The context of medicines’ use in benefit-risk evaluation

Abstract

In benefit-risk evaluations there is a trend towards a life cycle approach including continuous benefit-risk evaluation instead of a single benefit-risk assessment at a certain (fixed) point in time. The objective of this thesis is to unravel how the context in which a medicine is used adds to the assessment of the benefit-risk profile, and to gain more insight in the value of this information for both drug development and regulatory decision-making. One of the main findings was that it is important to select the appropriate patient population for randomised clinical trials (RCTs), especially if both the pharmacological working mechanism and the expected patient population outside the clinical trial setting indicate that safety issues will focus within a certain area. Otherwise, extrapolation of the results to the setting outside clinical trials will be difficult. In some studies (not well-controlled and/or progressive) diabetes has been associated with an increased risk of infections. It is therefore difficult to study the effect of antidiabetic medicines on the occurrence of infections. We performed two studies that both showed that infections were approximately two times more frequently reported for DPP-4 inhibitors compared to other antidiabetic drugs. These findings, in combination with the biologic plausibility may suggest a potential relation between DPP-4 inhibitors and infections. Several studies showed that psychiatric disease was almost twice as present among patients starting anti-obesity drugs (AODs) than patients not starting these drugs. In addition, differences in general health care utilisation between these groups increased gradually over the 3-year period before start of AODs but no specific point in time could be identified from where differences between began to appear. We found that the duration of anti-obesity drug use was limited. One study, for example, showed that patients with a history of psychiatric illness were at increased risk of early discontinuation of rimonabant therapy, most pronounced due to psychiatric events. This implies that patients discontinue treatment because of psychiatric events before the possible cardiovascular benefits could develop, thereby negatively affecting the benefit-risk profile of rimonabant. These findings urge us to be very careful in interpreting the benefits and risks of anti-obesity drugs, both in terms of preventing possible exposure of drugs associated with psychiatric events in susceptible patients and in the evaluation of causality when a possible drug induced problem occurs. Finally we showed that differences in health status exist between users of the two strengths of the anti-obesity drug orlistat which are available through different channels (available on prescription vs. available without prescription, only in pharmacies). This information illustrates that user information from observational studies is also valuable for benefit-risk evaluations of one active component that is being used in two different settings. We concluded that for an adequate benefit-risk evaluation of a medicinal product, information on the complete context in which medicines are being used is necessary. This consists of extensive information on the patient population in which medicines are being used, usage patterns including the duration of use, and the effect of the regulated availability of medicines.