OBJECTIVE: Before the availability of modern imaging studies the diagnosis of septic pelvic thrombophlebitis causing prolonged puerperal fever was difficult to confirm without surgical exploration. With the use of computed tomography infection-related pelvic phlebitis can now be confirmed, and this study was designed to determine its incidence after delivery. We also designed a randomized clinical trial to evaluate the efficacy of heparin added to antimicrobial therapy for treatment of women with septic phlebitis. STUDY DESIGN: We studied women who had pelvic infection and fever that persisted after 5 days despite adequate antimicrobial therapy with clindamycin, gentamicin, and ampicillin. After giving consent study participants underwent abdominopelvic computed tomographic imaging. Women with pelvic thrombophlebitis were randomly assigned to 1 of 2 management schemes that included continuation of antimicrobial therapy, either alone or with the addition of heparin, until the temperature was .5). The 54 women with persistent fever but without computed tomographic evidence of septic pelvic thrombophlebitis were hospitalized for a mean of 12.0 +/- 4.1 days, compared with 10.9 +/- 2.9 days for women in whom thrombosis was diagnosed (P =.14). These women were followed up for >/=3 months post partum and none showed evidence of reinfection, embolic episodes, or postphlebitic syndrome. CONCLUSIONS: The overall incidence of septic pelvic thrombophlebitis was 1:3000 deliveries. The incidence was about 1:9000 after vaginal delivery and 1:800 after cesarean section. Women given heparin in addition to antimicrobial therapy for septic thrombophlebitis did not have better outcomes than did those for whom antimicrobial therapy alone was continued. These results also do not support the common empiric practice of heparin treatment for women with persistent postpartum infection.