What is the etiology of this acute abdominal pain?
A 63-year-old Caucasian male presented to the emergency department (ED) with five days of fever (maximum 40°C), three days of jaundice, malodorous urine and altered mental status.
He had no significant past medical or surgical history. He had a 30-pack a year tobacco smoking history, but denied alcohol or drug use. Physical exam revealed a drowsy, ill-appearing, jaundiced male in no acute distress. Vital signs were: temperature 40.8°C, BP 153/73, HR 124, RR 24, and oxygen saturation 92% on room air. The patient exhibited dry mucus membranes, scleral icterus, and right upper quadrant (RUQ) tenderness without peritoneal signs.
Laboratory data was significant for a leukocytosis of 15.5 K/μL with an 80% neutrophil predominance, hyponatremia of 128 mmol/L, and creatinine of 1.23 mg/dL elevated above his baseline of 1.00 mg/dL. His International Normalized Ratio (INR) was 1.71. Aspartate transaminase (AST) was 141 IU/L, alanine aminotransferase (ALT) was 153 IU/L, alkaline phosphatase was 239 IU/L, total bilirubin was 16.1 mg/dL, and direct bilirubin was 10.0 mg/dL. Lipase was within normal limits at 46 IU/L. Lactate was minimally elevated at 2.8.
RUQ point-of-care ultrasound (POCUS) revealed a distended gallbladder, normal gallbladder wall thickness and normal common bile duct. Biliary sludge was noted without stones or polyps. The ultrasound team noted sonographic findings of pneumobilia versus portal venous gas.
Intravascular filling defects with absent Doppler flow in the vasculature were highly suggestive of portal vein (PV) thrombosis (figure 1-3).
Computed tomography (CT) scan demonstrated thrombus and gas in the inferior mesenteric vein extending into the main PV and left PV( Figure 4-6). A comprehensive liver duplex scan further confirmed complete occlusion of the left PV and nonocclusive thrombi in the peripheral branches of the right PV.
The patient was started on empiric broad-spectrum antibiotics and therapeutic low molecular weight heparin (LMWH) and admitted to the medical intensive care unit (MICU) with vasopressor support. Blood cultures grew pan-sensitive Escherichia coli.
Despite aggressive resuscitation measures, the patient’s lactate rose from 2.8 to 10.4 in the first 24-hour period. The interventional radiology (IR) team concluded that the risks of a technically challenging thrombectomy outweighed the benefits given the location of the thrombi and the severity of clot burden.
The patient was intubated on hospital day (HD) three for acute hypoxic respiratory failure. Heparin was discontinued on HD5 after the patient passed a large volume of bright red blood from his rectum. The patient was diagnosed with antiphospholipid syndrome (APLS) based on the detection of anti-nuclear antibodies and anti-cardiolipin antibodies and underwent plasma exchange on HD7.
Unfortunately, despite an increasingly aggressive antibiotic regimen, the patient remained bacteremic and required the addition of a second vasopressor on HD9. Interval imaging showed progression of thrombi. He subsequently developed disseminated intravascular coagulation and expired on HD11 from pylephlebitis secondary to catastrophic APLS.
Pylephlebitis is a suppurative thrombosis triggered by intraabdominal infections that ultimately drain into the portal venous system (PVS). The etiology of the root pyle- corresponds to the portal veins.
Most instances are associated with diverticulitis or appendicitis, but the disease also occurs as a complication of pancreatitis, inflammatory bowel disease, cholangitis, and spontaneous bacterial peritonitis. No infection was ever identified in our patient, which is true in less than 1% of documented cases.
Thrombophlebitis begins in the microvasculature at the source of infection and extends proximally to the PVS. More than 40% of patients have an underlying hypercoagulable disorder. Our patient had previously undiagnosed APLS. Hepatic abscesses are seen in 50% of cases and may require percutaneous or surgical drainage.
Other complications result from embolism, such as splenic infarct, and septic pulmonary embolism. Bowel infarct constitutes a large percentage of fatal outcomes, especially in cases involving mesenteric vein thrombosis that exacerbates the low perfusion state of septic shock.
Presentation and Work-Up
Fever and poorly localized abdominal pain are the two most common presenting symptoms, followed by nausea, vomiting, and jaundice. A fifth of patients present in shock. Although not always present, abdominal tenderness, hepatomegaly, and splenomegaly are important physical exam findings.
Laboratory studies often reveal leukocytosis with a left shift and elevations in hepatic enzymes, alkaline phosphatase, and bilirubin. In bacteremic patients, infection tends to be polymicrobial. Blood cultures are positive in 50% to 88% of cases, with microbial isolates being normal bowel flora such as Bacteroides fragilis and Escherichia coli.
Diagnosis of pylephlebitis is made primarily via radiographic studies. Abdominal ultrasonography and contrast-enhanced CT scan both can demonstrate the presence of the pathognomonic combination of gas and thrombus within the PVS.
Linear echogenic areas with posterior reverberations consistent with intraluminal gas can be seen on US. Color Doppler showed hyperechoic material within the lumen of the vein. US is fairly sensitive for PV thrombosis, however, diagnostic accuracy is operator-dependent and can be limited by the presence of bowel gas. CT scan provides the benefit of revealing underlying infectious processes as well as complications such as bowel ischemia or hepatic abscess. Interval CT scans, in this case, did demonstrate the progression of thrombus, but no source of infection.
Antimicrobials and Source Control
Once pylephlebitis is either suspected or diagnosed, treatment with broad-spectrum parenteral antibiotics should be initiated and can be tailored once culture and sensitivity results are available. A standard antibiotic regimen has not yet been established and may differ based on hospital protocol, but coverage should include both gram-negative and anaerobic organisms. Metronidazole, gentamicin, piperacillin, ampicillin, and imipenem have all been utilized successfully.
A minimum antibiotic course of four weeks is recommended for those without a hepatic abscess and six weeks for those with hepatic abscess. Our patient was started on empiric vancomycin and piperacillin-tazobactam in the ED. When cultures grew pan-sensitive E. coli, antibiotics were narrowed to amoxicillin-clavulanic acid and metronidazole. When source control could not be achieved and the patient decompensated, coverage was broadened again to vancomycin and meropenem.
Management of these patients often requires a multidisciplinary approach, involving gastroenterology, radiology, hematology, pharmacy and infectious disease. Although surgical intervention may not be pursued, consultation and involvement of surgical colleagues is prudent.
Patients may require procedures to achieve source control or to treat a complication. Some studies have shown surgical thrombectomy to be associated with a higher risk of the rate of recurrent thrombosis and therefore advise against it. In our case, IR concluded that the risks of thrombectomy outweighed the benefits.
Nearly half of patients with pylephlebitis are found to have an underlying hypercoagulable state, so early consultation with hematology is crucial. Nearly a week into his hospital course, our patient was diagnosed with APLS and started on definitive treatment including high-dose steroids and emergent plasma exchange. However, given his critical condition, these interventions had little time to counteract his prothrombotic state prior to his death.
Anticoagulation (AC) therapy remains controversial given the risk of bleeding. Some providers argue for AC use in patients with mesenteric vein thrombosis since these patients are at increased risk of death from bowel ischemia.
However, AC in these patients with poorly perfused bowel has a high risk of gastrointestinal bleed. Heparin therapy was discontinued in our patient after a large spontaneous bleed.
Outcomes data is underpowered due to the low incidence of pylephlebitis, but data suggests increased recanalization rates and improved mortality rates with the use of anticoagulation. Currently, there is no consensus on duration and no data on the use of thrombolytics.
https://www.amjmed.com/article/S0002-9343(14)00090-4/pdf (clinical presentation)
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3789899/ (clinical presentation)
https://www.hindawi.com/journals/crira/2013/627521/ (surgery indication)
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4882085/ (surgery may increase re-thromb)