Table 2

 Numbers of subjects withdrawn due to asthma and number of subjects with at least one moderate or severe exacerbation, and associated odds ratios (95% CI) between treatment with salmeterol and increased dose of inhaled corticosteroid

StudyWithdrawals due to asthmaModerate or severe exacerbations*
SalICSOR (95% CI)SalICSOR (95% CI)
OR, Odds ratio for withdrawal (greater than 1 means ICS more likely to withdraw and favours salmeterol). Note that a continuity correction was used for this analysis because the van Noord data set had a “zero” cell count.
Sal, group receiving added salmeterol; ICS, group receiving increased dose of inhaled steroid.
*Number of participants with one or more moderate or severe exacerbations of asthma.
Greening7/22014/2062.22 (0.88 to 5.61)19/22019/2061.07 (0.56 to 2.08)
van Noord(low dose)0/302/305.35 (0.25 to 116.32)
Vermetten6/11310/1201.62 (0.57 to 4.62)9/11317/1201.86 (0.81 to 4.29)
Busse14/28120/2771.48 (0.73 to 3.00)8/2816/2770.77 (0.27 to 2.17)
Bloom15/29522/2791.60 (0.81 to 3.15)6/2959/2791.56 (0.57 to 4.30)
Kelsen3/2396/2441.98 (0.49 to 8.02)38/23944/2441.16 (0.72 to 1.72)
Murray5/2605/2541.02 (0.29 to 3.58)43/26045/2541.09 (0.69 to 1.72)
Kalberg20/24632/2421.71 (0.95 to 3.06)
Condemi2/2216/2163.13 (0.62 to 15.68)21/22131/2161.58 (0.88 to 2.84)
FLTA40211/1181/1141.04 (0.06 to 16.75)7/1188/1141.19 (0.43 to 3.28)
FLTA40221/1131/1091.04 (0.06 to 16.79)4/11320/1095.57 (1.94 to 16.05)
Johansson3/1761/1730.34 (0.03 to 3.26)9/17610/1731.13 (0.46 to 2.80)
Pooled estimates of effect (95% CI)
Fixed effect1.58 (1.12 to 2.24)1.35 (1.10 to 1.66)
Random effect1.58 (1.12 to 2.24)1.37 (1.10 to 1.70)
Inconsistency measurements0 (0 to 10)10.4 (0 to 49.7)