A comparison of the cost-effectiveness of five strategies for the prevention of non-steroidal anti-inflammatory drug-induced gastrointestinal toxicity: a systematic review with economic modelling.

OBJECTIVES: To assess the relative effectiveness, patient acceptability, costs and cost-effectiveness of four strategies for the prevention of non-steroidal anti-inflammatory drugs (NSAIDs)-induced gastrointestinal (GI) toxicity: (1) Cox-1 NSAIDs plus histamine-2 receptor antagonist (H2RA), (2) Cox-1 NSAIDs plus proton pump inhibitors (PPIs), (3) Cox-1 NSAIDs plus misoprostol, and (4) Cox-2 NSAIDs (later expanded to 4a Cox-2 coxibNSAIDs and 4b Cox-2 preferential NSAIDs). DATA SOURCES: Electronic databases up to May 2002. REVIEW METHODS: Relevant studies were selected, assessed and analysed. Pooled relative risk ratios (RR) from the systematic review were combined with up-to-date UK resource use and unit costs data in an incremental economic analysis. A probabilistic decision-analytic model was designed and populated with data to carry out incremental economic analysis. Incremental cost-effectiveness ratios (ICERs) were generated for the outcome measure, endoscopic ulcer or serious GI event averted, against total cost, and non-parametric bootstrapping was used to simulate variance of these ICERs. RESULTS: Of 118 selected trials, including 125 relevant comparisons (which included 76,322 participants) only 138 deaths and 248 serious GI events were reported. Seven comparisons were judged to be at low risk of bias. Comparing the gastroprotective strategies against placebo, there was no evidence of effectiveness of H2RAs against any primary outcomes (few events reported), PPIs may reduce the risk of symptomatic ulcers [RR 0.09, 95% confidence interval (CI) 0.02 to 0.47], misoprostol reduces the risk of serious GI complications (RR 0.57, 95% CI 0.36 to 0.91) and symptomatic ulcers (RR 0.36, 95% CI 0.20 to 0.67), Cox-2 'preferentials' reduce the risk of symptomatic ulcers (RR 0.41, 95% CI 0.26 to 0.65) and Cox-2 'coxibs' reduce the risk of symptomatic ulcers (RR 0.49, 95% CI 0.38 to 0.62) and possibly serious GI events (RR 0.55, 95% CI 0.38 to 0.80). All strategies except Cox-2 'preferentials' reduce the risk of endoscopic ulcers. There were only 12 direct comparisons between gastroprotective strategies. All they suggest is that Cox-2 preferentials are better than misoprostol for preventing GI complications. Indirect comparisons suggested that PPIs may prevent symptomatic ulcers better than Cox-2 coxibs, but this is very weak evidence. For prevention of endoscopic ulcers PPIs and misoprostol appear more successful than H2RAs and misoprostol is better than Cox-2 preferentials. There were no UK head-to-head published economic analyses with regard to the main gastroprotective strategies. There were generally insufficient data with regards to cardiac or renal outcomes, serious GI outcomes or life-years gained to populate the mode. Mean (2.5th and 97.5th percentile) costs per endoscopic ulcer averted compared with Cox-1 NSAIDs alone were as follows: Cox-1 plus H2RAs, -186 pounds (-555 to 804); Cox-1 plus PPIs, 454 pounds (251 to 877); Cox-1 plus misoprostol, 54 pounds (-112 to 238); Cox-2 selective NSAIDs, 263 pounds (-570 to 1280), or Cox-2 specific NSAIDs, 301 pounds (189 to 418). With regard to the prevention of endoscopic ulcers, Cox-1 NSAID plus H2RA is a dominant option. Cost-effectiveness acceptability analysis showed a 95% probability that this combination was less costly and more effective. Cost-effectiveness acceptability frontiers showed that if the decision-maker is willing to pay up to 750 pounds to avoid an endoscopic ulcer, then Cox-1 plus H2RA is the optimal strategy. If the decision-maker is willing to pay over 750 pounds, the optimal strategy is NSAID plus misoprostol. Between 1900 pounds and 3750 pounds, Cox-2 selective inhibitors are optimal, and over 3750 pounds, Cox-2 specific inhibitors become optimal. NSAID plus PPI is never the optimal strategy. Sensitivity and subgroup analyses suggest that Cox-1 NSAID plus H2RA and Cox-1 NSAID plus misoprostol become more cost-effective in the older age group. Some conclusions were associated with high levels of uncertainty. CONCLUSIONS: Although there is a very large body of evidence comparing Cox-2 NSAIDs with Cox-1 NSAIDs, this is not matched by studies of the other types of gastroprotectors or by studies directly comparing active gastroprotective strategies. This lack of direct comparisons led to the use of indirect comparisons to help understand the relative efficacy of these strategies. Indirect evidence in itself is weak and was also hampered by lack of evidence in the underlying studies (where the gastroprotectors were compared with placebo). Economic modelling suggests that Cox-1 NSAID plus H2RA or Cox-1 NSAID plus PPI are the most cost-effective strategies for avoiding endoscopic ulcers in patients requiring long-term NSAID therapy. All strategies other than Cox-2 selective inhibitors reduce the rate of endoscopic ulcer compared with Cox-1 alone. The economic analysis suggests that there may be a case for prescribing H2RAs in all patients requiring NSAIDs. Misoprostol is more effective, but is associated with a greater cost and GI side-effects which may be unacceptable for patients. However, when assessing serious GI events, the economic analysis is sufficiently weakened by the data available as to render clear practice recommendations impossible. Further large, independent RCTs directly comparing various gastroprotective strategies are needed. These should report items such as major outcomes, primary data, adverse events, assessment of practice and patient preference.