International regulatory activity restricting COX-2 inhibitor use and deaths due to gastrointestinal haemorrhage and myocardial infarction

Purpose To investigate trends in mortality rates due to both myocardial infarction and gastrointestinal haemorrhage before and after rofecoxib withdrawal and the release of regulatory guidance regarding the use of other COX-2 inhibitors. Methods International ecological study of temporal trends in deaths from myocardial infarction and gastrointestinal haemorrhage around 2004 when regulatory activity restricted the use of COX-2 inhibitors. Mortality data in countries with low child and adult male mortality (WHO mortality stratum A) were analysed. Results Comparing, on a country-by-country basis, post-2004 mortality rates with those expected from a continuation of preceding trends, there was no evidence of a deviation from the earlier trends in mortality from gastrointestinal haemorrhage or acute myocardial infarction in 50-69 year olds. Amongst 70+ year olds however, there was evidence of lower gastrointestinal haemorrhage mortality (rate ratio 0.963, 95% confidence interval 0.948 to 0.977) and of lower acute myocardial infarction mortality (rate ratio 0.981, 95% confidence interval 0.977 to 0.986) after 2004. These associations were similar for males and females. Conclusions We did not find evidence of an increase in mortality due to gastrointestinal haemorrhage following the withdrawal of rofecoxib in 2004, and coincident concern amongst regulatory bodies about other COX-2 inhibitors. In fact in men and women aged 70 years or older there appeared to be reduced mortality due to gastrointestinal haemorrhage and acute myocardial infarction compared to what was expected from mortality trends before 2004. Copyright (C) 2010 John Wiley & Sons, Ltd.