MEDLINE, EMBASE and Cochrane Library were orderly searched for all randomized controlled trials (RCTs) designating R-PCA or EA to parturients and reporting at least one outcome of interest. The study included eight randomized trials of R-PCA vs. EA with 2351 patients. Measurement of maternal satisfaction was the primary outcome. Secondary outcomes were visual analog pain score (VAS at 1, 2, 3 h postoperatively), nausea, vomiting, pruritus, hypoxemia, acute respiratory depression or death (maternal or neonatal), need for Cesarean section, and neonatal Apgar score. 

The selected trials showed no specified variations between the R-PCA and EA groups for maternal satisfaction, VAS at 2 or 3 h, nausea, vomiting, need for cesarean section, respiratory depression, umbilical pH, and neonatal Apgar score at 1 min and 5 min. But, the degree of hypoxemia was found to be higher [OR 7.48, 95%CI 3.42–16.36] and VAS at 1 h was little higher [WMD 1.33, 95%CI 0.30–2.36] using R-PCA compared to EA. In R-PCA group, pruritus was less common. In any of the studies, acute respiratory failure and death were not described.

It is difficult to draw any conclusion from this meta-analysis as no significant differences were observed in case of maternal satisfaction or other clinical outcomes. R-PCA may present an option for analgesia in the peripartum period for obstetric patients who are not candidates for EA, but attention needs to assured mainly about hypoxemia. Moreover, there is a need to have well-powered randomized trials that focus on clinically-appropriate maternal and neonatal results so as to precisely define the related advantages and risks of R-PCA versus EA in this population.