other data is derived. The calculated extra-vascular lung water (EVLW) gives an indication of the water content of the lungs and is increased in left ventricular failure, pneumonia and sepsis. The normal range is 3-10ml.kg -1 and values greater than 14ml.kg-1 are associated with an increased mortality. The intra-thoracic blood volume index gives an indication of blood volume status (normal value 850-1000ml.m2). PiCCOplus replaced the original PiCCO machines in 2002 and has subsequently been replaced by PICCO2 with improved displays, automated features and the use of room temperature injectate for calibration.

Figure 3. Concentration-time graph for determining cardiac output.

Update in Anaesthesia | www.anaesthesiologists.org

page 55

Practical application A specialised arterial catheter, inserted into either the brachial artery or femoral artery is required, along with either a thoracic or femoral central line. Some centres use treatment algorithms based on these variables, to guide use of fluid and inotropes in an attempt to maximise intravascular filling, without increasing the EVLW and causing pulmonary oedema. The use of EVLW as an endpoint for resuscitation has not been validated. Advantages The arterial line can be simultaneously used for blood pressure monitoring and for blood sampling. The system is relatively easy to set up and calibrate. It can also be used to estimate preload using global-end-diastolic volume and index (GEDI), intra-thoracic blood volume (ITBV) and pulmonary vascular permeability index (PVPI) which gives a ratio of EVLW to pulmonary blood volume. Note that pleural effusions do not affect measurements. Disadvantages The arterial catheter is relatively large gauge and expensive, although few complications have been reported. Recalibration is required every 12 hours, or following a major change in the patients clinical condition. Variations in speed of injection and thermistor positioning may affect results. Results can be affected by arrhythmias, shunting, positive pressure ventilation and tricuspid regurgitation. Pulse contour analysis ProAQT (Pulsion), Vigileo (Edwards Lifesciences) and LIDCOrapid (LIDCO) are all similar, minimally invasive cardiac output monitors. They all work by pulse contour analysis using a specialised transducer on any arterial line. Parameters obtained may include: continuous cardiac output, stroke volume, stroke volume variation (SVV) and pulse pressure variation (PPV). In order to obtain SVRi (the systemic vascular resistance index), the patient needs CVP monitoring. To obtain values for PPV and SVV, the patient should be ventilated with a fixed tidal volume and so is less useful when weaning respiratory support in intensive care. dP max gives an indication of contractility. Pulmonary artery flotation catheters (PAFC) The use of PAFCs has been hotly debated in recent years and use in
Table 2. Interpretation of SvO2 readings.

the United Kingdom is currently low. The PAC-Man trial showed no improvement in survival for patients randomised to have a PAFC inserted, compared to those who were not.3 Theory of technique A flexible balloon-tipped, flow-directed catheter is inserted via a wide-bore catheter sited in a central vein. The catheter is floated through the right atrium and ventricle to enter the pulmonary trunk. From this position it can intermittently be wedged in one of the pulmonary arteries. The catheter allows a number of variables to be measured and others to be derived. The measured variables are pulmonary artery pressure, pulmonary capillary wedge pressure (PCWP), cardiac output and mixed venous oxygen saturation. Traditionally, cardiac output is measured by thermodilution of 10ml iced water, injected through the proximal lumen of the catheter. Measurement of the fall in blood temperature against time from injection, as the cooled blood passes the distal end of the catheter, allows calculation of the cardiac output of the right (and therefore the left ventricle). Semi-continuous cardiac output measurements are now available which use warming coils in the right ventricular portion of the catheter. A sequence of heating and recording gives an averaged cardiac output after a short delay. Practical application The catheter is inserted with reference to certain waveforms seen in the right atrium, right ventricle, pulmonary outflow tract and when wedged in the pulmonary artery. Insertion may take several attempts and is more difficult in patients with a low cardiac output. Advantages Measurement of cardiac output is probably the most reliable of the variables measured using a PAFC and is therefore a valuable guide to interventions introduced to increase cardiac output. The numerous assumptions made in interpretation of the PCWP as a measure of preload or ventricular filling make the PCWP a less reliable measurement. Some units use the mixed venous oxygen saturation, measured using a sample taken slowly from the pulmonary artery aperture of the catheter, as a further indicator of a patients overall tissue perfusion (see below).

SvO2 Interpretation >75% Increased O2 delivery e.g. high FiO2, or Decreased O2 utilization e.g. sepsis causing shunt Maybe normal or reflect compensation by increase O2 extraction by tissues O2 demand is greater than supply Implies tissue beyond maximal O2 extraction Severe lactic acidosis Cellular death

50-75% 30-50%

25-30% <25%

page 56

Update in Anaesthesia | www.anaesthesiologists.org

Disadvantages This invasive monitor is associated with a number of potential complications. The PAC-Man study recorded non-fatal complications in 10% of insertions. In addition to the usual complications of central venous access, PAFCs may cause arrhythmias, heart block, rupture of the right heart or pulmonary artery, thromboembolism, pulmonary infarction, valvular damage, endocarditis.3 Mixed venous oxygen saturation (SvO2) Mixed venous oxygen saturations can be used as a surrogate marker of the global balance between oxygen delivery and consumption. Oxygen delivery depends on cardiac output and the oxygen content of the blood. In the face of an increased demand for oxygen, there will be a greater degree of oxygen extraction. Occasionally SvO2 may be increased in severe sepsis due to decreased extraction resulting from shunting (where blood bypasses the tissues). SvO2 can be used as an early warning system where a sudden decrease in SvO2 of 1020% requires immediate assessment. SvO2 can also be used to assess treatment. Central venous oxygen saturation (ScvO2) Measurement of ScvO2 requires a central venous catheter rather than a pulmonary artery catheter. ScvO2 can be used as a surrogate marker of the regional balance between oxygen delivery and consumption in the head, neck and upper body. The value is usually 2-7% less than SvO2 partly due to mixing with returning venous blood. Under nonshock conditions, ScvO2 correlates well with SvO2. In shock states the
Table 3. Comparison of different cardiac output monitors.

difference from SvO2 increases and can be up to 7% higher than SvO2. ScvO2 trends with SvO2 in a parallel manner but should be used in combination with other markers of perfusion. CeVox (PULSION) is a system which monitors continuous SvO2 and can calculate oxygen delivery, consumption and oxygen extraction. It uses a fibreoptic probe that can be inserted through any central line. Thoracic bioimpedance Theory of technique The technique depends on the change in bioimpedance of the thoracic cavity during systole. Impedance is a measure of the opposition to alternating current. Baseline impedance reflects total thoracic fluid volume. Cardiac output is estimated by measuring changes in electrical resistance through the thorax, since blood volume within the aorta changes during systole and diastole. Magnitude and rate of change reflects LV contractility. Practical application A series of ECG type electrodes are placed on the thorax and neck. A small, non-painful current is passed and measurements made. Advantages Derived stroke volume is calculated and cardiac output computed. Thoracic fluid content is also measured. This is the least invasive method of cardiac monitoring and was initially conceived for space flight monitoring.

Method PAFC LiDCO

Technique Thermodilution

Invasiveness High

Cannulae PA catheter

Continuous Limitations Yes Shunts, arrhythmias. Requires regular injection speed and thermistor positioning Shunts, arrhythmias, haemodynamic instability, cannot be used if on lithium therapy or if pregnant, lithium can accumulate. PCA requires good quality waveform Shunts, arrhythmias, haemodynamic instability. PCA requires good quality waveform. Waveform dependant, useful for trend only.

Lithium-dilution + Moderate pulse contour analysis (PCA) Thermodilution + Moderate pulse contour analysis (PCA) Pulse contour analysis Low

Any venous Yes + arterial

PiCCO

Central venous + arterial

Yes

ProAQT Vigileo LIDCOrapid TOE TOD NICO

Arterial line Yes

Doppler / twodimensional imaging Doppler Partial CO2 rebreathing Fick principle

Moderate Low Nil (although requires intubation)

None None None

No Yes Yes

User dependent, needs sedation User dependent , needs sedation, may pick up interference from other vessels Needs intubation, poor accuracy in lung disease Inaccurate in the critically ill in general

Thoracic Bioimpedance

Measurement of Nil change of impedance

None

Yes

page 57 Update in Anaesthesia | www.anaesthesiologists.org

Disadvantages It is not useful with significant aortic regurgitation and open chest procedures. The correlation with PAFC in critically ill patients is inconsistent. Bioreactance The NICOM (non-invasive cardiac output monitor) measures the phase shift of pulses of alternating current, passed through the body using three electrodes. Early studies show promising correlation with passive leg raise, as an indicator of fluid responsiveness.4 SUMMARY At present no perfect system exists, but each of the monitors above, can aid the clinician when uncertain about the patients condition. The information gained must be understood in the context of how

it was gathered and interpreted alongside clinical evaluation of the patient. Only then can it be safely used to guide subsequent therapeutic strategies. rEfErENCEs
1. Vincent J-L et al. Sepsis in European intensive care units: Results of the SOAP study. Crit Care Med 2006; 34: 34453 2. Lees N, Hamilton M, Rhodes A. Clinical review: Goal-directed therapy in high risk surgical patients. Critical Care 2009; 13: 231 3. Harvey S, Harrison D, Singer M et al. Assessment of the clinical effectiveness of pulmonary artery catheters in the management of patients in the intensive care (PAC-Man): a randomised controlled trial. Lancet 2005; 366: 472-6.

4. Monnet X, Rienzo M et al. Passive leg raising predicts fluid responsiveness in the critically ill. Crit Care Med 2006; 34: 1402-4.

page 58

Update in Anaesthesia | www.anaesthesiologists.org

Acid-base disorders in critical care

Alex Grice Correspondence Email: [email protected]
INTRODUCTION Metabolic acidosis is a common component of critical illness. Evaluation of this component can aid diagnosis, CASE EXAMPLES
Case 1 A known diabetic patient presented with severe diabetic ketoacidosis. He was drowsy, exhibited classical Kussmaul respiration and proceeded to have a respiratory arrest whilst being admitted to the ICU. After immediate intubation the trainee ventilated the patient with the ventilators default settings (rate 12, tidal volume 500ml, PEEP 5, FiO2 0.5) and attempted to secure arterial and central venous access. Shortly after intubation the patient became asystolic and could not be resuscitated. Why did this patient have a respiratory arrest? Why did he deteriorate after intubation?

assess severity (and likely outcome) and allow the clinician to determine whether current treatments are working.
was dehydrated and hypotensive. Admission bloods revealed Na+ 134, K+ 2.5, Cl- 122, urea 15.4, creatinine 280, and blood gas analysis revealed: pH PaCO2 PaO2 HCO3- 7.21 2.9 kPa 19.5 kPa 6.4mmol.L-1

Summary
Disorders in acid-base balance are commonly found in critically ill patients. Clinicians responsible for these patients need a clear understanding of acid-base pathophysiology in order to provide effective treatment for these disorders. This article concentrates on aspects of metabolic acidosis often seen in intensive care, including poisoning with the alcohols (ethanol, methanol and ethylene glycol).

Describe the acid-base disorder present. What is the likely cause? Is her compensation adequate?

Case 2 A young woman was admitted to the surgical ward with a history of severe vomiting and increasing right iliac fossa abdominal pain. On examination she was found to have a rigid abdomen, with visible sub-diaphragmatic gas on chest Xray. She was cardiovascularly unstable and her admission bloods showed Hb 17.4, WCC 24.6, Plt 79, Na+ 125, K+ 4.3, Cl- 93, urea 20.4, creatinine 310 (mcmol.L-1) and blood gas analysis showed: pH 7.10 PaCO2 5.2kPa PaO2 29.3kPa HCO3- 14.3mmol.L-1 Interpret these results. What is the likely cause of her acidosis? How do you interpret her chloride level? Is her compensation adequate/maximal?

Case 4 A man was brought in to the emergency room heavily intoxicated. He was known to be alcohol dependent and attended regularly. Blood analysis confirmed normal biochemistry apart from a borderline low glucose (3.8mmol.l-1) and arterial gas analysis showed pH 7.43, PaCO2 4.8, PaO2 15.7, HCO3- 20. He was placed into an observation bed overnight with a diagnosis of alcohol intoxication but later became tachypnoeic and hypotensive. Repeat gas analysis showed pH 7.0, PaCO2 4.2, PaO2 24, and HCO3- 9. What is the cause of his deterioration? What other information would be useful?

Case 3 An elderly lady was admitted from a care home with a one week history of severe diarrhoea. She

Case 5 The same man re-presented a month later, again heavily intoxicated. Blood analysis confirmed normal biochemistry and arterial gas analysis showed pH 7.43, PaCO2 4.8, PaO2 15.7, and HCO3- 20. Toxicology was requested and minimal ethanol was measured and no methanol found. Urinalysis revealed ketones but no blood or protein. He became more deeply unconscious and on intubation his arterial gases showed pH 7.1, PaCO2 9.5, PaO2 21, HCO3- 27. Serum osmolarity was measured at 336 (calculated 284). What is the cause of his deterioration?

Alex Grice Consultant Anaesthetist Royal Devon and Exeter NHS Foundation Trust UK

Update in Anaesthesia | www.anaesthesiologists.org

page 59

Acid-base disorders

Anaesthesia

Update in

DEFINITIONS
An acid is a substance that has the ability to give up a proton (H+ - a positively charged hydrogen ion), and so when in an aqueous solution they have a low pH. pH is a format used to describe the proton concentration in a solution. It is the negative logarithm10 of the H+ concentration, so when the blood pH is normal (7.40) the H+ concentration in the blood is 40nmol.L-1. For every ten-fold increase in H+ concentration the pH goes down by 1 unit. A base is a substance that has the ability to accept a proton and has a high pH in solution. Metabolic acidosis (a low pH in the tissue) exists when there is an excess level of fixed or exogenous acids in the body. Fixed acids include hydrochloric acid, sulphuric acid, phosphoric acid, ketoacids and lactic acid. Examples of exogenous acids are salicylate and methanol. Metabolic acidosis is accompanied by a drop in plasma bicarbonate concentration (relative to the bicarbonate concentration present prior to the onset of the acidosis). This drop in bicarbonate can either be caused by bicarbonate loss or by the presence of extra acid. The body can accommodate significant alterations in acid levels through buffering. The primary buffer in the blood is bicarbonate, which combines with excess acid (hydrogen ions) to make carbon dioxide, which decreases the effect of the acid on the blood pH. Buffering means that metabolic acidosis (a low tissue pH) does not always lead to the presence of metabolic acidaemia (a low blood pH). Blood pH only falls appreciably when the buffering capacity of the blood becomes overwhelmed. A drop in bicarbonate concentration is the hallmark of metabolic acidosis

EVALUATION When evaluating a critically ill patient with a metabolic acidosis it is necessary to determine the type of acidosis in order to identify the cause of the acidosis. To classify metabolic acidosis it is useful to calculate the anion gap and, if present, the size of the osmolar gap. These concepts are explained below. The role of the anion gap The anion gap is defined as the concentration difference between the major measured cations (ions which are positively charged) and anions (ions which are negatively charged) within the plasma and is normally 12 to 18 mmol.l-1.2 Anionic proteins, phosphate, sulphate and low levels of organic acids, which are not measured, account for the difference (i.e. the gap). When examining the cause of a metabolic acidosis it is useful to calculate the anion gap. Anion gap = [Na+ + K+] - [HCO3- + Cl-] = 15(3)mmol.L-1 A normal anion gap implies that an acidosis is due to primary bicarbonate loss: Plasma bicarbonate is low (the hallmark of acidosis) and chloride concentration is raised. This bicarbonate loss may be: - - gastrointestinal (diarrhoea, fistula) renal (renal tubular acidosis, drug effect).

Also occurs with rapid intravenous infusion of normal saline (excess chloride) or intravenous nutrition rich in cationic amino acids (e.g. arginine). An increased anion gap implies that fixed acids are being retained or an abnormal organic acid is present.
Figure 1. Illustration of the anion gap (all figures are mmol.L-1)

page 60

Update in Anaesthesia | www.anaesthesiologists.org

 Plasma bicarbonate is low and chloride concentration is normal. Fixed acids may be retained in: - uraemia - ketoacidosis (diabetic, alcoholic) - lactic acidosis. If fixed acids are normal, exogenous acids should be considered: - salicylate (aspirin) poisoning - methanol poisoning - ethylene glycol poisoning.

develops. Similarly the osmolar gap is not sensitive in ethylene glycol poisoning as the large molecular weight of this substance determines the mortality, only causing a small rise in osmolar gap. COMPENSATION FOR METABOLIC ACIDOSIS When treating critically ill patients with metabolic acidosis, it is important to consider the adequacy of their ventilatory response to acidosis when deciding on treatment priorities. Buffering provides the main means of accommodating a metabolic acidosis. As buffering capacity is exceeded, acidaemia develops. Once this rise in hydrogen ion concentration has reached the CSF, it is detected by chemoreceptors and compensation occurs by reducing carbon dioxide levels through hyperventilation (first described by Kussmaul). Detection of low pH in CSF rather than blood explains the delay in this compensation; rapid onset acidosis (for example during convulsions) tends not to stimulate respiration in spite of a low blood pH. Even though respiratory compensation occurs relatively quickly, it can take up to twelve hours to reach maximal capacity. This maximal capacity can be calculated: PaCO2 (maximal change, in kPa) = 0.2 [HCO3-] + 1 If the patients PaCO2 lies within 0.5kPa of this calculated value, then the respiratory response is appropriate to the level of metabolic acidosis. If the PaCO2 is higher, then the compensation is in a very early stage or the patient has a superimposed respiratory acidosis. If this is the case then earlier intervention with respiratory support is indicated. When providing respiratory support in patients with a metabolic acidosis it is important to remember that respiratory compensation causes an increased minute volume. If you instigate controlled ventilation with a normal minute volume, then the PaCO2 level will rise rapidly towards normal, the acidaemia will worsen, and the patient will become acutely unstable. Young fit patients with severe diabetic ketoacidosis can generate huge minute volumes (20-30 l.min-1) and drop their PaCO2 to below 2kPa. PITFALLS IN ASSESSING METABOLIC ACIDOSIS It is impossible to interpret arterial gases accurately without considering the history and presentation first. Consider the following arterial gas: pH 7.1 PaCO2 10.5kPa PaO2
-

Limitations of the anion gap While the anion gap is useful in evaluating metabolic acidosis, it is not sensitive. The normal range is quoted as 12 to 18 mmol.L-1 which means it is possible for a patient with a normal anion gap of 12 to acquire a severe lactic acidosis (plasma lactate greater than 5mmol.L-1) without generating an anion gap outside the normal range. The anion gap is also affected by plasma albumin (an important unmeasured anion) and low albumin levels can significantly offset an anticipated rise in anion gap. Role of the osmolar gap The osmolar gap represents the difference between a samples measured osmolality (number of osmoles of solute per kilogram of solvent) and its calculated osmolarity (number of osmoles of solute per litre of solution). It is a useful calculation to perform if alcohol poisoning is suspected (see later in this article). Osmolality is measured in the laboratory with an osmometer that either assesses the depression of a samples freezing point or the depression of its vapour pressure. It is preferable to use the former as any volatile alcohols in the sample will evaporate as the sample is heated and the results from this method will be inaccurate. Osmolarity can be calculated using various formulae. One such formula is: Calculated osmolarity = 2 [Na+] + urea + glucose Osmolality and osmolarity differ according to whether the number of osmotically active particles is dissolved in a kilogram or a litre of solvent respectively. The calculated osmolarity utilises the plasma concentration of sodium, glucose and urea. Even though sodium and chloride represent the most important determinants of osmolarity in plasma, chloride concentration is not commonly available. The formula is simplified by taking into account the incomplete dissociation of sodium chloride in plasma. The measured and calculated values should lie within 10mmol.L-1 of each other (the difference being created by the inaccuracy of the calculation and the inaccuracy of the osmometer) and if the gap is larger it suggests the presence of unmeasured osmotically active species. It is important to realise that the osmolar gap also has significant limitations. When considering alcohol poisoning the osmolar gap is only present as an early feature and returns towards normal as the alcohol is metabolised and the associated metabolic acidosis

29.3kPa

HCO3 24.3mmol.L-1 The interpretation of this result will vary according to the clinical presentation: If the gas sample was taken from a young unconscious patient, admitted through the emergency room, with pinpoint pupils then you would interpret the gas as showing primary respiratory acidosis from opiate overdose.

If the gas was taken from an elderly man with severe COPD presenting with sepsis, then the gas interpretation will be different.

Update in Anaesthesia | www.anaesthesiologists.org

page 61

Full compensation for a chronic respiratory acidosis should raise the bicarbonate by 3 x (PaCO2 5.3). If the PaCO2 level is chronically raised then you would expect the bicarbonate to be 36mmol.L-1. This bicarbonate is significantly lower and this suggests a metabolic acidosis superimposed on the background of a compensated respiratory acidosis. This highlights the importance of basing gas interpretation on clinical assessment. Other pitfalls arise from failing to recognise the limitations of the anion gap and osmolar gap. METABOLIC ACIDOSIS DUE TO ENDOGENOUS ACIDS Lactic acidosis Lactic acid is a weak acid that is present in the blood in low levels (1-2mmol.L-1). It is produced from pyruvate, the end substrate in glycolysis, the process by which carbohydrates are broken down to produce energy. Since some tissues (such as skin) produce more pyruvate than their mitochondria can handle, excess pyruvate is converted to lactate, released into the blood and metabolised by the liver (60%, Cori cycle) or kidney (40%). A rise in the lactate level in the blood suggests increased lactate production or decreased lactate metabolism. As the livers capacity to metabolise lactate is large, a rise in blood lactate levels suggests that a degree of impaired liver lactate handling is present; however, increased lactate production is still the primary feature of lactic acidosis that is amenable to treatment. Lactic acidosis is categorised according to the state of oxygen delivery. If oxygen delivery is inadequate (type A lactic acidosis), then aerobic metabolism is impaired, pyruvate accumulates and lactate is produced. We know oxygen delivery is a product of cardiac output and blood oxygen content, but in lactic acidosis low cardiac output is invariably the most important consideration. Oxygen content is rarely low enough to create a lactic acidosis in isolation the haemoglobin would need to be less than 5g.dl-1 or the PaO2 less than 4kPa. Treatment aims concentrate on restoring correct distribution of cardiac output and, to a lesser extent, ensuring adequate blood oxygen content (this is one situation where the low transfusion threshold of 7g.dl-1 for the critically ill should not apply). Type B lactic acidosis occurs when oxygen delivery is normal and a problem in carbohydrate metabolism is present. Causes of type B acidosis are subdivided as follows: B1 Underlying disease, also called stress lactate (ketoacidosis, haematological malignancy) B2 Drug or toxin effect (e.g. salbutamol) B3 Inborn error of metabolism

produced (as the central conversion molecule of cellular metabolism). Acetyl coA normally binds to oxaloacetate (OAA) to enter the citric acid cycle and generate high energy substrates. However if inadequate levels of OAA are present, then acetyl coA is converted into acetoacetate. If adequate levels of NAD+ are present, then acetoacetate is subsequently converted into -hydoxybutyrate. Ketones can be used as energy sources by the brain and the heart. The main causes of ketoacidosis include: Starvation ketoacidosis Alcoholic ketoacidosis Diabetic ketoacidosis. Starvation ketoacidosis This occurs when glycogen levels in the liver have become exhausted and the liver attempts to make more glucose via the gluconeogenesis pathway. Gluconeogenesis requires OAA and the subsequent drop in OAA levels limits the ability of the citric acid cycle to utilise acetyl coA provided by lipid metabolism. The excess acetyl coA is converted into ketone bodies and ketoacidosis develops. The acidosis tends to be within buffering capacity and the anion gap rise is small. The situation is resolved by supplying glucose in a controlled fashion and allowing the liver to revert back to the usual metabolic pathways. Alcoholic ketoacidosis This condition develops when ethanol is taken without enough calories. The starvation response is now complicated by the livers effort to metabolise ethanol. The conversion of ethanol into acetaldehyde requires NAD+ (a proton carrier with a vital role in the generation of the fuel molecule, ATP) and the excess NADH generated inhibits gluconeogenesis. This exacerbates the glucose deficiency and the corresponding drop in insulin levels stimulates lipid metabolism and ketoacidosis. The anion gap in this instance will be raised and the acidosis more severe (pH approaches 7.0). Analysis of the acid-base balance can be complicated by an appropriate compensatory respiratory alkalosis and a metabolic alkalosis if the patient has been vomiting. If significant dehydration is present these patients can also get a lactic acidosis, amplified by the relative excess of NADH. Treatment involves restoration of adequate circulating volume and the administration of both insulin and glucose. With prompt treatment the acidosis should resolve rapidly. Diabetic ketoacidosis Diabetic ketoacidosis develops when inadequate amounts of insulin are available. The insulin deficit reduces available intracellular glucose and increases fat breakdown and free fatty acid levels. The liver responds, as if in a starving state, by increasing lipid metabolism (further encouraged by increased levels of stress hormones) and, as gluconeogenesis depletes available oxaloacetate, the acetyl CoA generated is converted into ketone bodies. Acetoacetic acid and -hydroxybutyric acid dissociate and the H+ ion released is buffered by bicarbonate. An increased anion gap acidosis develops.

Treatment in this situation depends on determining the cause from the history and clinical signs and addressing the root cause, rather than attempting to correct the acidosis directly. Ketoacidosis Ketone bodies include -hydroxybutyrate, acetoacetate and acetone. When lipids are metabolised by -oxidation, acetyl co-enzyme A is

page 62

Update in Anaesthesia | www.anaesthesiologists.org

This acid-base picture may be complicated by various factors. Patients are often severely dehydrated and this can cause lactic acidosis due to inadequate tissue perfusion. Ketoacidosis causes vomiting and the resulting loss of acid can cause the calculated anion gap to underrepresent the severity of the acidosis. In addition, initial resuscitation with chloride rich solutions (0.9% saline) will increase chloride levels and further decrease the anion gap. Renal acidosis The kidneys ability to regulate acid-base balance can be adversely influenced in numerous ways. It is useful to categorise these conditions according to their effect on glomerular filtration. Acidosis associated with decreased glomerular filtration The most common forms of renal acidosis seen in intensive care are associated with a profound drop in glomerular filtration. Acute kidney injury (commonly due to acute tubular necrosis), and acute exacerbation of chronic kidney disease, both cause a metabolic acidosis because the kidney is unable to excrete fixed acids. The acidosis is exacerbated by the associated tubule damage. This damage prevents bicarbonate production from CO2, and ammonia excretion and buffering capacity is reduced as a result. Bicarbonate levels drop and chloride tends to remain stable and as a result the anion gap rises. Treatment involves correction of the precipitating factors and supporting renal function (with renal replacement therapy if required). Acidosis associated with preserved glomerular filtration This is renal tubular acidosis (RTA), which is less common in the critically ill and usually associated with either inherited disorders or pre-existing renal disease. While the glomerular filtration rate may be depressed, the acidosis is disproportionate to this minor reduction and tends to exhibit a normal anion gap. Renal tubular acidosis (RTA) is subdivided according the site of the tubular defect: Type 1 Type 2 Distal tubular defect Proximal tubular defect Distal tubular resistance to aldosterone (or aldosterone deficiency)

is increased and the urine pH is raised. However, as the proportion of bicarbonate filtered by the kidney is proportional to the plasma bicarbonate concentration, the acidosis is less severe than with type 1 RTA. The condition tends to be self-limiting and the bicarbonate tends not to drop below 15mmol.L-1. Potassium loss is less marked than with distal RTA but can be a problem if bicarbonate supplements are given to correct the acidosis. Any supplements need to include both bicarbonate and potassium. Type 4 RTA is also rare and tends to occur while aldosterone is deficient or the distal tubule becomes resistant (papillary necrosis). Both acid and potassium secretion are reduced and the urine remains relatively alkali whilst a metabolic acidosis develops in the presence of raised potassium. If severe this condition can be treated with oral fludrocortisone (0.1mg.day-1). Other causes of normal anion gap acidosis Normal anion gap acidosis occurs due to primary bicarbonate loss and this can occur through the kidney or the gut. Renal loss occurs with renal tubular acidosis as discussed above but can also occur as a drug effect (acetalozamide) or when the ureters are diverted to the bowel (ureterosigmoidostomy). The latter causes a problematic acidosis that responds poorly to dietary supplements and can be difficult to treat. Gut losses occur with severe diarrhoea or via nasogastric aspirates in patients with small bowel obstruction. Pancreatic fistulae, biliary drains and some bowel tumours also lose bicarbonate and cause a hyperchloraemic normal anion gap acidosis. METABOLIC ACIDOSIS DUE TO EXOGENOUS ACIDS Alcohol poisoning Ethanol, methanol, ethylene glycol and isopropanol represent the main alcohols encountered in poisoning. Ethanol is by far the most common cause of alcohol poisoning. Specialist laboratories are able to measure plasma alcohols but this is often not immediately available, therefore the diagnosis of alcohol poisoning can be helped by estimation of the osmolar gap. In order to understand the patterns seen in alcohol poisoning it is necessary to discuss how the various alcohols are metabolised. Metabolism of alcohol (methanol, ethanol and ethylene glycol) Ethanol, methanol and ethylene glycol are metabolised by the same enzyme systems, but produce different metabolites. Collectively they are termed alcohols. With methanol and ethylene glycol poisoning, only the parent compounds and the first metabolites (formaldehyde and glycoaldehyde respectively) are osmotically active. The subsequent metabolites are weak acids that dissociate into electrically charged ions that become balanced by sodium, and so cease to exert an osmotic influence. In the initial stages, metabolism generates an osmolar gap with minimal acidosis, however, further metabolism produces formic and glycolic acid respectively. This is not the case with ethanol and accounts for the clinical progression seen in methanol and ethylene glycol poisoning. These metabolites generate a metabolic acidosis and as they are

Type 4

Type 1 RTA is the most common form and is caused by the distal convoluted tubule failing to excrete hydrogen ions when attempting to reabsorb sodium. A metabolic acidosis develops as a consequence and the urine fails to acidify. Potassium excretion is unaffected and potassium loss in the urine may be increased as a result. This type of RTA has numerous causes including drug induced damage (amphotericin), autoimmune disorders (lupus) and nephrocalcinosis. It is diagnosed by confirming a high urine pH (greater than 5.5) in the presence of a severe metabolic acidosis. The underlying disorder should be addressed and the episodes of acidosis prevented by giving adequate dietary bicarbonate. Type 2 RTA is much less common and is caused by a defect in the proximal convoluted tubule that prevents bicarbonate reabsorption. It can be inherited or associate with Fanconi syndrome (generalised defect of tubular amino acid reabsorption.) Urinary bicarbonate loss

Update in Anaesthesia | www.anaesthesiologists.org

page 63

Ethanol [CH3-CH2-CH]
Alcohol dehydrogenase

Acetaldehyde [CH3-C-CH]
Alcohol dehydrogenase

Acetic acid [CH3-COOH] CO2+ H2O

Ethylene glycol [HO-CH2-CH2-OH]

Alcohol dehydrogenase

Methanol [CH3-OH] Formaldehyde [H-CO-H]

Ethylene glycol poisoning Ethylene glycol is less potent than methanol with 100ml ingestion representing a severe overdose. Toxicity initially presents with intoxication - slurring of speech, drowsiness and nausea. This can progress to marked cerebral depression and convulsions. Twelve hours after ingestion significant metabolism to glycoaldehyde has occurred causing cardiorespiratory depression and acidosis. Aldehydes inhibit oxidative phosphorylation, mechanisms for cellular respiration, glucose metabolism, protein synthesis, nucleic acid replication and synthesis. Myocardial depression can be significant and pulmonary oedema is commonly encountered. Renal tenderness and oliguria may become evident as acute tubular necrosis becomes established. Metabolism of ethylene glycol to oxalic acid causes a demonstrable degree of oxalate crystalluria and accounts for the low plasma calcium seen (chelation to form calcium oxalate). Ethylene glycol poisoning is lethal with levels of 21mg.dl-1 but it must be remembered that this will only generate a late osmolality increase of 4mosm.L-1 (delayed presentation). The treatment approach is similar to methanol toxicity with emphasis on the early use of ethanol infusions and haemodiafiltration. These treatments should continue until ethylene glycol can no longer be detected in blood. Isopropanol poisoning Isopropanol should also be considered when patients present with alcohol toxicity. This alcohol forms a major component of rubbing alcohol and is used in windscreen cleaning preparations and de-icer. Unlike methanol or ethylene glycol, isopropanol is metabolised to acetone and excreted in the urine. Acetone is not metabolised further and organic acid production is minimal. Both the parent compounds and metabolites are osmotically active and significant osmolar gaps may be seen with ingestion. Isopropanol toxicity tends to present with intoxication, meiosis (small pupils), and brain stem depression with significant overdose. Isopropanol is irritant and causes marked gastritis, pancreatitis and, if aspirated, causes tracheitis and pulmonary oedema. It is rapidly absorbed from the stomach and gastric lavage is of little benefit. Ketosis is more marked with isopropanol ingestion and this can provide a useful clue to diagnosis. Treatment is supportive and no effort should be made to limit metabolism of isopropanol with ethanol infusions. Isopropanol is readily cleared by haemodiafiltration but this treatment is rarely required. Salicylate poisoning Aspirin (acetylsalicylic acid) poisoning causes over 200 deaths a year in the UK. Therapeutic doses of aspirin are absorbed rapidly and completely from the stomach and larger doses may be absorbed for up to 18 hours after ingestion as the tablets coalescence in the stomach. Metabolism Acetylsalicylic acid is hydrolysed to salicylic acid and further metabolised in one of three ways. 1. Conjugation with glycine to salicyluric acid 2. Hydroxylated to gentisic acid 3. Conjugation with glucuronic acid to either salicylacyl glucuronide or salicyl phenolic glucuronide.

Glycoaldehyde [HO-CH2-COH]
Alcohol dehydrogenase

Glycolic acid [HO-CH2-COOH]

Formic acid [H-CO-OH]

Oxalic acid [HO-CO-CO-OH]

Figure 2. Metabolism of ethanol, ethylene glycol and methanol.

produced the osmolar gap decreases. The enzyme responsible, alcohol dehydrogenase, metabolises both poisons at a lower rate than ethanol and this may result in toxicity being delayed for up to 30 hours for methanol, and twelve hours for ethylene glycol. Use of the osmolar gap may therefore be misleading. When blood is taken for these tests it is impossible to determine which substance has been taken and how far the metabolism of the alcohol has progressed. If the osmolar gap is normal then either the alcohol metabolism is advanced or the patient has not ingested a large quantity. However, patients usually present to hospital with the onset of symptoms and this normally coincides with intermediate alcohol metabolism and a raised osmolar gap. Methanol poisoning Methanol is a potent poison and serious toxicity is seen after an intake of 10ml. Methylated spirits contains 5% methanol and ingestion of over 200ml is required before serious toxicity is encountered. This toxicity presents after a latent period of twelve hours or more. Patients report headache, breathlessness and visual symptoms ranging from blurring to complete blindness. Severe abdominal pain and nausea are common and patients may present with a rigid abdomen. The cardiovascular system is initially stable until a severe acidosis develops, then marked myocardial depression and bradycardia is encountered. Treatment is directed at reducing the rate of organic acid production with intravenous ethanol and haemodialysis. Ethanol treatment consists of a 50g oral loading dose (125ml of spirits will suffice) followed by an intravenous infusion of 10-12g.h-1. Treatment should aim to achieve a plasma concentration of 1-2g.L-1. The indications for dialysis are not well defined but should be considered in the presence of a severe metabolic acidosis, marked visual or mental symptoms or in the presence of a high methanol plasma concentration (over 500mg.L-1). Folinic acid can also be given intravenously to help prevent ocular toxicity (30mg IV 6 hourly).

page 64

Update in Anaesthesia | www.anaesthesiologists.org

Conjugation with glucuronic acid is saturable so that levels of nonprotein bound salicylate rise disproportionately with increasing dose. The excretion of unchanged salicylate by the kidneys then becomes increasingly important and alkalisation of the urine will therefore maximise its excretion by ion trapping. Drug effects Central to the metabolic disturbances initiated by acetylsalicylic acid is the uncoupling of oxidative phosphorylation. The resulting increase in oxygen consumption and carbon dioxide production leads to a respiratory alkalosis that is worsened by direct stimulation of the respiratory centre. Bicarbonate excretion is enhanced and sodium, potassium and water loss also occurs. When this is combined with hyperpyrexia and sweating, then marked dehydration and electrolyte imbalance follow. Stimulation of the chemoreceptor trigger zone may induce vomiting and this will further exacerbate this imbalance. Uncoupled oxidative phosphorylation enhances glycolysis and increases the peripheral demand for glucose. This occurs mainly in muscle and may provoke hypoglycaemia. The brain is particularly sensitive to this and neuroglycopenia may then lead to depression of the respiratory centre. The metabolic acidosis seen in acetylsalicylic acid poisoning is caused by the stimulation of lipid metabolism (increasing the formation of ketoacids) and the inhibition of enzymes within the Krebs cycle (increasing levels of pyruvic and lactic acid). The acidosis is poorly tolerated due to the reduced buffering capacity following the initial respiratory alkalosis and bicarbonate excretion. Clinical presentation The clinical picture depends on the age of the patient and on the total dose ingested. Acute overdose in the setting of chronic use augments toxicity. Plasma salicylate levels 6 hours after ingestion can be classified as mild (300-500mg.L-1), moderate (500-750mg.L-1) and severe (>750mg.L-1) poisoning. Below ten years of age respiratory alkalosis tends to be a transient feature with metabolic acidosis predominating. However, respiratory alkalosis is more marked in adults, where it helps to keep salicylate in the ionised form, preventing it from crossing cell membranes. The following features are usually present regardless of the severity: sweating, vomiting, pyrexia, tinnitus and epigastric pain. As the metabolic acidosis worsens more salicylate enters the central nervous system and tremor, delirium, convulsions and eventually coma ensues. Non-cardiogenic pulmonary oedema and acute renal failure are also features of acetylsalicylic acid toxicity. Treatment involves vigorous gastric lavage to limit continued absorption. Dehydration and electrolyte disturbances (particularly hypokalaemia) should be corrected. Severe toxicity requires more than these basic manoeuvres and alkalisation of the urine should be performed. Urinary salicylate excretion is encouraged with intravenous infusions of bicarbonate, aiming to raise urinary pH >7.5 while avoiding a plasma pH >7.55. Haemodialysis has also been used to augment salicylate excretion.

CASE EXAMPLES - DISCUSSION Case 1 The patient presented with a metabolic acidosis caused by loss of diabetic control, relative insulin deficiency and consequently ketoacidosis. Kussmaul respiration is seen in respiratory compensation through hyperventilation - the rise in minute volume can be substantial and maintaining this compensation is very strenuous. As time passes patients tire, the minute volume decreases and the acidosis worsens rapidly. Young patients in particular can maintain full respiratory compensation, almost to the point of respiratory arrest, as in this example. Elderly patients tend to tire earlier. When taking over a patients ventilation, in the setting of a metabolic acidosis, it is vital that the minute volume used is raised appropriately. If standard ventilation settings are applied, as in this case example, the drop in alveolar ventilation postinduction will remove the respiratory compensation and the acidosis will deteriorate dramatically. This is the likely cause of this patients cardiovascular collapse. Case 2 This patient, presenting with peritonitis, has a raised haemoglobin, urea and creatinine suggesting dehydration and a likely lactic acidosis due to inadequate tissue perfusion. The arterial gases confirm that acidaemia is present (low pH) and the low bicarbonate level implies that the acidosis is metabolic. The anion gap is raised at 35 and this strongly suggests the presence of an increased anion gap acidosis. The vomiting that occurred in the days leading up to admission may have caused a metabolic alkalosis (loss of hydrogen ions and chloride) that has been masked by the superimposed lactic acidosis. Therefore a severely increased anion gap acidosis has only dropped the pH to 7.1, when you would expect the pH to be lower. Using the bicarbonate level (14.3mmol.L-1), the calculated PaCO2 during maximal compensation would be 3.8kPa ([14.3 x 0.2] +1). The fact that her PaCO2 is higher should serve as a warning of imminent fatigue and ventilation should be supported as quickly as possible in order to prevent cardiorespiratory collapse. Case 3 This woman presented with a low pH and low bicarbonate, implying that a metabolic acidosis is present. The extent of the dehydration raises the possibility of a lactic acidosis, however this would create an increased anion gap. The anion gap in is low (8.5), and this strongly supports the presence of a normal anion gap acidosis secondary to bicarbonate loss from the gut.

Update in Anaesthesia | www.anaesthesiologists.org

page 65

With a bicarbonate level of 6.4 you would expect the patient to drop her PaCO2 to 2.28 ([6.4 x 0.2] +1) and in this instance her respiratory compensation is appropriate. Case 4 This patient was initially felt to be intoxicated with alcohol and the low blood glucose raised the possibility of alcoholic ketoacidosis. The initial arterial gas does not support this as a low bicarbonate and an increased anion gap would be expected. The subsequent deterioration however is typical of methanol poisoning. Methanol metabolism is often delayed for 12 to 18 hours especially if taken with ethanol. The first stage in metabolism generates formaldehyde which is not an acidic species metabolic acidosis will not feature at this stage. When the formaldehyde is metabolised into formic acid then the increased anion gap acidosis develops. It would have been useful to determine whether there was an osmolar gap on initial presentation. If this was raised then it would have alerted the clinicians to the possibility of poisoning with an alcohol, and ethanol treatment would have prevented the deterioration. Interestingly the subsequent arterial gas shows a severe metabolic acidosis but the PaCO2 has only dropped minimally. The calculated PaCO2 for maximal compensation suggests the PaCO2 should have dropped to 3.0kPa. The relative respiratory acidosis arises from methanol induced respiratory depression and should prompt early ventilatory support. Case 5 In this example the same patient re-presents with an intoxicated picture but ethanol and methanol are not detected on this admission. In spite of this, and the normal arterial gases on admission he became comatose with arterial gases suggesting a primary respiratory acidosis (low pH, raised PaCO2). The rise in bicarbonate is appropriate for the rise in PaCO2 (expected bicarbonate rise = 0.75 [PaCO2 5.3] for an acute respiratory acidosis), so no metabolic effects are demonstrated.

This presentation is typical for isopropanol poisoning. Metabolic acidosis is not a feature of isopropanol poisoning; however the alcohol is heavily intoxicating and can easily cause respiratory depression as seen in this example. The key to aid diagnosis is the presence of the increased osmolar gap combined with the urine ketosis (from the metabolite, acetone). Treatment is supportive (the airway will need to be secured in this example and respiratory support provided) until the alcohol has been metabolised. SUMMARY Metabolic acidosis is commonly encountered in intensive care and correct management requires confirmation that an acidosis is present, determination of its nature and then determination of the probable cause. Identification of the root cause will then allow specific treatment with improved efficacy. FURTHER READING
1. Cohen RD, Woods HF. Disturbances of acid-base homeostasis. In: Weatherall DJ, Ledingham JGG, Warrell DA, Eds. Oxford Textbook of Medicine. Oxford: Oxford University Press, 1987; 9.174 2. Hoffman RS, Smilkstein MJ, Howland MA, Goldfrank LR. Osmol gaps revisited. Normal values and limitations. Journal of Toxicology Clinical Toxicology 1993; 31: 81-93. 3. Salem MM, Mujais SK. Gaps in the anion gap. Archives of Internal Medicine 1992; 152: 1625-9. 4. Stewart PA. How to understand acid-base. A quantitative acid-base primer for biology and medicine. Available at: http://www.acidbase. org/ 5. Vale JA, Meredith TJ Proudfoot AT. Poisoning by alcohols and glycols. In: Weatherall DJ, Ledingham JGG, Warrell DA, Eds. Oxford Textbook of Medicine. Oxford: Oxford University Press, 1987; 6.41-6.44. 6. Williams HE Alcoholic hypoglycaemia and ketoacidosis. Medical Clinics of North America 1984; 68: 33-8. 7. Almaghamsi A, Yeung C. Osmolal gap in alcoholic ketoacidosis. Clinical Nephrology 1997; 48:52-3.

page 66

Update in Anaesthesia | www.anaesthesiologists.org

Anaesthesia
Delirium in critical care
David Connor* and William English *Correspondence Email: [email protected]
INTRODUCTION Delirium is a common complication of critical illness. It has conventionally been regarded as an unavoidable and benign side effect of long-term sedation on an intensive care unit (ICU). However in recent years this pre-conception has been challenged by the publication of studies demonstrating poorer outcomes in ICU patients with delirium. This article will define delirium, summarise the risk factors, provide an overview of the current evidence for its detection and discuss its management. DEFINITON AND CLASSIFICATION The American Psychiatric Association defines delirium as a disturbance of consciousness, attention, cognition and perception which develops over a short period of time (usually hours to days) and tends to fluctuate during the course of the day.1 Delirium can be subclassified according to aetiology using the DSM IV criteria. This is difficult to apply to the critical care population in whom a multifactorial origin is likely. A more useful clinical classification system was first described in elderly patients by Lipowski in 1983.2 Three sub-types of delirium were described: Hypoactive delirium Patients appear subdued, withdrawn and have a poor response to stimulus. Hyperactive delirium Patients may display agitation or aggression and may experience delusions or hallucinations. Mixed delirium Patients fluctuate between hypo- and hyperactive subtypes. Ouimet et al first defined sub-syndromal delirium in a patient sub-group who displayed some features of delirium, but didnt meet the full diagnostic criteria. This introduced the concept of delirium as a spectrum of disease rather than a single entity.3 RISK FACTORS Numerous risk factors have been identified for the development of delirium on the ICU.4,-7 They are summarised in Table 1. DIAGNOSIS Delirium was traditionally diagnosed by a psychiatrist using DSM IV criteria. Whilst psychiatric referral can still be helpful, the development of specific delirium assessment tools, for use by the multi-disciplinary team, has greatly improved its recognition on intensive care. However delirium is probably still under-diagnosed, particularly in the hypoactive sub-type, where the more subtle features may be overlooked. The assessment tool most commonly employed in UK clinical practice is the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU).8 Both CAM-ICU and the Intensive Care Delirium Screening Checklist (ICDSC) have been specifically validated for use on the intensive care unit.4 Both tools are reproduced in the original version of this article (totw.anaesthesiologists.org). Both are easy and quick to perform and have good inter-observer reliability.4 CAM-ICU, performed once every 24 hours, directly assesses the patient performing tasks to command and can be used during mechanical ventilation. ICDSC, documented every 8 hours, is more subjective as it relies on data collected during routine nursing care, without direct assessment of the patient. Patients who are experiencing isolated hallucinations may be assessed as delirium negative by CAM-ICU, but delirium positive by ICDSC. Summary
Delerium is a common problem in ICU patients and results in longer hospital stays, and increased morbidity and mortality. Standardised assessment tools are available and have highlighted a higher incidence than previously believed. Haloperidol is the mainstay of treatment despite lack of a firm evidence base for its use. Benzodiazepines should be avoided unless alcohol withdrawal is the cause of delirium.

Update in

Both CAM-ICU and ICDSC have been shown to have a high sensitivity (97%9 and 99%10 respectively) but CAM-ICU has a much better specificity (99%9) than ICDSC (64%10). Another study, which directly compared the performance of the two scoring systems, David Connor suggested a good level of agreement between them.11 Registrar in Anaesthesia INCIDENCE For many years, the lack of a consistent definition for delirium, that could be applied to intensive care patients, hampered efforts to determine its incidence in this setting. The development of the two delirium William English Consultant in Intensive Care Medicine Royal Cornwall Hospital UK

Update in Anaesthesia | www.anaesthesiologists.org

page 67

Delirium

Table 1. Risk factors for delirium on ICU.

Medical history Age over 70 Hypertension Congestive cardiac failure Stroke Epilepsy Depression Dementia HIV Renal impairment Hepatic impairment Visual or hearing impairment Social history Smoker Alcohol abuse Malnutrition Environmental Physical restraints Rectal or urethral catheter Central venous catheter Sleep deprivation

Acute presentation Disease severity (APACHE II score) Metabolic derangement Thyroid function Glycaemic control Hyper/hyponatraemia Renal function

Peterson et al noted that the most common delirium subtypes were mixed (54.9%) and hypoactive (43.5%) whilst hyperactive was found to be relatively uncommon (1.6%).15 ICDSC Ouimet et al7 identified delirium in 31.8% of 764 patients in a mixed specialty intensive care unit using the ICDSC tool. Whatever the true incidence of delirium is, it appears to be much more common than previously thought and the introduction of validated assessment tools has improved the recognition of this important condition. PATHOPHYSIOLOGY Currently there is no comprehensive explanation for the mechanism by which delirium occurs in the critically ill. There are however numerous hypotheses and it seems likely that its pathophysiology is multifactorial. An excellent review by Girard et al16 covers several of the leading suggestions and these are summarised in Figure 1 (adapted from Figueroa-Ramos et al17): 1. 2. 3. 4. Increased levels of dopamine and reduced levels of acetylcholine are thought to increase neuronal excitability and precipitate delirium. These changes may be caused by changes in the synthesis, release and inactivation of these neurotransmitters. Whether other neurotransmitters (such as GABA, endorphins, glutamate or histamine) are also involved is unknown. Tryptophan is an amino acid which is actively transported across the blood brain barrier via LAT1 proteins. It is a precursor for serotonin and subsequently melatonin production. Low levels of tryptophan, and thus serotonin and melatonin, are hypothesised to cause hyperactive delirium. High levels of tryptophan, serotonin and melatonin may be responsible for hypoactive delirium.18 It is unclear whether these effects are due to serotonin, melatonin, the neurotoxic metabolites of tryptophan or all of the above. Phenylalanine is another amino acid which is actively transported across the blood brain barrier via the same transport channel as tryptophan. Consequently, high uptake of phenylalanine will compete with tryptophan and reduce levels of serotonin and melatonin. Once across the blood brain barrier, phenylalanine is converted into DOPA and subsequently dopamine, noradrenaline and adrenaline. High levels of phenylalanine have been associated with delirium,19 but it is unclear whether this effect is due to increased levels of noradrenaline and dopamine, reduced serotonin and melatonin, or all of the above. The inflammatory response to critical illness causes the release of cytokines into the circulation which results in a pro-thrombotic state. Animal studies suggest that this leads to reduced cerebral blood flow and it is possible that this could trigger delirium.

Thermoregulation Sepsis Hypoxaemia Anxiety Uncontrolled pain Medications Use of an epidural Opiates Benzodiazepines Propofol Anti-cholinergics

screening tools discussed has gone some way to address this issue. However, reported incidence still varies widely (16.1%-83.3%) depending on the patient demographics, illness severity and screening tool used.9,10 DSM IV One study in 2001 suggested that the incidence of delirium, when assessed by two independent psycho-geriatricians using DSM-IV criteria, was as high as 81.3% in the 48 study patients.12 During validation of the ICDSC, a psychiatrist identified delirium in 16.1% of 93 study patients using DSM IV criteria.10 CAM-ICU The pilot for the CAM-ICU assessment tool found a high incidence of 83.3% in 111 study patients.9 Subsequent studies using CAMICU suggest that the incidence varies between 41-74%.6,13 This is in comparison to the data from our local mixed surgical and medical ICU in which CAM-ICU screening detected delirium in 31.7% of patients at some point in their admission.14

5. Engel and Romano performed EEG recordings on delirious patients in the 1940s and concluded that the slow EEG appearance they observed was characteristic of a derangement in the general functional metabolism of the brain.20 Other investigators have suggested that this might result in delirium by reducing acetylcholine levels.21

page 68

Update in Anaesthesia | www.anaesthesiologists.org

PREVENTION A recent paper by Morandi et al introduces the concept of an ABCDE bundle which uses an evidence-based approach in the prevention of delirium.22 This is summarised in Figure 2. Awake and breathing The Awakening and Breathing Controlled Trial found that daily sedation breaks, paired with trials of spontaneous breathing, significantly improved outcome at 1 year.23 These findings have led to the adoption of this practice in many intensive care units, although in a survey of clinical practice, the majority of practitioners admit that sedation breaks are not performed as frequently as intended.24

Choice of sedation The mainstay of sedation on ICU has traditionally been propofol, benzodiazepines and opiates, all of which have been implicated in altering sleep patterns.25 Trials involving 2 receptor agonists (clonidine and shorter-acting dexmedetomidine) have reported a lower incidence of delirium and shorter time to extubation.26,27 Remifentanil is a shortacting pure receptor agonist. Its use as a sedative agent in intensive care has been shown to reduce the time to extubation,28 but further work is needed to assess its impact on the incidence of delirium. Interestingly, a Danish study randomised 140 mechanically ventilated patients to receive either no sedation or propofol sedation with daily sedation breaks.29 It reported shorter times to extubation and a lower

Abnormal tryptophan metabolism Decreased tryptophan Glutamate Histamine GABA Decreased serotonin Increased serotonin Increased tryptophan

Endorphins Mechanism of action unknown

Decreased melatonin

Increased melatonin

Hyperactive delirium

Hypoactive delirium

Cerebral ischaemia leading to diffuse brain injury

DELIRIUM

Endothelial damage Thrombin formation Microvascular compromise

Neuronal excitability increased

Neurotransmitters Increased noradrenaline Increased dopamine Reduced acetylcholine

Increased: IL1 IL2 TNF

Increased phenylalanine (precursor of dopamine & NA)

Inflammatory response

Figure 1. Pathophysiology of delirium17

Update in Anaesthesia | www.anaesthesiologists.org

page 69

Table 2. Modifiable factors in delirium.

Awakening and trial of breathing Choice of sedatives and analgesics

General factors Correct visual impairment with glasses Correct hearing impairment with hearing aids Medical factors Correct metabolic derangement Diagnose and treat sources of infection Achieve adequate tissue oxygen delivery Administer adequate analgesia Remove lines and catheters promptly Do not use physical restraints routinely but only use acutely to prevent harm

Reduced time to extubation Reduced length of stay in both hospital and ICU Reduced mortality Improved return of cognitive function Improved independence

Daily delirium monitoring

Early mobility Exercise

Medications Avoid deliriogenic drugs where possible Environmental factors Orientate the patient regularly Reduce noise Reduce sleep disturbance Mobilise where possible TREATMENT: PHARMACOLOGICAL There is a lack of randomised control trial evidence for pharmacological treatments for delirium on the intensive care unit. The mainstay of current therapy and that recommended by both the Intensive Care Society (UK) and the American College of Critical Care Medicine (level C recommendation) is haloperidol.25,34 Surveys of clinical practice in the US35 and the UK8 revealed that the majority of clinicians use haloperidol as their first line treatment for delirium. In the UK this remains an off-licence indication for haloperidol administration. Haloperidol Haloperidol is a dopamine receptor (D2) antagonist and acts centrally to reduce hallucinations and delusions. It is hepatically metabolised with an elimination half-life of 10-36 hours, secondary to active metabolites. Recognised adverse effects include extra-pyramidal side effects, prolonged QT interval (which can precipitate torsades de point tachycardia) and neuroleptic malignant syndrome. The optimum dosing schedule has not yet been established by trial evidence, but a commonly used schedule is 2.5-5mg intravenously every 6 hours. Doses may need to be reduced in the elderly. It has also been used as a continuous infusion in severe cases, but this does not represent routine practice.36 A retrospective study of 989 mechanically ventilated patients identified a significant reduction in hospital mortality in those patients who had received haloperidol during their intensive care stay. However, the study design meant that it was not possible to identify if the indication for commencing the haloperidol was delirium.37

Figure 2. ABCDE bundle22

incidence of delirium, without an increase in self-extubation in the group randomised to no sedation, but it is unlikely that this practice will become widely adopted. Daily delirium monitoring Daily screening for delirium is important as delirium is underdiagnosed without the use of assessment tools.30 Early mobility and exercise Schweickert et al demonstrated that if physical and occupational therapy was provided at the same time as a sedation break and trial of spontaneous breathing, then patients had shorter episodes of delirium and improved function at hospital discharge.31 Sleep It is unclear whether sleep disruption on intensive care is a cause or a consequence of delirium. Studies have shown that the total sleep time is unaffected by sedation, but that altered REM (rapid eye movement) patterns are observed, suggesting an impact on the quality of sleep.32 High levels of noise or ambient light, drugs, mechanical ventilation and routine patient care at inappropriate times of the day have all been associated with sleep disruption.33 TREATMENT: NON-PHARMACOLOGICAL The first stage in the management of delirium is to recognise its presence by use of an appropriate assessment tool. The next stage is to review the delirium risk factors in Table 1, looking for precipitant causes that may be correctable. Some of the risk factors listed are clearly more amenable to modification than others. The more important modifiable factors are seen in Table 2.

page 70

Update in Anaesthesia | www.anaesthesiologists.org

Atypical anti-psychotics Atypical anti-psychotics (such as olanzapine, quetiapine) are also dopamine receptor (D2) antagonists, but have additional antagonistic effects on serotonin receptors (5-HT2A). Enteral administration is required as there are no intravenous preparations available. They are generally metabolised in the liver and have active metabolites. Their half-lives vary according to the preparation, with quetiapine having the shortest half-life of 6 hours. The adverse effects that are most likely to be encountered include sedation and anti-cholinergic symptoms. A randomised, but unblinded, trial of enteral olanzapine versus haloperidol in 103 patients demonstrated improvement in daily Delirium Index scores and reduced benzodiazepine administration in both trial groups, without a significant difference between them.38 A randomised, double blinded trial of quetiapine against placebo, with rescue haloperidol if required, found that the quetiapine group had a faster resolution of delirium.39 The recently published MIND study randomly assigned 101 patients to haloperidol, ziprasidone (atypical anti-psychotic) or placebo. Doses were adjusted according to the level of delirium as assessed by CAM-ICU. There was no significant difference in the number of days patients survived without delirium or coma, in any of the 3 groups in this small pilot study. A further multi-centre placebo trial is planned.40 Benzodiazepines Benzodiazepines have a role in the management of delirium caused by alcohol withdrawal. However, their administration in other patient sub-groups has been identified as an independent risk factor for delirium development. Their use should therefore be avoided where possible in critically ill patients. An adapted summary of the delirium treatment guidance produced by the UK Clinical Pharmacy Association and the Intensive Care Society is in reference 25. PROGNOSIS Mortality A 6-month follow up study by Ely et al determined a statistically significantly higher 6-month mortality in ICU patients with delirium (34% v 15%, adjusted hazard ratio of 3.2).41 Another study of 102 mechanically ventilated patients determined that ICU mortality was higher for patients with delirium compared to those without (63.6% v 32.5%, hazard ratio of 2.5).42 Overall ICU mortality rates were lower in Ouimet et als study of 537 patients, but it was still significantly higher in patients with delirium compared to those without (15.9% v 2.4%).3 Morbidity Patients with delirium are more likely to self extubate and remove invasive medical devices.5 Length of stay A study of 48 patients demonstrated that delirium significantly increased both the hospital and ICU length of stay.12 A further study of 224 patients found that patients with delirium spent a median of

10 days longer in hospital than those without.41 These findings are supported by Ouimet et als study which demonstrated that even subsyndromal delirium significantly increased length of stay.3 Cost Milbrandt et al examined the cost of the hospital and ICU stays of 224 medical ICU patients in 2004.43 They reported that patients with delirium had a significantly higher cost of care than those without and that those costs were dependent on the severity of the delirium. Long-term cognitive impairment A long term cohort study of 77 ICU patients determined that 79% of survivors had cognitive impairment at 3 months and 71% at 12 months.44 A third remained severely impaired a year following ICU discharge. Delirium was identified as an independent predictor of cognitive impairment in this study. Duration of delirium also seems to be important. Patients who experienced delirium for 5 days scored almost 7 points fewer on cognitive testing, 1 year following discharge than those who experienced 1 day of delirium. SUMMARY Despite the surge of research activity into delirium over the past decade, the condition remains an important problem on intensive care. Standardised assessment tools validated for use in the ICU setting have been developed and have demonstrated a higher incidence of delirium than previously thought. Current treatments have a limited evidence base, particularly with respect to improving patient outcome. Whilst haloperidol currently remains the mainstay of pharmacological management, there is increasing interest in prevention of delirium by modification of its risk factors. Recent evidence suggests that delirium results in longer hospital stays, higher associated treatment costs and increased morbidity and mortality. WEB LINKS
www.icudelirium.org www.icudelirium.co.uk 1. Girard T, Pandharipande and Ely W. Delirium in the intensive care unit. Critical Care 2008; 12: S3. 2. King J and Gratrix A. Delirium in intensive care. BJA CEACCP 2009; 9: 144-7. 1. American Psychiatric Association practice guidelines for the treatment of psychiatric disorders: Compendium 2006: 72-73. 2. Lipowski Z. Transient cognitive disorders (delirium, acute confusional states) in the elderly. Am J Psychiatry 1983; 140: 1426-36. 3. Ouimet S, Riker R, Bergeon N, Cossette M, Kavanagh B & Skrobik Y. Subsyndromal delirium in the ICU: evidence for a disease spectrum. Intensive Care Med 2007; 33: 1007-13. 4. Devlin J, Fong J, Fraser G & Riker R. Delirium assessment in the critically ill. Intensive Care Med 2007; 33: 929-40. 5. Dubois M-J, Bergeron N, Dumont M, Dial S & Skrobik Y. Delirium in an intensive care unit: A study of risk factors. Intensive Care Med 2001; 27: 1297-304.

FURTHER READING

REFERENCES

Update in Anaesthesia | www.anaesthesiologists.org

page 71

6. Salluh J, Soares M, Teles J, Ceraso D, Raimondi N, Nava V et al. Delirium epidemiology in critical care (DECCA): an international study. Critical Care 2010; 14: R210. 7. Ouimet S, Kavanagh B, Gottfried S & Skrobik Y. Incidence, risk factors and consequences of ICU delirium. Intensive Care Med 2007; 33: 66 73. 8. MacSweeney R et al. A national survey of the management of delirium in UK intensive care units. QJM 2010; 103: 243-51. 9. Ely W, Inouye S, Bernard G, Gordon S, Francis J, May L et al. Delirium in mechanically ventilated patients: Validity and Reliability of the Confusion Assessment Method for the Intensive Care Unit (CAM ICU). JAMA 2001; 286: 2703-10.

(Awakening and Breathing Controlled trial): A randomised controlled trial. Lancet 2008; 371: 126-134. 24. Patel R, Gambrell M, Speroff T, Scott T, Pun B, Okahashi J et al. Delirium and sedation in the intensive care unit: Survey of behaviors and attitudes of 1384 healthcare professionals; Crit Care Med 2009; 37; 825-32. 25. Borthwick M, Bourne R, Craig M, Egan A and Oxley J. Detection, prevention and treatment of delirium in critically ill patients. 2006; UKCPA. 26. Riker R, Shehabi Y, Bokesch P, Ceraso D, Wisemandle W, Koura F et al. Dexmedetomidine vs midazolam for sedation of critically ill patients: A randomised trial. JAMA 2009; 301; 489-99. 27. Rubino A, Onorati F, Caroleo S, Galato E, Nucera S Amantea B et al. Impact of clonidine administration on delirium and related respiratory weaning after surgical correction of acute type-A aortic dissection: Results of a pilot study. Interactive Cardiovascular and Thoracic Surgery 2010; 10: 58-62. 28. Muellejans B, Matthey T, Scholpp J and Schill M. Sedation in the intensive care unit with remifentanil/propofol versus midazolam/ fentanyl: A randomised, open label, pharmacoeconomic trial. Crit Care Med 2006; 10: R91. 29. Strom T, Martinussen T and Toft P. A protocol of no sedation for critically ill patients receiving mechanical ventilation: A randomised trial. Lancet 2010; 375: 475-80. 30. Spronk P, Riekerk B, Hofhuis J and Rommes J. Occurrence of delirium is severely underestimated in the ICU during daily care. Intensive Care Med 2009; 35: 1276-80. 31. Schweickert W, Pohlman M, Pohlman A, Nigos C, Pawlik A, Esbrook C et al. Early physical and occupational therapy in mechanically ventilated, critically ill patients: A randomised controlled trial. Lancet 2009; 373: 1874-82. 32. Hardin K, Seyal M, Stewart T and Bonekat W. Sleep in critically ill chemically paralysed patients requiring mechanical ventilation. Chest 2006; 129: 1468-77. 33. Gabor J, Cooper A and Hanly P. Sleep disruption in the intensive care unit. Current Opinion in Critical Care 2001; 7: 21-7. 34. NICE, Delirium: Diagnosis, prevention and management; 2010; NICE clinical guideline 103. 35. Ely W et al. Current opinions regarding the importance, diagnosis and management of delirium in the intensive care unit: A survey of 912 healthcare professionals. Crit Care Med 2004; 32: 106-12. 36. Riker R, Fraser G and Cox P. Continuous infusion of haloperidol controls agitation in critically ill patients. Crit Care Med 1994; 22: 433-40. 37. Millbrandt E, Kersten A, Kong L, Weissfeld L, Clermont G, Fink M et al. Haloperidol use is associated with lower hospital mortality in mechanically ventilated patients. Crit Care Med 2005; 33: 226-9. 38. Skrobik Y, Bergeron N, Dumont M and Gottfried S. Olanzapine vs haloperidol: Treating delirium in a critical care setting. Intensive Care Med 2004; 30: 444-9. 39. Devlin J, Roberts R, Fong J, Skrobik Y, Riker R, Hill N et al. Efficacy and safety of quetiapine in critically ill patients with delirium: A prospective, multicentre, randomised, double-blind, placebo-controlled pilot study. Crit Care Med 2010; 38: 419-27.

10. Bergeron N, Dubois M, Dumont M, Dial S & Skrobik Y. Intensive Care Delirium Screening Checklist: Evaluation of a new screening tool. Intensive Care Med 2001; 27: 859-64. 11. Plaschke K, von Haken R, Scholz M, Engelhardt R, Brobeil A, Martin E et al. Comparison of the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) with the Intensive Care Delirium Screening Checklist (ICDSC) for delirium in critical care patients gives high agreement rate(s). Intensive Care Med 2008; 34: 431-6. 12. Ely W, Gautam S, Margolin R, Francis J, May L, Speroff T et al. The impact of delirium in the intensive care unit on hospital length of stay. Intensive Care Med 2001; 27: 1892-900. 13. Page V, Navarange S, Gama S & McAuley D. Routine delirium monitoring in a UK critical care unit. Critical Care 2008; 13: R16. 14. Vanstone R, Paddle J & Powell C. Delirium scoring and sedation in a UK ICU. Intensive Care Med 2009; 35: S105. 15. Peterson J, Pun B, Dittus R, Thomason J, Jackson J, Shintani A et al. Delirium and its motoric subtypes: a study of 614 critically ill patients. J Am Geriatr Soc 2006; 54: 479-84. 16. Girard T, Pandharipande & Ely W. Delirium in the intensive care unit. Critical Care 2008; 12: S3. 17. Figueroa-Ramos M, Arroyo-Novoa C, Lee K, Padilla G & Puntillo K. Sleep and delirium in ICU patients: A review of mechanisms and manifestations. Intensive Care Med 2009; 35: 781-95. 18. Pandharipande P, Morandi A, Adams J, Girard T, Thompson J, Shintani A et al. Plasma tryptophan and tyrosine levels are independent risk factors for delirium in critically ill patients. Intensive Care Med 2009; 35: 1886-92. 19. Van der Mast R, van den Broek W, Fekkes D, Pepplinkhuizen L & Habbema J. Is delirium after cardiac surgery related to plasma amino acids and physical condition. J Neuropsychiatry Clin Neurosci 2000; 12: 57-63. 20. Engel G & Romano J. Delirium, a syndrome of cerebral insufficiency. Journal of Chronic Diseases 1959; 9: 260-77. 21. Lipowski Z. Delirium (Acute confusional states). JAMA 1987; 258: 1789 92. 22. Morandi A, Brummel N & Ely W. Sedation, delirium and mechanical ventilation: The ABCDE approach. Current Opinion in Critical Care 2011; 17: 43-9. 23. Girard T, Kress J, Fuchs B, Thomason J, Schweickert W, Pun B et al. Efficacy and safety of a paired sedation and ventilator weaning protocol for mechanically ventilated patients in intensive care

page 72

Update in Anaesthesia | www.anaesthesiologists.org

40. Girard T, Pandharipande P, Carson S, Schmidt G, Wright P and Canonico A. Feasibility, efficacy and safety of antipsychotics for intensive care unit delirium: the MIND randomised, placebo-controlled trial. Crit Care Med 2010; 38: 428-37. 41. Ely W, Shintani A, Truman B, Speroff T, Gordon S, Harrell F et al. Delirium as a predictor of mortality in mechanically ventilated patients in the intensive care unit. JAMA 2004; 291: 1753-62. 42. Lin S-M, Liu C-Y, Wang C-H, Lin H-C, Huang C-D, Huang P-Y et al. The impact of delirium on the survival of mechanically ventilated patients. Crit Care Med 2004; 32: 2254-9.

43. Milbrandt E, Deppen S, Harrison P, Shintani A, Speroff T, Stiles R et al. Costs associated with delirium in mechanically ventilated patients. Crit Care Med 2004; 32: 955-62. 44. Girard T, Jackson J, Pandharipande P, Pun B, Thompson J, Shintani A et al. Delirium as a predictor of long-term cognitive impairment in survivors of critical illness. Crit Care Med 2010; 38: 1513-20.

Update in Anaesthesia | www.anaesthesiologists.org

page 73

Anaesthesia
Sedation in intensive care patients

Update in

Sedation

Gavin Werrett Correspondence Email: [email protected]

INTRODUCTION Sedation is an essential component of the management of intensive care patients. It is required to relieve the discomfort and anxiety caused by procedures such as tracheal intubation, ventilation, suction and physiotherapy. It can also minimise agitation yet maximise rest and appropriate sleep. Analgesia is an almost universal requirement for the intensive care patient. Adequate sedation and analgesia ameliorates the metabolic response to surgery and trauma. Too much or too little sedation and analgesia can cause increased morbidity, for example over sedation can cause hypotension, prolonged recovery time, delayed weaning, gut ileus, DVT, nausea and immunosuppression; under sedation can cause hypertension, tachycardia, increased oxygen consumption, myocardial ischaemia, atelectasis, tracheal tube intolerance and infection. Sedation in the ICU varies widely from producing complete unconsciousness and paralysis to being nursed awake yet comfortable. There are many components to the ideal regimen but key elements include recognition of pain, anxiolysis, amnesia, sleep and muscle relaxation. Although the mainstay of therapy is pharmacological, other approaches are just as important: 1. Good communication with regular reassurance, 2. Environmental control such as temperature, humidity, lighting and noise, 3. Explanation prior to procedures, 4. Management of thirst, hunger, constipation, full bladder, 5. Variety for the patient - e.g. radio, visits from relatives, washing/shaving, 6. Appropriate diurnal variation - gives pattern to days. ASSESSSING THE LEVEL OF SEDATION Gavin Werrett Derriford Hospital The dosage of commonly used sedative and analgesic Plymouth drugs varies widely between patients because of UK variations in pharmacokinetics and pharmacodynamics. A valid method for monitoring sedation would allow sedation to be tailored to the individual. Any scoring system needs to be simple, rapidly performed, non-invasive and, most importantly, reproducible. Physiological variables, serum concentrations of drugs and neurophysiological tools such as EEG, bispectral index and lower oesophageal contractility have all been used, but are expensive, unreliable and unavailable. The best systems are clinically based. Commonly used ones include the Richmond Agitation Sedation Scale and the Ramsay Scale.
Table 1. The Ramsay Scale - six levels of sedation are used.

Summary
Many critically ill patients require sedation in order to tolerate more invasive therapies, such as intubation and ventilation. The level of sedation should be assessed regularly and sedation holds may reduce the duration of ventilation. The available drugs are discussed with emphasis on their advantages and disadvantages.

1. Anxious and agitated 2. Cooperative, orientated and tranquil 3. Responds to verbal commands only 4. Asleep but brisk response to loud auditory stimulus/light glabellar tap (to the forehead) 5. Asleep but sluggish response to loud auditory stimulus/light glabellar tap 6. Asleep, no response. This should be completed hourly by the attending nurse, but can be reduced in frequency as the patient stablizes. Levels 2 to 5 can be considered suitable for patient in the ICU. An increase in the sedation score must prompt the physician to make a differential diagnosis between over sedation or neurological/biochemical disease. As a rule, the aim for the majority of patients is for them to be sleepy, although easily rousable and hence cooperative. There is a definite trend and increasing body of evidence towards less sedation and more analgesia, with daily sedation holds and spontaneous breathing trials.1 It is preferable to allow the patient to breathe as soon as possible with triggered ventilation, such as pressure support. Ventilators are becoming increasingly sophisticated to allow a patient to

page 74

Update in Anaesthesia | www.anaesthesiologists.org

Table 2. The Richmond Agitation Sedation Scale - the score ranges from +4 to -5.

Score Term Description +4 +3 +2 +1 0 -1 -2 -3 -4 -5 Combative Very agitated Agitated Restless Alert and calm Drowsy Light sedation Moderate sedation Deep sedation Unrousable Not fully alert, but has sustained awakening to voice (eye opening >10 seconds) Briefly awakens with eye contact to voice (<10 seconds) Movement or eye opening to voice but no eye contact No response to voice, but movement or eye opening to physical stimulation No response to voice or physical stimulation Overtly combative, violent, immediate danger to staff Pulls or removes tubes, catheters, aggressive Frequent non-purposeful movement, fights ventilator Anxious but movements not aggressive

This should be repeated regularly and a suggested aim should be for a score of 0 to -1. synchronise with the ventilator. Deep sedation with or without paralysis is reserved for severe head injury, critical oxygenation (reduces work of breathing and improves chest compliance) and diseases such as tetanus. DRUGS USED IN SEDATION The ideal sedative agent should possess the following qualities: Both sedative AND analgesic Minimal cardiovascular side effects Controllable respiratory side effects Rapid onset/offset of action No accumulation in renal/hepatic dysfunction Inactive metabolites Cheap No interactions with other ICU drugs. Such a drug does not exist and therefore typically drug combinations are required. Sedative drugs may be given as boluses or infusions. As a rule, infusions for maintenance are preferable, with boluses for procedures, although continuous infusion results in higher cumulative doses. Benzodiazepines These are particularly useful because they are anxiolytic, anticonvulsant, amnesic and provide some muscle relaxation in addition to their hypnotic effect. Their effects are mediated by depressing the excitability of the limbic system, via reversible binding at the gamma aminobutyric acid (GABA)-benzodiazepine receptor complex. They have minimal cardiorespiratory depressant effects and are also synergistic with opioids. However rapid bolus doses can cause both hypotension and respiratory arrest. They are all metabolised in the liver. The common drugs used in this class are diazepam, midazolam and lorazepam. Diazepam use has decreased because of concern about its active metabolites (especially nor-desmethyldiazepam), which has a long half life and can accumulate, particularly in the elderly and patients with hepatic impairment. It is safe to give in single boluses, if given sensibly. Midazolam is water soluble at pH 4, yet fat soluble at pH 7, thus avoiding the unnecessary solvents required with the other two drugs and hence causing less irritation at the injection site. It has three metabolites, one of which (1-hydroxymidazolam) can accumulate in the critically ill. The normal elimination half life is 2 hours but can be as long as a few days in the long term sedated, critically ill patient. Lorazepam undergoes glucuronidation and has metabolites that are thought to be inactive. Overdose or accumulation can be reversed by flumazenil, the benzodiazepine receptor antagonist. It should be given in small aliquots as large doses can precipitate seizures. It has a half life of only 1 hour so may need to be given as an infusion. There is wide inter-patient variability in the potency, efficacy and pharmacokinetics of benzodiazepines, so the dose must be titrated to the level of sedation. After long term administration the dose should be reduced gradually or a lower dose reinstated if there is withdrawal (symptoms include insomnia, anxiety, dysphoria and sweating.) Propofol (2,6-diisopropylphenol) The mode of action is via the GABA receptor, but at a different site to the benzodiazepines. It was first developed as an intravenous anaesthetic agent and has a rapid onset of action yet, because it is metabolised rapidly both hepatically and extra-hepatically, it is ideal for continuous infusion. Recovery usually occurs within 10 minutes but it can accumulate with prolonged use, particularly in the obese patient. It is solubilised as an emulsion and the formulation can cause thrombophlebitis and pain, so ideally it should be infused via a large or central vein. Prolonged infusions can lead to increased triglyceride and cholesterol levels and its use is not licensed in children because of associated deaths attributable to this fat load. A theoretical maximum recommended dose is thus 4mg.kg-1.h-1 to avoid propofol infusion syndrome, a rare syndrome leading to cardiac failure, rhabdomyolysis, metabolic acidaemia and renal failure. It is often fatal, treatment is supportive but early recognition reduces mortality.

Update in Anaesthesia | www.anaesthesiologists.org

page 75

Disadvantages also include cardiorespiratory depression, particularly in the elderly, septic or hypovolaemic patient. Infusions may cause the urine to turn green. Ketamine Ketamine acts at the N-methyl-D-aspartate (NMDA) receptor. In sub-anaesthetic doses ketamine is sedative and also analgesic. However it is generally not used because of the rise in blood pressure, ICP and pulse rate that may result. It also causes hallucinations but these can be avoided if administered concomitantly with a benzodiazepine. It appears not to accumulate and sometimes has a role in severe asthma given its bronchodilatory properties. Etomidate Historically was used in ICU as an infusion but is now no longer used as it has been shown to cause adrenal suppression, even after a single dose. Barbituates These, for example thiopentone, have been used, especially in the management of patients with head injuries and seizure disorders. They cause significant cardiovascular depression and accumulate during infusions leading to prolonged recovery times. Thiopentone is still used occasionally in severely raised ICP to induce a barbituate coma, and in intractable seizure activity. Its effect can be titrated using EEG burst suppression. Butyrophenones and phenothiazines Strictly these are classed as major tranquillizers but they remain useful in ICU, particularly in agitated/delirious patients. A sliding scale of haloperidol may be particularly useful in a patient with delirium to promote calmness i.e. increasing doses if no effect after 15 minutes until the desired response is achieved. Haloperidol in particular causes minimal respiratory depression and has less alpha blocking tendency than chlorpromazine and hence less hypotension. Other side effects include prolongation of the QT interval (caution when given with erythromycin), extrapyramidal effects or neuroleptic malignant syndrome. Clonidine This is the most well known of the alpha-2 agonists but also has alpha-1 agonistic properties. A more specific agonist is dexmedetomidine but this is expensive and rarely available at present. It is particularly useful in patients with sympathetic overactivity, such as alcohol withdrawal and tetanus, as it inhibits catecholamine release. It is also synergistic with opioids and acts at the spinal cord to inhibit nociceptive inputs, thus imparting analgesia. It is contraindicated in hypovolaemia and can cause hypotension, bradycardia and dry mouth. Chlormethiazole This is a vitamin B derivative widely used for treatment of delirium tremens. It is not a respiratory depressant and is an anticonvulsant. Chloral hydrate This is used in paediatric intensive care as an adjunct, usually to a

benzodiazepine such as midazolam. It is metabolised in the liver to the active compound trichloroethanol. Metabolites can accumulate in renal dysfunction. Volatile agents Isoflurane has been used in concentrations of up to 0.6% and produces good long term sedation with minimal cardiorespiratory side effects and yet rapid awakening. Scavenging and pollution are a problem as is incorporating the vaporiser into the ventilator. Free fluoride ions from metabolised methoxyfluorane can cause renal failure. More recently desflurane has been shown to be effective in sedation with rapid offset of effects. DRUGS USED FOR ANALGESIA (in combination with sedation) Opioids are the mainstay of treatment and possess sedative, antitussive (cough suppressant) and hypnotic effects, besides the obvious analgesic effects. They work at the opioid receptors, reclassified in the late 80s to OP1 (old delta), OP2 (old kappa), OP3 (old mu). Most of the recognised effects are mediated via the OP3 receptor. Undesirable effects include gastrointestinal stasis and respiratory depression. Newer opioids have fewer side effects and accumulate less. It is equally important however to remember other analgesic techniques such as non-steroidal anti-inflammatory drugs (NSAIDs), paracetamol and regional techniques (e.g. epidural infusions). Morphine This is a commonly used drug. All other opioids are measured against morphine, although some newer agents have specific advantages. The dose required for analgesia is very variable and it can be delivered as intermittent boluses (problems with peak and trough effects, but less accumulation) or as a continuous infusion. Morphine is metabolised mostly in the liver to two main products, morphine-3-glucuronide and morphine-6-glucuronide (M-6-G). Both are excreted renally and accumulate in renal dysfunction. The M-6-G metabolite also has independent long lasting, sedative activity. Morphine has minimal cardiovascular side effects, unless given as a large bolus to hypovolaemic patients or secondary to histamine release. It can be used in renal failure as long as the dosing interval is increased or the infusion rate reduced. Normal duration of action after a single dose is about 2 hours. Care should be taken, as with all opioids, in hepatic failure. Fentanyl Fentanyl is a potent synthetic opioid. It is presented as a short acting opioid, with a rapid onset. After prolonged infusion the duration of action approaches that of morphine, although it does not accumulate in renal failure. It does not cause histamine release and is suitable for analgesia in the haemodynamically unstable patient. Alfentanil Alfentanil is one of the newer synthetic opioids and has an onset of action about five times faster than fentanyl, due to the small volume of distribution, but is less lipid soluble so is not prone to accumulation. The duration of action is about a third that of fentanyl and it too

page 76

Update in Anaesthesia | www.anaesthesiologists.org

is safe in renal failure. It has minimal cardiovascular effects and is a potent antitussive agent. Remifentanil Remifentanil possesses many of the qualities desired of the ideal ICU analgesic and sedative agent. Remifentanil is an ultra short acting opioid metabolised by non-specific tissue and blood esterases. It has a rapid onset of action and does not accumulate after infusion even in organ dysfunction. It enables predictable recovery, facilitating patient interaction and assessment and therefore enables a shorter weaning time and potentially a reduction in the time spent on mechanical ventilation.2,3 Many claim that remifentanil could help control ICU costs, by reducing the time spent in ICU.2,3 It is however very expensive and each intensive care unit would need to determine its own cost saving analysis. DRUGS USED FOR MUSCULAR RELAXATION In some patients muscle relaxation may be needed in addition to sedation and analgesia. Such indications include: Early resuscitation (including intubation) Refractory hypoxaemia e.g. ARDS - will decrease oxygen consumption and optimise chest wall compliance Raised intracranial pressure - stops coughing and patients resisting ventilation Tetanus During patient transfer To allow inverse ratio/prone ventilation It is vital to remember that relaxants have no effect on conscious level or comfort and should be avoided if possible. There are no standard clinical techniques to monitor conscious level in the paralysed patient so it is necessary to give generous doses of sedative drugs. In the UK, use of relaxants has fallen from about 90% of patients in the 80s to 10% of patients in the 90s. Some relaxants used in anaesthesia are less suitable for use in the ICU. Suxamethonium is predominantly used during emergency tracheal intubation, but the resultant rise in serum potassium must be expected which makes it inappropriate for use in cases of renal failure. Excessive potassium release also occurs after 48hrs in extensive burns and spinal cord injury. Pancuronium is long acting, but it may cause tachycardia and accumulates in renal failure. Vecuronium is an analogue of the aminosteroid pancuronium, but causes minimal cardiovascular side effects. It is suitable for intubation and infusion. Atracurium is a benzylisoquinolinium and is metabolised by ester hydrolysis and Hoffman (spontaneous) elimination. Its metabolites are inactive and it doesnt accumulate in renal or hepatic dysfunction. Histamine release occasionally occurs with boluses, but recovery occurs predictably within one hour, regardless of duration of infusion. The intubating dose is 0.5mg.kg-1, infusion 4-12mcg.kg-1.min-1.

Monitoring should ideally be performed using a nerve stimulator (e.g. train-of-four count). Clinical monitoring such as cardiovascular reflexes to noxious stimuli should also be observed. Full surgical relaxation may not be necessary. Problems with relaxants 1. The patient may receive inadequate sedation and be aware. This can be checked by withdrawing muscle relaxants for a time to allow recovery of muscular function and assessment of sedation levels. 2. Accumulation (especially with aminosteroids) in acute renal failure. 3. Critical illness polyneuropathy and myopathy (esp. if steroids also used). 4. Tendency to over-sedate. 5. Enhanced paralysis from other common ICU problems such as hypokalaemia, aminoglycoside antibiotics, hypophosphataemia. RECOMMENDATIONS Non-ventilated patients Pain should be titrated with opioids to the desired level. Cooperative patients may benefit from patient-controlled analgesia. Regional techniques in selected patients are ideal. Always use simple analgesics in combination, and consider other causes for pain e.g. full bladder. Postoperative/short-term mechanical ventilation If available, then a combination of remifentanil or alfentanil and propofol allows a rapid wake up, but is only beneficial if used for less than 72 hours. Sometimes the high costs of short acting agents can be offset against the higher hidden costs of delayed weaning/ prolonged ICU stay. Alternatively a benzodiazepine/morphine combination can be used. Long term mechanical ventilation There is little logic in using very short acting substances in these cases. Kress et al performed a randomised controlled trial that showed that daily interruption of sedative infusions reduced the duration of mechanical ventilation and intensive care stay in the critically ill. Infusions were interrupted until the patient was awake and could follow instructions or became agitated or uncomfortable. 4 Morphine plus midazolam or propofol were the agents used and the daily wake up procedure helped prevent excessive administration of these agents. A policy of interruption of sedation should be considered in all patients every day. In some centres a newer technique of sedation is employed - patient controlled sedation - using increments of propofol, as opposed to morphine/fentanyl that is usually use