This study identified risk factors associated with readmission for postpartum endometritis. The study group consisted of 109 mothers (Group I) who were discharged after delivery and readmitted with endometritis. Control groups consisted of women who had endometritis immediately after delivery but who did not require readmission (Group II, n = 109), and women who had no intrapartum or puerperal infection and also were not readmitted (Group III, n = 109). Subjects in Groups II and III were matched to an index study subject for date of delivery and maternal age, race, and parity; and women in Groups I and III were also matched for route of delivery. Groups were compared in terms of demographic characteristics, intrapartum course, and clinical presentation. The data were analyzed with the t-test, chi2, and multiple logistic regression analyses, and a P value < .05 was considered significant. Women in Groups I and III delivered vaginally more often than mothers in Group II. In addition, mothers in Groups I and III had similar postpartum courses, no evidence of infection on discharge after delivery, and a similar period from delivery until postpartum discharge. Although women in Group I were more likely to have spontaneous rupture of membranes, a shorter latent period, and have fewer bilateral tubal ligations than mothers in Group II, multivariate analysis identified route of delivery as the only significant maternal variable associated with postpartum endometritis requiring readmission. Women who were readmitted for endometritis usually delivered vaginally, and the occurrence of late-onset postpartum endometritis was unrelated to the length of stay following delivery.