Fourteen relevant RCTs were identified (n=1,622 participants, range 47 to 247); four trials with a Jadad score of 4, eight with a Jadad score of 3, and two trials has a low Jadad score of 2. Follow-up was generally after eight weeks (range four to eight weeks).
There was no significant difference between the pooled Helicobacter pylori eradication rates with azithromycin (72.01%, 95% CI 58.09 to 85.93) and without azithromycin (69.78%, 95% CI 66.47 to 73.09), with an odds ratio of 1.17 (95% CI 0.64 to 2.14; 14 RCTs; I2=81%), with significant heterogeneity, using an intention-to-treat analysis.
There were significantly fewer total side effects with azithromycin (15.81%, 95% CI 12.50 to 19.12) than without azithromycin (25.20%, 95% CI 21.44 to 28.96), with an odds ratio of 0.58 (95% CI 0.41 to 0.82; 10 RCTs; I2=21%). There were also significantly lower incidences of diarrhoea (OR 0.33, 95% CI 0.12 to 0.96; four RCTs; I2=0%), nausea (OR 0.37, 95% CI 0.14 to 0.96; three RCTs; I2=19%) and taste disturbance (OR 0.28, 95% CI 0.11 to 0.70; four RCTs; I2=0%) with azithromycin than without azithromycin.
The subgroup analysis for eradication rate for duration of azithromycin showed no significant difference for a long-course (seven RCTs) or short-course of treatment (seven RCTs), with significant heterogeneity for both analyses. Subgroup analyses for eradication rate that compared the antibiotic given with azithromycin (amoxicillin, six RCTs; levofloxacin, three RCTs; metronidazole/tinidazole, five RCTs) also found no significant difference in the pooled eradication rates for each regime, and significant heterogeneity for each sub-analysis.
Egger’s test showed that there was no evidence of substantial publication bias.