Nineteen RCTs (n=4,626) were included. There were 14 paediatric studies and 5 studies in adults.
The mean Jadad quality score was 2.5 (standard deviation 0.8). Thirty-eight per cent of the studies were considered to be of relatively high quality (scored more than 2 on the Jadad scale).
Children (14 studies).
The bacterial cure rate was significantly increased with a 60-mg/kg total dose of azithromycin in comparison with a 10-day comparator regimen (OR 5.27, 95% CI: 3.34, 8.23, P<0.00001); conversely, the rate was significantly reduced with a 30-mg/kg total dose of azithromycin (OR 0.47, 95% CI: 0.24, 0.91, P=0.02).
Five-day azithromycin regimens significantly improved bacterial cure rates in comparison with 10-day comparator regimens (OR 4.37, 95% CI: 1.70, 11.17, P=0.002), but there was no statistically significant difference between 3-day azithromycin regimens and comparator regimens.
The clinical cure rate was significantly increased with 5 days' azithromycin compared with comparator regimens (OR 6.80, 95% CI: 3.30, 14.01, P<0.00001), but there was no statistically significant difference between 3 days' azithromycin and comparator regimens. There was significant heterogeneity among studies of children. Statistical heterogeneity persisted after a separate analysis of studies using penicillin as a comparator (P<0.0001). The jack-knife analysis eliminated statistical heterogeneity only after exclusion of the 3 studies with the widest CIs.
Adults (5 studies).
Five-day azithromycin regimens significantly reduced bacterial cure rates in comparison with 10-day comparator regimens (OR 0.41, 95% CI: 0.22, 0.78), but there was no statistically significant difference between either 3-day azithromycin regimens (all using 500 mg/day) or 30-mg/kg total dose azithromycin regimens and comparator regimens.
There was no statistically significant difference in clinical cure rates between 3- or 5-day azithromycin and comparator regimens.