Journal of Gastroenterology

, Volume 50, Issue 3, pp 350–360 | Cite as

Clinical characteristics, treatment, and prognosis of non-B, non-C hepatocellular carcinoma: a large retrospective multicenter cohort study

  • Ryosuke TateishiEmail author
  • Takeshi Okanoue
  • Naoto Fujiwara
  • Kiwamu Okita
  • Kendo Kiyosawa
  • Masao Omata
  • Hiromitsu Kumada
  • Norio Hayashi
  • Kazuhiko Koike
Open Access
Original Article—Liver, Pancreas, and Biliary Tract



The number of hepatocellular carcinoma (HCC) patients with non-viral etiologies is increasing in Japan. We conducted a nation-wide survey to examine the characteristics of those patients.


After we assessed the trend of patients who were first diagnosed with HCC at 53 tertiary care centers in Japan from 1991 to 2010, we collected detailed data of 5326 patients with non-viral etiology. The etiologies were categorized as autoimmune hepatitis, primary biliary cirrhosis, alcoholic liver disease (ALD), non-alcoholic fatty liver disease (NAFLD), unclassified, and other. Baseline characteristics at initial diagnosis, the modality of the initial treatment, and survival status were collected via a website. Survival of the patients was assessed by the Kaplan–Meier method and Cox proportional hazard regression.


The proportion of patients with non-viral etiologies increased from 10.0 % in 1991 to 24.1 % in 2010. Of the patients, 92 % were categorized as ALD, NAFLD, or unclassified. Body mass index (BMI) was ≥ 25 kg/m2 in 39 %. Diabetes was most prevalent in NAFLD (63 %), followed by unclassified etiology (46 %) and ALD (45 %). Approximately 80 % of patients underwent radical therapy, including resection, ablation, or transarterial chemoembolization. Survival rates at 3, 5, 10, 15, and 20 years were 58.2, 42.6, 21.5, 15.2, and 15.2 %, respectively. Multivariate analysis revealed that patients with BMI > 22 and ≤ 25 kg/m2 showed the best prognosis versus other BMI categories, after adjusting by age, gender, tumor-related factors, and Child-Pugh score.


Most cases of non-B, non-C HCC are related to lifestyle factors, including obesity and diabetes. Slightly overweight patients showed the best prognosis.


Hepatocellular carcinoma Non-alcoholic fatty liver disease Non-alcoholic steatohepatitis Alcoholic liver disease Retrospective study 


Hepatocellular carcinoma (HCC) is a typical example of an infection-associated malignancy [1]. The geographical distribution of the highly endemic area of HCC overlaps that of chronic hepatitis B and C [2]. Rigorous efforts to control horizontal transmission of hepatitis B virus (HBV) by vaccination since the mid-1980s succeeded in reducing hepatitis B-related HCC in children [3]. Screening for hepatitis C virus (HCV) and the ending of paid blood donations markedly reduced the incidence of transfusion-associated hepatitis [4]. In those with active chronic hepatitis B, long-term suppression using nucleotide analogs may reduce the incidence of HBV-related HCC [5, 6], and the eradication of HCV by interferon-based therapy can reduce HCV-related HCC [7, 8]. It can reasonably be concluded that hepatitis virus-related HCC will continue to decrease in the future [9, 10].

While HCC is a typical example of a virus-related cancer, it is also well known to be strongly related to life style. Chronic alcoholism is a classical risk factor [11]; more recently, obesity has been recognized to strongly affect HCC development in males, versus various other malignancies [12]. There is also growing evidence suggesting that type 2 diabetes increases the incidence of HCC [13, 14]. Due to the globally increasing proportion of the obese population over the past 30 years [15], obesity-related HCC will likely continue to increase.

Unlike virus-related HCC, in which the high-risk populations and surveillance programs are well established, little is known about the characteristics of virus-unrelated HCC. To reduce the forthcoming global burden of obesity-related HCC, to clarify its clinical features is quite important. The Non-B, Non-C Liver Cancer, Etiology, Prognosis and Treatment (NOBLESSE) study was conducted as a special project of the Inuyama Symposium, an assembly of 56 gastroenterology and hepatology units in university hospitals and tertiary care hospitals in Japan, to investigate the characteristics of non-B, non-C HCC patients.

Patients and methods


This retrospective study complied with the ethical guidelines for epidemiological research designed by the Japanese Ministry of Education, Culture, Sports, Science and Technology and Ministry of Health, Labour, and Welfare. The study protocol was approved by the University of Tokyo Medical Research Center Ethics Committee (approval number 3710) and the Institutional Review Board or Ethics Committee of each participating institution. Informed consent was waved because of the retrospective design. This study was registered with the University Hospital Medical Information Network (UMIN) Clinical Trial Registry (UMIN-CTR000007570).

First we collected the number of patients with HCC who were first diagnosed with HCC in the participating hospitals from 1991 to 2010 and categorized them as HBV-related, HCV-related, both HBV and HCV-related, and non-B, non-C to assess trends in the proportion of background etiologies. Next we collected detailed data of non-B, non-C HCC patients defined as negative for both hepatitis B surface antigen (HBsAg) and anti-HCV antibody. Patients who lost HBsAg before the diagnosis of HCC or who were positive for HBV DNA were excluded.

Diagnosis of HCC

The diagnosis of HCC was made by dynamic computed tomography (CT) or dynamic magnetic resonance imaging (MRI) with consideration of hyperattenuation in the arterial phase, with washout in the late phase as a definite sign of this disease [16] or pathology. In years when dynamic CT was not available, the diagnosis was also made by angiography.

Data collection

The patients were registered via a website specially designed by the investigators. The following characteristics at diagnosis were collected: age, gender, body height, body weight, etiology of background liver disease, daily alcohol consumption; comorbidities including liver cirrhosis, fatty liver by ultrasonography, hypertension, dyslipidemia, and diabetes; tumor factors including tumor size of the maximal nodule, number of tumor nodules, the presence of vascular invasion, and extrahepatic metastasis; symptoms including ascites and hepatic encephalopathy, laboratory data, including serum albumin, total bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (ALP), gamma-glutamyl transpeptidase (GGT), platelet count, prothrombin activity, alpha-fetoprotein (AFP), des-gamma-carboxyprothrombin (DCP), and lens culinaris agglutinin-reactive fraction of AFP (AFP-L3); and treatment modality for the first time, including hepatic resection, liver transplantation, ablation, transarterial chemoembolization (TACE), transarterial chemotherapy, systemic chemotherapy, radiation therapy, and supportive therapy. Body mass index (BMI), Child–Turcotte–Pugh (CTP) score, and Barcelona-Clinic–Liver-Cancer (BCLC) stages were calculated automatically using the data obtained above.

The etiology of background liver diseases was categorized as follows: autoimmune hepatitis (AIH), primary biliary cirrhosis (PBC), AIH–PBC overlap syndrome, alcoholic liver disease, non-alcoholic fatty liver disease (NAFLD), Budd-Chiari syndrome, hemochromatosis, Wilson disease, and others. The diagnosis of the background liver disease, hypertension, dyslipidemia, and diabetes was made by the attending physician, based on the Japanese clinical guidelines for each disease. Daily alcohol consumption was calculated from forms of alcohol and frequency. Alcoholic liver disease was defined as chronic liver injury with daily alcohol consumption ≥ 80 g/day without another definite etiology. NAFLD was defined as a history of fatty liver or who were diagnosed with fatty liver, radiologically or pathologically, with alcohol consumption ≤ 20 g/day. Those with cryptogenic chronic liver disease who did not meet the criteria described above for alcoholic liver disease or NAFLD were categorized as unclassified.

Patient survival status was also registered. Status was defined as alive, dead, or lost to follow-up. Observations were censored on 31 December 2011. In diseased patients, the cause of death was categorized according to the criteria of the Liver Cancer Study Group of Japan [17], as follows: liver cancer progression, liver failure, gastrointestinal bleeding, gastro-esophageal varices rupture, rupture of liver cancer, operative death, other, and unknown.

Statistical analysis

Data are expressed as medians with 25th to 75th percentiles, unless otherwise indicated. Numbers and percentages were used for qualitative variables. Student’s t test was used for comparisons of two continuous variables. Differences among groups were assessed with one-way analysis of variance (ANOVA) for continuous data, and with the Chi squared test for categorical data. The Cochran–Armitage trend test was used to evaluate increasing or decreasing trends in etiology. Survival time was defined as the interval between the day of the first diagnosis and death or the last visit to the hospital until 31 December 2011. Cumulative survival curves were constructed with the Kaplan–Meier method and compared with the log-rank test. To assess the hazard ratios of various factors on overall survival, the Cox proportional hazard model was used.

Statistical analyses were performed using the ‘R’ software (ver. 2.13.0; All tests were two-sided, and p values < 0.05 were considered to indicate statistical significance.


Patient profiles

Of 33,782 patients who were first diagnosed with HCC at the 53 participating hospitals from 1991 to 2010, 5326 (15.8 %) were categorized as non-B, non-C. A marked increase in the proportion of patients categorized as non-B, non-C was observed (p < 0.001 by Cochran–Armitage test; Fig. 1). The proportion of non-B, non-C patients was 24.1 % in 2010, whereas it was only 10.0 % in 1991. The distribution of background liver diseases among non-B, non-C patients was as follows: AIH in 161 (3.0 %), PBC in 164 (3.1 %), AIH–PBC overlap syndrome in 18 (0.3 %), alcoholic liver disease in 1423 (26.7 %), NAFLD in 596 (11.2 %), Budd-Chiari Syndrome in 20 (0.4 %), hemochromatosis in 9 (0.2 %), Wilson’s disease in 5 (0.1 %), unclassified in 2875 (54.0 %), and other in 53 (1.0 %). ‘Other’ included schistosomiasis japonica, suspicion of autoimmune liver diseases, and normal liver. As few patients were categorized as AIH–PBC overlap syndrome, Budd-Chiari syndrome, hemochromatosis and Wilson’s disease, they were combined with ‘others’ in Table 1. Among non-B, non-C patients, 31 and 10 % were diagnosed as HCC at the department of gastroenterology or hepatology and other department in the participating hospital, respectively. The remaining 59 % were diagnosed at other hospitals and referred to the participating hospitals. Forty-one percent of patients were followed by imaging modalities before the diagnosis of HCC.
Fig. 1

Trend in background liver disease in hepatocellular carcinoma in Japan. A marked increase in the proportion of patients categorized as non-B, non-C in the participating hospitals was observed (p < 0.001 by Cochran–Armitage test)

Table 1

Baseline characteristics of the HCC patients analyzed in this study (n = 5,326)





Alcoholic liver disease




Number of patients








Age (year)

















Male gender [n (%)]

4,022 (75.5)

43 (26.7)

52 (31.3)

1,327 (93.3)

348 (58.4)

2,188 (76.1)

64 (61.0)

Alcohol consumption (g/day)a

 ≤ 20 [n (%)]

2623 (50.9)

144 (90.0)

146 (90.7)


596 (100.0)

1661 (59.0)

80 (86.0)

 21–79 [n (%)]

1179 (22.9)

9 (5.6)

9 (5.6)


1154 (41.0)

7 (7.5)

 ≥ 80 [n (%)]

1351 (26.2)

7 (3.7)

6 (3.7)

1423 (100.0)


6 (6.5)

 Diabetes [n (%)]b

2345 (46.1)

48 (30.6)

27 (17.0)

621 (45.2)

359 (62.7)

1264 (46.4)

26 (27.1)

 Hypertension [n (%)]c

2063 (42.7)

51 (35.4)

42 (26.8)

493 (38.0)

313 (55.5)

1135 (44.1)

29 (31.9)

 Dyslipidemia [n (%)]d

720 (14.6)

26 (17.1)

12 (7.6)

171 (12.7)

125 (22.9)

374 (14.2)

12 (12.6)

 Fatty liver [n (%)]e

936 (24.0)

18 (15.5)

7 (5.5)

219 (20.7)

280 (64.4)

403 (19.3)

9 (13.4)

 Liver Cirrhosis [n (%)]f

3439 (67.0)

127 (80.9)

145 (87.9)

1115 (80.2)

368 (63.4)

1619 (59.0)

65 (67.0)

 Anti-HBcAb positive [n (%)]g

1501 (40.3)

27 (23.5)

35 (31.3)

410 (40.8)

159 (34.6)

837 (43.0)

33 (40.7)


















Platelet count (×109/μL)h

















Child-Pugh classi

 A [n (%)]

3500 (69.0)

89 (57.4)

83 (52.9)

843 (62.1)

439 (76.5)

1976 (72.4)

70 (72.2)

 B [n (%)]

1231 (24.3)

54 (34.8)

57 (36.3)

383 (28.2)

120 (20.9)

595 (21.8)

22 (22.7)

 C [n (%)]

338 (6.7)

12 (7.7)

17 (10.8)

131 (9.7)

15 (2.6)

158 (5.8)

5 (5.2)

Tumor characteristics

 Maximal tumor size (cm)j

















 Diffuse type [n (%)]

209 (4.0)

6 (3.7)

1 (0.6)

62 (4.4)

17 (2.9)

119 (4.2)

4 (3.8)

 Number of nodulesk

  Single [n (%)]

2700 (51.1)

87 (54.0)

110 (66.3)

664 (46.8)

340 (57.0)

1443 (50.8)

56 (53.8)

  2–3 [n (%)]

1368 (25.9)

46 (28.6)

40 (24.1)

402 (28.3)

156 (26.2)

697 (24.5)

27 (26.0)

  > 3 [n (%)]

1220 (23.1)

28 (17.4)

16 (9.6)

353 (24.9)

100 (16.8)

702 (24.7)

21 (20.2)

 Vascular invasion [n (%)]l

187 (3.5)

3 (1.9)

1 (0.6)

52 (3.7)

13 (2.2)

116 (4.1)

2 (1.9)

 Extrahepatic metastasis [n (%)]m

401 (7.6)

8 (5.0)

2 (1.2)

114 (8.0)

26 (4.4)

244 (8.6)

7 (6.7)

 AFP (ng/mL)n

  ≤20 [n (%)]

2908 (59.4)

80 (54.1)

71 (51.4)

827 (62.4)

361 (63.1)

1515 (58.0)

54 (55.7)

  21–200 [n (%)]

820 (16.8)

33 (22.3)

29 (21.0)

229 (17.3)

92 (16.1)

423 (16.2)

14 (14.4)

  >200 [n (%)]

1164 (23.8)

35 (23.6)

38 (27.5)

270 (20.4)

119 (20.8)

673 (25.8)

29 (29.9)

 DCP (mAU/mL)o

  ≤100 [n (%)]

2121 (45.8)

75 (53.6)

81 (59.1)

593 (46.8)

299 (53.9)

1032 (42.1)

41 (47.7)

  101–400 [n (%)]

787 (17.0)

23 (16.4)

25 (18.2)

227 (17.9)

95 (17.1)

400 (16.3)

17 (19.8)

  >400 [n (%)]

1727 (37.3)

42 (30.0)

31 (22.6)

448 (35.3)

161 (29.0)

1017 (41.5)

28 (32.6)

 AFP-L3 (%)p

  ≤10 [n (%)]

1765 (67.7)

53 (64.6)

39 (55.7)

498 (69.6)

263 (73.5)

881 (66.1)

31 (66.0)

  10.1–15 [n (%)]

74 (2.8)

3 (3.7)

4 (5.7)

17 (2.4)

7 (2.0)

43 (3.2)

0 (0)

  >15 [n (%)]

767 (29.4)

26 (31.7)

27 (38.6)

201 (28.1)

88 (24.6)

409 (30.7)

16 (34.0)

As few patients were categorized as having the AIH–PBC overlap syndrome, Budd-Chiari syndrome, hemochromatosis or Wilson’s disease, they were combined with ‘others’. Data were missing in a173, b241, c498, d388, e1434, f193, g1606, h61, i257, j42, k38, l28, m26, n434, o691, and p3677 patients

AFP alpha-fetoprotein, AFP-L3 lens culinaris agglutinin-reactive fraction of AFP, ALT alanine aminotransferase, Anti-HBcAb anti-hepatitis B core antibody, DCP des-gamma-carboxy prothrombin, IQR interquartile range

The median [interquartile range (IQR)] age in the entire cohort was 70.0 (63.0–75.0) years and approximately three-quarters were males. Patients with alcoholic liver disease were significantly younger than other etiologies (p < 0.001). The male to female ratio was different among the etiologies: females predominated in autoimmune liver diseases. The vast majority were non drinkers or light drinkers, except for those with alcoholic liver disease or unclassified etiology. Among those judged as unclassified, 41 % were moderate drinkers.

The distribution of BMI varied across the etiologies and gender. The median BMI was the highest in those with NAFLD. Females had significantly higher BMI than males in NAFLD and those unclassified (p = 0.01 and <0.001, respectively; Fig. 2).
Fig. 2

Body mass index according to background liver disease. Median (25th–75th percentiles) BMI values in all categories were 23.8 (21.6–26.3) kg/m2 in males and 24.4 (21.8–27.5) kg/m2 in females. Box plot ‘whiskers’ show the minimum and maximum values; the horizontal line in each box plot shows the median, and the colored segment shows the interquartile range. AIH autoimmune hepatitis, PBC primary biliary cirrhosis, ALD alcoholic liver disease, NAFLD non-alcoholic fatty liver disease

Nearly half of the patients were complicated with diabetes (Table 1, Supplementary Fig. 1). The proportion of those with diabetes was highest in NAFLD patients. A similar trend was observed in the proportions of hypertension and dyslipidemia. The presence of fatty liver, judged by ultrasonography at the diagnosis of HCC, varied across the etiologies. The proportion was approximately 20 % in alcoholic liver disease and unclassified etiology, while it was lower in autoimmune liver diseases, especially PBC. It was also suggested that fatty liver could not be detected by ultrasonography in approximately 30 % at the diagnosis of HCC in NAFLD.

Approximately two-thirds of the patients were complicated with cirrhosis. The proportion of those with cirrhosis was lower in those with NAFLD and unclassified etiology compared with other etiologies (p < 0.001). Reflecting the proportion of cirrhosis, platelet counts were highest in those with unclassified etiology, followed by those with NAFLD.

Regarding the diagnosis process, 30.3 % of the patients had their tumor pointed out for the first time in the participating department, 10.6 % in another department of the same hospital, and 59.1 % at other hospitals. Patients were diagnosed at more advanced stages in those with unclassified etiology; the tumor size was the largest and the proportion of patients with vascular invasion and extrahepatic metastasis was also the largest. The sensitivity of DCP was superior to that of AFP (54.2 vs. 40.6 % with cutoff values of 100 mAU/mL and 20 ng/mL, respectively).

Treatment and survival

Among 5058 patients in whom BCLC staging could be determined, 2533 (50.1 %), 1913 (37.8 %), 283 (5.6 %), and 329 (6.5 %) were categorized as stages A, B, C, and D, respectively (Table 2). The distribution of the initial treatment was as follows: resection in 1073 (20.3 %), ablation in 1060 (20.0 %), TACE + ablation in 470 (8.9 %), TACE in 1590 (30.1 %), transarterial chemotherapy with one-shot and continuous infusion in 99 (1.9 %), systemic therapy in 20 (0.3 %), radiation therapy in 20 (0.4 %), liver transplantation in 17, others in 30 (0.6 %), and supportive care in 429 (8.1 %).
Table 2

Distribution of treatments according to BCLC stage






Number of patients





Hepatic resection [n (%)]

616 (24.3)

398 (20.8)

30 (10.6)

3 (0.9)

Ablation [n (%)]

887 (35.0)

81 (4.2)

4 (1.4)

52 (15.8)

TACE + ablation [n (%)]

335 (13.2)

116 (6.1)

3 (1.1)

4 (1.2)

TACE [n (%)]

517 (17.1)

840 (43.9)

78 (27.6)

83 (25.2)

Transarterial chemotherapy [n (%)]

83 (3.2)

278 (14.5)

87 (30.7)

27 (8.2)

Systemic therapy [n (%)]

5 (0.2)

50 (2.6)

25 (8.8)

7 (2.1)

Radiation therapy [n (%)]

5 (0.2)

4 (0.2)

3 (1.1)

5 (1.5)

Liver transplantation [n (%)]

11 (0.4)

6 (0.3)


0 (0.0)

Others [n (%)]

12 (0.5)

5 (0.3)

2 (0.7)

4 (1.2)

Supportive therapy [n (%)]

64 (2.5)

135 (7.1)

51 (18.0)

144 (43.8)

BCLC stage could not be determined in 268 patients

TACE transarterial chemoembolization

During the mean follow-up period of 2.6 years, 2225 patients died and 670 patients were lost to follow-up. The causes of death were cancer progression in 1411 (58.0 %), liver failure in 359 (14.8 %), gastrointestinal bleeding, including varices rupture, in 87 (3.6 %), tumor rupture in 71 (2.9 %), operative death in 13 (0.5 %), and other in 284 (11.7 %). The cause of death was unspecified in 206 (8.5 %). Median survival time [95 % confidence interval (CI)] after the initial diagnosis of HCC was 4.03 (3.82–4.20) years. Overall survival rates at 1, 3, 5, 7, 10, 15, and 20 years were 80.1, 58.2, 42.6, 32.2, 21.5, 15.2, and 15.2 %, respectively (Fig. 3a). When stratified by BCLC stage, the median (95 % CI) survival times were 6.39 (5.96–6.85), 2.48 (2.34–2.68), 0.83 (0.61–1.03), and 0.80 (0.64–1.23) years in BCLC stages A, B, C, and D, respectively. There was a significant difference in survival among the stages (Fig. 3b, p < 0.001).
Fig. 3

Overall survival. A Overall survival of the entire patient cohort. Overall survival rates at 1, 3, 5, 7, 10, 15, and 20 years were 80.1, 58.2, 42.6, 32.2, 21.5, 15.2, and 15.2 %, respectively. B Overall survival according to BCLC stage. Survival rates at 1, 3, 5, 7, 10, 15, and 20 years were 94.5, 76.4, 58.7, 44.7, 30.7, 21.9, and 21.9 % in stage A, 71.1, 44.1, 29.1, 22.2, 13.0, 9.0, and 9.0 % in stage B, 44.6, 18.8, 15.5, 9.3, and 9.3 % in Stage C, and 48.0, 24.4, 12.3, 7.3, 3.1 %, respectively, in Stage D

Univariate Cox regression analysis revealed that the following factors were significantly related to poor survival: old age (p < 0.001), male gender (p = 0.003), alcohol consumption ≥80 g/day (p < 0.001), BMI (p = 0.001), Child-Pugh score (p < 0.001), maximal tumor size (p < 0.001), number of nodules (p < 0.001), the presence of vascular invasion (p < 0.001), the presence of extrahepatic metastasis (p < 0.001), AFP (p < 0.001), DCP (p < 0.001), and AFP-L3 (p < 0.001). The presence of diabetes was indicated as a better prognosis factor, though with marginal significance (hazard ratio, 0.93; 95 % CI, 0.86–1.01; p = 0.08). BMI showed a V-shaped hazard distribution: those with BMIs of 22.1–25 kg/m2 had the best outcomes, whereas those with higher and lower BMI showed worse prognoses. We plotted relative hazard against BMI using cubic splines. The V-shape hazard distribution was also observed in the plot (Supplementary Fig. 2).

We performed a multivariate analysis using the variables above, except that AFP-L3 was excluded because of missing values. The results showed that age, BMI, alcohol consumption, Child-Pugh score, tumor size, number of tumor nodules, extrahepatic metastasis, AFP, and DCP were significant factors related to poor prognosis (Fig. 4). The presence of diabetes again showed no statistical significance.
Fig. 4

Multivariate Cox proportional hazard regression analysis of survival. AFP alpha-fetoprotein, AFP-L3 lens culinaris agglutinin-reactive fraction of AFP, DCP des-gamma-carboxy prothrombin AFP alpha-fetoprotein, AFP-L3 lens culinaris agglutinin-reactive fraction of AFP, ALT alanine aminotransferase, Anti-HBcAb anti-hepatitis B core antibody, DCP des-gamma-carboxy prothrombin, IQR interquartile range


In the present study, a rapidly increasing proportion of HCC patients with non-viral etiologies was found. A similar trend was reported in a national survey by the Liver Cancer Study Group of Japan [18]. As the number of newly diagnosed HCC cases in Japan was almost at a plateau throughout the study period [19], not only the proportion, but also the number, of patients with non-viral etiologies was increasing. As a risk factor of HCC, alcohol consumption has not increased over the last two decades in Japan according to statistics from the Ministry Labour and Welfare in Japan [20]. In contrast, the size of the obese population is increasing rapidly due to changes in the diet in Japan. The proportion of patients with diabetes has also increased in the past three decades [21]. It seems reasonable that the rapidly increasing number of HCC patients with non-viral etiologies was largely due to the rapidly increasing obese population.

Among non-viral chronic liver diseases, the natural history of AIH, PBC, and alcoholic liver disease are well known compared with that of NAFLD. In these three, HCC ordinarily arises through cirrhosis after long-lasting chronic inflammation in the liver [22, 23, 24]. Indeed, liver cirrhosis was a complication in more than 80 % of those patients. In contrast, the proportion of cirrhosis was smaller and platelet counts were higher in NAFLD patients than those with AIH, PBC, and alcoholic liver disease. It has been reported that a significant proportion of patients (41.7 %) with both NAFLD and HCC are not complicated with cirrhosis [25]. That a significant proportion of patients with NAFLD or unclassified etiology were not complicated with cirrhosis suggests that to characterize a high-risk population within them would be difficult.

In this study, almost half of the patients were complicated with diabetes mellitus. According to a systematic review investigating the relationship between diabetes and HCC, the presence of diabetes is an approximately 2.5-fold risk of HCC [26]. Judging from the wide variation in the proportion of patients with diabetes among the etiologies, it seems that diabetes correlates more strongly with hepatocarcinogenesis in some chronic liver diseases, such as NAFLD, than others.

In this study, we defined NAFLD as a history of fatty liver and alcohol consumption of no more than 20 g/day. As shown in Table 1, fatty liver was not diagnosed by ultrasonography at the diagnosis of HCC in approximately 30 % of patients with NAFLD-related HCC. Those patients would be categorized as unclassified when a history of fatty liver was not confirmed. That is, a significant proportion of those categorized as unclassified could be burn-out non-alcoholic steatohepatitis (NASH). Similarly, alcohol-related HCC could be included in unclassified patients because approximately 40 % of the patients in the category were moderate drinkers. In the first place, it might be unreasonable to categorize those patients clearly, because moderate alcohol intake, obesity, and fatty liver are mutually correlated and may have a synergistic effect on hepatocarcinogenesis.

Occult infection with HBV represented by the presence of antibody to hepatitis B core antigen (anti-HBc) has been considered as a risk factor of non-B, non-C HCC defined as negative for both HBsAg and anti-HCV antibody [27, 28]. Indeed the prevalence of anti-HBc antibody was higher in this study compared to a previous report in blood donors [29]. It is also to be noted that those with anti-HBc antibody may include chronic HBV carriers with HBsAg loss before the diagnosis of HCC, who had significant risk for HCC [30].

Patients were diagnosed at less-advanced stages than we expected. This is partly because all participating hospitals were tertiary care centers. Those with terminal stages diagnosed in primary or secondary hospitals were unlikely to be referred to the participating hospitals. In addition, 41 % of patients were followed by imaging modalities before the diagnosis of HCC. As a result, a large majority of patients underwent radical therapies, such as hepatic resection, ablation, or TACE, as the initial treatment.

Prognostic factors for HCC have been investigated fully in previous studies [31]. However, to our knowledge, this is the first report of the detailed relationship between BMI and survival in HCC patients. Indeed, BMI showed a V-shaped hazard function for death. It is well known that the relationship between BMI and all-cause mortality is V-shaped, with a BMI around 22 kg/m2 showing the best prognosis. However in this study, the lowest relative hazard was observed at a slightly overweight BMI. We had expected that the best BMI would be around 22 kg/m2, because obesity is thought to affect hepatocarcinogenesis in this cohort and may affect recurrence after treatment. This would be because the relatively underweight patients included those with more advanced disease. However, the trend remained after adjustment for other significant factors, including those related to the tumor. Overweight patients may have some advantage versus underweight patients that we did not investigate.

The presence of diabetes did not affect survival in this study. One meta-analysis reported that the presence of diabetes increased the risk of all-cause mortality in HCC patients by 1.38-fold (95 % CI, 1.13–1.68) [32]. It is quite reasonable that those with diabetes had additional risk for death from cardiovascular, cerebrovascular, infectious or renal diseases. Some kind of biases might exist behind the fact that the presence of diabetes did not worsen the patients’ survival, which needs further investigation.

Most of the major limitations of this study relate to its retrospective design.

(1) Because the major data source was a database maintained by each participating hospital, some data were missing. Patients who were not registered in the database could not be entered into this study. However, the proportion of patients with missing data on important items, such as alcohol consumption, was less than 5 %; this would not affect the overall results. (2) As the amount of daily alcohol intake was self-reported, some patients might have underreported their alcohol intakes. Some should possibly have been categorized as having alcoholic liver disease. (3) Similarly, because the diagnosis of NAFLD was based on a past history or ultrasound examination at the diagnosis of HCC, undiagnosed burn-out NASH patients were included in those unclassified, especially when not followed in clinics or hospitals. Based on the high proportion of those with lifestyle diseases and moderate drinkers, at least a majority of those unclassified would be related to chronic alcoholism, obesity, or both.

In conclusion, the proportion of HCC patients without chronic viral hepatitis in Japan is increasing rapidly. Most had lifestyle disease-related backgrounds, especially related to obesity. Narrowing down a high-risk population would be difficult because one-third of the patients were non-cirrhotic, and obesity, fatty liver, and diabetes are prevalent in Japan.



This work was supported partly by Health Sciences Research Grants of The Ministry of Health, Labour and Welfare of Japan (Research on Hepatitis). Eisai Co. supported the maintenance fee of the website for the data registration. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Conflict of interest

The authors declare that they have no conflict of interest.

Supplementary material

535_2014_973_MOESM1_ESM.docx (375 kb)
Supplementary material 1 (DOCX 374 kb) Supplementary Figure 1. Proportions of life-style-related factors among background liver diseases. (A) The proportion of patients diagnosed with diabetes was 46.1 % in all categories. The proportion was the highest in NAFLD (62.7 %) and the lowest in PBC (17 %). (B) The proportion of patients diagnosed with hypertension was 42.7 % in all categories. The proportion was the highest in NAFLD (55.5 %) and the lowest in PBC (26.8 %). (C) The proportion of patients diagnosed with dyslipidemia was 14.7 % in all categories. The proportion was the highest in NAFLD (23.9 %) and the lowest in PBC (7.6 %). (D) The proportion of patients diagnosed with dyslipidemia was 24.0 % in all categories. The proportion was the highest in NAFLD (64.4 %) and the lowest in PBC (5.5 %). AIH, autoimmune hepatitis; PBC, primary biliary cirrhosis; ALD, alcoholic liver disease; NAFLD, non-alcoholic fatty liver disease. Supplementary Figure 2. Estimated hazard function of BMI. The relative hazard was lowest at a BMI of ~ 26 kg/m2


  1. 1.
    Parkin DM. The global health burden of infection-associated cancers in the year 2002. Int J Cancer. 2006;118(12):3030–44.CrossRefPubMedGoogle Scholar
  2. 2.
    Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin. 2011;61(2):69–90.CrossRefPubMedGoogle Scholar
  3. 3.
    Chang MH, Chen CJ, Lai MS, Hsu HM, Wu TC, Kong MS, et al. Universal hepatitis B vaccination in Taiwan and the incidence of hepatocellular carcinoma in children. Taiwan Childhood Hepatoma Study Group. N Engl J Med. 1997;336(26):1855–9 [Epub 1997/06/26].CrossRefPubMedGoogle Scholar
  4. 4.
    Yoshizawa H. Hepatocellular carcinoma associated with hepatitis C virus infection in Japan: projection to other countries in the foreseeable future. Oncology. 2002;62(Suppl 1):8–17.CrossRefPubMedGoogle Scholar
  5. 5.
    Liaw YF, Sung JJ, Chow WC, Farrell G, Lee CZ, Yuen H, et al. Lamivudine for patients with chronic hepatitis B and advanced liver disease. N Engl J Med. 2004;351(15):1521–31 [Epub 2004/10/08].CrossRefPubMedGoogle Scholar
  6. 6.
    Yokosuka O, Takaguchi K, Fujioka S, Shindo M, Chayama K, Kobashi H, et al. Long-term use of entecavir in nucleoside-naive Japanese patients with chronic hepatitis B infection. J Hepatol. 2010;52(6):791–9.CrossRefPubMedGoogle Scholar
  7. 7.
    Nishiguchi S, Kuroki T, Nakatani S, Morimoto H, Takeda T, Nakajima S, et al. Randomised trial of effects of interferon-alpha on incidence of hepatocellular carcinoma in chronic active hepatitis C with cirrhosis. Lancet. 1995;346(8982):1051–5 [Epub 1995/10/21].CrossRefPubMedGoogle Scholar
  8. 8.
    Yoshida H, Shiratori Y, Moriyama M, Arakawa Y, Ide T, Sata M, et al. Interferon therapy reduces the risk for hepatocellular carcinoma: national surveillance program of cirrhotic and non-cirrhotic patients with chronic hepatitis C in Japan. IHIT Study Group. Inhibition of Hepatocarcinogenesis by Interferon Therapy. Ann Intern Med. 1999;131(3):174–81.CrossRefPubMedGoogle Scholar
  9. 9.
    Yuen MF, Hou JL, Chutaputti A. Asia Pacific Working Party on Prevention of Hepatocellular C. Hepatocellular carcinoma in the Asia pacific region. J Gastroenterol Hepatol. 2009;24(3):346–53.CrossRefPubMedGoogle Scholar
  10. 10.
    Tanaka H, Imai Y, Hiramatsu N, Ito Y, Imanaka K, Oshita M, et al. Declining incidence of hepatocellular carcinoma in Osaka, Japan, from 1990 to 2003. Ann Intern Med. 2008;148(11):820–6 [Epub 2008/06/04].CrossRefPubMedGoogle Scholar
  11. 11.
    Yu MC, Tong MJ, Govindarajan S, Henderson BE. Nonviral risk factors for hepatocellular carcinoma in a low-risk population, the non-Asians of Los Angeles County, California. J Natl Cancer Inst. 1991;83(24):1820–6 [Epub 1991/12/18].CrossRefPubMedGoogle Scholar
  12. 12.
    Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med. 2003;348(17):1625–38 [Epub 2003/04/25].CrossRefPubMedGoogle Scholar
  13. 13.
    El-Serag HB, Tran T, Everhart JE. Diabetes increases the risk of chronic liver disease and hepatocellular carcinoma. Gastroenterology. 2004;126(2):460–8.CrossRefPubMedGoogle Scholar
  14. 14.
    Lai MS, Hsieh MS, Chiu YH, Chen TH. Type 2 diabetes and hepatocellular carcinoma: a cohort study in high prevalence area of hepatitis virus infection. Hepatology. 2006;43(6):1295–302 [Epub 2006/05/27].CrossRefPubMedGoogle Scholar
  15. 15.
    Finucane MM, Stevens GA, Cowan MJ, Danaei G, Lin JK, Paciorek CJ, et al. National, regional, and global trends in body-mass index since 1980: systematic analysis of health examination surveys and epidemiological studies with 960 country-years and 9.1 million participants. Lancet. 2011;377(9765):557–67.CrossRefPubMedGoogle Scholar
  16. 16.
    The committee for revision of the Clinical Practice. Guidelines for Hepatocellular Carcinoma. Clinical Practice Guidelines for Hepatocellular Carcinoma: The Japan Society of Hepatology 2009 update. Hepatol Res. 2010;40(s1):16–47.Google Scholar
  17. 17.
    Cancer Study Group of Japan. General Rules for the Clinical and Pathological Study of Primary Liver Cancer, Second English Edition. Tokyo: Kasahara; 2003.Google Scholar
  18. 18.
    Ikai I, Arii S, Okazaki M, Okita K, Omata M, Kojiro M, et al. Report of the 17th nationwide follow-up survey of primary liver cancer in Japan. Hepatol Res. 2007;37(9):676–91.CrossRefPubMedGoogle Scholar
  19. 19.
    Cancer Statistics in Japan. Foundation for Promotion of Cancer Research; 2012 [cited 2013 March 20th]. Available from:
  20. 20.
    Higuchi S, Matsushita S, Maesato H, Osaki Y. Japan: alcohol today. Addiction. 2007;102(12):1849–62 [Epub 2007/08/08].CrossRefPubMedGoogle Scholar
  21. 21.
    National Nutrition Survey in Japan. Ministry of Health Labour, and Welfare of Japan; 2012 [20th March]. Available from:
  22. 22.
    Liang Y, Yang Z, Zhong R. Primary biliary cirrhosis and cancer risk: a systematic review and meta-analysis. Hepatology. 2012;56(4):1409–17.CrossRefPubMedGoogle Scholar
  23. 23.
    Sola R, Alvarez MA, Balleste B, Montoliu S, Rivera M, Miquel M, et al. Probability of liver cancer and survival in HCV-related or alcoholic-decompensated cirrhosis. A study of 377 patients. Liver Int. 2006;26(1):62–72.CrossRefPubMedGoogle Scholar
  24. 24.
    Montano-Loza AJ, Carpenter HA, Czaja AJ. Predictive factors for hepatocellular carcinoma in type 1 autoimmune hepatitis. Am J Gastroenterol. 2008;103(8):1944–51 Epub 2008/06/20.CrossRefPubMedGoogle Scholar
  25. 25.
    Ertle J, Dechene A, Sowa JP, Penndorf V, Herzer K, Kaiser G, et al. Non-alcoholic fatty liver disease progresses to hepatocellular carcinoma in the absence of apparent cirrhosis. Int J Cancer. 2011;128(10):2436–43.CrossRefPubMedGoogle Scholar
  26. 26.
    El-Serag HB, Hampel H, Javadi F. The association between diabetes and hepatocellular carcinoma: a systematic review of epidemiologic evidence. Clin Gastroenterol Hepatol. 2006;4(3):369–80 [Epub 2006/03/11].CrossRefPubMedGoogle Scholar
  27. 27.
    Kondo R, Nakashima O, Sata M, Imazeki F, Yokosuka O, Tanikawa K, et al. Pathological characteristics of patients who develop hepatocellular carcinoma with negative results of both serous hepatitis B surface antigen and hepatitis C virus antibody. Hepatol Res. 2013.Google Scholar
  28. 28.
    Abe H, Yoshizawa K, Kitahara T, Aizawa R, Matsuoka M, Aizawa Y. Etiology of non-B non-C hepatocellular carcinoma in the eastern district of Tokyo. J Gastroenterol. 2008;43(12):967–74.CrossRefPubMedGoogle Scholar
  29. 29.
    Iizuka H, Ohmura K, Ishijima A, Satoh K, Tanaka T, Tsuda F, et al. Correlation between anti-HBc titers and HBV DNA in blood units without detectable HBsAg. Vox Sang. 1992;63(2):107–11.CrossRefPubMedGoogle Scholar
  30. 30.
    Simonetti J, Bulkow L, McMahon BJ, Homan C, Snowball M, Negus S, et al. Clearance of hepatitis B surface antigen and risk of hepatocellular carcinoma in a cohort chronically infected with hepatitis B virus. Hepatology. 2010;51(5):1531–7.CrossRefPubMedGoogle Scholar
  31. 31.
    Tandon P, Garcia-Tsao G. Prognostic indicators in hepatocellular carcinoma: a systematic review of 72 studies. Liver Int. 2009;29(4):502–10 [Epub 2009/01/15].CrossRefPubMedCentralPubMedGoogle Scholar
  32. 32.
    Yang WS, Va P, Bray F, Gao S, Gao J, Li HL, et al. The role of pre-existing diabetes mellitus on hepatocellular carcinoma occurrence and prognosis: a meta-analysis of prospective cohort studies. PLoS One. 2011;6(12):e27326.CrossRefPubMedCentralPubMedGoogle Scholar

Copyright information

© The Author(s) 2014

Open AccessThis article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.

Authors and Affiliations

  • Ryosuke Tateishi
    • 1
    Email author
  • Takeshi Okanoue
    • 2
  • Naoto Fujiwara
    • 1
  • Kiwamu Okita
    • 3
  • Kendo Kiyosawa
    • 4
  • Masao Omata
    • 5
  • Hiromitsu Kumada
    • 6
  • Norio Hayashi
    • 7
  • Kazuhiko Koike
    • 1
  1. 1.Department of Gastroenterology, Graduate School of MedicineThe University of TokyoTokyoJapan
  2. 2.Saiseikai Suita HospitalSuitaJapan
  3. 3.Ajisu Kyoritsu HospitalYamaguchiJapan
  4. 4.Shironishi HospitalMatsumotoJapan
  5. 5.Yamanashi Prefectural Hospital OrganizationKofuJapan
  6. 6.Toranomon HospitalTokyoJapan
  7. 7.Kansai Rosai HospitalAmagasakiJapan

Personalised recommendations