Objectives – Homeopathy is a popular treatment modality among patient, however there is sparse research about adverse effects of homeopathy. A concept unique for homeopathy, is homeopathic aggravation that is understood as a transient worsening of the patients’ symptoms before an expected improvement occurs. From a risk perspective it is vital that a distinction between homeopathic aggravations and adverse effects is established. There is a lack of systematic information on how frequent adverse effects and homeopathic aggravations are reported in studies. Therefore, a systematic review and meta-analysis were performed. Design and setting – Sixteen electronic databases were searched for Randomized Controlled Trials (RCTs). The searches were limited from the year 1995 to January 2011. Forty-one RCTs, with a total of 6.055 participants were included. A subtotal of 39 studies was included in the additional meta-analysis. Results – A total of 28 trials (68%) reported adverse effects and five trials (12%) reported homeopathic aggravations. The meta-analysis (including six subgroup comparisons) demonstrated that no significant difference was found between homeopathy and control with OR 0.99, 95% CI 0.86–1.14, I2  =   54%. More than two third of the adverse effects were classified as grade 1 (68%) and two third were classified as grade 2 (25%) and grade 3 (6%) according to the Common Terminology Criteria for Adverse Effects. Homeopathic aggravation was classified as grade 1 (98%) and grade 3 (2%), suggesting that homeopathic aggravations were reported to be less severe than adverse effects. The methodological quality according to a method recommended in the Cochrane handbook for RCTs, was high. Conclusion – Adverse effects including the concept of homeopathic aggravations are commonly reported in trials. The meta-analysis demonstrated that the proportion of patients experiencing adverse effects to be similar for patients randomized to homeopathic treatment compared to patients randomized to placebo and conventional medicine.