Hepatic artery pseudoaneurysm (HAP) is a rare entity and usually described after it ruptures and causes life-threatening gastrointestinal bleeding. Its real incidence is unknown with some reports stating incidence as low as 0.001% (1). HAP occurs when there is loss of integrity of the arterial wall and formation of a false lumen between the tunica media and adventitia of the hepatic artery. Most commonly, this occurs due to atherosclerosis, mechanical injury of the arterial wall due to abdominal trauma or iatrogenic trauma during hepatobiliary procedures, acquired inflammatory process such as vasculitis, or by visceral inflammation or infiltration to the arterial wall from malignancy and, infection, such as pancreatitis, cholangitis, or cholecystitis. Only a small number of cases are classified as spontaneous due to unidentified specific factors. HAP can be anatomically divided into intrahepatic and extrahepatic. HAP following liver transplantation are most commonly extrahepatic (2), while the less common intrahepatic HAP occurs following liver biopsy (3). Clinical presentation varies from asymptomatic and incidentally found to severe gastrointestinal hemorrhage leading to hemorrhagic shock. HAP rupture can cause bleeding into the retroperitoneal space, stomach, or duodenum, venous, portal, or biliary system (presenting as Quicke’s triad with jaundice, biliary colic, and gastrointestinal bleeding) (3). HAP requires elective or emergent hepatic artery repair or ligation irrespective of its symptoms, size or location due to its substantial risk of rupture.

This study’s aim was to describe HAP presentation, etiology, associated liver pathologies, treatment options and mortality at Mayo Clinic.

We conducted a retrospective chart review of patients seen at The Mayo Clinic (Jacksonville, Florida; Rochester, Minnesota; and Phoenix, Arizona) between September 1^st, 1998, and June 30th, 2022. This study was deemed minimal risk and exempt from IRB given its retrospective nature. All patients’ data were deidentified and collected in the excel spreadsheet. The Interventional Radiology department conducted a chart query to identify all patients who had a hepatic artery pseudoaneurysm (HAP) found on CT angiogram radiology reports that mentioned HAP with and without intervention. The final analysis included 27 cases of HAP. Each patient was counted only once, even if they had more than one admission with HAP. For these patients, the index event was the admission for the first HAP treatment.

Patients’ data were collected in excel spreadsheet and included the following: presenting symptoms, laboratory values (including white blood cell count, platelet count, liver function test, and INR) on the day of the index admission, hemodynamic stability (presence or absence of hemorrhagic shock), transfusion requirements as well as if patient was on anticoagulation when HAP was found. We further extracted data regarding the presence of liver necrosis, liver abscess, cholangitis, liver cirrhosis, portal venous thrombosis and, in patients with malignancy if there was tumor invasion causing HAP. Etiology was categorized as surgical (liver transplant or non-liver transplant), non-surgical procedure related, infection/tumor related, trauma or spontaneous. Spontaneous HAP was defined when there was no recent hepatobiliary procedure (liver biopsy, biliary drain placement, Whipple surgery), intrabdominal infection (e.g., cholangitis, pancreatitis), tumor (primary or metastatic), or trauma. HAP etiology was considered a surgical complication if it occurred within 90 days of any type of liver procedure or surgery, and we evaluated days from that procedure to the day HAP was found on the radiology report. HAP interventions and their complications were considered only for the index admission. We followed up the clinical status (alive or deceased) 1 year after the discovery of HAP. The presence of cancer, specified by type and stage, was also evaluated.

The literature search was done by searching PubMed/Medline database using the key words “hepatic artery pseudoaneurysm” to identify articles, including case reports and case series, previously published on this topic. By this strategy, we identified 55 cases from single case reports and case series as presented in Supplementary Table 1. All categorical variables were summarized with the number of patients and the percentage. Continuous variables were illustrated with a mean. Comparisons between patients alive and dead were completed using a Chi-square test for independence. Follow-up was at least 12 months, so a simple comparison was completed. We assume complete follow-up at 12 months. The numbers are too small for any adjustment for confounders. Analysis was completed with SAS version 9.

HAP is a rare condition and is associated with high mortality after its rupture. Due to its rarity, publications are limited to case presentations, case series or reviews related to specific subgroups as illustrated in Supplementary Table 1. Our study is the largest retrospective study in the literature describing 27 patients with HAP from various etiologies, their clinical manifestations, management, and outcome.

HAP is commonly diagnosed after it ruptures and causes GI bleeding, while asymptomatic cases are usually discovered on routine post trauma screening or as a part of comprehensive post-transplant evaluation and monitoring. Although our study found that the need for transfusion in the setting of HAP led to increased mortality risk, hemorrhagic shock following HAP rupture occurred only in a small number of patients and resulted in favorable outcome following treatment with IR coil embolization. Most common location of HAP was found to be intrahepatic (85.2%), and only 14.8% were extrahepatic. These findings are different from what Marshall at co-authors described in 13 liver transplant recipients who found extrahepatic location to be more common (2). Our study points out that the most frequent location of HAP was on the RHA and that was the most frequent site of HAP rupture as well. Prior studies have not reported on this specific data. Due to the small sample size, we were unable to perform subgroup analysis regarding the outcome in relation to the HAP location.

Etiology of HAP is usually related to hepatobiliary and pancreatic procedures, infections, penetrating or blunt abdominal trauma, atherosclerosis, and an exceedingly small fraction are considered spontaneous (4–19). Possible underlying causes of spontaneous HAP could be atherosclerosis, remote history of abdominal trauma and lack of reporting those during history or during publication process, undiagnosed collagen disorders, variance in hepatic artery anatomy and arterial branching. HAP is a subclassification of splanchnic artery aneurisms (SSA), and its risk factors include diabetes mellitus, hypertension, hyperlipidemia, coronary insufficiency, smoking history, pregnancy and previous pancreatitis history (20). It is possible that these same risk factors play a role in the development of HAP especially in cases labeled as spontaneous. Due to the rarity of HAP and the ununified reporting and investigative process to determine the exact causes of spontaneous HAP in previous publications, future reporting should include the presence or absence of these risk factors and a detailed history. A retrospective study from Denmark, which included 259 non-operatively managed patients who sustained blunt abdominal trauma and had appropriate CT follow up, reported a 4% incidence of HAP following abdominal trauma which was not related to the degree and severity of liver injury (21). We describe comparable results with 89.9% of HAP being post-procedural, while the remaining 10.1% categorized as spontaneous. Spontaneous HAP is exceedingly rare but has been reported in several previous cases and tends to have an excellent prognosis (1, 3, 22). Similarly, all patients with spontaneous HAP in our institution had an excellent outcome.

HAP occurrence in patients with hepatobiliary and pancreatic malignancy is associated with biliary stent placement, liver biopsy, percutaneous biliary drainage, tumor invasion, or regionally applied chemotherapy which increase the risk of arterial wall damage, infection, and HAP development (4, 6–12, 15, 17–19). HAP can occur from 5 days to 2 years after biliary plastic stent placement as illustrated in the literature review on the topic of HAP rupture following biliary stent placement (16). Mucosal necrosis from cancer, intraluminal compression of the stent and effects of chemotherapy before and after stenting are thought to be the risk factors for HAP rupture in this setting. In our institution half of the patients with HAP had hepatobiliary or pancreatic malignancy, but that was not associated with increased mortality. Only in 5 cases (18%) HAP occurred after biliary stent placement.

HAP is a rare complication of liver transplantation (LT) procedures with an incidence of 0.3–2.6% (2, 13). Although a rare post-transplant complication, it comprised 26% of all HAP cases in our institution, which may be related to routine surveillance imaging required by this patient population leading to an increased rate of detection. HAP is usually diagnosed within 2 to 3 weeks, and up to 2 months after liver transplantation (2, 13). At Mayo Clinic HAP was diagnosed on an average 22 days (about 3 weeks) post LT. Incidental finding of HAP in post LT screening was higher in our institution, compared to previously reported data (43% vs. 23%) (2). Interestingly, compared to the reported high mortality rate of HAP in LT recipients 69–75% (2, 13), we observed significantly lower mortality of 15%. This observed difference might be related to the rates of infections, which were higher in previous studies compared to our findings. Concomitant infection in HAP in LT recipients was documented to be as high as 85% (2) with substantial number of fungal infections. Contrary to these findings, in our experience at Mayo Clinic, HAP associated with infection was seen in only one LT recipient (15%) which was also the only patient that died. These observed differences in outcome might be related to the expertise of the transplant center itself, improvement in IR techniques and improved surgical techniques and infection prevention over the years. Also, in our institution many incidental HAP diagnosed in routine post-transplant screening led to early and safe interventional radiology treatments which might have contributed to a significant decrease in mortality.

Treatment options for HAP are embolization and stent graft placement. Various embolic agents can be considered depending on anatomic considerations, coagulation profile, and operator expertise. When abscess is present caution should be exercised with stent graft placement to prevent long-term graft infection. Super-selective embolization with microcatheter technique is recommended to prevent non-target embolization. These are typically complex patients, and close follow-up is recommended to monitor for complications and HAP recurrence or incomplete treatment. Ruptured HAP has higher chances of post treatment complications. Out of 25 interventions, we had 3 initial treatment failures that required reintervention, and 2 patients had suspected procedural complications. Despite a few unsuccessful first attempts at HAP treatment, all of them had successful IR reintervention.

The mortality rate after HAP rupture may be as high as 70% (3). HAP with intraperitoneal hemorrhage has been shown to have the highest mortality rate of 82% (3). Most of the previously published data were case presentations or case series and the two largest studies of HAP were related to LT recipients. Previously published data describing HAP related to various etiologies including LT, trauma, biliary stent placement, pancreaticoduodenectomy, percutaneous biliary drainage, splenectomy, gastrectomy, hemi-pancreatectomy, liver resection, idiopathic, reported an overall mortality rate of 34.5% (Supplementary Table 1) (23–30). The overall mortality rate of HAP in Mayo Clinic was 18.5%, despite almost 65% of cases having ruptures. In our study we did not find that HAP rupture, therapeutic anticoagulation, gastrointestinal bleeding, and hemorrhagic shock were associated with increased mortality. In contrast, risk factors associated with increased mortality in our study were need for transfusion, HAP associated with tumor invasion, presence of PVT and liver necrosis. Of the 5 patients (8.5%) who died within 1 year of HAP diagnosis none of the deaths were attributed to HAP. Four of the five deaths were attributed to progression of the underlying malignancy, and while two of these four cases experienced GI bleeding, this was not directly related to HAP. The remaining patient with traumatic HAP died due to pulmonary complications related to the traumatic event. Overall infection rate in our cohort was low, 3%, which is the most prominent difference from previously reported data on patients with HAP and might have had an impact on a higher one-year survival of our cohort compared to previous studies.