[A meta-analysis of HIV seroprevalence in pregnant women with syphilis and the impact of syphilis infection on mother-to-child HIV transmission].

Objective: To estimate HIV seroprevalence among pregnant women with syphilis and evaluate the influence of syphilis infection on mother-to-child-transmission (MTCT) of HIV by meta-analysis. Methods: We conducted a systematic literature search for 1 678 articles related to maternal syphilis and HIV infection published until October 1(st) 2015 using the PubMed, Web of Science, Chinese Web of Knowledge, Wanfang, Weipu, and SinoMed databases and evaluated the quality of each papers using the STROBE checklist, and the keywords were " pregnant women/maternal/pregnancy" , "syphilis/AIDS" , "HIV/human immunodeficiency virus" , "mother- to-child transimission/vertical transmission" . Excluding studies with the special subgroups of HIV-positive pregnant women as the research objects, review or meeting abstract, impossibility of full-text acquisition, sample size <50, duplication or impossibility of data extraction, finally, 16 studies were included. Random-effects meta-analysis was used to estimate HIV seroprevalence among pregnant women with syphilis and the RR of MTCT for women infected with both syphilis and HIV. Subgroup analyses were undertaken by study location, sample size, use of anti-retroviral therapy and study quality. Results: Sixteen studies with a combined sample of 110 573 pregnant women were included in the analysis. Of these, ten reported HIV seroprevalences among pregnant women with syphilis and six studies evaluated the influence of syphilis infection on MTCT of HIV. Pooled estimates yielded a HIV seroprevalence of 11.6% (95% CI: 6.7%-19.5%) among pregnant women with syphilis. We estimated that the risk of MTCT of HIV was 1.86 times (RR=1.86, 95% CI: 0.89%-3.89%) higher among pregnant women with syphilis compared with those only infected with HIV-although this effect was not statistically significant. Cochran's Q test showed a high degree of heterogeneity in estimates of HIV seroprevalence and the effect of syphilis infection on MTCT of HIV across studies (I(2)=89.4% and 86.2%, respectively, P<0.10). Subgroup analysis estimated HIV seroprevalences of 24.9% (95%CI: 17.4%-34.3%) in Africa, 2.8% (95% CI: 1.4%-5.6%) in Asia and 2.2% (95% CI: 0.7%-6.7%) in South America. While studies with a large sample size (≥100) or of higher quality estimated overall seroprevalence at 15.2% (95%CI: 9.0%-24.7%), this was 2.2% (95%CI: 0.7%-6.7%) for lower-quality or smaller studies. Meanwhile, subgroup analyses of the RR of MTCT of HIV in pregnant women infected both with HIV and syphilis gave estimates of 1.19 (0.62-2.29) for the higher quality studies, 4.76 (2.65-8.53) for the lower-quality studies, 1.47 (0.77-2.81) for studies with a large sample size, 5.82 (3.16-10.74) for studies with a small sample size, 4.76 (2.65-8.53) for studies in which participants received antiretroviral treatment and 1.19 (0.62-2.29) for studies in which they did not. While Begg's test showed evidence of publication bias in studies of HIV seroprevalence estimates in pregnant women with syphilis (t=-2.48, P=0.038), no evidence of publication bias was found in studies on the influence of syphilis infection on MTCT of HIV (t=-0.22, P=0.835). Conclusion: HIV seroprevalence is higher among pregnant women with syphilis than uninfected women. Further research is warranted to verify whether syphilis infection can increase the risk of MTCT of HIV.