Risk of mother-to-child transmission of HIV among women on triple antiretroviral drugs in Sub-Saharan Africa: systematic review of observational studies and explorative meta-analysis
METHODS: A systematic review and explorative meta-analysis was undertaken. PUBMED and EMBASE were searched using relevant terms. Inclusion criteria were as followed: population, HIV-exposed infants and their mothers; intervention, triple ARVs treatment/prophylaxis; outcome, MTCT rate (proportion); and study design, observational studies. The New-Castle Ottawa scale was used to assess the methodological quality. Random-effect modelling was used to combine individual MTCT rates. Forest plots were used to summarize the results, and funnel plot was used to assess publication bias
RESULTS: Thirty studies were reviewed. Most were facility-based or child-testing data of intermediate quality. They were relatively similar in terms of exposure, but the outcome, sample size (10-5141), and type/number of covariates reported varied considerably. Competing risk such as weaning and bias due to loss to follow-up were not usually considered in individual studies, leading to an incorrect estimation of transmission rate. Twelve studies were included in the meta-analysis leading to a pooled estimate of 4% (95% CI 1.6-7.3) with large variations in individual MTCT rates (0.9-11.9%); heterogeneity was high (Q-statistic p<0.000; I2-index 95% CI 92.9-96.5), and the funnel plot was asymmetric. Subgroup analyses by infants’ age at testing showed no differences, but the prophylaxis group had a slightly lower pooled estimate than the treatment group (3.7 vs. 4.9%)
CONCLUSIONS: The preventive effect of triple ARVs in SSA is context specific. MTCT rate estimation in individual studies need to be improved. Major methodological differences prevent meaningful comparisons, and limit the value of the pooled estimates