Va s c u l a r a n d I n t e r ve n t i o n a l R a d i o l o g y • O r i g i n a l R e s e a r c h Woo et al. Liver Abscess After Transarterial Chemoembolization
Downloaded from www.ajronline.org by National Cancer Center on 12/29/13 from IP address 49.50.1.229. Copyright ARRS. For personal use only; all rights reserved
Vascular and Interventional Radiology Original Research
Liver Abscess After Transarterial Chemoembolization in Patients With Bilioenteric Anastomosis: Frequency and Risk Factors Sungmin Woo1 Jin Wook Chung Saebeom Hur Seung-Moon Joo Hyo-Cheol Kim Hwan Jun Jae Jae Hyung Park Woo S, Chung JW, Hur S, et al.
OBJECTIVE. The purpose of this study was to clarify the frequency of and risk factors for liver abscess formation after transarterial chemoembolization (TACE) in patients with hepatocellular carcinoma or metastatic hepatic tumors after undergoing bilioenteric anastomosis. MATERIALS AND METHODS. From January 1996 to June 2012, 25 patients (21 men, four women; age range, 34–74 years) with hepatocellular carcinoma (n = 12) or metastatic hepatic tumors (n = 13) with an underlying bilioenteric anastomosis underwent 65 TACE procedures. The incidence of liver abscess, predisposing factors (diabetes, Child-Pugh class, leukopenia, tumor number, tumor size, tumor burden, tumor type, portal vein thrombus, lipiodol dose, particulate embolization, embolization selectivity, oily portogram, antibiotic prophylaxis, and occurrence of liver abscess at initial TACE), and clinical outcome were evaluated. Statistical analysis for relations between liver abscess and predisposing factors was performed by Fisher exact test and linear-by-linear association. RESULTS. Liver abscess developed after 17 of 65 (26.2%) TACE procedures performed on 12 of 25 (48%) patients. Two patients died of progression of liver abscess into sepsis. Univariate and multivariate analyses showed that leukopenia (p = 0.029), occurrence of liver abscess at initial TACE (p = 0.082), and particulate embolization or oily portogram (grade 2) (p = 0.001) were associated with a higher incidence of liver abscess. CONCLUSION. The incidence of liver abscess was high among patients with bilioenteric anastomoses who underwent TACE. Leukopenia, occurrence of liver abscess at initial session of TACE, and particulate embolization or oily portogram (grade 2) were associated with the development of liver abscess.
T
Keywords: bilioenteric anastomosis, liver abscess, risk factor, technique, transarterial chemoembolization DOI:10.2214/AJR.12.9630 Received July 16, 2012; accepted after revision August 24, 2012. 1 All authors: Department of Radiology, Seoul National University College of Medicine, 101 Daehak-ro, Jongnogu, Seoul 110-744, Korea. Address correspondence to J. W. Chung (
[email protected]).
AJR 2013; 200:1370–1377 0361–803X/13/2006–1370 © American Roentgen Ray Society
1370
ransarterial chemoembolization (TACE) has become a popular treatment of unresectable hepatocellular carcinoma (HCC) and metastatic hepatic tumors [1–4]. It is associated with a variety of hepatic and biliary damage, including hepatic failure, liver abscess, and biloma formation [5]. Although the reported incidence of liver abscess formation is low (0.1–4.5%) [6–9], bilioenteric anastomosis significantly increases risk [7]. An odds ratio up to 894 was reported by Kim et al. [8]. Moreover, high mortality rates, ranging from 13.3% to 50% have been reported after an abscess has complicated TACE [7, 10, 11]. Therefore, avoidance of TACE in the care of these patients has been recommended, but if TACE is performed, broad-spectrum antibiotics and bowel preparation should be administered [12]. There has been controversy, however, over the role of antibiotics in preventing abscess formation after TACE [8, 13–15].
Furthermore, no consensus has been reached with regard to patient selection. To the best of our knowledge, no study has been focused on risk factors, especially those concerning the technical methods of TACE procedures, for the development of liver abscess after TACE in patients with a history of bilioenteric anastomosis. We believe that limiting the study population to only patients with a history of bilioenteric anastomosis would facilitate identification of risk factors and would contribute to improving the safety of chemoembolization. The purpose of this study was to clarify the frequency of and risk factors for liver abscess formation after TACE in patients with HCC or metastatic hepatic tumors who previously underwent bilioenteric anastomosis. Materials and Methods Study Population This retrospective study was approved by our institutional review board. Informed consent was
AJR:200, June 2013
Downloaded from www.ajronline.org by National Cancer Center on 12/29/13 from IP address 49.50.1.229. Copyright ARRS. For personal use only; all rights reserved
Liver Abscess After Transarterial Chemoembolization
A
B
C
D
E
F
Fig. 1— 73-year-old man with hepatocellular carcinoma who had undergone pylorus-preserving pancreaticoduodenectomy because of distal common bile duct cancer 3 years earlier. A, Gadoxetic acid–enhanced MR image shows 3.2-cm mass with hepatobiliary phase defect in liver segment VII that proved to be hepatocellular carcinoma at ultrasound-guided percutaneous liver biopsy (not shown). B, Celiac angiogram shows mass with hypervascular tumor staining. C, Transarterial chemoembolization (TACE) was performed with subsegmental level of catheter tip placement (not shown). Fluoroscopic image obtained during TACE shows grade 2 oily portogram. Additional particulate embolization was performed until complete arterial flow stasis was achieved. D, CT image obtained 12 days after TACE shows large liver abscess adjacent to mass compacted with iodized oil. E, Fluoroscopic image shows appearance at percutaneous catheter drainage with 10.5-French catheter performed 13 days after TACE. F, CT image obtained 2 months after TACE shows liver abscess has resolved. Air is present in both intrahepatic ducts because of underlying pylorus-preserving pancreaticoduodenectomy.
AJR:200, June 2013 1371
Downloaded from www.ajronline.org by National Cancer Center on 12/29/13 from IP address 49.50.1.229. Copyright ARRS. For personal use only; all rights reserved
Woo et al. waived for this records-based study. A computerized search was performed of medical records from January 1996 to June 2012 for patients with a history of bilioenteric anastomosis who underwent TACE procedures at our institution. A total of 87 TACE procedures were performed for hepatic tumors in 35 patients. Six patients who underwent 18 TACE procedures were excluded because TACE was performed before bilioenteric anastomosis. Another four patients who underwent four procedures were excluded from analysis because embolization was performed for bleeding control (three patients) or as empirical embolization without definite evidence of tumor at CT or angiography (one patient). Ultimately, a total of 25 patients (21 men, four women; mean age, 57.8 ± 12.5 [SD] years; range, 34–74 years) who underwent 65 TACE procedures were included in our study. The types of bilioenteric anastomosis were Whipple operation (eight patients), pylorus-preserving pancreaticoduodenectomy (nine patients), and liver resection involving hepaticojejunostomy or choledochojejunostomy (eight patients). The diagnoses of 12 HCCs and 13 metastatic hepatic tumors were made as follows: Four HCCs and six metastatic hepatic tumors were diagnosed with percutaneous biopsy. The other HCCs were diagnosed clinically on the basis of risk factors and radiologic features of the tumor on dynamic contrast-enhanced CT scans or dynamic MR images according to the clinical guidelines proposed by the American Association for the Study of Liver Diseases [16] in eight patients. For the other seven patients with metastatic tumors, diagnosis was based on a recent history of surgery- or biopsy-proven extrahepatic primary malignancy, elevated concentrations of tumor markers, and consistent imaging findings. The metastatic hepatic tumors were six pancreatic neuroendocrine tumors, two duodenal gastrointestinal stromal tumors, two duodenal leiomyosarcomas, two adenocarcinomas of the ampulla of Vater, and one gastric neuroendocrine tumor.
Methods of Chemoembolization and Follow-Up Informed consent was obtained from all patients before each TACE procedure. First, 2–15 mL (mean, 5.6 ± 3.1 mL) of iodized oil (Lipiodol, Andre Guerbet) and 10–50 mg of doxorubicin hydrochloride (ADM, Dong-a) emulsion was intraarterially infused. In preparation of an iodized oil–anticancer drug emulsion, 10 mg of doxorubicin hydrochloride was dissolved in 0.5 mL of water-soluble contrast medium (iopamidol, Iopamiro 300, Bracco). Iodized oil and dissolved doxorubicin hydrochloride were then mixed by a pumping technique that entailed two syringes connected by a three-way stopcock [7]. Depending on the operator’s discretion, particulate embolization was performed with absorbable gelatin sponge particles (1–2 mm in diam-
1372
eter; Gelfoam, Upjohn) soaked in 20 mL of iopromide (Ultravist 370, Bayer Schering Pharma). Further TACE sessions were planned individually according to tumor response and the patient’s tolerance of the procedure. For example, when follow-up evaluation with CT and alpha-fetoprotein measurement showed liver function was severely deteriorated or tumor had undergone complete remission, additional TACE was not performed. After each procedure, we periodically assessed patients for treatment-related complications. When liver abscess was suspected, the patient underwent contrast-enhanced CT of the liver. Otherwise, iodized oil CT or contrast-enhanced CT was performed routinely 2–3 weeks after the procedure.
Investigation of Liver Abscess The medical records of all 25 patients were retrospectively reviewed for the development of liver abscess after each TACE session. Liver abscess was defined as a hypoattenuating lesion with peripheral rim enhancement on CT images if the patient met any of the three following conditions: blood or aspirate culture revealed bacteria; aspirate from percutaneous drainage had a purulent appearance; or temperature higher than 38.5°C lasted more than 5 days with leukocyte count greater than 12 × 109/L without another cause [17].
Investigation of Predisposing Factors We defined several variables regarding patient and tumor characteristics and methods of TACE as risk factors for the development of liver abscess after TACE. They were diabetes mellitus, ChildPugh class, leukopenia, tumor number, maximum tumor size, tumor burden, tumor type (HCC vs metastasis), portal vein thrombus for patient and tumor characteristics and lipiodol dose, particulate embolization, degree of embolization selectivity, achievement of oily portogram, and antibiotic prophylaxis for methods of TACE. In addition, whether a liver abscess occurred at initial TACE was evaluated for the repeated sessions of TACE. An interventional radiologist with 20 years’ experience and a radiologist with 2 years’ experience retrospectively reviewed the medical records and radiographs of the TACE procedures for predisposing factors. Leukopenia was defined as a WBC count less than 4000/µL. Particulate embolization was graded as follows: 0, no embolic material used; 1, embolic materials used without achievement of arterial stasis; 2, embolic materials used with achievement of arterial stasis. Visualization of an oily portogram was categorized according to the classification suggested by Miyayama et al. [18]: grade 0, no obvious visualization; 1, slight visualization; 2, marked visualization. Degree of embolization selectivity was categorized as follows according to the location
of the catheter tip on embolization: lobar or sectional, segmental, and subsegmental.
Statistical Analysis Statistical analysis was performed with software (SPSS version 17.0, SPSS). Fisher exact test or linear-by-linear association was used to evaluate the relation between the risk factors and liver abscess. Variables with p < 0.1 were taken into consideration for multivariate linear logistic regression analysis. Finally, p < 0.05 at multivariate analysis was considered to indicate a significant difference.
Results Frequency of Predisposing Factors Among the patients who underwent TACE, 12 patients with HCC underwent 33 TACE procedures. The other 13 patients with metastatic hepatic tumor underwent 32 TACE procedures. Seven (28%) patients had diabetes mellitus. The Child-Pugh class was A in 58 (89.2%) and B in seven (10.8%) procedures. WBC count ranged from 2220 to 11,900/µL (mean, 5830 ± 2280/ µL), and leukopenia was observed before 13 (20%) procedures. A single tumor was treated in 26 (40%) sessions, two tumors in nine (13.8%) sessions, three tumors in three (4.6%) sessions, and four or more tumors in 27 (41.5%) sessions. Maximum tumor size ranged from 0.1 to 12.1 cm (mean, 2.5 ± 2.1 cm) in greatest diameter. Tumor burden was less than 10% in 57 (87.7%) cases and 10% or more in eight (12.3%) cases. Portal vein thrombosis was found in three (4.6%) procedures at the right secondorder branch. The dose of iodized oil ranged from 2 to 15 mL (mean, 5.6 ± 3.1 mL). Particulate embolization was performed with grades 1 and 2 in 14 (21.5%) and 20 (30.8%) procedures. Oily portogram after TACE was obtained with grades 1 and 2 in six (9.2%) and 14 (21.5%) procedures. The selective degree of embolization was lobar or sectional in 28 (43.1%), segmental in 19 (29.2%), and subsegmental in 18 (27.7%) sessions. Antibiotic prophylaxis was administered before TACE in 17 (26.2%) sessions. Incidence of Liver Abscess Liver abscess developed after 17 of 65 (26.2%) TACE procedures performed on 12 of 25 (48%) patients (Fig. 1). The clinical profiles and methods of TACE for these 12 patients are summarized in Table 1. In 7 of 25 (28%) patients liver abscess developed after the first session of TACE. These seven patients underwent seven additional TACE procedures and had four (57.1%) liver abscesses. The other 18 of 25 (72%) patients without liver abscess after the initial TACE procedure
AJR:200, June 2013
73
74
61
71
66
69
53
51
51
51
6, procedure 1
6, procedure 2
7
8
9
10
11
12, procedure 1
12, procedure 2
12, procedure 3
M
M
M
M
M
M
M
F
M
M
F
M
M
M
M
M
M
HCC
HCC
HCC
HCC
HCC
HCC
Pancreatic neuroendocrine tumor
Adenocarcinoma of ampulla of Vater
HCC
HCC
Pancreatic neuroendocrine tumor
Duodenal GIST
Pancreatic neuroendocrine tumor
Pancreatic neuroendocrine tumor
HCC
HCC
HCC
Diagnosis
B
B
B
A
A
A
A
A
A
A
B
A
A
A
A
A
A
No
No
No
No
Yes
No
No
No
Yes
Yes
Yes
No
No
No
No
Yes
Yes
2990
3840
3150
4085
11,900
4700
6700
6000
6400
5700
3800
3220
4910
6000
3810
10,700
11,330
WBC Count (/µL)
Right lobectomy hepaticojejunostomy
Right lobectomy hepaticojejunostomy
Right lobectomy hepaticojejunostomy
Left lobectomy hepaticojejunostomy
Right lobectomy hepaticojejunostomy
PPPD
PPPD
PPPD
PPPD
PPPD
Whipple operation
Whipple operation
Whipple operation
Whipple operation
Whipple operation
Central bisegmentectomy, choledochojejunostomy
Central bisegmentectomy, choledochojejunostomy
Bilioenteric Anastomosis Type
1
1
1.8
2.4
2.3
2.6
2.3
7.8
3
3.2
3.8
1.7
3.1
1.3
1.7
1.5
2.8
1
1
4
1
1
2
2
1
1
1
20
2
3
5
1
1
6
aTotal number of TACE sessions before liver abscess formation.
3
2
1
7
1
1
1
1
2
1
4
2
6
4
6
2
1
0
0
0
1
2
2
2
2
2
2
2
2
2
2
2
2
2
0
0
0
0
2
2
2
2
2
2
1
1
2
0
2
0
0
Segmental
Lobar
Lobar
Segmental
Subsegmental
Subsegmental
Subsegmental
Segmental
Subsegmental
Subsegmental
Segmental
Lobar
Subsegmental
Subsegmental
Subsegmental
Segmental
Segmental
2
5
7
4
2
4
6
8
6
4
10
10
4
3
6
3
3
10
30
40
30
10
20
30
50
30
30
50
50
30
20
30
20
20
Yes
Yes
No
No
No
Yes
No
No
Yes
No
Yes
No
No
No
No
No
Yes
Particulate Oily Tumor No. of No. of Embolization Portogram Iodized Oil Doxorubicin Grade Size (cm) Tumors Sessionsa Grade TACE Selectivity Dose (mL) Dose (mg) Antibiotics
Note—HCC = hepatocellular carcinoma, GIST = gastrointestinal stromal tumor, PPPD = pylorus-preserving pancreaticoduodenectomy.
72
56
5
43
3, procedure 2
4
80
43
3, procedure 1
68
1, procedure 2
2
68
Age (y) Sex
1, procedure 1
Patient. No.
Child-Pugh Diabetes Class Mellitus
TABLE 1: Characteristics of Patients With Liver Abscess After Transarterial Chemoembolization (TACE)
Downloaded from www.ajronline.org by National Cancer Center on 12/29/13 from IP address 49.50.1.229. Copyright ARRS. For personal use only; all rights reserved
Liver Abscess After Transarterial Chemoembolization
AJR:200, June 2013 1373
Woo et al. TABLE 2: Relation Between All Variables and Abscess Formation in 25 Initial Transarterial Chemoembolization Procedures for Hepatocellular Carcinoma and Metastatic Hepatic Tumors in 25 Patients Liver Abscess Formation Downloaded from www.ajronline.org by National Cancer Center on 12/29/13 from IP address 49.50.1.229. Copyright ARRS. For personal use only; all rights reserved
Variable
Without Predisposing Factor
With Predisposing Factor
p
Child-Pugh class B
6/22 (27.3)
1/3 (33.3)
1.000
Diabetes mellitus
4/18 (22.2)
3/7 (42.9)
0.355
Leukopenia
6/21 (28.6)
1/4 (25.0)
1.000
Antibiotics
2/7 (28.6)
5/18 (27.8)
1.000
Single tumor
4/16 (25.0)
3/9 (33.3)
0.673
Tumor size ≤ 3 cm
5/18 (27.8)
2/7 (28.6)
1.000
Tumor burden ≤ 10%
6/20 (30.0)
1/5 (20.0)
1.000
Type of tumor (hepatocellular carcinoma)
2/13 (15.4)
5/12 (41.7)
0.202
Portal vein thrombosis
7/24 (29.2)
0/1 (0.0)
1.000
Embolization (grade 2)
1/15 (6.7)
6/10 (60.0)
0.007
Oily portogram (grade 2)
2/18 (11.1)
5/7 (71.4)
0.007
1/15 (6.7)
6/10 (60.0)
0.007
Iodized oil dosea
0.215
Selectivitya
0.051
Embolization or portogram (grade 2) Note—Numbers in parentheses are percentages. aLinear-by-linear association.
TABLE 3: Relation Between All Variables and Abscess Formation in 40 Repeated Transarterial Chemoembolization (TACE) Procedures for Hepatocellular Carcinomas and Metastatic Hepatic Tumors in 15 Patients Liver Abscess Formation With Predisposing Factor Variable
No
Yes
p (Univariate)a
Child-Pugh class B
7/36 (19.4)
3/4 (75.0)
0.042
Diabetes mellitus
7/24 (29.2)
3/16 (18.8)
0.711
Leukopenia
5/31 (16.1)
5/9 (55.6)
0.029
Antibiotics
4/10 (40.0)
6/30 (20.0)
0.232
Single tumor
4/23 (17.4)
6/17 (35.3)
0.274
Tumor size ≤ 3 cm
7/29 (24.1)
3/11 (27.3)
1.000
Tumor burden ≤ 10%
9/37 (24.3)
1/3 (33.3)
1.000
Type of tumor (HCC)
p (Multivariate)b
Odds Ratio
90% CI for Odds Ratio
0.020
17.476
2.363–129.253
4/19 (21.1)
6/21 (28.6)
0.721
10/38 (26.3)
0/2 (0.0)
1.000
Embolization or portogram (grade 2)
2/21 (9.5)
8/19 (42.1)
0.028
0.033
16.493
1.900–143.183
Abscess at first TACE
6/33 (18.2)
4/7 (57.1)
0.052
0.082
5.471
1.098–27.256
Portal vein thrombosis Iodized oil dosec
0.868
Note—Numbers in parentheses are percentages. ap for Fisher exact test analyzed for all variables. bp for multivariate linear logistic regression analysis analyzed for variables with p < 0.1 in univariate analysis. cLinear-by-linear association.
underwent 33 additional TACE procedures and had six (18.2%) liver abscesses. Cultures of blood or aspirate revealed bacteria in 14 of 17 (82.4%) cases. Blood culture results were positive in nine cases (52.9%). The causative organisms were Klebsiella pneumoniae in four cases, Bacteroides fragi-
1374
lis in one case, Escherichia coli in four cases, and Enterococcus faecium in four cases. Results of aspirate cultures were positive in seven cases (41.2%). The causative organisms were K. pneumoniae in two cases, E. coli in three cases, E. faecium in one case, and Pseudomonas aeruginosa in one case.
Relation Between Liver Abscess Formation and Predisposing Factors The incidences of liver abscess for each potential risk factor compared with the corresponding control group are shown in Tables 2–4. Of the 25 initial TACE procedures, grade 2 embolization (6/10, p = 0.007), grade 2
AJR:200, June 2013
Liver Abscess After Transarterial Chemoembolization TABLE 4: Relation Between All Variables and Abscess Formation in All 65 Transarterial Chemoembolization Procedures for Hepatocellular Carcinoma and Metastatic Hepatic Tumors in 25 Patients
Downloaded from www.ajronline.org by National Cancer Center on 12/29/13 from IP address 49.50.1.229. Copyright ARRS. For personal use only; all rights reserved
Liver Abscess Formation With Predisposing Factor Variable
No
Yes
p (Univariate)a
Child-Pugh class B
13/58 (22.4)
4/7 (57.1)
0.070
Diabetes mellitus
11/42 (26.2)
6/23 (26.1)
1.000
Leukopenia
11/52 (21.2)
6/13 (46.2)
0.085
Antibiotics
6/17 (35.3)
11/48 (22.9)
0.348
Single tumor
8/39 (20.5)
9/26 (34.6)
0.255
Tumor size ≤ 3 cm
12/47 (25.5)
5/18 (27.8)
1.000
Tumor burden ≤ 10%
15/57 (26.3)
2/8 (25.0)
1.000
Type of tumor (HCC)
8/32 (25.0)
9/33 (27.3)
0.835
Portal vein thrombosis
17/62 (27.4)
0/3 (0.0)
0.561
3/36 (8.3)
14/29 (48.3)
< 0.001
Iodized oil dosec Embolization or portogram (grade 2)
p (Multivariate)b
Odds Ratio
95% CI for Odds Ratio
0.029
17.476
1.216–40.510
0.001
16.493
3.097–85.215
0.458
Note—Numbers in parentheses are percentages. ap for Fisher exact test analyzed for all variables. bp for multivariate linear logistic regression analysis analyzed for variables with p < 0.1 in univariate analysis. cLinear-by-linear association.
oily portogram visualization (5/7, p = 0.007), and higher selectivity of TACE (p = 0.051) correlated with higher incidence of liver abscess formation (Table 2). With regard to the TACE technique at our institution, where we tend to perform higher degrees of particulate embolization (p = 0.001) and to achieve higher grades of oily portogram visualization (p = 0.002) with higher selectivity, we integrated the three variables into a single factor: grade 2 embolization or oily portogram. This integrated factor significantly correlated with liver abscess formation (6/10, p = 0.007). Of the 40 repeated sessions of TACE, the following were associated with higher incidence of liver abscess (Table 3): Child-Pugh class B (3/4, p = 0.042), leukopenia (5/9, p = 0.029), grade 2 embolization or oily portogram (8/19, p = 0.028), and abscess at initial TACE (4/7, p = 0.052). In multivariate linear regression analysis (Child-Pugh class was excluded because all four patients with Child-Pugh class B had leukopenia), leukopenia (p = 0.033) and embolization or oily portogram (grade 2) (p = 0.020) were statistically significant. Experience of abscess at initial TACE had borderline association (p = 0.082; 90% CI, 1.1–27.3) Of the total of 65 TACE procedures, ChildPugh class B (4/7, p = 0.070), leukopenia (6/13, p = 0.085), and grade 2 embolization or oily portogram (14/29, p < 0.001) were associated with liver abscess development (Table 4). In multivariate linear regression analysis (Child-Pugh class was excluded because all
four patients with Child-Pugh class B disease in whom a liver abscess developed had leukopenia), leukopenia (6/13, p = 0.029) and grade 2 embolization or oily portogram (14/29, p = 0.001) were statistically significant. Treatment and Clinical Outcome All patients underwent treatment with parenteral antibiotics. Percutaneous catheter drainage (PCD) with 10.5- to 14-French catheters was performed in nine patients (10 sessions), who recovered. The other seven patients (seven sessions) did not undergo PCD insertion owing to lack of safe access (patient 1), death before PCD could be performed (patient 3), and surgeon’s specific request for diagnostic aspiration only (patient 9). The reason for not performing PCD could not be identified from the medical records of the other four patients. Of the seven patients who did not undergo PCD, five patients recovered without further event. However, two patients (patients 1 and 3) died of progression of liver abscess to sepsis 23 days after and on the day of liver abscess diagnosis. Thus the mortality rate directly related to hepatic abscess was 11.8% (2/17). The duration of hospitalization ranged from 7 to 378 days (mean, 52.5 ± 86.0 days). Discussion In our study, as in the literature, liver abscess formation after transarterial chemoembolization was not rare when underlying
bilioenteric anastomosis was present. The overall incidence in our series, 17/65 (26.2%) per procedure and 12/25 (48.0%) per patient, was comparable to previous reported values. De Baère et al. [10] reviewed the cases of 561 patients who underwent local treatment of liver tumors. Although the overall rate of liver abscess was 0.7%, liver abscess developed in three of four (75%) patients who underwent bilioenteric anastomosis. In another study, Kim et al. [8] performed 397 TACE procedures and reported an abscess rate of 2.0%. However, the incidence was 6/7 (85.7%) in a subgroup of patients with a history of pancreaticoduodenectomy. Although the study population was small, Ishikawa et al. [19] also reported a high rate (2/3, 66.7%) of liver abscess from TACE after pancreaticoduodenectomy. Song et al. [7], however, reported that only 4 of 55 (7.3%) patients with type 2 biliary abnormality had a liver abscess. However, the definition of type 2 biliary abnormality included not only bilioenteric anastomosis but also other conditions, such as endoscopic papillotomy and percutaneous transhepatic biliary drainage. In this study, we identified several other potential risk factors that correlated with the development of liver abscess. Regarding patient and tumor characteristics, the incidence was significantly higher among patients with leukopenia and those in whom a liver abscess developed at the first session of TACE. With regard to technical aspects of TACE, grade 2
AJR:200, June 2013 1375
Downloaded from www.ajronline.org by National Cancer Center on 12/29/13 from IP address 49.50.1.229. Copyright ARRS. For personal use only; all rights reserved
Woo et al. embolization with absorbable gelatin sponge or grade 2 visualization of oily portogram correlated with a significantly elevated incidence of liver abscess. The higher incidence of liver abscess after grade 2 visualization of an oily portogram may be explained by the fact that grade 2 oily portogram indicates adequate blockage of the portal supply to tumors [18]. This is attributed to the semifluid nature of iodized oil, which allows it to flow into the surrounding portal veins, allowing visualization of the oily portogram via the peribiliary plexus, vasa vasorum of the portal vein, and drainage routes from HCC when TACE is performed in a selective manner [20]. Furthermore, if particulate embolization is performed, additional blockage of arterial supply is expected. It is well known that gelatin particles occlude the artery temporarily, recanalization occurring within 2 weeks [21]. Therefore, when the portal blood supply is obstructed owing to achievement of a grade 2 oily portogram, when arterial flow is blocked owing to particulate embolization, or when both of these conditions are present, massive necrosis of the tumor and atrophy of surrounding hepatic parenchyme can be accomplished within this limited area. In the literature, this is referred to as a subsegmentectomy-like effect [22, 23]. Our results are in agreement with those of a previous study [7] that showed that although not independently significant, embolization with gelatin sponge was associated with an increased incidence of abscess when portal vein obstruction was present. Still, such stronger necrosis alone does not yield a high rate of abscess formation [22]. With an underlying bilioenteric anastomosis, however, in which bile ducts are contaminated by enteric bacteria in a retrograde manner, various disruptions, including TACE with grade 2 particulate embolization or oily portogram visualization, may transform this contamination into obvious infection [19, 24]. In regard to tumor characteristics, there was no statistically significant relation between tumor type and liver abscess formation. In the literature, however, liver abscess has been reported to be more frequent after embolization of metastatic tumors [7, 8, 10]. In our review of the literature, we found that two of three (66.7%), six of seven (85.7%), and three of three (100%) patients in the studies by Song et al. [7], Kim et al. [8], and de Baère et al. [10] with metastasis who had an abscess after TACE had previously undergone bilioenteric anastomosis. We agree with Song et al. [7] that
1376
tumor type in itself is not a significant factor in the development of abscess. In our study, the use of prophylactic antibiotics did not significantly decrease the incidence of liver abscess. However, antibiotics were not administered according to a standardized regimen. The antibiotics given usually were a third-generation cephalosporin with or without metronidazole. In the literature, controversy persists over the role of antibiotics in preventing abscess [8, 13–15]. A few studies, however, have shown promising results that vigorous antibiotic prophylaxis may be helpful in preventing liver abscess. Khan et al. [15] observed no development of abscess after TACE in patients with previous biliary interventions when moxifloxacin, an antibiotic with a high degree of biliary secretion, was administered. Geschwind et al. [14] reported that prophylaxis with tazobactampiperacillin and bowel preparation resulted in no abscess formation in patients who underwent bilioenteric anastomosis followed by chemoembolization. In addition, Hoffmann et al. [25] suggested that at least 10 days of antibiotic therapy to cover anaerobic bacteria may be helpful in preventing abscess formation in patients with bilioenteric anastomosis who undergo percutaneous radiofrequency ablation. Therefore, although our results did not show usefulness of routine antibiotic prophylaxis in preventing abscess formation after TACE in patients with bilioenteric anastomosis, we carefully speculate that use of vigorous antibiotic prophylaxis as suggested in more recent literature may be effective. We believe that our study results may assist in establishing a guideline for TACE in patients with a history of bilioenteric anastomosis. In our study, liver abscess developed after approximately 50% of the procedures on patients with risk factors such as leukopenia, liver abscess occurring at first session of TACE, and use of high-risk TACE technique. Therefore, at the initial session of TACE, precaution is required for those with leukopenia, and vigorous prophylactic antibiotic administration should be undertaken in advance. When chemoembolization is performed with high-risk technique (grade 2 particulate embolization or oily portogram), close observation is needed for early detection and timely management with antibiotics and PCD. Furthermore, the patient should be observed for liver abscess formation after the first session of TACE. If abscess formation occurs after TACE with a high-risk technique, additional sessions should be tried with a low-risk tech-
nique. If an abscess develops even though TACE is performed without a high-risk technique, the risk of abscess may be considered high in the corresponding patient regardless of TACE technique, and therefore special attention should be paid. The main limitation of our study was the long period of data collection and the corresponding lack of standardization of the methods of chemoembolization and periprocedural care. However, TACE is a technique with a great deal of flexibility that must be tailored to tumor and patient characteristics. Therefore, the heterogeneity was essential for analyzing risk factors for the development of liver abscess. Furthermore, TACE was performed with conventional methods with iodized oil. TACE is currently being performed with more diverse therapeutic agents, such as drug-eluting beads, which have been suggested as being associated with increased risk of liver and biliary injuries. Further studies for comparison of different therapeutic agents with respect to risk of liver abscess are warranted. In addition, this study was retrospective, and there may have been selection bias in how the patients were treated. Moreover, the sample size was relatively small owing to the paucity of patients who underwent bilioenteric anastomosis followed by TACE. Nevertheless, we do not believe statistical power was limited for elucidating several risk factors for liver abscess formation. Conclusion The incidence of liver abscess formation was high among patients with bilioenteric anastomosis who underwent TACE. Leukopenia, occurrence of liver abscess at initial session of TACE, and grade 2 particulate embolization or oily portogram were found to be risk factors for development of liver abscess. Therefore, in patients with such risk factors, TACE should be performed with even more caution than usual, and close follow-up is required for early detection and timely management of liver abscess. References 1. Cammà C, Schepis F, Orlando A, et al. Transarterial chemoembolization for unresectable hepatocellular carcinoma: meta-analysis of randomized controlled trials. Radiology 2002; 224:47–54 2. Llovet JM, Bruix J. Systematic review of randomized trials for unresectable hepatocellular carcinoma: chemoembolization improves survival. Hepatology 2003; 37:429–442 3. Ho AS, Picus J, Darcy MD, et al. Long-term out-
AJR:200, June 2013
Downloaded from www.ajronline.org by National Cancer Center on 12/29/13 from IP address 49.50.1.229. Copyright ARRS. For personal use only; all rights reserved
Liver Abscess After Transarterial Chemoembolization come after chemoembolization and embolization of hepatic metastatic lesions from neuroendocrine tumors. AJR 2007; 188:1201–1207 4. Therasse E, Breittmayer F, Roche A, et al. Transcatheter chemoembolization of progressive carcinoid liver metastasis. Radiology 1993; 189:541–547 5. Sun Z, Li G, Ai X, et al. Hepatic and biliary damage after transarterial chemoembolization for malignant hepatic tumors: incidence, diagnosis, treatment, outcome and mechanism. Crit Rev Oncol Hematol 2011; 79:164–174 6. Chung JW, Park JH, Han JK, et al. Hepatic tumors: predisposing factors for complications of transcatheter oily chemoembolization. Radiology 1996; 198:33–40 7. Song SY, Chung JW, Han JK, et al. Liver abscess after transcatheter oily chemoembolization for hepatic tumors: incidence, predisposing factors, and clinical outcome. J Vasc Interv Radiol 2001; 12:313–320 8. Kim W, Clark TW, Baum RA, Soulen MC. Risk factors for liver abscess formation after hepatic chemoembolization. J Vasc Interv Radiol 2001; 12:965–968 9. Xia J, Ren Z, Ye S, et al. Study of severe and rare complications of transarterial chemoembolization (TACE) for liver cancer. Eur J Radiol 2006; 59: 407–412 10. de Baère T, Roche A, Amenabar JM, et al. Liver abscess formation after local treatment of liver tumors. Hepatology 1996; 23:1436–1440 11. Chen C, Chen PJ, Yang PM, et al. Clinical and microbiological features of liver abscess after transarterial embolization for hepatocellular car-
cinoma. Am J Gastroenterol 1997; 92:2257–2259 12. Vanderwalde AM, Marx H, Leong L. Liver abscess as a complication of hepatic transarterial chemoembolization: a case report, literature review, and clinical recommendations. Gastrointest Cancer Res 2009; 3:247–251 13. Plentz RR, Lankisch TO, Basturk M, et al. Prospective analysis of German patients with hepatocellular carcinoma undergoing transcatheter arterial chemoembolization with or without prophylactic antibiotic therapy. J Gastroenterol Hepatol 2005; 20: 1134–1136 14. Geschwind JF, Kaushik S, Ramsey DE, Choti MA, Fishman EK, Kobeiter H. Influence of a new prophylactic antibiotic therapy on the incidence of liver abscesses after chemoembolization treatment of liver tumors. J Vasc Interv Radiol 2002; 13:1163–1166 15. Khan W, Sullivan KL, McCann JW, et al. Moxifloxacin prophylaxis for chemoembolization or embolization in patients with previous biliary interventions: a pilot study. AJR 2011; 197:[web] W343–W345 16. Bruix J, Sherman M. Management of hepatocellular carcinoma: an update. Hepatology 2011; 53:1020–1022 17. Shibata T, Yamamoto Y, Yamamoto N, et al. Cholangitis and liver abscess after percutaneous ablation therapy for liver tumors: incidence and risk factors. J Vasc Interv Radiol 2003; 14:1535–1542 18. Miyayama S, Matsui O, Yamashiro M, et al. Ultraselective transcatheter arterial chemoembolization with a 2-f tip microcatheter for small hepatocellular carcinomas: relationship between local
tumor recurrence and visualization of the portal vein with iodized oil. J Vasc Interv Radiol 2007; 18:365–376 19. Ishikawa O, Ohigashi H, Yasuda T, et al. Transarterial chemoembolization in metastatic liver tumor after curative resection of pancreatic cancer. Gan To Kagaku Ryoho 1988; 15:2505–2509 20. Terayama N, Matsui O, Gabata T, et al. Accumulation of iodized oil within the nonneoplastic liver adjacent to hepatocellular carcinoma via the drainage routes of the tumor after transcatheter arterial embolization. Cardiovasc Intervent Radiol 2001; 24:383–387 21. Coldwell DM, Stokes KR, Yakes WF. Embolotherapy: agents, clinical applications, and techniques. Radiographics 1994; 14:623–643; quiz 645–626 22. Matsui O, Kadoya M, Yoshikawa J, et al. Small hepatocellular carcinoma: treatment with subsegmental transcatheter arterial embolization. Radiology 1993; 188:79–83 23. Matsui O, Kadoya M, Yoshikawa J, Gabata T, Takashima T, Demachi H. Subsegmental transcatheter arterial embolization for small hepatocellular carcinomas: local therapeutic effect and 5-year survival rate. Cancer Chemother Pharmacol 1994; 33(suppl):S84–S88 24. Spies JB, Rosen RJ, Lebowitz AS. Antibiotic prophylaxis in vascular and interventional radiology: a rational approach. Radiology 1988; 166:381–387 25. Hoffmann R, Rempp H, Schmidt D, Pereira PL, Claussen CD, Clasen S. Prolonged antibiotic prophylaxis in patients with bilioenteric anastomosis undergoing percutaneous radiofrequency ablation. J Vasc Interv Radiol 2012; 23:545–551
AJR:200, June 2013 1377
