Ten RCTs (1,483 patients) were included.
Only 5 RCTs indicated the method of randomisation; 4 RCTs defined the inclusion and/or exclusion criteria in detail; and 7 RCTs included a sample size estimate. The mean methodological quality score was 0.66. There was a significant correlation (rho) between the year of publication and methodological quality (rho=0.76, p=0.02). However, there were no significant correlations between the effect size and year of publication (rho=0.29, p=0.4) or methodological quality (rho =0.42, p=0.2).
Pulmonary function at 90 minutes: the addition of ipratropium to beta-agonist therapy led to a significant benefit. The pooled effect size was 0.14 (95% CI: 0.04, 0.24, p=0.008). The fail-safe N was 36.4. There was no evidence of heterogeneity (p>0.5).
Sensitivity analyses: when the 5 lower quality studies were excluded, the findings were practically unchanged; the results were not reported. For the 4 studies of patients with FEV1 or PF less than 35% of the predicted value, the addition of ipratropium to beta-agonist therapy led to a significant benefit (pooled effect size 0.38, 95% CI: 0.09, 0.67). Systemic steroids had a moderate effect on outcome (pooled effect size 0.14, 95% CI: 0.00, 0.28).
Publication bias: the funnel plot was statistically symmetrical (p=0.58).
Admission rates (5 RCTs, 1,186 patients): the addition of ipratropium to beta-agonist therapy led to a significant reduction in the admission rates. The odds ratio was 0.62 (95% CI: 0.44, 0.88, p=0.007) and the number-needed-to-treat was 18 (95% CI: 11, 77). No evidence of heterogeneity was found (p>0.3). Most papers did not clearly describe their criteria for admission or discharge.
Side-effects (6 RCTs): there were no significant differences between the treatment groups in terms of tremor, heart rate, blood-pressure, respiratory rate, anxiety, dry mouth, or oxygen saturation.