Clin Rheumatol (2007) 26:1134–1135 DOI 10.1007/s10067-006-0445-5
BRIEF REPORT
Meta-analysis is no substitute for a comprehensive national registry Magdalena Dziadzio & Roy Smith
Received: 27 June 2006 / Accepted: 11 September 2006 / Published online: 18 November 2006 # Clinical Rheumatology 2006
Abstract Anti-tumor necrosis factor alpha (TNF-α) agents have become an established treatment option for rheumatoid arthritis (RA), but are not without risks. As TNF-α has a role in tumour surveillance, anti-TNF-α blockade could potentially increase the risk of malignancy. Recent meta-analysis by Bongartz et al. (JAMA 295:2275–2285, 2006) reported an increase in the incidence of malignancy attributed to anti-TNF-α antibodies, and the information has quickly and uncritically reached several secondary sources with huge effects on public perception of risks related to anti-TNF-α therapy. In contrast, the results from the British Society for Rheumatology Biologics Register show that in clinical practice, anti-TNF-α therapy in RA does not appear to increase the risk of malignancy in those patients with low risk of malignancy. We discuss the issues emerging from these published data and suggest caution against giving weight to meta-analysis of short-term drug studies. Keywords Malignancy . Meta-analysis . National registry . Safety data . TNF-α blockers Anti-tumor necrosis factor alpha (TNF-α) agents have become an established treatment option for rheumatoid arthritis (RA), but are not without risks. As TNF-α has a role in tumour surveillance, anti-TNF-α blockade could potentially increase the risk of malignancy.
M. Dziadzio (*) Arthritis Centre, Northwick Park Hospital, Watford Road, HA1 3UJ Middlesex, London, UK e-mail:
[email protected] R. Smith Medical Physics, Royal Free Hospital, London, UK
This was reported recently by Bongartz et al. [1]. The authors performed a meta-analysis of nine clinical trials selected out of 144 potentially relevant publications of antiTNF-α antibodies (infliximab and adalimumab) in RA patients. The study showed an increased risk of infection (odds ratio, OR: 2.0), which is in keeping with the recent report from the German biologic register [2]. However, an increase in the incidence of malignancy attributed to antiTNF-α antibodies was also reported (overall pooled OR 3.3, CI 1.2–9.1; high-dose anti-TNF-α, OR 4.3; low-dose anti-TNF-α, OR 1.2). We are critical of the study. First, the meta-analysis did not include etanercept (a human soluble TNF-α receptor– IgG1 fusion protein). Although both infliximab and adalimumab are monoclonal antibodies, their dosages, modes, frequency of administration, and need for concurrent methotrexate differ. The rationale for pooling results for infliximab and adalimumab while excluding etanercept seems to weaken rather than strengthen the study. Secondly, follow-up was limited to the duration of the trials (maximum duration was 54 weeks). Indeed, six lymphomas were noted in adalimumab patients after their respective trial periods, but were not included in the meta-analysis. Thirdly, the authors mention that their results differ from those of Askling et al. [3], but do not discuss these differences. In that population-based Swedish study, no increased risk of malignancy in 4,160 RA patients treated with anti-TNF-α agents (mean follow-up of 2.3 years) was found. Fourthly, Bongartz et al. [1] chose as an example of etanercept’s link with malignancies a study of patients with Wegener’s granulomatosis treated with either cyclophosphamide alone or with etanercept and cyclophosphamide [4]. In this paper solid cancers were reported in 6 out of 89 patients treated with two drugs, but no cancers were observed in the cyclophosphamide-only arm. The work of
Clin Rheumatol (2007) 26:1134–1135
Fleischmann et al. [5], where the incidence of overall malignancies including lymphoma was comparable with that expected in the RA population and was found stable over time, was not cited in the paper by Bongartz et al. Finally, a published correction to the authors’ disclosure section [1] has led the editors of JAMA to request that the Mayo Clinic College of Medicine conduct an investigation. The British Society for Rheumatology Biologics Register (BSRBR) was established in 2001 to document long-term safety of biologic agents in rheumatic conditions. Recent results from the BSRBR presented at the BSR 2006 meeting [6] show that in clinical practice anti-TNF-α therapy in RA does not appear to increase the risk of malignancy in those patients with low risk of malignancy [6, 7]. The authors compared cancer incidence in 9,999 first exposure, antiTNF-α treated RA patients with 1,877 biologic naïve patients taking disease-modifying antirheumatic drugs (DMARD). They state that the incidence of new malignancy was not increased compared to the DMARD cohort (incidence rate ratio, IRR 0.7). However, there was an increased risk of developing a further malignancy in those patients who had a previous malignancy before starting anti-TNF-α therapy (IRR 2.5), and the authors recommend extreme caution in the use of these drugs in patients with prior malignancy. The current published data does not yet differentiate between the licensed anti-TNF-α agents, and we look forward to the time when there is sufficient data to allow such results to be adduced. Already Bongartz et al.’s [1] headline finding of an increase in the incidence of malignancy attributed to antiTNF-α antibodies has been widely reported in secondary sources such as the BMJ [8], the Nursing Times [9], Reuters Health [10] and medpagetoday.com [11]. Inconsistency and errors have appeared. For example “Rare risks tied to antiTNF α agents for RA” [11] might suggest that the results applied to all three agents currently licensed. Meanwhile, “Drugs for arthritis triple cancer risk” [9] is marvelously vague, but, appearing as it does in a nursing journal, may lead to further concern within health professional groups. One online source has inflated an OR from 4.3 to 44.3 [10]. Patients are becoming more informed and take an active and informed involvement in therapeutic choices. They are able to seek information from the Internet but are more likely to stop at summaries and other secondary sources, leaving analysis of primary sources to clinicians. This can clearly make consultation more complex and frontline clinicians must be up-to-date and able to discuss, and where necessary,
1135
refute evidence patients present them with so that inflated concerns and misapprehensions are addressed. The speed with which the paper of Bongartz et al. [1] uncritically reached secondary sources is instructive. Such secondary sources may alter public perception of the risk of anti-TNF-α agents: we believe this would be unmerited in the light of the findings of the BSRBR and cause unwarranted problems in clinical practice. Furthermore, this emphasizes the importance of practitioners reading entire papers and not relying on summaries in secondary sources. We caution against giving weight to meta-analysis of short-term drug studies when longer follow-up is held by registers like the BSRBR [7] or the German Biologics Register [2].
References 1. Bongartz T, Sutton AJ, Sweeting MJ, Buchan I, Mattison EL, Montori V (2006) Anti-TNF antibody therapy in rheumatoid arthritis and the risk of serious infections and malignancies. JAMA 295:2275–85 (correction JAMA 295:E1) 2. Listing J, Strangfeld A, Kary S et al (2005) Infections in patients with RA treated with biologic agents. Arthritis Rheum 52:3403– 3412 3. Askling J, Fored CM, Brandt L et al (2005) Risks of solid cancers in patients with rheumatoid arthritis and after treatment with tumour necrosis factor antagonists. Ann Rheum Dis 64:1421– 1426 4. Wegener’s Granulomatosis Etanercept Trial (WGET) Research Group (2005) Etanercept plus standard therapy for Wegener’s granulomatosis. N Engl J Med 352:351–361 5. Fleischmann R, Baumgartner SW, Weisman MH, Liu T, White B, Peloso P (2006) Long term safety of etanercept in elderly subjects with rheumatic diseases. Ann Rheum Dis 65:379–384 6. Watson KD, Dixon WG, Hyrich KL, Lunt M, Symmons DPM, Silman AJ (2006) Influence of anti-TNF therapy and previous malignancy of cancer incidence in patients with rheumatoid arthritis: results from the BSR biologics register. Rheumatology (Oxford) 45(Suppl 1):i10 7. Symmons DP, Silman AJ (2006) The world of biologics. Lupus 15:122–126 8. Tonks A (2006) Meta-analysis finds rare but serious side effects. BMJ 332:1264 9. In Brief (2006) Drugs for arthritis triple cancer risk. Nurs Times 102(21):8 10. Reuters Health Information (2006) Anti–TNF antibody therapy for RA increases cancer and infection risk. Available at http:// www.medscape.com/viewarticle/532660 Accessed on 29 May 2006 11. Groch J (2006) Rare risks tied to anti-TNF agents for RA. Medpage Today. Available at http://www.medpagetoday.com/ Rheumatology/GeneralRheumatology/tb/3308 Accessed on 29 May 2006
