Lupus (2006) 15, 122–126

www.lupus-journal.com

REVIEW

The world of biologics

DPM Symmons* and AJ Silman
arc Epidemiology Unit, University of Manchester, Manchester, UK

In March 2002 the National Institute for Health and Clinical Excellence (NICE) published guidelines
for the use of anti-TNF therapy for patients with rheumatoid arthritis (RA). The guidelines
recommended that all RA patients treated with these drugs should be enrolled on a national register
which had been established by the British Society for Rheumatology (the BSRBR). A comparison
cohort of RA patients treated with traditional disease modifying drugs (DMARDs) is also being
recruited. The main role of the BSRBR is to study the long-term safety of biologic drugs. Up to the
end of March 2005, 9508 patients with RA had been enrolled on the BSRBR. Four thousand, three-
hundred and six had been treated with etanercept, 3561 with infliximab, 1500 with adalimumab and
141 with anakinra. With regards to anti-TNF drugs, 79% remained on their original drug at six
months, 65% of whom could be classified as responders. Co-prescription with methotrexate was
associated with a 70% response rate. Patients with a high baseline level of disability were less likely
to respond. Overall the rates of serious infection were not increased in the anti-TNF versus the
comparison cohort. However the rates of skin and soft tissue infection and of intracellular infections
(eg, salmonella, listeria, legionella) were increased. There were 11 cases of tuberculosis (seven extra-
pulmonary). There was concern about the high mortality rates among patients with baseline
pulmonary fibrosis treated with anti-TNF therapy. It is unclear whether this is related to the drug or
to the underlying disease. The rates of malignancy and mortality were not increased compared to the
DMARD treated group in the short term. Further follow-up is needed to determine the long term
safety of these drugs. Lupus (2006) 15, 122–126.

Key words: anti-TNF therapy; infection; mortality; pharmacovigilance; rheumatoid arthritis

This paper concerns the impact which anti-TNF drugs (at present Abbott Laboratories, Amgen, Schering
therapy has made in the management of patients with Plough and Wyeth Laboratories). Rheumatologists
rheumatoid arthritis (RA). The paper is written from an throughout the UK are requested to register all patients
epidemiological perspective and the data presented with rheumatic diseases who start treatment on a bio-
here are largely based on results emerging from the logic agent for the first time. After providing informed
British Society for Rheumatology Biologics Register consent, the patient is registered. Both the consultant
(BSRBR) up to March 2005. We describe the way the and the patient complete questionnaires which cover
register works, information we have learned about the information on demographic, lifestyle and co-morbid-
efficacy of anti-TNF therapy in routine clinical use and ity factors. In addition, information is collected about
early information on adverse events, in particular the disease activity at the time the biologic agent was
infection and lymphoma. started. For patients with RA, the disease activity
The BSRBR was established at the University of measure used is the DAS-28.1 In March 2002, the
Manchester in 2001. It is funded via a grant to National Institute for Clinical Excellence (NICE)
the British Society for Rheumatology (BSR) by the published guidelines for the use of anti-TNF therapy in
pharmaceutical companies who manufacture biologic patients with RA.2 Enrolling patients on the BSR
Biologics Register was one of the recommendations
made by NICE. The patients enrolled on the register
are followed every six months for the first three years.
Questionnaires are sent both to the consultant and to
*Correspondence: Deborah PM Symmons, Professor of Rheumatology and
Musculoskeletal Epidemiology, arc Epidemiology Unit, Stopford Building, the patient at the end of each six-month period. They
University of Manchester, Oxford Road, Manchester, M13 9PT, UK. seek information about disease activity, whether the
E-mail: [email protected] drug has been discontinued and whether the patient has
© 2006 Edward Arnold (Publishers) Ltd 10.1191/0961203306lu2278rr
Downloaded from lup.sagepub.com at University of Manchester Library on April 13, 2015
 The world of biologics
DPM Symmons and AJ Silman
123
experienced any serious adverse events. Consultant We have examined a number of demographic and
questionnaires only are then sent annually for the fourth disease specific factors to try and identify predictors of
and fifth years. In addition, all patients are flagged response.4 Age, gender and co-morbidity have no influ-
with the Office for National Statistics for England and ence on response rates to either etanercept or infliximab.
Wales and General Register Office for Scotland. This Smoking has no influence on response rate to etanercept
means that the register is automatically informed of but, in patients treated with infliximab, smokers are
any malignancies which develop in this cohort and will much less likely to respond than non-smokers (adjusted
also be notified when each patient dies and provided odds ratio (OR) ⫽ 0.77, 95% CI 0.60, 0.99). Among the
with a copy of the death certificate. disease specific factors, disease duration and rheuma-
Up to the end of March 2005, 9508 patients had toid factor status did not influence response to either
been enrolled on the BSRBR who had RA. Of these etanercept or infliximab. Patients with a high baseline
4306 (44%) had been treated with etanercept, 3561 had disease activity were more likely to respond to either
been treated with infliximab (38%), 1500 had been etanercept (adjusted OR ⫽ 1.47 (95% CI 1.25, 1.71) or
treated with adilimumab (16%) and 141 had been infliximab (adjusted OR 1.20 (95% CI 1.05, 1.38) than
treated with anakinra (an IL1 receptor antagonist) (2%). those with lower activity. Conversely, patients with a
HAQ score ⬎2.0 were less likely to respond to
etanercept (adjusted OR ⫽ 0.48 95% CI 0.34, 0.68) or
Efficacy infliximab (adjusted OR ⫽ 0.73 95% CI 0.55, 0.96)
than patients who had a lower HAQ.
Response in the patients with RA is defined according
to the EULAR response criteria (see Figure 1).3
After six months, 79% of patients who have been Adverse events
started on an anti-TNF therapy, remain on their origi-
nal drug. By 12 months, this has fallen to 70% and by It is important to remember that anti-TNF therapy is
18 months to 61%. At six months, 16% are good currently being used in patients with very severe RA.
responders, 49% moderate responders and 35% poor This cohort of patients will have a high background
responders. The response rate is similar between the risk of premature mortality and of co-morbidity.
three main anti-TNF drugs. Co-prescription with Mortality studies of RA in UK populations over the
methotrexate improves the overall response (moderate last four decades consistently show a higher mortality
plus good) rate to 70%. Use of another DMARD in than the general population5–8 with standardized
place of methotrexate is also of benefit. The patients mortality ratios (SMRS) varying from 1.4 to 2.7. The
with the poorest response rate are those in whom no majority of excess deaths are due to cardiovascular
other DMARD is prescribed together with their anti- disease but deaths from respiratory disease (predomi-
TNF therapy. nantly infections) and, in some series, deaths from
malignancy are also increased.9 A number of hospital
linkage studies have shown that the overall incidence
of cancer in patients hospitalized with RA is not
increased.10
However, studies consistently find an increased risk
of non-Hodgkin’s lymphoma (NHL) and of lung
cancer amongst patients with RA. Conversely, the risk
of colorectal cancer is reduced. It is the excess risk of
NHL which is of most interest, since this is a malig-
nancy which may occur with increased frequency in
patients treated with biologic agents. Work from
Sweden, however, has shown that the risk of develop-
ing NHL in patients with RA is related to cumulative
disease activity.11 Patients with the highest cumulative
disease activity had an standardised incidence ratio
(SIR) of 25.8 (95% CI 3.1, 213.0) compared to those
with the lowest cumulative disease activity.
Interestingly, steroid use in this cohort reduced the risk
of NHL by around 50% – suggesting that treatment
Figure 1 Adapted from van Gestel et al; Arthritis Rheum 1998; may actually be beneficial. Work from the Mayo Clinic
41: 1845–1850. has shown that patients with RA have an increased risk
Lupus
Downloaded from lup.sagepub.com at University of Manchester Library on April 13, 2015
 The world of biologics
DPM Symmons and AJ Silman
124
of being hospitalized with infection compared to Anti-TNF therapy and infection
the remainder of the population who do not have
RA.12 The risk of being hospitalized with respiratory TNF␣ plays a number of important roles in the control
infections is increased by 88% Rate ratio 1.88; of infection in the normal individual. For example,
95% CI 1.41, 2.53. The rate ratio for skin and soft TNF␣ is involved in macrophage activation and
tissue inflection was 3.28 (95% confidence intervals endothelial cell activation and cell recruitment, and in
2.67, 4.07). the control of granulomas and latent infection. In addi-
Because of the known high levels of background co- tion, TNF␣ acts as a pro-coagulant which limits the
morbidity and mortality in patients with severe RA, it spread of infection. The three TNF␣ inhibitors fall into
was felt important to recruit a comparison cohort for two distinct drug groups: monoclonal antibodies
the BSR Biologics Register. The comparison cohort (infliximab and adalimumab) and a soluble TNF␣
comprises patients with active RA who have not been receptor (etanercept). Both of these drug groups bind
treated with biologic agents. The aim of recruiting the to soluble TNF␣ (although the monoclonal antibodies
comparison cohort was to provide background data on do so with greater affinity). However, the monoclonal
long-term safety such as malignancy and mortality. antibodies (but not soluble TNF␣ receptor), are able
By the end of March 2005, 8455 patients with RA to bind to transmembrane TNF␣. Conversely, soluble
treated with anti-TNF had been recruited by the TNF␣ receptors, but not the monoclonal antibodies,
BSRBR and 1199 controls. Patients treated with bio- have the capacity to bind lymphotoxin ␣ (also known
logic agents were younger, had higher disease activity, as TNF␤), a cytokine which has a role in the control
a higher previous number of DMARDs and higher of intracellular infections such as tuberculosis.
HAQs than patients in the control cohort (Table 1) Finally, the monoclonal antibodies, but not the
Conversely, more of the control cohort were smokers. soluble receptors, are able to direct complement
With regards to comorbidity, there was a higher fre- mediated, or antibody mediated cell lysis of cells
quency of coronary heart disease, pulmonary disease, expressing transmembrane TNF␣. There are, there-
demyelination and previous malignancy in the control fore, reasons to suppose that the different TNF␣
cohort (Table 2). However, there was a higher frequency inhibitors may predispose individuals to different
of pulmonary fibrosis in the anti-TNF treated group. infection profiles.
For the purpose of this analysis, serious infections
were defined as those which resulted in death, hospi-
Table 1 Baseline characteristics of RA patients treated with talization or required the use of intravenous antibiotics.
anti-TNF agents or on the control register Tuberculosis was also defined as a serious inflection. A
Anti-TNF Controls preliminary analysis of the overall rate of serious
Characteristics n ⫽ 8455 n ⫽ 1199 infection was 52.6 per 1000 patient years (pyr) in
patients first treated with etanercept, 52.1 per 1000s
Age, years (mean, SD) 56 (12) 60 (12)
Female (n, %) 6483 (77) 842 (70) pyr in infliximab treated patients and 62.4 in adali-
Current smokers 1676 (21) 253 (26) mumab treated patients.13 The most common site of
Disease duration infection was in the lower respiratory tract, followed
(mean, SD) 13.7 (9.4) 9.1 (9.9)
DAS28 (mean, SD) 6.6 (1.0) 5.1 (1.4) by skin and soft tissue, bone and joint, and then urinary
Previous no. DMARDS tract infections. For all of these sites, there was no dif-
(mean, SD) 4 (2) 2 (2) ference in the rate of infection across the three agents.
HAQ score (mean, SD) 2.1 (0.5) 1.6 (0.8)
The rates of intracellular infection were three per 1000
pyr years. There were nine tuberculosis cases in
patients treated with infliximab and two in patients
Table 2 Baseline co-morbidity (RA) in RA patients treated with treated with etanercept. However, these results are not
anti-TNF or on the control register
statistically significant due to small numbers. As in
Anti-TNF Controls OR other series, there were more cases of extra-pulmonary
Co-morbidity n (%) n (%) (95% CI)*
tuberculosis than would otherwise have been expected.
Any comorbidity 4505 (58) 606 (62) 0.91 (0.79–1.05) In this series of eleven patients, seven were extra pul-
Cardiovascular disease 1964 (25) 299 (31) 0.93 (1.80–1.08) monary – two cervical lymph nodes, one colitis, one
Coronary heart disease 470 (6) 101 (10) 0.77 (0.61–0.98)
Pulmonary disease 1196 (15) 211 (22) 0.70 (0.59–0.83)
omentum, one meningitis and two disseminated. The
Pulmonary fibrosis 249 (3) 17 (2) 2.42 (1.47–4.00) mean time from starting biologic to the diagnosis of
Demyelination 14 (0.2) 5 (0.5) 0.32 (0.11–0.90) tuberculosis was five months. There were also six
Past tuberculosis 152 (2) 21 (2) 1.09 (0.74–1.61)
Past malignancy 231 (3) 47 (5) 0.69 (0.51–0.93)
other cases of intracellular infections – three salmo-
nella (included two salmonella septic arthritis), one
*Adjusted for age, gender and smoking history. mycobacterium fortuitum, one listeria and one
Lupus
Downloaded from lup.sagepub.com at University of Manchester Library on April 13, 2015
 The world of biologics
DPM Symmons and AJ Silman
125
legionella. This raises the question as to whether malignancy in the previous 10 years. It is, therefore,
patients treated with anti-TNF therapy should, for perhaps not surprising, that patients treated with bio-
example, be warned to avoid foods containing raw logic agents have a lower rate of malignancy compared
eggs and soft cheeses made from unpasteurized milk. with the control subjects in the first few months of
When compared to controls, these initial data follow-up. In the BSR Biologics Register, the inci-
suggest that the rates of serious infection were not dence of lymphoma is not currently increased in the
increased in the biologic treated group. There are a biologic treated patients compared to the controls.7
number of important potential confounders that could However, follow-up time is relatively short and we
effect such comparisons which will be considered in estimate that it will require a further three years of
future analyses. The rates of lower respiratory tract follow-up to answer this question robustly.
infection were, in fact, lower in those treated with
biologic agents than in the controls. This is probably
explained by residual confounding as the controls had Biologic agents and mortality
higher rates of smoking and background pulmonary
disease. However, the rates of skin infection (ie, By October 2004, there had been 150 deaths in
cellulitis) were substantially higher in the biologic patients treated with biologic agents. The most
treated group than in the controls. Up to March 2005, common cause of death was circulatory (29%).15
there had been only one control skin infection Compared to the general population, the mortality
requiring hospital admission compared to 156 in rate in patients with RA treated with anti-TNF was
the biologic treated group. This may be related to increased for both men (SMR 1.5 (95% CI 1.0, 2.2))
inhibition by the biologic drugs of the normal TNF␣ and women (SMR 1.4 (95% CI 1.1, 1.8)). Compared
mediated mechanism which limits the spread of to the controls, the mortality rate was not increased.
infection. There was no difference in the relative risk of
mortality in the three main anti-TNF agents. Again
longer follow-up and more attention to the role
Biologic agents and pulmonary fibrosis of confounders will be needed to assess whether
the biologic agents themselves, by suppressing
Early in 2004, we were made aware of a patient with inflammatory activity can have a beneficial effect
RA and mild interstitial lung disease who had been on all cause mortality, or whether this is counter-
treated with a combination of infliximab and azathio- balanced by an increase in severe drug-related
prine. This patient died four months after starting treat- adverse events.
ment with infliximab from rapidly progressive In summary, our early experience in routine clinical
pulmonary fibrosis.14 Review of the register at that practice shows that around two-thirds of patients with
time showed that there had been nine patients with RA RA treated with an anti-TNF therapy experience either
and pulmonary fibrosis who had been treated with a a moderate or a good EULAR response. Response
combination of infliximab plus azathioprine. Three of rates are enhanced by co-prescription of methotrexate.
these had died, all from lung disease. However, it was Patients treated with anti-TNF experience an
clear that patients with pulmonary fibrosis on the increased risk of skin and intracellular infection
register had a substantially higher mortality than compared to other RA patients. Further follow-up
those without pulmonary fibrosis, regardless of time is needed to understand the long-term safety of
whether they were treated with azathioprine or these drugs.
methotrexate as the DMARD combined with inflixi-
mab. The question which could not be answered at
this stage was whether the biologic agents them- Acknowledgements
selves were contributing to this excess mortality or
whether it was simply the pulmonary fibrosis. We acknowledge the contributions of Kath Watson,
Information beginning to accumulate from the controls Kimme Hyrich, Will Dixon, Yvonne King, Lesley
suggests that pulmonary fibrosis per se carries a high Munroe, Katie McGrother and Pat Creighton in running
mortality. and analysing the data from the Biologics Register.
We also acknowledge the continued support of UK
consultant rheumatologists, rheumatologist specialist
Biologic agents and malignancy nurses and patients in providing data to the register. The
BSR Register is funded by Abbot Laboratories,
It is recommended that biologic agents should not AMGEN, Schering-Plough and Wyeth Laboratories via
be used in patients who have been treated for a a grant to the British Society for Rheumatology.
Lupus
Downloaded from lup.sagepub.com at University of Manchester Library on April 13, 2015
 The world of biologics
DPM Symmons and AJ Silman
126
References 8 Chehata JC, Hassell AB, Clarke SA et al. Mortality in rheumatoid
arthritis: relationship to single and composite measures of disease
activity. Rheumatology (Oxford) 2001; 40: 447–452.
1 Prevoo ML, van’t Hof MA, Kuper HH, van Leeuwen MA, van de Putte 9 Symmons DP, Jones MA, Scott DL, Prior P. Longterm mortality
LB, van Riel PL. Modified disease activity scores that include twenty- outcome in patients with rheumatoid arthritis: early presenters continue
eight-joint counts. Development and validation in a prospective longitu- to do well. J Rheumatol 1998; 25: 1072–1077.
dinal study of patients with rheumatoid arthritis. Arthritis Rheum 1995; 10 Symmons DP, Silman AJ. Anti-tumor necrosis factor alpha therapy and
38: 44–48. the risk of lymphoma in rheumatoid arthritis: no clear answer. Arthritis
2 National Institute for Clinical Excellence. Guidance on the use of Rheum 2004; 50: 1703–1706.
etanercept and infliximab for rheumatoid arthritis. NICE Technology 11 Baecklund E, Ekbom A, Sparen P, Feltelius N, Klareskog L. Disease
Appraisal London, NICE 2002; 36. activity and risk of lymphoma in patients with rheumatoid arthritis:
3 van Gestel AM, Haagsma CJ, van Riel PL. Validation of rheumatoid nested case-control study. Brit Med J 1998; 317: 180–181.
arthritis improvement criteria that include simplified joint counts. 12 Doran MF, Crowson CS, Pond GR, O’Fallon WM, Gabriel SE.
Arthritis Rheum 1998; 41: 1845–1850. Frequency of infection in patients with rheumatoid arthritis compared
4 Hyrich H, Watson K, Symmons D, Silman A. Predicting response to with controls: a population-based study. Arthritis Rheum 2002; 46:
anti-TNF alpha therapy among patients with rheumatoid arthritis. 2287–2293.
Rheumatology (Oxford) 2005; 44 (Suppl 1): i24. 13 Dixon W, Hyrich K, Watson K, Silman A, Symmons D. Serious
5 Duthie JJ, Brown PE, Truelove LH, Baragar FD, Lawrie AJ. Course and infection rates in patients receiving biologic therapy in the United
prognosis in rheumatoid arthritis. A further report. Ann Rheum Dis 1964; Kingdom: results from the BSR Biologics Register. Rheumatology
23: 193–204. (Oxford) 2005; 44 (Suppl 1): i11.
6 Lewis P, Hazleman BL, Hanka R, Roberts S. Cause of death in patients 14 Ostor AJ, Crisp AJ, Somerville MF, Scott DG. Fatal exacerbation of
with rheumatoid arthritis with particular reference to azathioprine. Ann rheumatoid arthritis associated fibrosing alveolitis in patients given
Rheum Dis 1980; 39: 457–461. infliximab. Brit Med J 2004; 329: 1266.
7 Prior P, Symmons DP, Hawkins CF, Scott DL, Brown R. 15 Watson K, Hyrich K, Silman A, Symmons D. Mortality among RA
Cancer morbidity in rheumatoid arthritis. Ann Rheum Dis 1984; 43: patients receiving anti-TNF alpha therapy in the United Kingdom: results
128–131. from the BSR Bologics Register. Rheumatolog 2005; 44 (Suppl 1): i23.

Lupus
Downloaded from lup.sagepub.com at University of Manchester Library on April 13, 2015