An estimated 15% of clinically recognized pregnancies abort spontaneously. Recurrent spontaneous abortion (RSA) is defined as three or more consecutive miscarriages conceived with the same partner in the absence of uterine, genetic or autoimmune abnormalities. Evidence points to human leucocyte antigens (HLA) as playing a role in the successful development of the foetus. In particular, HLA compatibility is more prevalent in couples experiencing reproductive failure, especially RSA couples, compared to fertile couples. According to the immunological hypothesis, an adequate immune response is necessary for proper implantation of the embryo; conversely, a depressed response of maternal lymphocytes to the stimulation by paternal antigens because of HLA sharing can result in disorders, such as RSA. The genetic hypothesis implicates homozygosity for recessive lethal alleles in linkage disequilibrium with specific HLA haplotypes. The specificity of HLA alleles or haplotypes responsible for or linked to other RSA susceptibility genes remains unclear. In this study, we identified 40 observational studies (32 case-control, five cohort, one cross-sectional, one case series and one basic science) that examined the associations between HLA and RSA, focusing on HLA allele couple and maternal-foetal sharing, and the special role of HLA-G. We sought to identify consistent findings among studies examining similar questions. Evidence remains divided concerning the role of HLA allele couple sharing. Of major concern is the focus of many studies on couple sharing as a proxy measure of maternal-foetal sharing. Therefore, adequately powered studies are needed, which employ standard case definitions and reproducible methodologies to directly assess the role of maternal-foetal HLA sharing on the risk of RSA.