Transarterial chemoembolization (TACE) is an effective method for the treatment of unresectable malignant liver tumors (1, 2). With the development of embolization concepts and advances in embolization materials, TACE has become the preferred non-surgical treatment for unresectable primary liver cancer and liver metastases (3). It can embolize multiple lesions in a single course of treatment and can be applied repeatedly to the same patient (4).

However, TACE has a series of complications, including post-embolization syndrome (PES), hepatic impairment, and leukopenia, which are common and mild (5). In contrast, a liver abscess is a rare and serious complication. Its incidence is approximately 0.22%–4.46% (5, 6), and it may prolong hospital stay, delay tumor treatment, and even lead to death from severe infection (7–9). The risk factors for liver abscess after TACE include diabetes, biliary abnormalities, large tumor size, and portal vein occlusion (10–13).

Antibiotics combined with percutaneous catheter drainage (PCD) is the mainstream method for the treatment of liver abscess, but there are only a few studies on the therapeutic effect of liver abscess after TACE (8, 9, 12, 13). In addition, the relationship between TACE-induced liver abscess formation and tumor prognosis remains poorly understood. Abscess formation may promote tumor necrosis, but its inflammatory microenvironment may trigger tumor proliferation, migration, and other malignant processes.

Thus, this study aimed to retrospectively examine the outcomes of patients who developed liver abscesses following TACE to elucidate the impact of liver abscess formation on the prognosis of these patients.

A total of 1,387 patients with malignant liver tumors underwent 3,341 TACE sessions at our hospital between January 2017 and January 2022. Among them, 15 (1.1%) cases were complicated with liver abscess, resulting in a total of 16 (0.5%) liver abscess cases after 3,341 TACE sessions (including one patient with two events) ( Figures 1A–C ). The patients’ baseline characteristics are presented in Table 1 .

Eight (50%) TACE sessions were “first” sessions; the mean number of sessions received was 2.31 ± 2.2 (range, 1–9). The mean maximum tumor diameter was 7.09 ± 2.86 (range, 2.8–12.4) cm. Five, one, and 10 TACE sessions were performed using iodized-oil emulsion (IOE), IOE and gelatin sponge particles, and drug-eluting beads (100–300 μm), respectively. The mean TACE duration was 109.06 ± 37.87 (range, 60–180) min. The mean interval from TACE to the diagnosis of liver abscess formation was 14.19 ± 9.21 (range, 3–40) days. The most common symptom was fever (93.8%), followed by chills (43.8%) and abdominal pain (31.3%). One patient received antibiotics only, while 14 patients received PCD combined with antibiotic treatment ( Figures 1D, E ). The infection caused by liver abscesses was effectively controlled in all patients after treatment. Enterococcus faecalis (31.25%) was the most common bacterial species detected in blood/pus cultures ( Table 2 ).

Follow-up assessments after PCD revealed six (40%) abscess cases were healed, nine (60%) abscess cases did not heal completely, eight patients died before drainage tube removal, one patient retained the drainage tube until the end of the follow-up, and five patients underwent drainage tube removal. The mean drainage tube removal time was 149.17 ± 134.19 (range, 73–420) days. A total of zero, three (18.75%), six (37.5%), and seven (43.75%) cases showed complete response (CR), partial response, stable disease, and progressive disease (PD), respectively ( Figure 1F ). Eleven patients died during the follow-up period owing to causes unrelated to infections caused by liver abscesses ( Table 3 ). The survival rates at 3 months, 6 months, 1 year, and 5 years were 86.7%, 50.9%, 25.5%, and 17%, respectively.

PCD is an effective treatment for liver abscess formation after TACE to control infection (13, 16); however, our study showed that its cure rate was low, mainly due to the inability to remove the drainage tube. In addition, the drainage tube removal time in this study was longer than that in patients with community-acquired liver abscesses (17, 18). Delayed or failed extubation may be the result of the connection between abscesses and bile ducts or even biloma formation (19).

One study has argued that complete liquefactive necrosis of tumor lesions may be indicative of a good prognosis and that the discharge of necrotic tumor tissue fluid via the drainage tube may have positive implications for tumor treatment (13). However, the findings of our study were not consistent with previous findings. In this study, no cases were evaluated as CR, while cases with PD accounted for 43.75%. Furthermore, 11 patients died during the follow-up period. The formation of a liver abscess delayed subsequent treatment and was bound to affect the prognosis, although the infection caused by the abscess could be controlled. Liver abscess was not directly related to death, but it led indirectly to poor prognosis in these patients.

History of biliary-enteric anastomosis is an important risk factor for liver abscess formation (6, 9, 10). In this study, 53.3% of the patients had a history of biliary surgery. Some studies have found that Oddi sphincter dysfunction or incision permitted retrograde intestinal bacteria entry into and colonization of the bile duct (8, 20). Meanwhile, the local toxicity of chemoembolic agents and the embolization of bile duct feeding vessels resulting from TACE may lead to bile duct injury (21). In this study, patients had received an average of 2.31 ± 2.24 TACE sessions at the time of liver abscess formation; among them, one patient developed liver abscesses after the second and ninth TACE sessions. Bile duct injury following TACE may enable opportunistic pathogens to colonize the bile duct and enter the liver parenchyma, where they undergo rapid proliferation within the local ischemic and hypoxic microenvironment after TACE, thus leading to liver abscess formation (6, 22, 23).

Diabetes is also thought to be a predisposing factor for liver abscess formation after TACE.

Diabetes is also thought to be a predisposing factor for liver abscess formation after TACE (9, 10). On the one hand, the chronic inflammatory state that arises in diabetes leads to continuous reactive oxygen species (ROS) production, and ROS stimulate the expression of pro-inflammatory cytokines by activating transcription factors such as nuclear factor-kappa B (24). In addition, in diabetic patients, excess adipose tissue secretes pro-inflammatory cytokines, further amplifying oxidative stress (25). On the other hand, hyperglycemia in diabetes is thought to cause dysfunction of the immune response, which fails to control the spread of invading pathogens in diabetic subjects, through the following mechanisms: impairment of cytokine production, leukocyte recruitment inhibition, defects in pathogen recognition, neutrophil dysfunction, macrophage dysfunction, natural killer cell dysfunction, and inhibition of antibodies and complement effector (26).

In this study, 87.5% of the tumors had a maximum diameter of >5 cm. Large tumor size and extensive use of embolic materials may have simultaneously increased the extent of intrahepatic necrosis and the risk of abscess formation (9). In addition, portal vein tumor thrombosis and gelatin sponge use have been associated with liver abscess formation (8). Owing to the small sample size, we could not examine the association between these risk factors and liver abscess formation.

It has been suggested that patients with liver metastases are more likely to develop liver abscesses after TACE compared with hepatocellular carcinoma (HCC) (27). In HCC, high concentrations of chemoembolic agents in the tumor tissue after infusion can be achieved because the arteries that feed the tumor and intratumoral blood space are mostly dilated. On the contrary, the feeding artery and the intratumoral blood space of metastatic liver tumors are not usually dilated, thus decreasing the intratumoral concentration of chemoembolic agents. This may result in higher concentrations of the chemoembolic agents in the surrounding liver parenchyma and initiation of biliary epithelial damage. Additionally, HCC is mostly noted in cirrhotic livers that are known to have dilation of the perivascular plexus, which can serve as a porto-arterial shunt and compensate for the decreased arterial flow (27, 28). Therefore, they may be the reasons for the higher incidence of liver abscesses in patients with liver metastases.

Prevention of the occurrence of liver abscesses after TACE has been the focus of clinical attention. Incidentally, it is unreasonable and unnecessary to use antibiotics prophylactically for all patients with TACE. It has been shown that prophylactic antibiotics reduce the incidence of liver abscess formation in patients with bile duct injury (29, 30). Moreover, previous studies have shown that the incidence of liver abscess formation does not increase among patients without bile duct injury who do not receive prophylactic antibiotics (30, 31). Therefore, stratification according to risk factors is more reasonable for the administration of prophylactic antibiotics before TACE.

In this study, 15 of 1,387 (1.1%) patients undergoing TACE subsequently developed liver abscesses; this rate was similar to that previously reported (5–13, 16, 19, 27, 32–36) ( Table 4 ). Although the incidence of liver abscess formation after TACE is low, it is necessary to grasp its clinical symptoms. In our study, the most common symptom reported by patients was fever (93.8%), followed by chills (43.8%) and abdominal pain (31.3%). However, PES characterized by fever, right upper quadrant pain, nausea, and vomiting is the most common adverse event, affecting approximately 60%–80% of patients. PES generally occurs immediately after TACE, with fever peaking within 2 days after TACE, and it is often self-limiting and does not require antibiotic treatment (37–40). In contrast, in this study, the mean interval from TACE to the discovery of liver abscess was 14.19 ± 9.21 (range, 3–40) days, suggesting that the onset of liver abscess formation occurred later than PES at the time point after TACE. Furthermore, chills are frequently observed among patients with liver abscesses but not in those with PES, which may help differentiate between these outcomes.

This study has some limitations. First, as this was a single-center study, the results cannot be generalized. Second, as this was a retrospective study, our data integrity may not be as high as that in prospective studies or clinical trials. Third, as some patients after TACE were subsequently lost to follow-up, this study may have underestimated the incidence of liver abscess formation following TACE. Fourth, the low incidence of post-TACE liver abscess and small sample size precluded comparative analyses of prognosis in patients without liver abscess formation. Moreover, we could not use univariate and multivariate Cox regression statistical methods to reach more robust conclusions.

In conclusion, patients with liver abscess formation following TACE for malignant liver tumors had a long drainage tube removal time after PCD and even required continuous canalization. Although liver abscess formation was not directly related to death, it indirectly led to poor prognosis in these patients.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

This study protocol was reviewed and approved by the Ethics Committee of Shengjing Hospital of China Medical University (protocol number 2022PS1175K). The studies were conducted in accordance with the local legislation and institutional requirements. Owing to the retrospective nature of this study, the informed consent requirement was waived.

YW: Writing – original draft. ZC: Data curation, Funding acquisition, Writing – review & editing. JHZ: Formal Analysis, Writing – review & editing. ZL: Data curation, Writing – original draft. JZ: Writing – review & editing.