The risk of coronary events with non-steroidal anti-inflammatory drugs has been the subject of much debate since the original trial of rofecoxib raised the issue. Since then, over almost 20 years, such risks have been shown in clinical trials of long-term high-dose users, and in observational studies comparing users with non-users. The roles of cyclooxygenase (COX)-2/COX-1 selectivity and COX-2 inhibitory potency have been proposed to explain this increased risk of myocardial infarction (MI). Among NSAIDs, diclofenac appeared to be associated with a relatively higher risk of MI, similar to that of rofecoxib, compatible with the drug's high COX-2 inhibitory potency. Recent studies have resulted in further information being available. A study in the Danish healthcare system using active comparators found a slightly increased risk of MI in healthy persons. However, risk decreased with increasing baseline cardiovascular risk, to the point that in patients at high cardiovascular risk, there was no additional risk associated with diclofenac compared with paracetamol or other NSAIDs. The other major study, from the SOS project, studied several million persons in four countries in Europe, comparing the use of many NSAIDs with non-use. That study found a slightly increased risk with diclofenac compared with non-use, but this was not different from other NSAIDs. Comparing risks with selectivity or potency found no effect of either. These studies refute the main hypotheses to explain the coronary risk of NSAIDs. Finding risk in healthy low-risk patients only questions the reality of a link between the use of the drugs and the occurrence of MI in these conditions. Biases or confounding may be the major reason for small increases in cardiovascular risks in healthy users of NSAIDs in real life.