Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

NSAIDs and COXIBs: overall risk


Many folk are spending many hours trying to come up with an equation allowing the overall calculation of gastrointestinal benefit and cardiovascular risk of NSAIDs and coxibs for particular patients. This is something of a forlorn hope given the complexities and arguments over what constitutes the best evidence. A large observational study [1] provides a little more meat.


This retrospective cohort study was based on prescriptions and hospital admission, covering persons aged 65 years or older in Quebec. Those filling a prescription for a coxib or nonselective (ns)-NSAID over 2.5 years ending in 2002 were included in the cohort if they had one year of prescription data and complete coverage for inpatient services. Exclusions were those with prescriptions for both classes of drugs.

The main outcomes were acute myocardial infarction and gastrointestinal bleeding, obtained from hospital discharge summaries using coded information. Calculations were based on the hazard ratio for drugs with or without low dose aspirin compared with paracetamol without aspirin. Crude results were adjusted in the statistical analysis.


The study included 511,000 patients, with 112,000 receiving aspirin. About 22% of patients not receiving aspirin were using gastroprotective agents at some time, with slightly more among aspirin users, but at the time of an event gastroprotective agents were used by about 10% or less.

For both gastrointestinal bleeding and acute myocardial infarction celecoxib was the drug with the lowest hazard ratio. Results for the combined outcome are shown in Figure 1. Without aspirin, naproxen and rofecoxib were associated with significantly greater risk than paracetamol without aspirin. With aspirin, diclofenac and naproxen showed no significantly increased risk.

Figure 1: Combined gastrointestinal bleeding and myocardial infarction rates compared with paracetamol without aspirin


What the analyses show is that no drug had significantly increased risk of myocardial infraction compared with paracetamol without aspirin. With aspirin any increased risk had bare statistical significance with low absolute risk increases. Gastrointestinal risk was significantly increased with most drugs except celecoxib, ibuprofen and diclofenac without aspirin. In patients with osteoarthritis, results were generally similar if numerically somewhat higher.

In the overall risk analysis the gastrointestinal risk predominated, in accord with what we know from other observational studies and randomised trials, though, as usual, we have no information about drug doses. The good thing about these observational studies is that they look at patients like ours. The bad thing is that they cannot provide the whole answer because there can be large differences in patients treated with different drugs, and because there can be confounding factors of which we know little.


  1. E Rahme, H Nedjar. Risks and benefits of COX-2 inhibitors vs non-selective NSAIDs: does their cardiovascular risk exceed their gastrointestinal benefit? A retrospective cohort study. Rheumatology 2007 doi:10.1093/rheumatology/kel428

previous or next story