A prospective study regarding the complications of transcatheter intraarterial lipiodol chemoembolization in patients with hepatocellular carcinoma




Hepatocellular carcinoma (HCC) is a common cause of cancer death throughout the world. The majority of patients are not suitable for curative resection either because of the advanced stage of the disease at the time of presentation or because of underlying cirrhosis. Transcatheter intraarterial lipiodol chemoembolization (TACE) has been reported to be one of the most effective palliative measures for HCC. However, its severe side effects continue to make its use controversial.


In the current study, the authors prospectively evaluated 197 sessions of TACE performed in 59 patients with HCC.


Acute hepatic decompensation occurred in 20% of the 197 sessions with 3% of cases being irreversible. Significant elevation of bilirubin was associated with the dosage of cisplatin used (P = 0.0001), basal bilirubin level (P = 0.0001), basal prothrombin time (P =0.004), basal aspartate aminotransferase (AST) level (P = 0.013), and stage of cirrhosis (P < 0.0001). Patients with irreversible hepatic decompensation were more likely to have higher pre-TACE bilirubin levels (P = 0.009), more prolonged prothrombin time (P = 0.015), received a higher dose of cisplatin (P = 0.033), and more advanced cirrhosis (P < 0.0001). The majority of the other side effects were self-limiting with the exception of one patient who died of liver and splenic abscesses. Approximately 36% of the patients achieved a tumor response, 39% achieved stable disease, and 29% developed progressive disease.


The results of the current study identified factors that appeared to predispose patients to irreversible hepatic decompensation after TACE. Despite the high percentage of patients who developed hepatic decompensation after TACE, irreversible damage occurred in only a minority. Cancer 2002;94:1747–52. © 2002 American Cancer Society.

DOI 10.1002/cncr.10407

Hepatocellular carcinoma (HCC) is a common cause of cancer death throughout the world. Despite the availability of screening procedures for the detection of early HCC (i.e., ultrasonography and serum α-fetoprotein levels), the majority of patients still are not considered suitable for curative resection.1

Transcatheter intraarterial lipiodol chemoembolization (TACE) is considered to be one of the most effective palliative measures for patients with inoperable HCC. However, the severity of its side effects makes its use controversial. The majority of earlier series published to date reported relatively few side effects, with TACE being well tolerated.2–5 A later study from Europe6 reported that approximately 60% of patients had at least 1 episode of acute liver failure within a few days of TACE. In contrast, in a series of 132 patients from the University of Hong Kong, the incidence of liver failure was reported to be only 1.5%.7

The objective of the current study was to investigate prospectively the side effects and complications of TACE in Chinese patients with unresectable HCC.


This was a single-center, prospective study, recruiting 59 consecutive patients with unresectable HCC who required TACE. These individuals were patients who were admitted to the Department of Medicine at Queen Mary Hospital between January 1, 1998 and August 31, 1999. The diagnosis of HCC was made either histologically (n = 19) or by hepatic arteriogram and computed tomography scan combined with an elevated α-fetoprotein level (n = 40).

TACE procedures were performed by experienced radiologists. After conventional hepatic angiography, a vascular catheter was inserted superselectively into the branch of the hepatic artery that was believed to feed the tumor. Chemoembolization then was performed. Ten milliliters of lipiodol was mixed with 10 mg of cisplatin and emulsified to create a milky solution. The emulsion slowly was infused into the liver. The dose given depended on the size of the HCC tumor, the position of the catheter, the patient's liver function, and the response to previous courses of TACE. Gelatin pellets of 1 mm × 1 mm mixed with contrast medium and 40 mg of gentamicin were used for embolization. The embolization procedure would be abandoned if the liver function tests deteriorated and approached predetermined limits. Each patient received repeated courses of TACE at intervals of 8–12 weeks if they fulfilled the following criteria: bilirubin level ≤ 50 umol/L, prothrombin time ≤ 5 seconds above the control value, a serum creatinine concentration < 200 umol/L, no portal vein thrombosis, and no extensive arteriovenous shunting. TACE would be stopped when there was angiographic evidence of complete resolution of the HCC or when the liver function deteriorated irreversibly.

A second-generation cephalosporin, usually cefuroxime, was given 1 day before TACE was initiated and continued for 2 days after the procedure or until the patient became afebrile. Blood was taken for liver biochemistry, prothrombin time, and α-fetoprotein level before and on Day 2 and Day 5 after the procedure. Blood cultures were taken if the patient's temperature became > 38 °C. If the fever persisted for > 7 days, the antibiotics were withdrawn and the sepsis workup was repeated. This included blood, sputum, and urine cultures, as well as chest X-ray. Ultrasonography and/or computed tomography were performed for patients in whom intraabdominal sepsis was suspected. Antibiotics were changed to a third-generation cephalosporin after the sepsis workup. Ultrasonography of the femoral artery was performed to detect hematoma, pseudoaneurysm formation, or arterial thrombosis if clinically suspected. Emergency upper endoscopy was performed if clinical gastrointestinal bleeding and/or evidence of a decreased hemoglobin level was detected. The patient was discharged only when he/she was afebrile.

Acute hepatic decompensation was defined by the occurrence of any one of the following: the development of encephalopathy, increasing ascites, increase in the prothrombin time by > 3 seconds of the level before TACE, increase in the serum bilirubin level to twice the upper limit of normal (i.e., 38 umol/L) if the pre-TACE level was normal, or an increase to twice the basal level if the pre-TACE level was abnormal. Irreversible hepatic decompensation was defined as the failure of bilirubin to return to either the normal level or the pre-TACE level before the next TACE procedure, the failure of the prothrombin time to return to the pre-TACE level, or the inability to withdraw antiencephalopathy treatment.

Data were analyzed using the Statistical Package for the Social Sciences 10.0 for Windows software (SPSS, Inc., Chicago, IL). Comparisons were performed using the independent sample Student t test and the chi-square test. The relation between two continuous variables was analyzed by correlation. The statistical level was considered to be significant at P ≤ 0.05.


Forty-eight men and 11 women were recruited into the current study. The mean age of the patients was 61 ± 11 years. Forty-seven patients (79.7%) were carriers of hepatitis B, 9 patients (15.3%) were carriers of hepatitis C, and 2 patients (3.4%) were chronic alcohol drinkers. Only one patient had HCC of an unknown cause. Approximately 49% (29 of 59), 27% (16 of 59), and 24% (14 of 59) of patients, respectively, had Child A, B, and C cirrhosis.

The total number of TACE sessions was 197. The average number of TACE sessions received by each patient was 3.4 (range, 1–7 sessions). TACE was the first-line treatment in 52 patients. In the remaining seven patients, TACE was the second-line treatment either because of HCC recurrence after surgery or failure of percutaneous ethanol injection. The mean duration of hospitalization was 3.12 ± 3.7 days.

The mean dosage of lipiodol and cisplatin injected were 7.7 ± 5.2 mL and 7.7 ± 5.2 mg, respectively. Embolization was performed in 44 of 197 procedures. Embolization was not performed in the remaining sessions either because the liver function was approaching the upper limit of the entry criteria or the blood flow in the hepatic artery was sluggish. The frequency and nature of the side effects are listed in Table 1. The changes in the biochemical parameters after TACE are shown in Table 2. The most frequent side effects encountered were fever, nausea/emesis, and abdominal pain. The mean duration of fever was 3.12 ± 3.7 days (range, 0–30 days) and the mean highest temperature was 37.62 ± 2.8 °C. The total duration of fever was found to correlate positively with the initial tumor size (P = 0.007; correlation coefficient [r] = 0.43), the total dosage of cisplatin (P < 0.0001; r = 0.4), and pretreatment AST (P = 0.002; r = 0.22). The duration of fever was reported to be significantly longer in patients undergoing gel foam embolization compared with those patients not undergoing embolization (6.6 ± 4.8 days vs. 2.2 ± 2.7 days; P ≤ 0.0001). The total duration of fever was found to be correlated negatively with the number of TACE sessions performed (P < 0.0001; r = -0.3). Nausea/emesis and abdominal pain were self-limiting in the majority of patients and were controlled by the administration of antiemetics and painkillers.

Table 1. Complications from the Chemoembolization Procedure
ComplicationFrequency per 197 procedures
  1. GI: gastrointestinal.

Fever146 (74%)
Positive blood culture3 (1.5%)
Abdominal pain89 (45.2%)
Nausea/emesis116 (58.9%)
Bleeding or hematoma13 (6.6%)
Variceal bleeding2 (1%)
GI bleeding from peptic ulcer or gastritis6 (3%)
Acalculous cholecystitis1 (0.5%)
Liver abscess1 (0.5%)
Splenic abscess1 (0.5%)
Development of ascites2 (1%)
Hepatic encephalopathy2 (1%)
Table 2. Change in Biochemical Parameters after the Chemoembolization Procedure
Biochemical changeNo. of episodes
  1. AST: aspartate aminotransferase; ALT: alanine aminotransferase; TACE: transcatheter intraarterial chemoembolization; AFP: α-fetoprotein.

> 50% increase in AST106/197 (53.8%)
> 50% increase in ALT108/197 (54.8%)
> 50% increase in creatinine1/197 (0.5%)
> 3 seconds increase in prothrombin time12/177 (6.8%)
Bilirubin > 38 umol/L if pre-TACE level was normal, or ≥ twice the basal level if pre-TACE level was abnormal29/197 (14.7%)
≥ 25% decrease in AFP43/188 (22.9%)
≥ 50% decrease in AFP13/188 (6.9%)

Acute hepatic decompensation was observed in 17 patients after 39 sessions of TACE (20% of the total given); 3 patients developed ascites, 2 developed hepatic encephalopathy, 12 demonstrated an increase in prothrombin time, and 29 demonstrated an increase in the bilirubin level. In six of these sessions, more than one of the criteria for hepatic decompensation was met. Among the 39 sessions of TACE that were complicated by acute hepatic decompensation, gel foam was used in 13 courses (33.3%), the mean dosage of cisplatin was 9.5 ± 5.9 mg, the median basal bilirubin level was 23 umol/L (range, 5–45 umol/L), and the mean basal prothrombin time was 14 ± 2.1 seconds. In six of these episodes, liver function tests did not return to baseline level at the time of the next TACE session. Therefore, irreversible hepatic decompensation was determined to have occurred after 6 of 197 TACE sessions (3%). None of the patients died within 1 month of TACE. The bilirubin levels of all the patients with reversible hepatic decompensation returned to baseline during the post-TACE follow-up period (mean duration of 56 days; range, 4–88 days).

The effects of procedure-related factors and patient factors on hepatic decompensation were analyzed by univariate analysis (Table 3). It was found that the dosage of cisplatin, the basal bilirubin level, the basal prothrombin time, the basal AST level, and the Child stage of cirrhosis were associated significantly with the post-TACE increase in bilirubin. When comparing patients with irreversible hepatic decompensation with those patients without hepatic decompensation, the former group had significantly higher pre-TACE bilirubin levels, had a longer prothrombin time, had received larger doses of cisplatin, and had a more advanced stage of cirrhosis (Table 4).

Table 3. Univariate Analysis of Factors Leading to Hepatic Decompensation
FactorsIncrease in bilirubin according to the definition of hepatic decompensationIncrease in PT > than 3 secondsDevelopment of ascites after TACEDevelopment of encephalopathy after TACE
  • PT: prothrombin time; TACE: transcatheter arterial chemoembolization; NA: not available; AST: aspatate aminotransferase; ALT: alanine aminotransferase.

  • a

    A P value < 0.1 but > 0.05 was considered to be approaching statistical significance.

  • b

    Boldface type indicates a statistically significant P value of < 0.05.

Initial tumor sizeP = 0.15P = 0.29NAP = 0.8
Total no. of TACEP = 0.11P = 0.47P = 0.20P = 0.60
Presence of gelfoamP = 0.26P = 0.078aP = 0.30P = 0.16
Total dosage of cisplatinP= 0.0001bP = 0.54P = 0.72P = 0.77
Prebilirubin levelP= 0.0001bP = 0.44P = 0.48P= 0.031b
Preprothrombin time levelP= 0.004bP = 0.84P= 0.036bP = 0.24
Pre-TACE Childs stagingP< 0.0001bP= 0.002bP = 0.071aP = 0.25
Pre-ASTP= 0.013bP = 0.63P = 0.023bP = 0.26
Pre-ALTP = 0.13P = 0.21P = 0.064aP = 0.11
AgeP = 0.54P = 0.51P = 0.11P = 0.30
GenderP = 0.70P = 0.61P = 0.85P = 0.84
Table 4. Comparing Patient Factors and TACE Factors in Patients with Irreversible Hepatic Decompensation with Those without Hepatic Decompensation
FactorsIrreversible hepatic decompensation vs. no hepatic decompensation
  • TACE: transcatheter arterial chemoembolization; ALT: alanine aminotransferase; AST: aspartate aminotransferase.

  • a

    A P value < 0.1 but > 0.05 was considered to approach Statistical significance.

  • b

    A P value < 0.05 was considered to be statistically significant.

Basal bilirubin levelP = 0.009a
Basal prothrombin timeP = 0.015b
Basal ALTP = 0.381
Basal ASTP = 0.239
Pre-TACE Child stageP < 0.0001b
Dosage of cisplatinP = 0.033b
The use of gelfoam embolisationP = 0.83
Total no. of TACE receivedP = 0.416

One of the patients in the current study who had previously underwent a partial gastrectomy developed both liver and splenic abscesses followed by hepatic encephalopathy after TACE. He died 30 days after the TACE session. His death was attributable to both the side effects of TACE and advanced malignancy. Another patient had a positive Murphy sign and was confirmed by ultrasound to have acalculous cholecystitis. His disease was managed conservatively. His symptoms and signs subsided completely after 3 days. During the study period, 20 patients were withdrawn from further courses of TACE; 6 patients (30%) developed arteriovenous shunting, 5 patients (25%) developed progressive local disease, 5 patients (25%) developed progressive liver failure, and 4 patients (20%) developed extrahepatic metastasis.

At the time of last follow-up, 15 patient had died, 15 patients had progressive disease, 25 patients had improved or stable disease, and 4 patients had been lost to follow-up. One of the patients who died was the patient who died of liver and splenic abscesses after 30 days of chemoembolization. Four patients died of progressive malignancy. Eight patients died of progressive liver failure. The other two patients died of unrelated medical problems. The median survival time was 34.4 weeks (range, 13.4–267.3 weeks). There was no statistically significant difference with regard to the median survival time between the patients with and the patients without irreversible liver decompensation after TACE (36.2 weeks vs. 34 weeks; P = 0.8). The 1-month mortality rate was 1.7%. The median follow-up for the survivors from the time of the first TACE session was 1.95 years (range, 0.56–7.5 years). The tumor response rate was evaluated in 58 patients during the most recent follow-up. Approximately 36% (21 of 58 patients) of the patients achieved a tumor response, 35% (20 of 58 patients) had stable disease, and 29% (17 of 58 patients) had progressive disease. Among the patients who responded to TACE, 19% (4 of 21 patients) had a 25–50% reduction in tumor size, 52% (11 of 21 patients) had a 50–75% reduction in tumor size, and 29% (6 of 21 patients) had a ≥ 75% reduction in tumor size. Three patients achieved a complete response with undetectable tumor.


Over the past 20 years, TACE has become the treatment of choice for patients with inoperable HCC. TACE has been shown to improve survival in uncontrolled studies.3, 8–13 However, this survival benefit may be offset by the worsening of liver function. Liver function tests often worsen slightly after TACE but the majority of studies have demonstrated a return to baseline function within 1 week.14–16 However, a significant number of cases of hepatic failure were reported by a French study group in 1995.6 The current study demonstrated that reversible hepatic decompensation developed in a significant proportion of sessions of TACE (20%). Only 3% of these sessions resulted in irreversible liver failure. This incidence is much lower than that reported in the French study.6 One of the patients in the current study died because of hepatic decompensation but his treatment was complicated by liver and splenic abscesses. In the majority of patients, the return of liver function to pretreatment levels was gradual but complete.

The current study identified the factors that appear to predispose patients to the development of irreversible acute hepatic decompensation after TACE, namely a high dosage of cisplatin, high basal levels of bilirubin, prolonged prothrombin time, and advanced cirrhosis. Pretreatment liver function and the stage of cirrhosis always have been the chief considerations for patients receiving TACE. The results of the current study demonstrated that when a strict entry criterion of a bilirubin level < 50 μmol/L was used, only 3% of TACE sessions would result in irreversible liver failure. Even in the patients with irreversible hepatic decompensation, survival was similar to that for patients without irreversible hepatic decompensation.

Of the 40 sessions of TACE that resulted in hepatic failure, embolization was given in 13 sessions (33.3%) whereas chemotherapy without embolization was performed in the remaining 27 sessions. The use of embolization was not shown to be a significant factor in causing irreversible hepatic failure. However, embolization was given in only 44 of 197 sessions of TACE in the current study. The number of sessions with embolization might be too small to reach statistical significance. Postembolization syndrome, comprised of fever, nausea, and abdominal pain, was the most common side effect experienced by patients in the current study and was similar to those reported in the literature.2–5, 8, 10 The duration of fever was found to be correlated significantly with the use of gel foam, the initial tumor size, total dosage of cisplatin used, and basal AST level, and was found to be related inversely to the number of TACE sessions. Although the syndrome was self-limiting, it did prolong the period of hospitalization.

Although all the patients recruited in the French study had Child-Pugh Class A cirrhosis whereas some of the current study patients had more advanced cirrhosis, a much lower rate of hepatic failure was observed in the patients in the current study. There are two possible reasons for this finding. One is the low dosage of cisplatin used in the current study. Earlier studies suggest that the dosage of cisplatin should be 2 mg/kg body weight.2, 3 The results of the current study demonstrated that even with a relatively low maximum dose of 10 mg of cisplatin per session, those patients receiving higher doses tended to have higher morbidity. The dosage of cisplatin used in the current study patients was much smaller than the median dose of 70 mg used in the French study. Hepatic decompensation could result from incidental damage caused by the chemotherapuetic agent to the nontumorous part of the already cirrhotic liver. The second possible reason for the difference in survival noted between the French study and the current study is that superselective embolization was employed in the current study, thus limiting any collateral liver damage, whereas nonselective embolization was performed in the French study.

With regard to the patient who died 30 days after the TACE session, the previous gastrectomy might have been a risk factor for the development of a liver abscess.17 Liver abscess has been reported to be one of the serious side effects of TACE, but is reported to occur at a low incidence of 1.1%.18

Upper gastrointestinal bleeding was one of the most common complications in the patients in the current study. This is in agreement with the findings of other reported series. It may be a result of ischemic changes to the gastric mucosa induced by TACE. The incidence of nonbleeding occult peptic ulcer is unknown. Further studies regarding the frequency of gastric ulceration in patients receiving TACE will be worthwhile. Prophylactic protection of the gastric mucosa may be the best solution.

The acalculous cholecystitis reported in one patient could have been due to unintentional blockage of the cystic artery during hepatic artery embolization, thus causing ischemic change to the gallbladder. Although this has been reported to be a common complication by Takayasu et al.,19 to our knowledge it has been reported rarely in other series. Treatment is conservative in the majority of patients. However, surgical cholecystectomy or percutaneous cholecystostomy may be necessary in situations such as gallbladder perforation or emphysematous cholecystitis.20, 21

The results of the current study demonstrated that 20% of patients developed acute hepatic decompensation after TACE. However, in the majority of patients, the liver function returned to its pretreatment level before the next course of TACE was initiated. Only a minority of patients eventually developed irreversible liver failure. Other side effects also were reported to be usually self-limiting. The use of high-dose cisplatin, high basal bilirubin and prothrombin time, and advanced stage of cirrhosis were predisposing factors for hepatic decompensation. Further randomized studies may be needed to define the criteria for patients who may benefit from TACE and to assess the optimum dosage of the chemotherapeutic agent used. Careful patient selection should enhance the effectiveness and safety for TACE in patients with inoperable HCC.