A meta-analysis of association studies was performed to assess whether the reported genetic polymorphisms in cytokine genes are risk factors for recurrent miscarriage (RM). The electronic PubMed database was searched for case-control studies on immunity-related genes in RM. Investigations of a single polymorphism/gene involvement in RM reported more than five times were selected. Aggregating data from seven case-control studies on -308/tumour necrosis factor-alpha polymorphism, the odds ratio (OR) for RM was 1.1 (0.87-1.39) if the polymorphism was considered under a dominant genetic model. In six studies on -1082/interleukin-10 (IL-10) polymorphism, the OR under a dominant model was 0.76 (0.58-0.99), and under a recessive model the OR was 0.90 (0.71-1.15). In five case-control studies on -174/IL-6 polymorphism, the OR for RM under a recessive model was 1.29 (0.69-2.40). The results show a statistically significant association with RM for the -1082/IL-10 genotype.