Cefazolin compared with cefoxitin for cesarean section prophylaxis: the use of a two-stage study design

The authors used a two-stage design to compare the risk of endometritis in women undergoing non-elective cesarean section who received cefazolin prophylaxis (n = 481) with those who received cefoxitin prophylaxis (n = 1799). The primary data source for this study was an automated record linkage system which allowed the ascertainment of exposure(antibiotic prophylaxis) and preliminary ascertainment of outcome (post-partum endometritis) on a consecutive sample of women undergoing cesarean section between 1 April 1987 and 30 September 1989. Potentially important covariates not available in the automated data source were sampled by review of complete medical records of a random sample of each exposure-disease category of the cohort. Of the 2280 women studied, 99 (4.3%) developed postpartum endometritis. After control for age, race, anemia, presence of ruptured membranes, parity, labor, number of vaginal examinations and payor status the adjusted odds ratio for cefazolin compared to cefoxitin was 0.95 (95% C.I. 0.5-1.9). The cost of prophylaxis was significantly higher for women who received cefoxitin prophylaxis ($56/patient vs $9.55/patient). These results suggest that cefazolin prophylaxis should be favored over cefoxitin due to lower cost and similar efficacy. This study also demonstrates the efficiency of a two-stage design in the setting where exposure and outcome are available for an entire cohort but information about important covariates is more difficult to obtain.