Cesarean section to reduce perinatal transmission of human immunodeficiency virus. A metaanalysis.

OBJECTIVE: Individual epidemiologic investigations into the association between type of delivery and perinatal HIV transmission have been suggestive but inconclusive. Metaanalysis was used in an attempt to establish if there is, at present, adequate evidence concerning the effectiveness of cesarean section in reducing vertical HIV transmission rates. METHODS: The MEDLINE data retrieval system and other sources were used to identify studies containing data on the relationship between type of delivery and vertical HIV transmission. No randomized control trials were located. Six cohort studies identified were included in the metaanalysis. Crude and, in the only study in which these were available, adjusted data were extracted and pooled. RESULTS: The overall weighted risk of perinatal HIV infection was 20.2% and 14.0% after vaginal and cesarean delivery, respectively. Pooling data of all studies showed a statistically significant difference of HIV perinatal transmission rates between cesarean and vaginal delivery (odds ratio 0.65; 95% CI, 0.43 to 0.99; P = 0.044) (Random effects model: DerSimonian and Laird method). Approximately 16 (95% CI, 76 to 9) HIV-infected women must deliver by cesarean in order to prevent 1 case of HIV perinatal infection. CONCLUSIONS: Results of this study show that performing elective cesarean section in HIV-infected women is potentially an effective procedure. However, the nonexperimental nature of the available studies leads us to conclude that randomized control trials are indicated before setting specific guidelines for mode of delivery in HIV-infected women.