|Vascular and upper gastrointestinal effects of non-steroidal anti-inflammatory drugs: meta-analyses of individual participant data from randomised trials
|Coxib and Traditional NSAID Trialists' (CNT) Collaboration
This review concluded that the vascular risks with high-dose diclofenac, and possibly ibuprofen, were comparable to those with cyclooxygenase-2 inhibitors, but they were less with high-dose naproxen. The size of these risks could be predicted. The review appears to have been well conducted, and the authors' conclusions are likely to be reliable.
To characterise and quantify the cardiovascular and gastrointestinal risks of nonsteroidal anti-inflammatory drug (NSAID) regimens for different types of patients, particularly those at an increased risk of vascular disease.
MEDLINE and EMBASE were searched to January 2009, without language restrictions; a search strategy was provided in an appendix. Trial registers, such as ClinicalTrials.gov and Clinical Trial Results, were searched and reference lists of reviews were examined; manufacturers of cyclooxygenase-2 inhibitors, and experts on this topic were contacted.
Randomised trials were eligible if they assessed at least four weeks of daily treatment, comparing a NSAID with placebo or another NSAID, or comparing two different NSAID regimens. Trials had to be of adults, aged 18 years or older. The primary vascular outcome was major vascular events, defined as non-fatal myocardial infarction, non-fatal stroke, or death from a vascular cause. The primary gastrointestinal outcome was upper gastrointestinal complications, defined as an upper gastrointestinal perforation, obstruction, or bleed.
In the included trials, the NSAIDs were traditional ones (diclofenac, ibuprofen and naproxen), or cyclooxygenase-2 inhibitors (celecoxib, rofecoxib, lumiracoxib, etoricoxib and valdecoxib); their doses varied. In those trials that provided individual patient data, their mean age was 61 years, around two thirds of them were female, and their mean body mass index was 29kg/m². About one fifth of patients were on aspirin at randomisation, and most patients had rheumatoid arthritis or osteoarthritis.
Two authors selected studies for inclusion.
Assessment of study quality
The methods of randomisation, blinding and publication status were recorded. To be eligible for inclusion trials had to randomise with robust allocation concealment.
The number of reviewers involved in the assessments was not stated.
Individual patient data were requested from trial investigators. Where these were not available, aggregate data were sought from investigators or extracted, by two authors, from the publications. Intention-to-treat data were sought, and the data on the numbers of events and patients were extracted to produce contingency tables.
Methods of synthesis
Direct and indirect rate ratio estimates were calculated. For direct treatment comparisons, pooled rate ratios, with 95% confidence intervals, were calculated using fixed-effect meta-analyses, based on calculating log-rank statistics. The rate ratios, for indirect treatment comparisons, were estimated from the differences between the relevant direct comparisons.
Subgroup analyses were performed to investigate possible heterogeneity, and Bonferroni corrections were made.
Results of the review
Six hundred and thirty-nine trials were included; the total sample included over 300,000 patients.
Compared with placebo, major vascular events were increased by around a third for patients on a cyclooxygenase-2 inhibitor (rate ratio 1.37, 95% CI 1.14 to 1.66) and for those on diclofenac (rate ratio 1.41, 95% CI 1.12 to 1.78). Ibuprofen significantly increased major coronary events (rate ratio 2.22, 95% CI 1.10 to 4.48), but not major vascular events (rate ratio 1.44, 95% CI 0.89 to 2.33). Naproxen did not significantly increase major vascular events (rate ratio 0.93, 95% CI 0.69 to 1.27). The proportional effects on major vascular events were independent of baseline characteristics, including vascular risk.
Deaths due to vascular events were increased significantly by cyclooxygenase-2 inhibitors (rate ratio 1.58, 99% CI 1.00 to 2.49) and diclofenac (rate ratio 1.65, 99% CI 0.95 to 2.85), increased non-significantly by ibuprofen (rate ratio 1.90, 99% CI 0.56 to 6.41), and not increased by naproxen (rate ratio 1.08, 99% CI 0.48 to 2.47).
All NSAID regimens significantly increased upper gastrointestinal complications (mostly bleeds), compared with placebo; cyclooxygenase-2 inhibitors rate ratio 1.81 (95% CI 1.17 to 2.81), diclofenac rate ratio 1.89 (95% CI 1.16 to 3.09), ibuprofen rate ratio 3.97 (95% CI 2.22 to 7.10), and naproxen rate ratio 4.22 (95% CI 2.71 to 6.56).
The risk of hospitalisation, due to heart failure, was approximately doubled by all NSAIDs. Further results were reported.
The vascular risks with high-dose diclofenac, and possibly ibuprofen, were comparable to those with cyclooxygenase-2 inhibitors, but less with high-dose naproxen. The size of these risks could be predicted, which could help guide clinical decision making.
This individual patient data review and meta-analysis addressed a clear question and was supported by reproducible eligibility criteria. Attempts were made to identify all relevant trials in any language. No data validation procedures, nor risk of bias assessment, were reported, but all trials had to have adequate randomisation methods, and many were placebo controlled, so the risks of some biases were low. As the review assessed adverse effects, an evaluation of the attrition rates and reasons would have been helpful.
Individual trial details were not provided, but this was understandable given that there were 639 included trials; summary details were presented. Appropriate methods were used to pool the data and to assess and investigate heterogeneity. Some results were based on indirect comparisons of treatments, which might be less reliable than direct comparisons.
Overall, the review appears to have been well conducted, and the authors' conclusions are likely to be reliable.
Implications of the review for practice and research
Practice: The authors did not state any implications for practice, beyond those stated in their conclusions.
Research: The authors did not state any implications for research.
Funded by the UK Medical Research Council, and the British Heart Foundation.
Coxib and Traditional NSAID Trialists' (CNT) Collaboration. Vascular and upper gastrointestinal effects of non-steroidal anti-inflammatory drugs: meta-analyses of individual participant data from randomised trials. Lancet 2013; 382(9894): 769-779
Subject indexing assigned by CRD
Gastrointestinal Diseases; Gastrointestinal Tract; Anti-Inflammatory Agents, Non-Steroidal; Humans; Peptic Ulcer; Dyspepsia; Vascular Diseases
Date bibliographic record published
Date abstract record published
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.