Monitoring drug safety with registries: useful components of postmarketing pharmacovigilance systems

Journal of Clinical Epidemiology 65 (2012) 121e125

COMMENTARY

Monitoring drug safety with registries: useful components of postmarketing pharmacovigilance systems Cameron D. Willisa,*, John J. McNeilb, Peter A. Cameronc, Louise E. Phillipsc a

School of Population Health and Clinical Practice, Mail Drop DX 650 550, The University of Adelaide, Adelaide, South Australia 5005, Australia b Department of Epidemiology and Preventive Medicine, School of Public Health and Preventive Medicine, Monash University, Australia c Department of Epidemiology and Preventive Medicine, School of Public Health and Preventive Medicine, Monash University, Australia Accepted 21 June 2011; Published online 8 October 2011

Abstract Objective: At the time of licensing by regulatory agencies, the full range of risks and possible adverse drug reactions associated with a medication is rarely fully realized. This commentary aims to describe the role of registries as useful components of postmarketing pharmacovigilance systems for monitoring highly specialized medications associated with significant financial costs. Study Design and Setting: We consider the limitations of traditional pharmacovigilance programs and discuss the strengths, limitations, and uses of registries in postmarketing pharmacovigilance systems. Results: Registries have become increasingly appealing in postmarketing surveillance of medications; however, their exact role continues to evolve. Key registry projects, including the Prospective Immunogenicity Surveillance Registry, British Society for Rheumatology Biologics Register, Australian Rheumatology Association Database, the Haemostasis Registry, and the Bosentan Patient Registry highlight the value of registries for monitoring the incidence of rare adverse events. Conclusion: Although often limited by lack of a control group and the need for complete case ascertainment to maintain data integrity, registries are a useful component of postmarketing pharmacovigilance systems for monitoring highly specialized medications associated with significant financial costs. Ó 2012 Elsevier Inc. All rights reserved. Keywords: Registries; Product surveillance; Postmarketing; Adverse drug reaction reporting systems; Safety; Safety management

1. Introduction At the time of licensing by regulatory bodies, such as the Medicines and Healthcare products Regulatory Agency, not all adverse events or risks associated with a medication may have been identified. Therefore, postmarketing surveillance is essential for providing data on drug safety through monitoring the use of a medication, the types of patients receiving treatment, the outcomes of treatment, and the occurrence of adverse drug reactions (ADRs). The ‘‘science and activities relating to the detection, assessment, understanding and prevention of adverse events or any other drug-related problem’’ has been termed ‘‘pharmacovigilance’’ [1]. In September 2008, the European Medicines Agency (EMA) introduced specific pharmacovigilance guidelines that supersede many of the provisions contained

Conflicts of interest: There are no conflicts of interest to declare. * Corresponding author. Tel.: þ61-8-8313-0612; fax: þ61-8-83036899. E-mail address: [email protected] (C.D. Willis). 0895-4356/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. doi: 10.1016/j.jclinepi.2011.06.017

in the United Kingdom’s Medication Act (1968). These guidelines highlight the importance of traditional postmarketing surveillance methods, such as the Yellow Card scheme and encourage the introduction of systematically collected data using registries. Although disease or drug exposure registries may have an important function in postmarketing surveillance, their precise role in pharmacovigilance remains unclear. 1.1. Traditional pharmacovigilance methods Postmarketing drug surveillance has historically relied on data from clinical trials or epidemiological studies. Postmarketing randomized controlled trials (RCTs), such as the Adenomatous Polyp Prevention on Vioxx (APPROVe) trial that prompted the global withdrawal of Vioxx (rofecoxib) [2], are expensive and are not routinely powered to detect rare ADRs, those with a long latency or events occurring in specific (and often excluded) patient groups [3e6]. As such, epidemiological methods, such as spontaneous reporting systems (SRS), have long been recognized as useful

122

C.D. Willis et al. / Journal of Clinical Epidemiology 65 (2012) 121e125

What is new? For postmarketing surveillance of pharmaceuticals, registries offer a more representative picture of the range of patients receiving a drug, their additional medications, and existing medical conditions than is contained in other clinical investigations. Registries are useful for providing reliable estimates of the incidence of adverse events across defined populations. However, as registries do not routinely include control groups of similar patients not receiving the medication, they are not able to address questions of relative efficacy or treatment effect and cannot serve as replacements for clinical trials. Therefore, because of their high operational costs and the routine requirement for external support, registries are of most value for the surveillance of highly specialized medications that are used in significant clinical contexts.

tools in drug-safety monitoring. The Yellow Card scheme adopted in the United Kingdom more than 40 years ago has become an important detection system for prescribers, and now patients, to voluntarily report ADRs. Spontaneous reporting has proven useful for capturing ADRs occurring in the initial period after treatment; however, this method does not define the population from which reports arise (the ‘‘denominator’’), leading to poor estimation of the incidence of ADRs especially in long-term drug use. Despite previous success of SRS schemes in identifying rare ADRs, some have questioned the future legitimacy of this method, citing its low position in the hierarchy of evidence and persisting problems of underreporting [3,7]. 1.2. A role for registries in pharmacovigilance A clinical registry is a register of patients in a particular clinical setting, recording treatment and outcomes of all patients within the institution/units involved and collecting information on variables that facilitates risk adjustment. Registries have become increasingly appealing in postmarketing surveillance of medications; however, their exact role continues to evolve. Registries are of most value for providing drug safety data through monitoring the incidence of rare adverse events particularly for high cost and/or highly specialized drugs. In the United Kingdom and Europe, for licensing new medications, pediatric administrations, or when changes are made to a drug’s manufacturing process, the new EMA guidelines now require detailed risk management plans, in which clinical registries may find greater inclusion as tools to detect ADRs. The changes made to the EMA guidelines provide a mechanism for the detection of rare adverse events after even

subtle modifications in the manufacturing of medications or biological agents. A manufacturing change that replaced human serum albumin with polysorbate 80 and glycine in the formulation of Eprex (epoetin alphada medication licensed for use in patients with chronic renal failure) was reported to be associated with an increased incidence of pure red cell aplasia (PRCA): an extremely rare adverse outcome undetected by premarketing studies [8]. Whereas findings from spontaneous reports were used to estimate the incidence of PRCA in patients receiving Eprex, the need to prospectively study the rate of PRCA in chronic renal failure patients treated with erythropoiesis-stimulating agents, such as Eprex, has resulted in the establishment of the Prospective Immunogenicity Surveillance (PRIMS) registry [9]. Under the new EMA guidelines, registries, such as PRIMS may be useful ADR detection tools in situations where RCTs have not identified rare or latent ADRs, such as PRCA in this population. As such, the PRIMS registry provides a novel method for recruiting and following a large cohort of patients (a total of 20,000 person years exposure to Eprex) receiving erythropoiesis-stimulating agents to continually monitor the safety of this agent in real world populations. In postmarketing surveillance settings, registries offer a number of advantages over other approaches for investigating the safety of therapeutics. In contrast to the tightly controlled patient selection approach adopted by clinical trials, registries provide a more representative picture of the range of patients receiving a drug, their additional medications, and existing medical conditions [10,11]. For monitoring the safety of medications used in the treatment of Rheumatoid arthritis (RA), a number of registries have been established that aim to provide data across the full range of medicated patients. Two such registriesdthe British Society for Rheumatology Biologics Register (BSRBR) and the Australian Rheumatology Association Database (ARAD)dwere established to monitor the use and safety of therapeutics in RA patients. Launched in 2001, the BSRBR has provided important data on the long-term safety of antitumor necrosis factor therapies and their relationships with serious infections, myocardial infarction, and congenital malformations in pregnant patients [12]. The ARAD was established to obtain more representative data on the use of biological disease-modifying antirheumatic drugs (DMARDS) in Australia and provide better estimates of the incidence of ADRs in RA patients treated with DMARDS [13]. This national registry has become a useful tool for monitoring quality of life, health care utilization, and the impact of therapy switching in a clinically representative cohort of patients, of whom just 5% meet the inclusion criteria of published RCTs [14]. In acute care pharmacovigilance, registries have been developed to monitor the safety of highly specialized medications. The Haemostasis Registry was initiated to study the use of Recombinant Activated Factor VII (rFVIIa, Novo Nordisk Pharmaceuticals Pty. Ltd.) in off-licence settings in Australia and New Zealand, with a focus on detecting the

C.D. Willis et al. / Journal of Clinical Epidemiology 65 (2012) 121e125

incidence of thromboembolic adverse events (TAEs). This medication, approved for use in patients with Hemophilia A and B with inhibitors, has gained widespread use in other settings of critical bleeding, such as trauma and cardiac and liver surgery [15e17]. As the Haemostasis Registry includes all patients receiving off-licence rFVIIa at participating hospitals, reliable estimates of the incidence of TAEs across this defined population can be made. Recently, a study from the Registry assessed the adverse outcomes of Registry cardiac surgery patients in comparison with cardiac surgery patients included in the Australasian Society for Cardiothoracic Surgery (ASCTS) Cardiac Surgery Database and found no statistically significant difference in the rate of TAEs [18]. As a component of pharmacovigilance systems, clinical registries offer a means of providing credible, reliable, and epidemiologically sound data elements. The collection of a core set of data items is done through a standardized process, allowing reliable comparison of patients and outcomes treated at different centers and over time. These standardized data are critical for ensuring registry reliability [19]; however, a narrow scope of data collection may result in missed ADRs attributable to unanticipated drug reactions. Although a registry cannot collect all information relating to a patient’s treatment, the reliable and standardized collection of predefined data items allows registries to provide a sufficient level of risk adjustment. Registries’ methods are therefore distinct from other databases’ methods that have been used for pharmacovigilance, such as data-linkage techniques between population level administrative database (including health insurance databases) and prescribing records used to examine the use of medications in pregnancy [20]. Improvements in data-linkage techniques through the development of unique identifiers, coupled with greater access to routine data, have the capacity to add a cost-effective surveillance method to future pharmacovigilance initiatives. Interest in registries as postmarketing surveillance tools for specialized or high-cost pharmaceuticals continues to gain attention from manufacturers, academics, regulatory authorities, and subsidizing schemes, such as the Pharmaceutical Benefits Scheme (PBS) in Australia. The introduction of the Bosentan Patient Registry (BPR) demonstrates an innovative risk-sharing arrangement between a manufacturer and the government [21]. In 2004, Bosentan was listed by the PBS for the treatment of patients with idiopathic pulmonary arterial hypertension (PAH): a disease considered to carry a poor prognosis [22]. The subsidy provided by the PBS for Bosentan was based on the improvements in life expectancy resulting from treatment; however, longterm clinical trials had failed to confirm this finding. As a result, the BPR was established to collect outcome data from all patients receiving Bosentan in Australia to assess the accuracy of internationally conducted clinical trials. In this setting, licensing of Bosentan for the treatment of PAH in Australia was dependent on the establishment and maintenance of the BPR.

123

The experiences of PRIMS, Haemostasis Registry, BSRBR, ARAD, and BPR highlight the use and potential role of large scale, prospectively collected, epidemiologically sound data sources in pharmacovigilance systems. However, registries are not without limitations. A number of pregnancy registries have been established, such as the Lamotrigine (LTG) Pregnancy Registry, to investigate the use of anticonvulsant medications in pregnant women because of the potential for drug exposure early in pregnancy, the possible link with congenital malformations, and the risk of teratogenicity associated with polytherapy [23]. Data from the registry appear to indicate that the risks associated with LTG are similar to those in the general population; however, more definitive conclusions will only be possible with greater case numbers [24]. Criticisms have been made that such data collections are prone to selection bias when based on volunteering women, resulting in exclusion of the most at-risk patient groups and therefore serious underestimation of ADR rates [25]. Therefore, minimizing patient dropout (particularly of certain subgroups, such as those with greater illness severity) and ensuring data completeness are critical for enabling registries to detect ADRs with long latency periods [10]. ‘‘Opt-off’’ consent procedures, whereby patient data are included on the registry unless a patient lodges an objection, have been recommended as one method for ensuring complete case ascertainment and minimizing selection bias [26]. Collection of additional variables beyond core parameters adds considerable costs to the financial inputs required to maintain standardized data collection, training of data collectors, access to medical records, complete case ascertainment, and database expenses associated with registries. For the Haemostasis Registry, the BSRBR, and the BPR, funding for all operations has been provided by the drug manufacturers whereas the management of the registries and the publication of results are the responsibility of independent academic institutions. Similar to other registries, such as the Consortium of Rheumatology Researchers of North America, the independence of these registries from industry influence is essential in providing objective and unbiased research findings [27]. In Europe, this distinction is even clearer, with pharmaceutical companies mandated to support postmarketing independent data collections [27]. Because of the significant costs associated with registries and the common need for external support, registries are of most value for the postmarketing surveillance of expensive or highly specialized medications in highly significant clinical contexts. Unlike RCTs or case-control studies, drug registries are often limited to analyses of specific agents or medications, such as rFVIIa or Bosentan. As a result, registries do not routinely include control groups of similar patients not receiving the medication and are therefore not able to address questions of relative efficacy or treatment effect [28]. Registries, such as the ARAD and the BSRBR have taken steps to recruit control samples of patients whereas other

124

C.D. Willis et al. / Journal of Clinical Epidemiology 65 (2012) 121e125

registries have initiated data-linkage efforts with other existing sources, such as the Haemostasis Registry’s link with ASCTS. Health insurance databases offer another useful resource for reliably capturing large numbers of patients, monitoring the course of disease, assessing drug utilization, and following patient groups over long time horizons [11]. Because of a lack of clinical detail in these databases, linkage to clinical registries is becoming increasingly useful, facilitating a linked-data source which minimizes selection bias, allows identification of a control group, and contains detailed clinical data for sufficient risk adjustment. The questions that may be investigated using registry data are therefore often restricted to those concerned with rates and patterns of potentially linked ADRs and events that may not occur immediately after drug administration. Therefore, registries are not designed to provide conclusive evidence of treatment efficacy or effectiveness and cannot serve as replacements for clinical trials. As tools for monitoring the safety and usage of pharmaceuticals and other therapeutic agents, registries are superior to all other postmarketing surveillance methods for identifying patterns of use and outcomes as well as applications that may benefit from further exploration and possible clinical trials.

2. Conclusion At the time of licensing by regulatory agencies, the full range of risks and possible ADRs associated with a medication is rarely fully realized. Postmarketing surveillance of pharmaceuticals is therefore essential for monitoring the use of medications and identifying rates of adverse reactions across actual populations. A postmarketing pharmacovigilance framework needs to identify true signals of ADRs using a range of interventional and observational approaches that broaden the scope of premarketing clinical trials that are often restricted by stringent inclusion criteria. Established methods, such as SARs, continue to be valuable for all medications in flagging novel cases worthy of review, whereas systematic, prospective data collections that are capable of tracking medication use and patient outcomes as well as providing key covariates to enable risk adjustment are useful tools for amplifying signal detection in highly specialized medications. Although often limited by a lack of a control group and the reliance on complete case ascertainment to maintain data integrity, registries (and their linkage to other data sources) are a useful component of postmarketing pharmacovigilance systems for monitoring highly specialized medications associated with significant financial costs.

Acknowledgments This work was supported by an unrestricted educational grant from Novo Nordisk Pharmaceuticals Pty. Ltd.

References [1] World Health Organisation Collaborating Centre for International Drug Monitoring. The importance of pharmacovigilance: safety and monitoring of medicinal products. Uppsala: WHO; 2002. Contract no.: ISBN 9241590157. [2] Bresalier R, Sandler R, Quan H, Bolognese J, Oxenius B, Horgan K, et al. Cardiovascular events associated with rofecoxib in a colorectal adenoma chemoprevention trial. New Engl J Med 2005;352: 1092e102. [3] Aagaard L, Holme Hansen E. Information about ADRs explored by pharmacovigilance approaches: a qualitative review of studies on antibiotics, SSRIs and NSAIDs. BMC Clin Pharmacol 2009;9:4. [4] Ray W. Population-based studies of adverse drug effects. New Engl J Med 2003;349:1592e4. [5] Hrmark L, van Grootheest AC. Pharmacovigilance: methods, recent developments and future perspectives. Eur J Clin Pharmacol 2008;64:743e52. [6] Luo X, Cappelleri J, Frush K. A systematic review on the application of pharmacoepidemiology in assessing prescription drugrelated adverse events in pediatrics. Curr Med Res Opin 2007;23:1015e24. [7] Waller P, Evans SJW. A model for the future conduct of pharmacovigilance. Pharmacoepidemiol Drug Saf 2003;12:17e29. [8] Schellekens H, Jiskoot W. Eprex-associated pure red cell aplasia and leachates. Nat Biotechnol 2006;24:613e4. [9] Sharma B, Ryan M, Boven K. Reactions to Eprex’s adverse reactions. Nat Biotechnol 2006;24:1199e200. [author reply 1200]. [10] Schneeweiss S, Avorn J. A review of uses of health care utilization databases for epidemiologic research on therapeutics. J Clin Epidemiol 2005 Apr;58:323e37. [11] Smeets HM, de Wit NJ, Hoes AW. Routine health insurance data for scientific research: potential and limitations of the Agis Health Database. J Clin Epidemiol 2011;64:424e30. [12] Hyrich KL, Watson KD, Isenberg DA, Symmons DPM. The British Society for Rheumatology Biologics Register: 6 years on. Rheumatology 2008;47:1441e3. [13] Buchbinder R, March L, Lassere M, Briggs AM, Portek I, Reid C, et al. Effect of treatment with biological agents for arthritis in Australia: the Australian Rheumatology Association Database. Intern Med J 2007;37:591e600. [14] Sokka T, Pincus T. Eligibility of patients in routine care for major clinical trials of anti-tumor necrosis factor alpha agents in rheumatoid arthritis. Arthritis Rheum 2003;48:313e8. [15] Carr M, Martin E. Recombinant Factor VIIa: clinical applications for an intravenous hemostatic agent with broad-spectrum potential. Expert Rev Cardiovasc Ther 2004;2:661e74. [16] Willis C, Bird R, Mullany D, Cameron P, Phillips L. Use of rFVIIa for critical bleeding in cardiac surgery: dose variation and patient outcomes. Vox Sang 2010;98:531e7. [17] Willis CD, Cameron PA, Phillips L. Variation in the use of recombinant activated factor VII in critical bleeding. Intern Med J 2010;40:486e93. [18] Mitra B, Phillips L, Cameron PA, Billah B, Reid C. The safety of recombinant factor VIIa in cardiac surgery. Anaesth Intensive Care 2010;38:671e7. [19] Willis CD, Jolley DJ, McNeil JJ, Cameron PA, Phillips LE. Identifying and improving unreliable items in registries through data auditing. Int J Qual Health Care 2011;23:317e23. [20] Colvin L, Slack-Smith L, Stanley F, Bower C. Pharmacovigilance in pregnancy using population-based linked datasets. Pharmacoepidemiol Drug Saf 2009;18:211e25. [21] Owen A, Spinks J, Meehan A, Robb T, Hardy M, Kwasha D, et al. A new model to evaluate the long-term cost effectiveness of orphan and highly specialised drugs following listing on the Australian Pharmaceutical Benefits Scheme: the Bosentan Patient Registry. J Moral Educ 2008;11:235e43.

C.D. Willis et al. / Journal of Clinical Epidemiology 65 (2012) 121e125 [22] Department of Health and Ageing. December 2003 PBAC outcomes e positive recommendations. Canberra: Pharmaceutical Benefits Scheme, Commonwealth of Australia; 2004. [23] GlaxoSmithKline. The Lamotrigine Pregnancy Registry interim report: 1 September 1992 through 31 March 2007. Wilmington, NC: 2007. [24] Cunnington M. The International Lamotrigine pregnancy registry update for the epilepsy foundation. Epilepsia 2004;45:1468. [25] Briggs G, Polifka J. Better data needed from pregnancy registries. Birth Defects Research Part A. Clin Mol Teratol 2009;85: 109e11.

125

[26] Australian Commission on Safety and Quality in Healthcare. Operating principles and technical standards for Australian clinical quality registries. Canberra: Australian Commission on Safety and Quality in Healthcare; 2008. Contract no: TRIM Record 20308. [27] Kremer J, Gibofsky A, Greenberg J. The role of drug and disease registries in rheumatic disease epidemiology. Curr Opin Rheumatol 2008;20:123e30. [28] O’Connor R, Gilmore A, Manjunath R, Stanford R, Legorreta A, Jhingran P. Comparing outcomes in patients with persistent asthma: a registry of two therapeutic alternatives. Curr Med Res Opin 2006;22:453e61.