Efficacy and harm of pharmacological prevention of acute mountain sickness: quantitative systematic review

Abstract

Objective: To quantify efficacy and harm of pharmacological prevention of acute mountain sickness.

Data sources: Systematic search (Medline, Embase, Cochrane Library, internet, bibliographies, authors) in any language, up to October 1999.

Study selection: Randomised placebo controlled trials.

Data extraction: Dichotomous data on efficacy and harm from 33 trials (523 subjects received 13 different interventions, 519 a placebo).

Data synthesis: At above 4000 m the mean incidence of acute mountain sickness with placebo was 67% (range 25% to 100%); incidence depended on the rate of ascent, but not on the altitude or the mode of ascent. Across all ascent rates, dexamethasone 8-16 mg prevented acute mountain sickness (relative risk 2.50 (95% confidence interval 1.71 to 3.66); number needed to treat (NNT) 2.8 (2.0 to 4.6)), without evidence of dose responsiveness. Acetazolamide 750 mg was also efficacious (2.18 (1.52 to 3.15); NNT 2.9 (2.0 to 5.2)), but 500 mg was not. In two trials, adverse reaction (including depression) occurred after dexamethasone was stopped abruptly (4.45 (1.08 to 18); NNT 3.7 (2.5 to 6.9)). With acetazolamide, paraesthesia (4.02 (1.71 to 9.43); NNT 3.0 (2.0 to 6.0)) and polyuria (4.24 (1.92 to 9.37); NNT 3.6 (2.5 to 6.2)) were reported. Data were sparse on nifedipine, frusemide (furosemide), dihydroxyaluminium-sodium, spironolactone, phenytoin, codeine, phenformin, antidiuretic hormone, and ginkgo biloba.

Conclusions: At above 4000 m, with a high ascent rate, fewer than three subjects need to be treated with prophylactic dexamethasone 8-16 mg or acetazolamide 750 mg for one subject not to experience acute mountain sickness who would have done so had they all received a placebo. Acetazolamide 500 mg does not work.

Figures

Figure 1

Incidence of acute mountain sickness in subjects receiving placebo (control event rate) in relation to final altitude (top), rate of ascent (bottom), and mode of ascent. Each symbol represents one trial; symbol sizes do not take into account trial sizes

Figure 2

Incidence of acute mountain sickness with interventions and with placebo (18 trials, 22 comparisons). Symbol sizes are proportional to trial sizes. “Others”=codeine, phenformin, nifedipine, spironolactone, and dexamethasone plus acetazolamide. The dotted line represents equality

Figure 3

Relation between ascent rate and efficacy of pharmacological prevention of acute mountain sickness (17 trials; in the bottom graph one trial is omitted as event rates with placebo and with intervention were similar and thus the number needed to treat could not be calculated). Each symbol represents one trial; symbol sizes do not take into account trial size