Eleven studies were included: 2 RCTs (908 children) and 9 prospective cohort studies with a control group (about 940 children).
One RCT described an inadequate method of randomisation and the second RCT gave no details of the randomisation method. The losses to follow-up were considerable (range: 18 to 50%) in 6 studies. None of the studies analysed data on an intention-to-treat basis.
There was no significant difference in AOM between swimming without ear protection and not swimming; the OR (7 studies, 943 children) was 0.78 (95% CI: 0.42, 1.44). Significant statistical heterogeneity was found (P=0.0093). The results were similar after omitting the 3 low-quality studies; the OR was 0.73 (95% CI: 0.31, 1.70).
There was no significant difference in AOM between swimming with and without ear protection (swimming cap and/or earplugs); the OR (4 studies, 233 children) was 0.75 (95% CI: 0.38, 1.48). No significant statistical heterogeneity was found (P=0.26).
There was no difference in AOM between the use of antibiotic ear drops after swimming and swimming with ear protection (swimming cap and/or earplugs); the OR (2 studies, 961 children) was 0.76 (95% CI: 0.56, 1.02). No significant statistical heterogeneity was found (P=0.93).
Swimming plus the use of antibiotics after swimming significantly increased AOM compared with no swimming; the OR (2 studies, 249 children) was 3.14 (95% CI: 1.40, 7.05). No significant statistical heterogeneity was found (P=0.37). Both studies were of a low quality (Jadad score of 0).