HYPERTENSION IN PREGNANCY

Vol. 23, No. 1, pp. 121–127, 2004

Septic Pelvic Thrombophlebitis and Preeclampsia AreRelated Disorders#

Christy M. Isler, M.D.,1,* Brian K. Rinehart, M.D.,2

Dom A. Terrone, M.D.,2 J. Holt Crews, M.D.,2

Everett F. Magann, M.D.,2 and James N. Martin Jr., M.D.2

1Department of Obstetrics and Gynecology, East Carolina University Brody School of

Medicine, Greenville, North Carolina, USA2Department of Obstetrics and Gynecology, University of Mississippi Medical Center,

Jackson, Mississippi, USA

ABSTRACT

Objective: To elicit factors associated with the postpartum development of septic

pelvic thrombophlebitis in a single large referral tertiary patient population. Methods:

A nine-year single institution retrospective case review of all patients with enigmatic

fever and septic pelvic thrombophlebitis was analyzed. Results: A total of 55 patients

with septic pelvic thrombophlebitis were provided care during the study interval. The

average gestational age at delivery was 36.8±4.3 weeks. The most prevalent

concurrent medical complication of pregnancy was preeclampsia (45%) while

chorioamnionitis affected only 13%. The average length of ruptured membranes

was 22.8±56.8 hours (median 10.5, 95% confidence interval [CI] 7.0–38.7 hours),

with 22% of patients undergoing amnion rupture at the time of cesarean delivery.

Prolonged (>24 hours) amnion rupture occurred in only 9% of patients. Most affected

patients were delivered abdominally (91%) but a minority delivered vaginally (9%).

#Presented at the Twentieth Annual Meeting of the Society for Maternal-Fetal Medicine, Miami,

Florida, February 1–5, 2000.*Correspondence: Christy M. Isler, M.D., OB-GYN Publication Office, Department of Obstetrics

and Gynecology, University of Mississippi Medical Center, 2500 North State St., Jackson, MS

39216-4505, USA; Fax: (601) 984-6773; E-mail: islerc@mail.ecu.edu.

121

DOI: 10.1081/PRG-120029858 1064-1955 (Print); 1525-6065 (Online)

Copyright D 2004 by Marcel Dekker, Inc. www.dekker.com

ORDER REPRINTS

Antibiotic therapy for presumed infection was initiated at 27.4 ±24.6 hours

postpartum. Subsequently intravenous heparin therapy was initiated 128.9±54.2

hours thereafter enigmatic fever defervesed 37.2±36.8 hours later (median 34.0, 95%

CI 27.2–47.3 hours). Patients received 6.3±1.8 days of heparin therapy. Conclusion:

In this series, septic pelvic thrombophlebitis was frequently preceded by cesarean

delivery and commonly associated with preeclampsia. Unexpectedly, a small number

of patients suffered prolonged rupture of membranes or chorioamnionitis. We

speculate that the cesarean delivery of a population of at-risk patients with

preeclampsia may predispose them to develop septic pelvic thrombophlebitis.

Key Words: HELLP syndrome; Preeclampsia; Pregnancy; Septic pelvic

thrombophlebitis.

INTRODUCTION

Septic pelvic thrombophlebitis is a disorder characterized by inflammation of the

pelvic veins with infected thromboses, most often following an obstetrical procedure

such as cesarean section or septic abortion. Optimal diagnosis and management of this

disorder remains elusive. Schulman et al. (1) described the diagnosis as one of

exclusion, and recommended an empiric trial of heparin after failure of antimicrobial

agents to produce defervescence of fever in these patients with no obvious site of

infection. Although historically there was much debate concerning medical (2,3) versus

surgical (4,5) management of this disease, it seems clear that the modern management

of septic pelvic thrombophlebitis primarily involves medical management with surgical

therapy reserved for patients who fail primary therapy. Current controversies concern

how clinical versus radiologic diagnosis of this condition differ, whether anti-

coagulation therapy is a necessary adjunct to broad spectrum antibiotic therapy, and

how long to continue antibiotic and/or anticoagulation therapy after presumptive

diagnosis is made (6–8).

The purpose of this study was to elicit factors associated with the postpartum

development of septic pelvic thrombophlebitis in a large, single institution, tertiary

referral center which provides medical care primarily for a population of underserved

socioeconomically disadvantaged patients in central Mississippi.

MATERIALS AND METHODS

A retrospective chart review was performed for all women discharged from the

University of Mississippi Medical Center between January 1990 and December 1998

with the diagnosis of septic pelvic thrombophlebitis. Each identified chart was

individually scrutinized to ascertain that each patient satisfied the diagnostic criteria

for septic pelvic thrombophlebitis which included a persistent puerperal enigmatic

febrile course (>72 hours) despite broad spectrum antibiotic therapy (>48 hours) and

absent evidence by culture or radiologic study of an infectious etiology (determined

by urine culture and chest x-ray supporting physical examination findings). In view of

the controversy regarding efficacy for radiologic techniques to support diagnosis of

122 Isler et al.

ORDER REPRINTS

septic pelvic thrombophlebitis and an institutional bias by the Department of

Radiology against such, imaging studies were infrequently undertaken and not

required for diagnosis.

Medical records were individually reviewed for demographic information including

maternal age, race, parity, body mass index, and gestational age. All concurrent medical

complications of pregnancy were noted. Both mild and severe preeclampsia were

diagnosed using the criteria established by The American College of Obstetricians

and Gynecologists. The diagnosis of hemolysis, elevated live enzymes, low platelets

(HELLP) syndrome was based upon the documentation of hemolysis (lactate

dehydrogenase >600 IU/L or progressive anemia), hepatic dysfunction (elevated aspartate

aminotransferase >70 IU/L), and thrombocytopenia (platelet count <100,000 cells/mL).

Labor characteristics including length of membrane rupture, the presence of chorioam-

nionitis, and route of delivery were specifically targeted. The diagnosis of chorioamnio-

nitis was based upon an elevated temperature accompanied by uterine tenderness and fetal

tachycardia. The postpartum course was examined for timing and length of febrile course

as it related to delivery and institution of antibiotic and heparin therapy. The use of blood

cultures and imaging studies were reviewed and their results noted.

RESULTS

During the study period, there were 32,542 deliveries at University Medical Center

with 7453 of these patients delivered abdominally. The medical records of 55 patients

met the criteria for diagnosis of septic pelvic thrombophlebitis (1.7 per 1000

deliveries). The average maternal age was 21.5±5.1 years and the racial distribution

included 85% African–American, 11% Caucasian, and 4% Native-American. Most

patients (78%) were nulliparous. Patients generally were obese with an average body

mass index of 43.0 kg/m2 which corresponds to an average weight of 185±47 pounds.

Average gestational age at delivery was 36.8±4.3 weeks. The most prevalent medical

complication of pregnancy was preeclampsia, occurring in 45% compared to an

institutional incidence of 13.1% during the study period (see Table 1). Of the 25

patients with preeclampsia, 11 patients (44%) had mild preeclampsia, and 14 patients

(56%) had severe preeclampsia, with 3 of these 14 patients having HELLP syndrome.

The patients with HELLP syndrome received intravenous high dose dexamethasone for

therapy (10 mg every 12 hours until evidence of disease resolution was present).

Diabetes mellitus was present in 2%, while chronic hypertension was present in 5%.

Table 1. Comparison of study and institutional rates of pregnancy complications.

Pregnancy complication Study rate (%) Institutional rate (%)

Preeclampsia 45 13

Chronic hypertension 5 8

Diabetes mellitus 2 2

Chorioamnionitis 13 6

Cesarean section (total) 91 25

Repeat cesarean section 9 10

Septic Pelvic Thrombophlebitis 123

ORDER REPRINTS

The average time interval between amnion rupture and delivery in study subjects

was 22.8±56.8 hours (median 10.5, 95% CI 7.0–38.7 hours), with 22% of patients

undergoing rupture of membranes at cesarean and 9% of patients experiencing

prolonged (>24 hours) rupture of membranes. A diagnosis of chorioamnionitis was

made in 13% of patients. Cesarean was the mode of delivery in 91% of the study

population with the two most common indications being arrest of labor (44%) and

nonreassuring fetal status (20%). All patients undergoing cesarean delivery received

one dose of prophylactic intraoperative antibiotics or, if chorioamnionitis were pre-

sent, therapeutic peripartum antibiotic therapy was begun preoperatively and con-

tinued postpartum.

The average time after delivery for onset of new febrile morbidity (>100.4�F)

necessitating institution of antibiotic therapy was 27.4±24.6 hours. All patients

received triple antibiotic coverage with ampicillin, gentamicin, and clindamycin for

persistence of fever greater than 100.4�F. Four patients received vancomycin instead of

ampicillin due to a history of penicillin allergy. Antibiotic therapy was discontinued

after a patient eventually became afebrile for >24 hours. Thirty-eight patients had

blood cultures performed, of which 7 (18%) were positive. The most common

organisms isolated were gram negative rods (two patients with Escherichia coli, one

patient with Enterobacter, one patient with Citrobacter), although gram positive cocci

(one patient with Group B Streptococcus and one patient with Staphylococcus aureus)

and anaerobes (one patient with Bacteroides) were also isolated.

After a presumptive diagnosis of septic pelvic thrombophlebitis was made, heparin

therapy was begun. An intravenous heparin bolus followed by continuous intravenous

infusion sufficient to maintain an activated partial thromboplastin time 1.5 to 2 times

normal was initiated 128.9±54.2 hours after the first antibiotics were started. Patients

defervesced to become afebrile 37.2±36.8 hours later (median 34.0, 95% CI 27.2–

47.3 hours). Defervescence occurred at less than 24 hours after heparin institution in

41.8% of patients, 72.7% within 48 hours, 89.8% within 72 hours, and 100% became

afebrile within 96 hours. Collectively, the patients studied received 6.3±1.8 days of

heparin therapy before discontinuation and discharge from the hospital. Five patients

received less than 7 days of heparin infusion because they left the hospital against

medical advice.

Imaging studies were obtained at the discretion of the physician staff. Only two

of 34 patients in the series who underwent abdominal/pelvic CT scanning were

reported to have significant findings. Both of those studies revealed a right ovarian

vein thrombosis.

All patients were afebrile for >24 hours prior to hospital discharge. No patient was

discharged on antibiotic or anticoagulation therapy. There were no known recurrent

febrile episodes, although not all patients received follow-up care at our institution. No

patient required surgical therapy.

DISCUSSION

Septic pelvic thrombophlebitis is reported to occur in 1 in 2000 deliveries (3).

Gibbs et al. found this disorder to be more common after postcesarean endometritis,

affecting 2% of patients with this diagnosis (9). Classically, the disorder is divided into

124 Isler et al.

ORDER REPRINTS

two distinct syndromes. The first scenario has been called ‘‘enigmatic fever’’ and

includes patients with persistent spiking fevers despite appropriate antibiotic coverage

who are otherwise asymptomatic (10). These patients do not have severe abdominal

pain or sepsis, and typically do not have a clot present on diagnostic studies. These

patients have been hypothesized to have multiple small clots present in small diameter

vessels. In contrast, patients with ‘‘ovarian vein thrombosis’’ have persistent spiking

fevers despite appropriate antibiotic coverage but are clinically ill. They typically have

severe abdominal pain, radiologically evident thrombosis of the vascular system, and

often exhibit signs of sepsis (11).

Despite the differences in clinical presentation between these two syndromes,

treatment has traditionally been similar. Patients are continued on broad spectrum

antibiotic coverage while intravenous continuous infusion heparin is administered.

Surgical therapy with ligation or excision of the thrombosed vessel is instituted if the

patient does not improve with medical therapy. Recently Brown and coworkers have

questioned whether anticoagulation is necessary (8) given the imprecision of the

therapy. Patients in their series had radiologic documentation of pelvic vein thrombosis

with some patients responding to antibiotics alone while other patients had protracted

febrile courses despite heparin anticoagulation and antibiotic therapy.

This investigation used clinical criteria to establish the diagnosis of septic pelvic

thrombophlebitis and found radiologic studies not particularly helpful in guiding

management. Many investigations have examined the efficacy of radiologic studies in

diagnosing septic pelvic thrombophlebitis by visualization of clot in the ovarian vessels.

When imaging studies are used to clarify the diagnosis, it appears that computed

tomography scanning and magnetic resonance imaging are superior to ultrasonography

(12). Our radiology department at the medical center typically does not recommend

routine imaging of a patient with suspected septic pelvic thrombophlebitis due to the

low sensitivity and specificity of radiologic diagnosis for this condition.

Traditionally, septic pelvic thrombophlebitis has been associated with antecedent

cesarean delivery, especially when the cesarean is performed in the presence of

chorioamnionitis. Our review confirms that abdominal delivery is overwhelmingly the

route of delivery for these patients (91% compared with an institutional rate of 24%),

but, chorioamnionitis was detected in only 13% of cases. In fact, a greater number of

patients had rupture of membranes performed at the time of cesarean section (22%)

than those who experienced prolonged rupture of membranes >24 hours (9%).

An unexpected finding during this review was the association of septic pelvic

thrombophlebitis with preeclampsia. Compared with our institutional rate of 13–15%,

the rate of preeclampsia in this series was 45%. Other vascular diseases, such as

diabetes mellitus and chronic essential hypertension, did not occur at an increased rate

in this patient population. As preeclampsia has also been associated with an increased

risk of deep venous thrombosis, we speculate that endothelial dysfunction and damage

are the predisposing factors. This would correlate with the presumed pathologic

mechanism of septic pelvic thrombophlebitis which Collins (5) defined and verified

histologically. There is damage to the intimal lining of affected veins by bacterial

invasion and the clotting process is activated, only to have the formed clots

subsequently invaded by microorganisms. Any preexisting endovascular damage, such

as surgery or preeclampsia, presumably can predispose the patient to this pathologic

sequence. An alternate hypothesis addressing the association of septic pelvic

Septic Pelvic Thrombophlebitis 125

ORDER REPRINTS

thrombophlebitis and preeclampsia is the high incidence of thrombophilias among

preeclamptic patients predisposing them to thrombophlebitis.

The average time to defervescence after initiation of heparin therapy in our patient

population was 37.2 hours. This response is more consistent with investigators who

used clinical criteria to diagnose septic pelvic thrombophlebitis (2.5 days) (3,11) than

those who used clinical criteria with radiologic confirmation for diagnosis (134 hours)

(8). This may be a reflection on the diagnostic degree of certainty or may represent the

extent of disease. Dunihoo et al. (13) proposed a spectrum of disease with initial milder

disease, represented by septic thrombophlebitis in the small veins, progressing to

ovarian vein thrombophlebitis. This concept of a disease spectrum has not been

supported by some investigators (6,7). It is possible that patients with visible clot on

radiologic studies may represent the syndrome of acute ovarian vein thrombosis, as

opposed to enigmatic fever. Thrombosis in small pelvic veins, consistent with

enigmatic fever, would not be easily visualized with CT scan or magnetic resonance

imaging (14,15).

In conclusion, this review confirms the association of septic pelvic thrombophle-

bitis and cesarean section, though not the association with chorioamnionitis. The

optimal management of this disorder has yet to be determined. The frequent

coexistence of preeclampsia observed in this study needs to be investigated further.

Perhaps there is a subset of patients with preeclampsia who also have a genetically

based incompetence of vascular systems to compensate adequately for the combined

stresses of cesarean delivery, parturition, and preeclampsia. We support the idea of

further follow-up studies using more rigorous diagnostic criteria to clarify this poorly

understood condition.

ACKNOWLEDGMENT

Supported in part by the Vicksburg Hospital Medical Foundation, Vicksburg,

Mississippi.

REFERENCES

1. Schulman H. Use of anticoagulants in suspected pelvic infection. Clin Obstet

Gynecol 1969; 12:240–246.2. Ledger WJ, Peterson EP. The use of heparin in the management of pelvic

thrombophlebitis. Surg Gynecol Obstet 1970; 131:1115–1121.3. Josey WE, Staggers SR Jr. Heparin therapy in septic pelvic thrombophlebitis: a

study of 46 cases. Am J Obstet Gynecol 1974; 120:228–233.4. Miller CJ. Ligation or excision of the pelvic veins in the treatment of puerperal

pyaemia. Surg Gynecol Obstet 1917; 25:431.5. Collins CG. Suppurative pelvic thrombophlebitis: a study of 202 cases in which the

disease was treated by ligation of the vena cava and ovarian vein. Am J Obstet

Gynecol 1970; 108:681–687.6. Brown CE, Lowe TW, Cunningham FG, Weinreb JC. Puerperal pelvic

126 Isler et al.

ORDER REPRINTS

thrombophlebitis: impact on diagnosis and treatment using x-ray computed

tomography and magnetic resonance imaging. Obstet Gynecol 1986; 68:789–794.7. Witlin AG, Sibai BM. Postpartum ovarian vein thrombosis after vaginal delivery: a

report of 11 cases. Obstet Gynecol 1995; 85:775–780.8. Brown CE, Stettler RW, Twickler D, Cunningham FG. Puerperal septic

thrombophlebitis: incidence and response to heparin therapy. Am J Obstet Gynecol

1999; 181:143–148.9. Gibbs RS, Jones PM, Wilder R. Antibiotic therapy of endometritis following

cesarean section. Obstet Gynecol 1977; 62:31.10. Dunn LJ, Van Voorhis LW. Enigmatic fever and pelvic thrombophlebitis. Response

to anticoagulants. N Engl J Med 1967; 276:265–268.11. Munsick RA, Gillanders LA. A review of the syndrome of puerperal ovarian vein

thrombophlebitis. Obstet Gynecol Surv 1981; 36:57–66.12. Twickler DM, Setiawan AT, Evans RS, Erdman WA, Stettler RW, Brown CE,

Cunningham, F. Imaging of puerperal septic thrombophlebitis: prospective com-

parison of MR imaging, CT, and sonography. Am J Roentgenol 1997; 169:1039–1043.

13. Dunihoo DR, Gallaspy JW, Wise RB, Otterson WN. Postpartum ovarian vein

thrombophlebitis: a review. Obstet Gynecol Surv 1991; 46:415–427.14. Twickler D, Setiawan A, Harrell R, Brown CE. CT appearance of the pelvis after

cesarean section. Am J Roentgenol 1991; 156:523–526.15. Woo G, Twickler D, Stettler R, Erdman WA, Brown CE. The pelvis after cesarean

section and vaginal delivery: normal MR findings. Am J Roentgenol 1993;161:1249–1252.

Septic Pelvic Thrombophlebitis 127

Copyright of Hypertension in Pregnancy is the property of Taylor & Francis Ltd and its content may not be

copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.