Septic pelvic thrombophlebitis

INTRODUCTION — Septic pelvic thrombophlebitis (SPT) can occur in the setting of pelvic vein endothelial damage, venous stasis, and hypercoagulability. It is usually associated with postpartum endometritis/parametritis following cesarean deliveries in the setting of chorioamnionitis but can also occur with other conditions, such as pelvic surgery or underlying malignancy. SPT was first described in the late 1800s by Von Recklinghausen [1] and was further elucidated in the 1950s by a series of reports describing a cohort of 70 women who had fever following obstetric or gynecologic procedures and had grossly palpable intravenous thrombi and seropurulent fluid in the pelvis on exploratory laparotomy [2-5]. Surgical excision or ligation of the thrombosed vein was the initial treatment of choice, although medical therapy has since become the preferred approach.

The clinical manifestations, diagnosis, and treatment of SPT are discussed here. Suppurative thrombophlebitis of other veins is discussed elsewhere. (See "Catheter-related septic thrombophlebitis" and "Pylephlebitis" and "Septic dural sinus thrombosis".)

EPIDEMIOLOGY

Incidence — Septic pelvic thrombophlebitis (SPT) is a rare complication of pregnancy. One survey in the United States suggested that the incidence of SPT was 1 in 3000 deliveries (1 in 9000 vaginal deliveries and 1 in 800 cesarean deliveries) [6]. Similarly, in one study of over 73,000 women who underwent cesarean delivery or vaginal delivery after prior cesarean, only 89 (0.1 percent) had suspected or documented SPT [7].

Risk factors — SPT is largely a condition of postpartum women. The risk is higher with cesarean compared with vaginal delivery. Women with peripartum or postpartum pelvic infections, such as endometritis or chorioamnionitis, are also at higher risk for SPT. As an example, in one case control study of over 73,000 women in a pregnancy registry, cesarean delivery and chorioamnionitis were each independently associated with SPT (adjusted odds ratios 6.3 and 4.8, respectively) [7].

Other pregnancy-related risk factors include [4,7-14]:

●Induced abortion

●Maternal age <20 years

●Multiple gestation

●Preeclampsia, although the high rates of cesarean delivery in this setting make it difficult to distinguish between these risk factors

SPT has more rarely been reported in nonpregnant women or in men [6,13,15,16]. In these cases, SPT is often associated with other pelvic infections (eg, pelvic inflammatory disease), pelvic surgery (eg, hysterectomy), uterine fibroids, hormonal stimulation with gonadotropins, or underlying malignancy.

PATHOGENESIS — The physiologic conditions in the postpartum setting fulfill Virchow's triad for the pathogenesis of thrombosis (ie, endothelial damage, venous stasis, and hypercoagulability), resulting in an increased risk for septic pelvic thrombophlebitis (SPT) (see "Overview of the causes of venous thrombosis"):

●Endothelial damage can occur as a result of intrapartum trauma to vascular structures, as a result of uterine infection, or from pelvic surgery itself. The pathophysiology of SPT was first described in the 1950s in a series of reports describing a cohort of 70 women with fever following obstetric or gynecologic procedures [2-5]. Histopathologic evaluation showed perivascular and intimal inflammatory exudate and frequent microabscesses, but rare bacteria.

●Venous stasis occurs as a result of pregnancy-induced ovarian venous dilatation and low postpartum ovarian venous pressures [17]. These changes can lead to retrograde (left to right) ovarian venous flow, perhaps explaining why ovarian vein thrombophlebitis is observed more frequently on the right than on the left [18-20].

●The hypercoagulable state of pregnancy completes Virchow's triad for the pathogenesis of thrombosis [1,17,21,22].

MICROBIOLOGY — Because blood cultures are frequently negative in septic pelvic thrombophlebitis (SPT) and tissue for culture is not typically obtained, the microbiology is not well defined. However, based on limited data, the spectrum of pathogens associated with SPT appears similar to those for other pelvic infections and includes streptococci, Enterobacteriaceae, and anaerobes.

In a review of 158 cases of postpartum ovarian vein thrombophlebitis (a type of SPT), bacteria were isolated in 34 cases (22 percent), of which 14 were Streptococcus spp [23]. Sites of isolation included the thrombosed ovarian vein, cervix, endometrium, urine, and blood.

Other organisms have also been reported with SPT, including methicillin-resistant Staphylococcus aureus [24].

CLINICAL FEATURES

Presentation — Patients with SPT can present on a spectrum, with two somewhat distinctive syndromes:

●Ovarian vein thrombophlebitis – Some patients present acutely ill, with fever and abdominal pain within one week after delivery or pelvic surgery [8,20,25]. Abdominal pain localizes to the side of the affected vein (usually the right) but can be felt in the flank or back. The pelvis is tender to palpation, and some patients may have a tender rope-like mass on examination that extends centrally from the uterus to the upper lateral abdomen. Nausea, ileus, and other gastrointestinal symptoms may occur but are usually mild. This presentation is mostly associated with SPT involving the ovarian vein (ie, ovarian vein thrombophlebitis), which can often be visualized radiographically.

●Deep septic pelvic thrombophlebitis – Some patients present more subtly, with fever in the early postpartum or postoperative period (usually within three to five days, but onset may be delayed to up to three weeks following delivery) [9-11]. Patients are usually not acutely ill; fever or chills may be the only symptoms, and patients appear clinically well between fever spikes [26]. Abdominal or pelvic tenderness is notably absent. Radiographic imaging typically does not identify thrombosis of a specific vein. These patients are usually identified because of persistent fever despite appropriate antibiotics for presumed endometritis or other pelvic infection and subsequent response to anticoagulation for SPT. (See 'Treatment' below.)

Leukocytosis of >12,000/microL occurs in 70 to 100 percent of patients with SPT [8,20]. However, this is not a specific finding, and leukocytosis can be seen in the uncomplicated postpartum state. Blood cultures are frequently negative in the setting of SPT. The rate of positive blood cultures ranges from 29 percent in the earliest report to 3 percent in later series [1,8].

Imaging findings are discussed elsewhere. (See 'Imaging' below.)

Complications — Pulmonary emboli in the setting of SPT occur uncommonly, in about 2 percent of cases; when they do occur, they tend to be small and rarely cause hypoxia [1,4,5,8,10,11,26].

Case reports have described other potentially morbid thrombotic complications, including retrograde migration of thrombosis from the iliofemoral system into the legs, extension proximally to the diaphragm and distally to the femoral vein, free-floating thrombosis in the vena cava, renal vein thrombosis, and ureteral obstruction [27-32].

DIAGNOSTIC APPROACH

Initial evaluation — Septic pelvic thrombophlebitis (SPT) should be suspected in patients who have persistent fever of at least three to five days duration, despite antibiotic therapy (usually for presumptive endometritis or other pelvic infection), and no evidence of abscess in the few weeks following vaginal delivery, cesarean delivery, or pelvic surgery.

Physical examination should focus on the abdominal examination to assess for fundal tenderness, lower abdominal tenderness, pelvic tenderness, or mass. A palpable cord in the abdomen (an uncommon finding) should heighten the suspicion for SPT. Pelvic examination (examination of external genitalia, speculum examination, and bimanual examination) is usually poorly tolerated immediately postpartum.

Because SPT is a rare condition, the initial evaluation should also include history and examination to assess for other common causes of postpartum or postoperative infection, including infections at other sites, such as wound or surgical site infection, respiratory tract infections, urinary tract infection, and catheter-associated infection. (See "Postpartum endometritis", section on 'Differential diagnosis' and "Fever in the surgical patient", section on 'Clinical evaluation'.)

Basic laboratory testing to evaluate postpartum or postoperative fever usually includes a complete blood count with differential, urinalysis with culture, and blood cultures. Although blood cultures are usually negative in the setting of SPT, positive results can lead to identification of an alternative diagnosis and/or guide antimicrobial therapy.

Unless this initial evaluation reveals an evident cause of persistent fever, imaging is generally performed to evaluate for SPT as well as other potential complicated infections, such as pelvic abscess, as discussed in further detail below.

Imaging — Some UpToDate contributors obtain abdominopelvic imaging in all patients with suspected SPT. The optimal imaging modality is uncertain. In general, those contributors use computed tomography (CT). Magnetic resonance imaging (MRI) is a reasonable alternative. These modalities are most useful for identification of ovarian vein thrombophlebitis (OVT) specifically. Imaging also has the benefit of providing information on alternate potential diagnoses. A negative imaging study cannot exclude SPT, however, as smaller deep pelvic vessels, such as uterine and cervical branches, are not well visualized on either CT or MRI. When imaging is performed, it is also important to obtain appropriate views to optimize sensitivity [33]. Ultrasonography should not be used for the diagnosis of SPT if CT or MRI is available [34,35].

●If CT is performed, it should be with contrast and a dedicated venous phase in order to optimize visualization of the vessels. Typical CT findings associated with SPT include enlargement of the involved vein, sharp enhancement of the vessel wall, and a low-density vessel lumen or filling defect from a thrombus (image 1) [36-39]. The venous wall is often thickened and enhancing, with perivenous stranding.

The ideal MRI study is gadolinium-enhanced magnetic resonance venography (MRV) to highlight any pelvic vein-filling defect, although nonenhanced MRI can also be useful [39]. Thrombosed vessels appear bright on MRI while vessels with normal blood flow appear dark [40].

OVT is most often identified on the right side, as that vein is longer and is more likely to be compressed by the uterus. Many patients with ovarian thrombophlebitis also have unrecognized involvement of other pelvic veins. In the original series of 70 SPT cases, for example, only 16 (23 percent) had isolated ovarian vein involvement [4].

Rigorous evaluation of imaging studies for diagnosis of SPT is challenging because diagnostic confirmation with laparotomy is usually not performed, so there is no gold standard in most studies evaluating the efficacy of imaging. One retrospective study of postpartum with persistent fever despite 72 hours of antibiotic therapy highlighted the utility of CT imaging; among 238 women, 60 percent had abnormal findings, and CT resulted in a management change in 30 percent [41]. Another large study compared CT, MRI, and ultrasonography in 64 women who met clinical criteria for SPT [34]. Both ovarian veins were visualized by CT and MRI in all patients, and a diagnosis of OVT was made in 12 patients with excellent agreement between CT and MRI. Ultrasonography with Doppler color flow was less useful (only 52 right ovarian veins and 23 left ovarian veins were visualized) and identified only 6 of these 12 cases.

Pelvic imaging studies should be interpreted with caution, however, as findings suggestive of pelvic venous thrombosis can be seen in the absence of suspected SPT, and clinical correlation is thus essential. In a study of 30 asymptomatic women at low risk for venous thromboembolism who agreed to undergo screening pelvic MRV within four days following a spontaneous vaginal delivery, two-thirds had evidence of nonobstructing pelvic vein thrombosis (definite in nine, probable in eight, and possible in three women) [42]. One hypothesis is that the natural course of uterine involution following delivery could involve pelvic vein thrombosis, which would account for the high detection rate found in this study despite the absence of symptoms or clinical suspicion [43].

Establishing the diagnosis — For patients who present with early postpartum or postoperative fever and abdominal pain, the diagnosis of SPT can be made relatively easily if they have a palpable cord-like mass on exam and/or have an ovarian vein (or other deep pelvic vein) thrombus on imaging. However, because a palpable cord is an uncommon finding, imaging is often negative, and there is no definitively diagnostic laboratory test, the diagnosis of SPT can be challenging and is often one of exclusion.

For patients who present following vaginal or cesarean delivery or pelvic surgery with persistent fever despite antibiotic therapy (eg, for presumed endometritis) and no other apparent cause, we treat for possible SPT with the addition of anticoagulation, even if no thrombosis is identified on imaging. The diagnosis of SPT can then be presumed if they defervesce within 48 hours of empiric systemic anticoagulation [8-12,20,26,27,44].

The diagnosis of SPT can be definitively made by finding a palpable intravenous thrombus and seropurulent fluid on exploratory laparotomy, but this is rarely performed.

Evaluation for hypercoagulability — The optimal approach to evaluation of hypercoagulability in the setting of SPT is not known. We do not routinely perform a hypercoagulability workup (ie, test for acquired and inherited risk factors for thrombosis) in patients with pregnancy-related SPT who do not have any other history of thromboembolic disease. The physiologic changes associated with pregnancy and the immediate postpartum state are sufficient on their own to precipitate SPT [45]. (See 'Pathogenesis' above.)

The approach to evaluation for hypercoagulability in patients with non-pregnancy-related SPT is the same as that for other adult patients with venous thromboembolic disease and is discussed elsewhere. (See "Evaluating adult patients with established venous thromboembolism for acquired and inherited risk factors".)

DIFFERENTIAL DIAGNOSIS — Most women who present with postpartum fever, abdominal pain, and pelvic tenderness will be diagnosed with endometritis in the absence of any other evident cause. The possibility of septic pelvic thrombophlebitis (SPT) is usually entertained after fever persists despite antibiotic therapy. (See "Postpartum endometritis".)

Otherwise, the main differential diagnoses for persistent postpartum or post-pelvic surgery fever despite antibiotics include:

●A complicated pelvic infection (such as a pelvic abscess or infected hematoma in the pelvis) that is not adequately treated with antibiotics alone. This can be distinguished on pelvic imaging. (See "Posthysterectomy pelvic abscess", section on 'Diagnosis'.)

●Endometritis with an organism resistant to the chosen antibiotic regimen. Persistent fever should prompt reevaluation of any culture results and assessment of the adequacy and dosing of the antibiotics. In the absence of culture data, the antibiotic regimen can be broadened empirically. (See "Postpartum endometritis", section on 'Persistent postpartum fever'.)

●An alternative source of infection not adequately treated by the chosen antibiotic regimen (because of a different microbial spectrum or drug resistance). This includes skin and soft tissue infection, pneumonia, pyelonephritis, appendicitis, Clostridioides difficile infection, urinary tract infection, and catheter-associated infection. Some of these can be diagnosed or excluded by history and physical examination alone; for others, laboratory studies (eg, urinalysis and culture, stool testing, blood cultures) and/or imaging studies (eg, chest radiograph) can clarify the diagnosis.

TREATMENT — In the early reports of septic pelvic thrombophlebitis (SPT), surgical excision or ligation of the thrombosed vein was the treatment of choice [46-48]. Antibiotic therapy in conjunction with systemic anticoagulation has since become the most common treatment for this disorder.

Site of care — Patients with suspected or documented SPT warrant inpatient evaluation and management.

Antibiotics — For patients with SPT, we suggest parenteral antibiotic therapy that is active against streptococci, Enterobacteriaceae, and anaerobes. Many patients will already be on antibiotics, such as gentamicin plus clindamycin or ampicillin-sulbactam, to treat endometritis or other presumed pelvic infections, which usually involve the same microbial spectrum.

Other appropriate options include:

•Piperacillin-tazobactam 4.5 g every 8 hours

•A carbapenem (eg, ertapenem 1 g every 24 hours, imipenem 500 mg every 6 hours, or meropenem 1 g every 8 hours)

•Ceftriaxone 1 g every 24 hours plus metronidazole 500 mg every 8 hours

•Cefepime 2 g every 8 hours plus metronidazole 500 mg every 8 hours

For patients who have allergies that preclude the use of a beta-lactam (such as those with an IgE-mediated allergy to penicillin when skin testing or test dose administration is unavailable), alternative regimens include a fluoroquinolone (eg, ciprofloxacin 400 mg every 12 hours or levofloxacin 500 mg every 24 hours) plus metronidazole 500 mg every 8 hours. (See "Choice of antibiotics in penicillin-allergic hospitalized patients", section on 'Management of reactions WITH features of immediate allergy (some risk)'.)

Antibiotic selection is similar to that for intraabdominal infections in general. (See "Antimicrobial approach to intra-abdominal infections in adults", section on 'Approach to empiric antibiotic selection'.)

The optimal duration of antibiotic therapy is unknown. We continue antibiotics until the patient has clinically improved, with no fever for at least 48 hours, and leukocytosis has resolved. This usually takes several days to a week.

Complications such as septic emboli or positive blood cultures generally warrant longer therapy; a minimum of one to two weeks is typical, although the precise duration is not well defined and depends on the clinical course. For patients with positive blood cultures, the extended regimen can be tailored to the specific identified organism.

Anticoagulation — We suggest systemic anticoagulation in addition to antibiotic therapy for the management of SPT. Anticoagulation is thought to prevent further thrombosis and reduce the spread of septic emboli [1,8-11,49,50].

Initial anticoagulation is with either unfractionated heparin or with low-molecular-weight heparin. There have been no studies comparing the different formulations of heparin for the treatment of SPT, so the choice between them is at the discretion of the clinician and patient [1]. Standard dosing of unfractionated heparin for management of SPT is an initial bolus of 5000 units followed by continuous infusion of 16 to 18 units/kg for a goal partial thromboplastin time of 1.5 to 2.0 times the patient's baseline [8,10]. Low-molecular-weight heparin dosing is standard (eg, enoxaparin 1 mg/kg subcutaneously every 12 hours). (See "Use of anticoagulants during pregnancy and postpartum".)

The optimal duration of anticoagulation therapy in SPT is uncertain. We individualize the duration depending on clinical features:

●For patients who do not have documented thromboses or a known underlying hypercoagulable state, we discontinue anticoagulation 48 hours after the resolution of fever.

●For patients who do have documented thromboses, the minimum duration of anticoagulation depends on the extent of thrombosis:

•For patients who have radiographic evidence of pelvic branch vein thromboses, we continue anticoagulation with low-molecular-weight heparin for at least two weeks.

•For patients who have radiographic evidence of extensive pelvic thromboses (eg, thrombosis involving the ovarian vein, iliac veins, or vena cava) or septic emboli, we continue anticoagulation with low-molecular-weight heparin or an oral anticoagulant for at least six weeks. Six weeks is likely adequate for pelvic vein thrombosis related to pregnancy or other transient processes. Anticoagulation for longer than six weeks may be appropriate for patients with embolic disease outside the pelvis or with more chronic pro-thrombotic risk factors (including factor V Leiden trait or hypercoagulability related to malignancy); we typically consult with a hematologist regarding the optimal duration. (See "Venous thromboembolism: Anticoagulation after initial management", section on 'Duration of treatment'.)

Not all clinicians favor anticoagulation for SPT [12,20]. There are very few studies that have evaluated it, and one small randomized trial did not demonstrate a benefit. In that trial, 14 women with persistent postpartum fever for five days despite antibiotics and radiographically confirmed SPT were randomly assigned to treatment with heparin plus antibiotics (n = 6) or antibiotics alone (n = 8) until they were afebrile for 48 hours [12]. There were no differences in the time to defervescence (five to six days) or duration of hospitalization between the two groups. However, the study was powered to assess fever resolution only; it did not address the potential benefit of anticoagulation with regard to risk of thrombus extension, pulmonary embolus, or associated thrombophilias [51].

Other observational data suggest that anticoagulation is safe. In a study of 46 patients with SPT treated with heparin, 42 (91 percent) patients (including 11 with pulmonary emboli) had resolution of clinical symptoms within one week; there were no relapses and no adverse effects related to anticoagulation [9].

Further study is needed to address the full role and treatment duration of anticoagulation definitively.

PROGNOSIS — The mortality due to septic pelvic thrombophlebitis (SPT) is very low and has usually been attributable to overwhelming systemic infection in cases complicated by septic emboli. In a study including 69 cases of SPT among nearly 45,000 deliveries, no deaths were observed [12].

The rate of recurrent SPT is approximately 3 per 100 patient-years. The recurrence can involve the contralateral ovarian vein, left renal vein, or inferior vena cava [6]. A history of SPT does not appear to confer increased maternal or fetal risk in subsequent pregnancies [2,10,12].

SUMMARY AND RECOMMENDATIONS

●Septic pelvic thrombophlebitis (SPT) is a rare condition that occurs in the setting of pelvic vein endothelial damage, venous stasis, and hypercoagulability (usually postpartum or following pelvic surgery). (See 'Epidemiology' above and 'Pathogenesis' above.)

●The microbiology is not well defined but generally reflects the polymicrobial spectrum of other pelvic infections and includes streptococci, Enterobacteriaceae, and anaerobes. (See 'Microbiology' above.)

●Patients with SPT can present with two somewhat distinctive syndromes.

•Patients with ovarian vein thrombophlebitis are usually acutely ill, with fever and abdominal pain within one week after delivery or surgery; thrombosis of the ovarian vein (usually on the right side) can sometimes be visualized radiographically.

•Patients with deep SPT present more subtly with isolated fever in the early postpartum or postoperative period and, notably, do not have abdominal or pelvic tenderness, and radiographic imaging typically does not identify thrombosis of a specific vein.

Blood cultures are infrequently positive. Pulmonary emboli are an uncommon complication, only occurring in about 2 percent of cases. (See 'Clinical features' above.)

●SPT should be suspected in patients who have persistent fever despite antibiotic therapy (usually for presumptive endometritis or other pelvic infection) in the few weeks following vaginal delivery, cesarean delivery, or pelvic surgery. If initial history and physical examination do not reveal an evident cause of fever, we perform abdominopelvic imaging to evaluate for SPT and/or other potential processes. Some UpToDate contributors typically use computed tomography (CT); magnetic resonance imaging (MRI) is a reasonable alternative. These are most useful for identification of ovarian vein thrombophlebitis and alternative causes of fever, but a negative imaging study does not exclude SPT. (See 'Initial evaluation' above and 'Imaging' above.)

●For patients who present with early postpartum or postoperative fever and abdominal pain, the diagnosis of SPT can be made relatively easily if they have a palpable cord-like mass on exam (an uncommon finding) and/or have an ovarian vein (or other deep pelvic vein) thrombus on imaging. Otherwise, we empirically treat cases of suspected SPT even if no thrombosis is identified on imaging; the diagnosis of SPT can be presumed if they defervesce within 48 hours of systemic anticoagulation. (See 'Establishing the diagnosis' above.)

●We suggest parenteral antimicrobial therapy with activity against streptococci, Enterobacteriaceae, and anaerobes (Grade 2C). Many patients will already be on antibiotics to treat endometritis or other presumed pelvic infections. We discontinue antibiotics 48 hours following clinical improvement (ie, defervescence and resolution of leukocytosis). A longer course of antibiotics is appropriate in patients who have complications such as septic emboli or positive blood cultures. (See 'Antibiotics' above.)

●We suggest systemic anticoagulation for patients with suspected, presumptive, or documented SPT (Grade 2C). The duration of anticoagulation depends on the clinical presentation. For patients who do not have documented thromboses or underlying hypercoagulable state, we discontinue anticoagulation 48 hours after the resolution of fever. For patients who have radiographic evidence of pelvic branch vein thromboses, we continue anticoagulation for at least two weeks. For patients who have radiographic evidence of extensive pelvic thromboses (eg, thrombosis involving the ovarian vein, iliac veins, or vena cava) or septic emboli, we continue anticoagulation for at least six weeks. Six weeks is likely adequate for pelvic vein thrombosis related to pregnancy or other transient processes. A longer period of anticoagulation may be appropriate for those with embolic disease outside the pelvis or more chronic pro-thrombotic risk factors. (See 'Anticoagulation' above and "Venous thromboembolism: Anticoagulation after initial management".)

●Morbidity and mortality associated with SPT are low. There does not appear to be a risk of recurrent SPT in subsequent pregnancies. (See 'Prognosis' above.)