Nephrol Dial Transplant (2006) 21: 88–92

doi:10.1093/ndt/gfi163
Advance Access publication 11 October 2005

Original Article

The unrecognized prevalence of chronic kidney disease in diabetes

Rachel J. Middleton1, Robert N. Foley2,3, Janet Hegarty1, Ching M. Cheung1, Patrick McElduff 4,
J. Martin Gibson5, Philip A. Kalra1, Donal J. O’Donoghue1 and John P. New5

1
Department of Renal Medicine, Hope Hospital, Salford, UK, 2Chronic Disease Research Group and University of

Downloaded from https://academic.oup.com/ndt/article/21/1/88/1819015 by guest on 04 October 2021

Minnesota, USA, 3University of Minnesota, Minneapolis, Minnesota, USA and 4Evidence of Public Health Unit,
School of Epidemiology and Health Sciences, University of Manchester, UK and 5Department of Diabetes and
Endocrinology, Hope Hospital, Salford, UK

Abstract Conclusions. Undiagnosed CKD is common in

Background. Diabetes mellitus and chronic kidney diabetes. Current screening strategies, based on creati-
disease (CKD) are common and exhibit synergistic nine or albuminuria, fail to identify a considerable
associations with premature mortality. Current number of subjects with CKD. Incorporating eGFR
diabetes guidelines in the UK recommend annual into screening for CKD would identify individuals
urinary albumin and serum creatinine determinations earlier in the natural history of the disease and enable
to screen for diabetic kidney disease. The aim of this early effective treatment to delay progression of CKD.
study was to estimate the burden of CKD in patients
with diabetes and examine the ability of serum Keywords: serum creatinine; estimated glomerular
creatinine and albuminuria to detect clinically filtration rate; modification of diet in renal disease
meaningful CKD compared with estimated glomerular (MDRD) study equation; diabetic kidney disease;
filtration rate (eGFR). sensitivity
Methods. All adults known to have diabetes in
primary and secondary care in Salford, UK, alive
with independent renal function on 1 January 2004
were included in this observational study (n ¼ 7596).
Demographic and laboratory parameters were Introduction
obtained from the Electronic Patient Record. eGFR
was determined using the 4-variable modification of Diabetes mellitus is the most common cause of end-
diet in renal disease (MDRD) formula. Clinically stage renal disease (ESRD) in the UK and it has been
meaningful CKD was defined as an eGFR estimated that 366 million people worldwide will have
<60 ml/min/1.73 m2. diabetes mellitus by 2030 [1,2]. Furthermore, diabetes
Results. Creatinine and albuminuria were measured and chronic kidney disease (CKD) exhibit synergistic
in the preceding 2 years in 82.3 and 55.2% of subjects, associations with cardiovascular disease and premature
respectively. In patients with CKD, normoalbuminuria mortality.
was present in 48.8%, and serum creatinine was Guidelines for diabetes care in the UK recommend
normal (
120 mmol/l) in 54.7%. An abnormal serum annual urinary albumin and serum creatinine determi-
creatinine (120 mmol/l) had a sensitivity and specifi- nations, and nephrology referral when serum creatinine
city of 45.3 and 100%, respectively, to identify CKD. levels exceed 150 mmol/l [3,4]. Serum creatinine produc-
The combination of abnormal creatinine and albumin- tion is dependent on lean body mass and therefore may
uria had an improved performance but still failed to not be an accurate reflection of glomerular filtration
detect a large number with CKD (sensitivity 82.4%, rate (GFR). At identical GFR levels, older subjects
specificity 75.4%). Serum creatinine failed to identify and females have lower muscle mass, lower rates of
CKD more often in females (OR 8.22, CI 6.56–10.29). creatinine production and lower serum creatinine
levels [5]. Isotope GFR is the definitive measure of
the level of renal function; however, it is logistically
unrealistic for population screening. Formulae based
Corresponding and offprint requests to: Rachel J. Middleton,
Specialist Registrar in Nephrology, Department of Renal
on age, sex, race and serum creatinine attempt to
Medicine, Hope Hospital, Stott Lane, Salford M6 8HD, UK. correct for differences in muscle mass and have been
Email: [email protected] shown to be comparable to isotope GFR in CKD [6].
ß The Author [2005]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.
For Permissions, please email: [email protected]
The unrecognised prevalence of chronic kidney disease in diabetes 89
Table 1. Prevalence of CKD according to K/DOQI classification

Stage Description GFR N Prevalence % Age (years)

(ml/min/1.73 m2) Mean (SD)

0–2 GFR 60 ml/min/1.73 m2 without albuminuria dataa 60 2254 36.0 64.7 (13.4)
0 Normal 90 1719 27.5 54.1 (12.6)
1 Kidney damageb with normal GFR 90 168 2.7 48.0 (12.7)
2 Kidney damageb with mild #GFR 60–89 391 6.3 59.7 (11.9)
3 Moderate # GFR 30–59 1547 24.8 72.1 (10.8)
4 Severe # GFR 15–29 156 2.5 71.9 (12.9)
5 Kidney failure <15 12 0.2 63.2 (13.9)
a
Subjects with GFR 60 ml/min/1.73 m2 without albuminuria data may have no kidney disease or stage 1–2 CKD.
b
Kidney damage is defined as structural or functional abnormalities of the kidney manifest by pathological abnormalities or marker of
kidney damage such as abnormal urine analysis or renal imaging.

Downloaded from https://academic.oup.com/ndt/article/21/1/88/1819015 by guest on 04 October 2021

Recent American Diabetes Association guidelines sug- diet in renal disease (MDRD) formula (eGFR) [9]:
gest the use of estimated GFR as a superior measure of
eGFR ¼186ðsCr=88:4Þ1:154
renal function compared with serum creatinine levels [7].
Studies have shown that a GFR <60 ml/min/1.73 m2  age0:203 ð  0:742 if female;  1:21 if blackÞ:
is a harbinger of premature cardiovascular death [8].
The Salford population is predominantly Caucasian (96.1%
The detection of CKD is therefore important because
of the population are Caucasian, 0.6% of black origin).
it identifies patients at high risk of cardiovascular Therefore all subjects were assumed to be non-black for the
disease, and not merely because it identifies those at risk purposes of GFR calculation. Subjects were classified as
for progressive kidney disease and its complications. having microalbuminuria if the albumin–creatinine ratio
The aim of this study was to establish the burden (ACR) was >2.5 mg/mmol in males and >3.5 mg/mmol in
of clinically meaningful, but unrecognised, CKD in females and proteinuria if ACR was >30 mg/mmol or 24 h
patients with diabetes and to evaluate the effectiveness proteinuria >0.3 g/24 h. The term ‘albuminuria’ includes
of serum creatinine and albuminuria to detect CKD. patients with both micro and macroalbuminuria.
The US National Kidney Foundation’s Kidney Disease
Outcomes Quality Initiative (K/DOQI) classification was
Methods used to define the stage of CKD (Table 1). For the purposes
of this study, clinically significant CKD was defined as
K/DOQI CKD stages 3–5 (eGFR <60 ml/min/1.73 m2). The
Subjects
sensitivity and specificity of serum creatinine and albumi-
Salford is an urban district of Greater Manchester in the nuria alone or in combination were compared with eGFR
UK with a population of 216 000 served by a single teaching <60 ml/min/1.73 m2.
hospital. The Salford diabetes Electronic Patient Record This study addresses key audit criteria outlined in the UK
(EPR) was introduced in 1992 and is a continuously updated NICE (National Institute of Clinical Excellence) guidelines
electronic diabetes health care system that incorporates for management of diabetes [3]. All patients gave verbal
data from primary and secondary care. The EPR was used consent to the storage and use of anonymized data to monitor
to define the cohort of all adults with diagnosed diabetes and improve clinical care.
in Salford district. The EPR is part of the hospital electronic
patient record system, allowing real-time collection of
clinical, biochemical, haematological, clinic visit and hospital Statistics
admission data. Normally distributed variables are summarized by their
All patients on the EPR in 2002 were included in this means and SDs, median and range are used for skewed data.
study provided they were still alive on 1 January 2004, except Analysis of variance and chi-square analysis, respectively,
for those on dialysis or with a functioning renal transplant. were used for between-group comparisons of continuous and
The population with diabetes have a skewed distribution categorical variables, respectively. Logistic regression was
with respect to age with a greater number of older subjects used to calculate adjusted odds ratios for (i) the presence of
compared with the general population. Therefore the pre- CKD and (ii) the presence of CKD failed to be diagnosed by
valence of CKD (GFR<60 ml/min/1.73 m2) was examined serum creatinine >120 mmol/l and albuminuria. SPSS version
according to age;  or <70 years old. 11.5 was used to perform all the analyses.

Definitions of CKD and albuminuria Results

The most recent serum creatinine and urinary albumin level
in the period January 2002 to December 2003 were used for 7596 subjects were studied. Table 2 shows their charac-
the analysis. Creatinine measurements were performed using teristics on 1 January 2004, as well as the propor-
an uncompensated modified Jaffé reaction using a Roche tions receiving measurements of serum creatinine,
diagnostics analyser, Integra 700. GFR was calculated using albuminuria, total cholesterol and blood pressure in
a validated GFR estimate, the 4-variable modification of the preceding 2 years. Table 1 shows the prevalence of
90 R. J. Middleton et al.
Table 2. Baseline characteristics

Number evaluated (%) %/Mean (SD)

Type 2 diabetes mellitus (%) 7596 (100) 87.7

Male sex (%) 7596 (100) 54.2
Age (years) 7596 (100) 62.6 (14.8)
Primary care (%) 7596 (100) 64.2
HbA1c (%) 6109 (80.4) 7.9 (1.7)
Blood pressure (mmHg) 5797 (76.3) 138/75 (20/11)
Urinary albumin-to-creatinine ratio ratio (mg/mmol) 2898 (38.2) 1.2 (0.5, 3.5)a
Urinary albumin-to creatinine ratio (secondary care only)
(mg/mmol) 1910 (70.2) 1.3 (0.6, 4.1)a
24-hour protein excretion (mg/day) 341 (4.5) 0.3 (0.2, 0.6)a
Albuminuria present(%) 3034 (39.9) 30.5
Cholesterol (mmol/l) 5964 (78.5) 4.7 (1.0)
Creatinine (mmol/l) 6247 (82.2) 88 (77, 103)a

Downloaded from https://academic.oup.com/ndt/article/21/1/88/1819015 by guest on 04 October 2021

Creatinine >150 mmol/l (%) 6247 (82.2) 4.9
a
Non-parametric data expressed as median and inter-quartile range.

Table 3. Diagnostic test performance of serum creatinine >120 mmol/l, albuminuria and proteinuria to detect estimated GFR
<60 ml/min/1.73 m2

Creatinine >120 mmol/l Albuminuriaa Creatinine >120 mmol/l Creatinine >120 mmol/l Proteinuriab
or albuminuria and albuminuria

Sensitivity 45.3% 51.2% 82.4% 21.0% 19.4%

Specificity 100% 75.5% 75.4% 100.0% 96.1%
Positive predictive value 100% 38.5% 61.7% 100.0% 57.8%
Negative predictive value 82.8% 83.8% 89.9% 80.8% 81.4%
a
Albuminuria ¼ all proteinuria including microalbuminuria.
b
Proteinuria ¼ macroproteinuria.

CKD within Salford according to the K/DOQI Creatinine >120 mmol/l had a superior predictive value
classification 27.5% (n ¼ 1715) of the population to detect subjects with an eGFR <60 ml/min/1.73 m2 in
had an eGFR <60 ml/min/1.73 m2 (stage 3–5 CKD); the presence of albuminuria than alone; however, it still
of these 19.4% (n ¼ 333) had normoalbuminuria; failed to identify all subjects with eGFR <60 ml/min/
20.4% (n ¼ 350) had albuminuria, the remainder not 1.73 m2. This analysis was also performed for patients
having had albuminuria determined. Serum creatinine attending primary care (low capture of albuminuria
was normal (
120 mmol/l) in 54.7% (n ¼ 938) of those quantification) and secondary care (high capture of
with eGFR <60 ml/min/1.73 m2 i.e. moderate to severe albuminuria quantification) producing comparable
CKD and
150 mmol/l in 82.2% (n ¼ 1409). The pre- sensitivities and specificities (data not shown).
valence of eGFR <60 ml/min/1.73 m2 was 16% in Unidentified CKD, defined as the presence of an
people <70 years old and 49% if 70 years old. eGFR <60 ml/min/1.73 m2 but without any evidence
An increased risk of CKD (eGFR <60 ml/min/ of an abnormal creatinine (i.e. serum creatinine
1.73 m2) was seen with: female sex (adjusted OR 2.11,
120 mmol/l) was significantly greater in females com-
CI 1.87–2.40); older age (adjusted OR per year 1.09, CI pared with males adjusting for age, type of diabetes and
1.08–1.09) duration of diabetes (adjusted OR per year secondary care setting (OR 8.22, CI 6.56–10.29). Using
1.03, CI 1.02–1.04) and follow-up in a secondary care albuminuria as a screening test also failed to identify
setting (adjusted OR 1.55, CI 1.34–1.78). The type of CKD in females (OR 2.22, CI 1.63–3.03). The presence
diabetes was not associated with an increased incidence of abnormal serum creatinine and albuminuria to
of CKD on multivariate logistic regression. identify CKD continued to display a significant bias
Patients with clinically significant CKD (eGFR against females (OR 7.58, CI 5.44–10.57).
<60 ml/min/1.73 m2) had better glycaemic control
(HbA1c 7.8% vs 7.9%; P<0.001), better cholesterol
(total cholesterol 4.6 mmol/l vs 4.7 mmol/l; P ¼ 0.015), Discussion
higher systolic blood pressure (BP 140 mmHg vs
137mmHg; P<0.001) and lower diastolic blood pres- This study shows that 27.5% of the population with
sure (BP 72 mmHg vs 75 mmHg; P<0.001). diabetes have clinically significant CKD, as defined
Table 3 shows the performance characteristics of by an eGFR <60 ml/min/1.73 m2. Current screening
an abnormal serum creatinine (>120 mmol/l) and techniques based upon albuminuria and/or abnormal
albuminuria to detect an eGFR <60 ml/min/1.73 m2. serum creatinine would fail to detect a significant
The unrecognised prevalence of chronic kidney disease in diabetes 91
2
number of subjects with an eGFR <60 ml/min/1.73 m . determine that CKD is biased against females; this may
The sensitivity of abnormal serum creatinine levels in translate to late diagnosis of kidney disease, late presen-
identifying eGFR <60 ml/min/1.73 m2 is 45.3%, albu- tation to renal services, and subsequently crash landing
minuria is 51.2% and either an abnormal serum onto dialysis. The combination of ACR and eGFR
creatinine or albuminuria is 82.4%. In this study 60.6% would lead to earlier detection of kidney disease in
have CKD stage 3 with a normal serum creatinine. diabetes and eliminate this gender bias. Early detection
Therefore, without eGFR reporting the clinician may of kidney disease enables intensive blood pressure con-
not be alerted to the presence of CKD and be falsely trol, institution of an angiotensin converting enzyme
reassured that renal function is normal. This study inhibitor or angiotensin 2 receptor blocker and meti-
suggests that current screening for CKD could be culous control of glycaemia; factors which have all been
improved by incorporating eGFR reporting. shown to slow progression of CKD in diabetes [12].
Serum creatinine is the most commonly recognized Whilst the patients with CKD had statistically better
marker of renal function; it is quick and easy to glycaemic and cholesterol control, in reality the clinical
perform and inexpensive. Creatinine is produced by significance of this is negligible, with the difference in
skeletal muscle and its plasma concentration is there- HbA1c being 0.2% and cholesterol being 0.1 mmol/l.

Downloaded from https://academic.oup.com/ndt/article/21/1/88/1819015 by guest on 04 October 2021

fore proportional to muscle mass. Formulae such as the It is, however, worth noting that the metabolic control
Cockcroft–Gault equation and the 4-variable MDRD achieved in both groups was sub-optimal with respect
equation attempt to correct for factors affecting the to glycaemic control and blood pressure targets, whilst
muscle mass, such as age, body size, gender and race. acceptable for total cholesterol. This emphasizes the
Microalbuminuria assessment by an albumin creati- need for improvement in diabetes management in all
nine ratio (ACR) remains an essential component of patients and especially in those with CKD [3].
diabetes care as an indicator of the development Late referral to renal services may have serious
of diabetic nephropathy and the rate of progression consequences for the individual patient, having been
of chronic kidney disease. This study highlights the shown to be associated with greater use of temporary
practical difficulty in quantifying albuminuria in dialysis access, increased hospitalization, mortality and
population based diabetes care. Albuminuria was access to transplantation waiting lists [13,14]. In the
determined in only 39.8% of subjects with an eGFR United States third National Health and Nutrition
<60 ml/min/1.73 m2 over the 2-year period of our study Examination survey (NHANES III), less than 10% of
despite current recommendations in the UK for annual subjects with CKD stage 3 had any awareness of having
screening. A greater proportion of subjects (70%) ‘weak or failing kidneys’ with women having less
receiving diabetes management in a secondary care awareness than men, which may reflect the bias in CKD
setting had albuminuria quantified and this figure is detection using serum creatinine [15]. Mortality is high
comparable with other diabetes centres. However, this in diabetic kidney disease. According to the UKPDS
leaves a significant number failing to have any measure- study group, patients with macroalbuminuria have a
ment [10]. When the sensitivity and specificity of greater annual risk of dying than of progression of
albuminuria for detecting CKD was calculated for nephropathy or development of ESRD (4.6% vs 2.3%)
those patients in secondary care, similar results were [16]. Comparable data from the US observed Medicare
obtained, suggesting that the low ascertainment of system demonstrate that patients with diabetes and
albuminuria does not appear to influence the sensitivity CKD are five times more likely to die than develop
or specificity of this test. Furthermore, albuminuria was ESRD in a 2-year follow-up period compared with
absent in 52.6% of subjects with CKD stage 3 in whom those without CKD [17]. Diagnosing CKD is the first
a measurement was obtained, reinforcing the need for step to improving the poor prognosis.
a simple functional measurement of GFR, as provided The 4-variable MDRD formula for estimated GFR
using the eGFR. was used in this study, as opposed to the Cockcroft–
Diabetic nephropathy represents only a proportion Gault formula suggested in the ADA guidelines, as
of the aetiology of chronic kidney disease in diabetes; it relies on age, sex, race and serum creatinine only.
a number of subjects will have ischaemic renal disease Consequently, this formula is quick and easy to
or other renal pathologies [11]. The low positive pre- calculate on all patients using the demographic data
dictive value of albuminuria to detect eGFR <60 ml/ routinely provided when requesting a serum creatinine
min/1.73 m2 is unsurprising given that the presence of measurement and can be automatically reported by
microalbuminuria does not necessarily correlate with clinical biochemistry laboratories.
impaired renal function. Proteinuria, a marker of overt The 4-variable MDRD formula was derived from
nephropathy, had an improved positive and negative isotope GFR measurements in 1628 patients with CKD.
predictive value in detecting eGFR <60 ml/min/1.73 m2 The equation has been validated on large populations
compared with albuminuria, however, with a greater with CKD, including subjects with diabetic nephro-
number of false negative results. pathy, however, it may underestimate GFR in subjects
In our study population, the prevalence of CKD is with normal renal function or GFR 60 ml/min/
higher in females and older age groups as well as in 1.73 m2 [6,18]. The Cockcroft–Gault formula requires
patients receiving their diabetes management in a sec- the knowledge of the patient’s weight and there-
ondary care setting. This study shows that the current fore is a less convenient method. Furthermore the
practice of using serum creatinine and albuminuria to Cockcroft–Gault formula has also been shown to be
92 R. J. Middleton et al.
less accurate than the MDRD formula in CKD patients 2. The UK Renal Registry. The Sixth Annual Report 2003.
with diabetes [6]. The MDRD formula is widely used in 3. National Institute for Clinical Excellence. Management of
type 2 diabetes: Renal disease–prevention and early manage-
population-based studies and the recently published
ment. 2002.
National Service Framework for Renal Services in the 4. Scottish Intercollegiate Guidelines Network. Management of
UK supports the routine use of eGFR in clinical Diabetes: a national clinical guideline. 2001.
practice [19]. 5. Perrone RD, Madias NE, Levey AS. Serum creatinine as
This is an observational study of diabetes care within an index of renal function: new insights into old concepts.
Salford, therefore as one would expect the entire local Clin Chem 1992; 38: 1933–1953
population with diabetes did not undergo serum 6. Poggio ED, Wang X, Greene T, Van Lente F, Hall PM.
Performance of the modification of diet in renal disease and
creatinine measurements and other relevant investiga- the Cockroft–Gault equations in the estimation of GFR in
tions. Consequently, the exact prevalence of CKD in health and in chronic kidney disease. J Am Soc Nephrol 2005;
diabetes may differ slightly from our estimates, how- 16: 459–466
ever, at least 23% of the population would have CKD 7. American Diabetes Association. Standards of medical care in
stages 3–5 (eGFR <60 ml/min/1.73 m2). We have not diabetes. Diabetes Care 2004; 27 [Suppl 1]: S15–S21
attempted to determine the number of subjects who 8. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu C-Y.

Downloaded from https://academic.oup.com/ndt/article/21/1/88/1819015 by guest on 04 October 2021

Chronic kidney disease and the risk of death, cardiovascular
may have had transient deterioration in kidney func- events, and hospitalization. New Engl J Med 2004; 351:
tion rather than chronic kidney disease and the kinetics 1296–1305
and clinical significance of this warrant further 9. Levey AS, Greene T, Kusek JW, Beck GJ. MDRD Study
investigation. In addition, in line with the observational Group. A simplified equation to predict glomerular filtration
nature of this study, not all the serum creatinine and rate from serum creatinine [Abstract]. J Am Soc Nephrol 2000;
albuminuria tests will have been concurrent, allowing 11: A0828
10. Craig KJ, Donovan K, Munnery M, Owens DR, Williams JD,
a potential source of error. The number of subjects who
Phillips AO. Identification and management of diabetic
have had their albuminuria successfully treated with nephropathy in the diabetes clinic. Diabetes Care 2003; 26:
angiotensin converting enzyme inhibitors or angio- 1806–1811
tensin 2 receptor blockers were not examined and 11. Keller CK, Bergis KH, Fliser D, Ritz E. Renal findings in
potentially could impact on our results. patients with short-term type 2 diabetes. J Am Soc Nephrol
Population prevalence studies in the United States 1996; 7: 2627–2635
have reported the prevalence of stage 3–5 CKD to be 12. Gæde P, Vedel P, Parving HH, Pederson O. Intensified
multifactorial intervention in patients with type 2 diabetes
15.1% in subjects with diabetes which is significantly mellitus and microalbuminuria: the Steno type 2 randomised
higher than those without diabetes (3.9%), but less study. Lancet 1999; 353: 617–622
than our reported prevalence [20]. The difference in 13. Cass A, Cunningham J, Snelling P, Ayanian JZ. Late referral
reported prevalence may be due to differences in case- to a nephrologist reduces access to transplantation. Am J
mix or differences in creatinine assays and calibration. Kidney Dis 2003; 42: 1043–1049
The diabetes and hospital electronic patient registers 14. Roderick P, Jones C, Drey N et al. Late referral for
end-stage renal disease: a region-wide survey in the
in Salford are well established and validated to capture south west of England. Nephrol Dial Transplant 2002; 17:
all subjects with diabetes in the local population, 1252–1259
including investigations and consultations initiated out- 15. Coresh J, Byrd-Holt D, Astor BC et al. Chronic kidney disease
side of the diabetes care setting. It provides a true picture awareness, prevalence, and trends among U.S. adults, 1999 to
of current practice in screening for CKD in diabetes. 2000. J Am Soc Nephrol 2005; 16: 180–188
This study demonstrates the burden of chronic 16. Adler AI, Stevens RJ, Manley SE, Bilous RW,
Cull CA, Holman RR. Development and progression of
kidney disease in a large population-based cohort
nephropathy in type 2 diabetes: the United Kingdom
with more than 25% having moderate to severe CKD Prospective Diabetes Study (UKPDS 64). Kidney Int 2003;
(eGFR <60 ml/min/1.73 m2). Current recommenda- 63: 225–232
tions for screening using serum creatinine and presence 17. Collins AJ, Li S, Gilbertson DT, Liu J, Chen SC,
of albuminuria fail to identify subjects with moderate Herzog CA. Chronic kidney disease and cardiovascular
to severe CKD, and in particular discriminate against disease in the Medicare population. Kidney Int Suppl 2003;
women and older age groups. Incorporating eGFR S24–S31
18. Rigalleau V, Lasseur C, Perlemoine C et al. Estimation of
into screening for CKD in diabetes would enable glomerular filtration rate in diabetic subjects: Cockcroft
clinicians to identify individuals with CKD earlier formula or modification of diet in renal disease study equation?
in the natural history of the disease, enabling early Diabetes Care 2005; 28: 838–843
effective treatment to delay the progression of CKD. 19. Department of Health. The National Service Framework for
renal services—Part two: chronic kidney disease, acute renal
failure and end of life care. 2005.
References 20. Coresh J, Astor BC, Greene T, Eknoyan G, Levey AS.
Prevalence of chronic kidney disease and decreased kidney
1. Wild S, Roglic G, Green A, Sicree R, King H. Global function in the adult US population: third national health
prevalence of diabetes: estimates for the year 2000 and and nutrition examination survey. Am J Kidney Dis 2003;
projections for 2030. Diabetes Care 2004; 27: 1047–1053 41: 1–12

Received for publication: 15.4.05

Accepted in revised form: 24.8.05