Module 3 - Strategic case studies in practice

medications may be categorized as non-exposed. Misclassification of exposure may bias the results leading to an underestimation or overestimation of the risk. Therefore, the use of dispensing data is preferred to the use of prescription data, as in the latter case it is uncertain whether the prescription was also filled. However, it should be considered that even if it is known that a prescription has been filled, information is missing whether the patient has actually taken the drug or not. In addition, the impact of individual contaminated batches is hardly observable, as information on dispensed batch numbers is not routinely collected in most healthcare databases. If possible, stratified analyses should be performed according to predefined categories of cumulative exposure to examine a potential dose-response relationship. The cumulative amount of applied contaminated medication and the treatment duration should be considered to check the impact of both long-term exposure and treatment intensity. However, in case of slightly increased risks and/or low precision of the point estimates, or settings where only minimum thresholds need to be reached to observe risks, a dose-response relationship may not be observable. In order to get an impression of the clinical relevance of the observed risk estimates, the population attributable risk (i.e. an estimate of the excess risk) should be calculated to determine the proportion of cancer cases attributable to N- nitrosamines contamination. Finally, sensitivity analyses are recommended to check the robustness of the primary findings. Further epidemiological studies to assess the association between the intake of potentially N- nitrosamines-contaminated drugs and risk of cancer are desirable, but their conduct is challenging. This is mainly due to the difficulty of reliably determining the exposure. Irrespective of the study design applied, potential data sources should contain sufficient patient numbers, should cover a sufficiently long observation time and should contain the required variables to answer the study question, i.e. information on exposure, outcome and important covariates. Prospective study designs using primary data collection are obviously not feasible for the examination of the association of N- nitrosamines-contaminated drugs and cancer risk. Therefore, using nationwide registries or large healthcare database might be the most promising approach for the conduct of further studies, despite the mentioned limitations. If a single data source does not contain all the necessary information, data linkage to other data sources that may contain the missing information should be checked prior to study initiation. Furthermore, the possibility of a meta-analytical approach may be considered in case of insufficient patient numbers in a given data source. 2.7. Relevance for the CHMP opinion on medicinal products containing sartans with a tetrazole ring As a final outcome of the Art 31 referral on medicinal products containing sartans with tetrazole ring, CHMP required the following actions: 1. Obligatory risk assessments to be performed for manufacturing processes of the drug substances in order to evaluate the theoretical risk of N-nitrosamine formation and contamination 2. Implement a control strategy to detect and control N -nitrosamine impurities in the API. Specifically, CHMP considered that NDMA and NDEA long-term limits in the API should be as low as technically possible. In this regard, a limit of quantification of 0.03 ppm for NDMA and NDEA was considered achievable according to the available data on analytical methods. Limits for NDMA and NDEA for the API based on AI calculated according to ICH M7 were considered acceptable for a transitional period of 2 years. Thereafter, a technical limit for NDMA and NDEA of 0.03 ppm should be implemented.

EMA/369136/2020

Page 58/90