Skip to main content

Safety Of non-Steroidal anti-inflammatory drugs

Final Report Summary - SOS (Safety of non-steroidal anti-inflammatory drugs)

NSAIDs are widely used drugs for treatment of pain, fever and inflammatory diseases. However, traditional non-steroidal anti-inflammatory drugs (tNSAIDs) have been associated with a three- to five-fold increased risk in serious upper gastrointestinal (GI) complications such as bleedings, obstructions and perforation, which is related to the inhibition of the enzyme cycloogenase I (COX). After recognition of this mechanism a new class of NSAIDs was developed which did not inhibit COX-I but more selectively COX-II which meant they kept the beneficial effects but less gastro-intestinal toxicity. In 2004, however, one of the most widely used COX-II inhibitors (rofecoxib) was withdrawn due to its association with an increased risk of cardiovascular (CV) events. In consequence, two comprehensive review processes were initiated by the European Medicines Agency (EMA). As a result, in the European Union (EU) the use of coxibs has been contraindicated in patients with established coronary heart disease, cerebrovascular disease and peripheral arterial disease and a number of warning statements concerning CV, GI and skin toxicity have been introduced in the coxib product information. For the tNSAIDs, information was lacking on the cardiovascular effects which stimulated the request from EMA to ask the European Commission (EC) to fund research into the evaluation of the gastro-intestinal and cardiovascular effects of traditional NSAIDs and COX-II selective NSAIDs (coxibs). This was the main objective of the SOS project.

In the SOS project, we have used various resources for the assessment of the GI and CV effects of tNSAIDs and coxibs.

The meta-analyses were published and showed that clinical trials were mostly not large enough to demonstrate NSAID GI and CV safety, especially not prior to the COXIB era, and not at all in children. The meta-analyses from observational studies yielded estimates for GI, acute myocardial infarction and stroke but none for heart failure. Information on dose and duration effects was missing and several methodological gaps were identified, as well as a lack of data in children.

The SOS healthcare database study allowed us to study the association between 13 individual and all study outcomes through a nested case control study in a cohort of 8.5 million new NSAID users, while addressing many of the gaps that had been identified in the literature. Extensive efforts were put in harmonising data from the different databases and utilising common definitions, codes, analyses and protocols. Analyses were distributed across partners by using a central datawarehouse with secure access. All SOS data were communicated to EMA to allow for regulatory decision making. This was done by a predefined SOS communication process that allows pre-publication sharing of data.

A decision model for choosing the least GI/CV toxic NSAID was developed for clinicians, based on the empiric evidence in the consortium and data from the literature.

SOS was the first project that addressed an EMA defined drug safety issues and has broken grounds and has set the stage because of its scale, the type of distributed collaboration across databases and sites and the feedback and communication with EMA.