Pharmacovigilance of biologicals: dynamics in post-approval safety learning

Regulatory decisions to allow new drugs on the market by definition have to accept a certain level of uncertainty about the full benefit-risk balance. Pre-approval studies typically provide information on a limited number of patients over a relative short follow-up period, and handle strict inclusion criteria with respect to patients’ comorbidities and use of concomitant medications. Therefore, uncertainties exist regarding risks, particularly regarding rare and long-term adverse effects, as well as with regard to the benefit-risk profile in specific patient populations. Pharmacovigilance activities are essential in gaining knowledge on previously identified uncertainties, as well as in the detection of unexpected adverse effects after approval. These activities thereby support a continuous assessment of the benefit-risk balance of marketed drugs.

This thesis focuses on the pharmacovigilance of biologicals, a group of drugs that are produced by or extracted from a biological source. Biologicals have over the last decades become increasingly important in the treatment of various serious diseases, including several types of cancer, rheumatoid arthritis, and multiple sclerosis. Biologicals have, however, due to their specific characteristics been associated with unique risks and challenges in pharmacovigilance. These challenges include, among others, the very specific type of safety concerns associated with biologicals, their sensitivity to changes in manufacturing conditions, and the increased number of uncertainties identified for biologicals at approval, as found by previous research.

The studies presented in this thesis provide new insights into the process of knowledge gain on adverse effects and previously identified uncertainties for biologicals after approval. First, it is shown that spontaneous reports for the same suspected adverse drug reaction (ADRs) may differ between two biologicals with distinct regulatory paths and used under diverse clinical conditions, with regard to quality (data completeness) and temporality of reporting (lag time to recognition of the adverse reaction, and time between reaction and reporting). These factors may impact on the timely identification and characterization of (unexpected) adverse effects. Furthermore, it is shown that healthcare professional should remain aware of false positive signals in pharmacovigilance, including those caused by drug interference in clinical laboratory testing.

A unique characteristic of biologicals is that new risks, including previously unobserved ADRs, may emerge at any point in the life-cycle as a result of manufacturing variability within or across products over time. The studies in this thesis show that the necessary product-, and particularly batch-specific, exposure information is not always available in major spontaneous reporting systems. Furthermore, the robustness of pharmacovigilance systems to the potential for exposure misclassification (i.e. attribution to incorrect product) was explored, and recommendations were provided to improve the availability of detailed exposure information, by exploring the individual processes involved in ensuring traceability.

Lastly, this thesis explored the rate at which knowledge gaps in the safety profile that were identified for new drugs at approval (between 2006 and 2009) are filled after approval. Overall, it was found that 20.7% of the uncertainties (potential risks and missing information) documented in the risk management plan at approval had been resolved 5 years after approval. The rate at which uncertainties were resolved did not differ between small-molecule drugs and biologicals.