INT J TUBERC LUNG DIS 17(12):1590–1595 © 2013 The Union http://dx.doi.org/10.5588/ijtld.13.0368
Risk of tuberculosis in rheumatoid arthritis patients on tumour necrosis factor-alpha inhibitor treatment in Taiwan W-M. Ke,* L-S. Chen,† I-M. Parng,† W-W. Chen,* A. W. F. On* * Division of Drug Safety, Taiwan Drug Relief Foundation, Taipei, † Office for HealthCare Quality Measurement, Department of Health, Taipei, Taiwan SUMMARY OBJECTIVES:
To quantify the incidence of tuberculosis (TB) in rheumatoid arthritis patients undergoing treatment with tumour necrosis factor-alpha inhibitors (TNFi). D E S I G N : In a retrospective cohort study conducted using data from Taiwan’s National Health Insurance claims databases, rheumatoid arthritis patients notified during the period 2006–2008 were recruited and classified based on types of TNFi treatment received. Active TB was the primary outcome. TB risk was estimated using Cox’s proportional hazard model. The TB screening rate within 30 days of initiating treatment with TNFi was examined. R E S U LT S : Respectively 5079 and 829 patients were included in the non-TNFi and TNFi groups. Active TB rates were respectively 1411.3 and 679.5 events per
100 000 person-years in patients treated with adalimumab and etanercept. Significant TB risk was noted in patients treated with TNFi (aHR 4.87, 95%CI 2.14– 11.06). No significant difference in active TB was observed between the TNFi subgroups (etanercept as reference, aHR 1.89, 95%CI 0.40–6.04). Only 8.7% (n = 9) of TNFi users underwent screening for TB before the first dose of TNFi. C O N C L U S I O N S : Patients on TNFi have a significantly greater risk of active TB than non-TNFi patients in the Taiwanese population. No difference in TB risk between the two available TNFi groups was noted. Screening for TB before initiating treatment with TNFi should be implemented. K E Y W O R D S : TB screening; pharmacovigilance; claims database
TUBERCULOSIS (TB), a disease caused by the Mycobacterium tuberculosis bacillus, has posed a significant burden on and threat to public health in human history. More than 9.4 million incident TB cases were estimated worldwide in 2009, with most cases occurring in Asia and Africa.1 In Taiwan, around 15 000 new cases have been identified annually in recent years. In 2007, Taiwan had much higher incidence and prevalence rates (respectively 63 and 111 per 100 000 population)2 than countries in Europe and America.3 Various conditions, such as recent tuberculous infection, human immunodeficiency virus (HIV) infection, drug use, immunosuppressive treatment and malnutrition, might facilitate the development of active TB in infected patients.1 Tumour necrosis factor-alpha inhibitors (TNFi) are immunomodulators, and include the soluble receptor, etanercept (Enbrel®; Amgen, Thousand Oaks, CA, USA) and the monoclonal antibody, adalimumab (Humira®; AbbVie Inc, North Chicago, IL, USA). Etanercept and adalimumab were approved by the US Food and Drug Administration in 1998 and 2002 respectively, for the treatment of several rheumatic diseases.4 However, several safety issues have been
linked to TNFi treatment, such as risk of active TB or TB reactivation,5 serious infections and neoplasms.6,7 Although the association between TNFi and risk of TB has been confirmed in several observational studies,8 registries9–14 and meta-analyses,15,16 most of these studies were conducted in Europe, which has a lower TB prevalence than other regions. Active TB rates in adalimumab and etanercept users were respectively 270 and 19 per 100 000 person-years (py) in clinical trial settings, and were reported to be 141–500 and 9.3–233/100 000 py in post-marketing observational studies.10,11,14,17 In the Western Pacific Region, a higher background TB prevalence was observed, along with higher TNFi-associated TB incidence.18 The Taiwan National Adverse Drug Reaction Reporting Center receives approximately 10 000 reports of adverse drug reactions from health care professionals and market authorisation holders annually. Etanercept and adalimumab have been available in Taiwan since respectively 2003 and 2004. On detecting an abnormal surge in the reporting rate of TNFi-associated TB in 2011, we were interested in ascertaining whether the risk profile for TB had changed in Taiwanese society. To better understand
Correspondence to: Wei-Ming Ke, Division of Drug Safety, Taiwan Drug Relief Foundation, Taipei, Taiwan. Tel: (+886) 2 2358 7343, extn 146. Fax: (+886) 2 2358 4100. e-mail:
[email protected] Article submitted 22 May 2013. Final version accepted 20 July 2013.
TB among TNF-𝛂 users in Taiwan
this issue, we conducted a population-based cohort study to examine the relationship between the use of TNFi treatment and TB risk among rheumatoid arthritis patients in Taiwan.
METHODS Data source The Taiwan National Health Insurance (NHI) claims databases and death certification data file were used to conduct this study. The NHI system is a mandatory single-payer social health insurance system covering 99% of the Taiwanese population and 92.7% of the hospitals and clinics in Taiwan.19 NHI claims data consist of reimbursement records of in-patient, out-patient and pharmacy services, professional details of hospital and health care and catastrophic illness registry data sets. Catastrophic illness registries include 30 illnesses or injuries, such as cancer, endstage renal disease, chronic psychotic disorder and rheumatoid arthritis, which pose a significant burden on beneficiaries. All disease-associated copayments can be waived for registered patients; all diagnosisrelated documentation must be submitted for review by the Bureau of National Health Insurance on application of certification. Laboratory results and socioeconomic information are not recorded. Diagnoses are recorded using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes for a maximum of respectively three and five diagnoses in each out-patient and inpatient claim. Pre-defined reimbursement codes for specific chronic illness are also used as diagnosis indicators. Privacy related information is encrypted in all data sets. Study design and population This is a retrospective cohort study that recruited rheumatoid arthritis patients (ICD-9-CM code 714.0) who received catastrophic illness certification during 2006–2008 in Taiwan. The date of the first prescription of non-biological disease-modifying antirheumatic drugs (DMARDs), including methotrexate, hydroxychloroquine, sulfasalazine, leflunomide, azathioprine, d-penicillamine and minocycline, was considered as the index day. Baseline medical conditions were determined by tracking all reimbursement records from 1 January 2003 up to the index day. Patients commencing DMARDs therapy before 1 July 2005, with no diagnosis of rheumatoid arthritis or DMARDs prescription throughout the observation period, age >110 years or
