ObjectiveTo determine the effect of prophylactic antibiotics on endometrial and endocervical microflora upon diagnosis of postcesarean endometritis.MethodsThe medical records of patients enrolled in open-label comparative trials of therapeutic antibiotics for postpartum endometritis between 1989 and 1994 were reviewed (n= 682). Endometritis was diagnosed by a standard definition that included fever and localizing signs. Endometrial cultures were obtained by a sheathed injection/aspiration technique. Aerobes and anaerobes were isolated by standard microbiologic techniques. The primary outcome, endometrial and endocervical microflora, was compared in women who received intravenous ampicillin (2 g every 6 hours for 1-3 doses), cephalosporin (2 g every 6 hours for 1-3 doses), or no prophylaxis. Secondary outcomes included the cure of endometritis and the prevalence of wound infection in the three groups.ResultsFour hundred sixty-five of 682 patients (67%) had a cesarean delivery. One hundred fifty-one patients received ampicillin prophylaxis, 100 patients received cefazolin prophylaxis, 18 patients received extended-spectrum antibiotics, and 196 patients received no prophylaxis. Patients who received cefazolin prophylaxis had a significant increase in enterococcus (P< .05) and a significant decrease inProteusspecies (P< .05) from endometrial samples. Patients who received ampicillin prophylaxis had a significant increase ofMycoplasmaspecies (P< .05),Klebsiella pneumoniae(P< .0001),Escherichia coli(P= .04), and any aerobic gramnegative rod (P= .003) from endometrial samples. Ampicillin prophylaxis was associated with a decrease inPrevotella bivia(P< .05) and any anaerobe (P< .01). Endometritis cure rates were similar between prophylaxis groups and between prophylaxis and treatment groups. However, the cefazolin prophylaxis followed by cephalosporin treatment was associated with more wound infections (19%) than other prophylaxis and treatment groups, (6%,P< .01).ConclusionAmpicillin and cefazolin prophylaxis alters endometrial and endocervical microflora of women who develop endometritis. Whereas these changes had no effect on endometritis cure rate (power > 80%), the association between cefazolin prophylaxis followed by an extendedspectrum cephalosporin and wound infection may warrant caution against the use of the combination.