Seventeen double blinded RCTs (n=13,645 patients) were included in the review. Quality was variable across included trials; allocation concealment was adequate in four trials (unclear in 13), withdrawal rates ranged from 6.1% to 42% and data on loss to follow-up was available for six trials (range 0% to 3.4%).
Main outcomes: Overall, inhaled anticholinergics significantly increased the risk of cardiovascular death, myocardial infarction or stroke (RR 1.60, 95% CI 1.22 to 2.10). Inhaled anticholinergics significantly increased the risk of myocardial infarction (RR 1.52, 95% CI 1.04 to 2.22) and cardiovascular death (RR 1.92, 95% CI 1.23 to 3.00) but did not significantly increase the risk of stroke (RR 1.46, 95% CI 0.81 to 2.62). No statistical heterogeneity was observed. Inhaled anticholinergics did not significantly increase the risk of all-cause mortality (RR 1.29, 95% CI 1.00 to 1.65). The number-needed-to-harm for myocardial infarction was estimated to be 174 per year (95% CI 75 to 1,835 per year) and for cardiovascular death 40 per year (95% CI 18 to 185 per year)
Sensitivity analyses: Results remained significant when a random-effects model was used; when limited to the six long-term trials (lasting over six months); when inhaled tiotropium and inhaled ipratropium were analysed separately in the long term trials; when five trials were excluded for which endpoint data was missing; and after excluding one trial which contributed to more than 50% of the weight in the fixed-effect model. Short-term trials did not show significantly increased risk.
Fail safe N: Sixteen non-significant long-term trials, each with a sample size of approximately 1,450, would be required to reverse the main outcomes.
It should be noted that the figures in this review are corrected from the original publication (cited in the Bibliographic Details field) using subsequently published errata (see Other Publications of Related Interest field).