The Effect of Coffee Consumption on the Development of Hepatocellular Carcinoma in Hepatitis B Virus Endemic Area

Eun Sun Jang; Sook-Hyang Jeong; Sang Hyub Lee; Sung Ho Hwang; So Yeon Ahn; Jaebong Lee; Young Soo Park; Jin Hyoek Hwang; Jin-Wook Kim; Nayoung Kim; Dong Ho Lee; Hyun Young Kim


Liver International. 2013;33(7):1092-1099. 

In This Article


Baseline Characteristics of the Study Population

Of the 1364 subjects, 258 patients were diagnosed as HCC and therefore comprised the case group in this case–control design. The mean age of the entire study population was 51.3 ± 12.9 years, and HCC patients were significantly older than health-check examinees or chronic liver disease patients without HCC (median 61.0 years in HCC cases vs. 46.5 years in HCE group vs. 51.0 years in CLD group, P < 0.001, Table 1). The frequency of male gender was highest in HCC group (77.1%, Table 1). However, mean BMI of HCC group was lower than CLD group and higher than HCE group (P < 0.001, Table 1).

CHB patients were more frequent in HCC group (67.8%) rather than in CLD group (55.0%, P < 0.001), but frequencies of chronic hepatitis C patients did not differ significantly among 3 groups (19.2% in HCC group vs. 23.4% in CLD group). Of 175 HCC patients with HBV, 34 (19.4%) had HBeAg and 79 (46.5%) showed high serum HBV DNA level (>105 copies/ml). Although the proportion of patients with HBeAg was significantly smaller (19.4% in HCC group vs. 32.3% in CLD group, P = 0.002), subjects with high HBV DNA level were found more frequently in HCC group compared to CLD group (46.5% in HCC group vs. 37.5% in CLD group, P = 0.054, Table 2). Among 519 patients infected with HBV, 163 patients (31.4%) had received antiviral therapy and the proportion of those who underwent antiviral treatment was higher (34.3%) in CLD group rather than HCC group (25.7%, P = 0.046, data not shown).

Behavioural Characteristics: Coffee Consuming, Alcohol Drinking and Smoking

A total of 1042 (76.4%) of the study population were current coffee drinkers and their mean duration of coffee drinking was 18.0 (SD 14.7) years. As shown in Table 1, current coffee drinkers were least frequent in the HCC group (64.7% in HCC group vs. 88.5% in HCE group vs. 71.9% in CLD group, P < 0.001, Table 1). However, 21 of 35 (61.1%) HCC patients who had quit drinking coffee stopped it after the first diagnosis of HCC. Thus, the frequency of coffee drinking at the time of HCC diagnosis was not different between HCC and CLD groups (72.9% vs. 71.9%). Nevertheless, high coffee consumers (≥3 cups/day) were fewer in the HCC group (12.8%) compared with HCE group (24%) and CLD group (22.2%, P < 0.001, Table 1). We defined high lifetime coffee consumers as those consumed more than 20 000 cups (3 cups per day*mean coffee consuming duration of our study population). The frequency of high lifetime coffee consumers (>20 000 cups) did not differ among the three groups (Table 1).

There were 828 current or past alcohol drinkers (60.7%) and their mean duration of drinking was 22 (SD 11.4) years. Although current alcohol drinkers were rare in HCC group (8.5% vs. 62.1% in HCE group vs. 30.7% in CLD group, P < 0.001, Table 1), the proportion of heavy drinkers (over 60 g of alcohol per day) was highest in HCC group (15.5%) compared with HCE (4.8%) or CLD (7.8%) groups (P < 0.001, Table 1). Subjects consuming a large lifetime amount of alcohol (>150 × 103 g) were also most frequently found in the HCC group (38.4% in HCC group vs. 21.3% in HCE group vs. 19.8% in CLD group, P < 0.001, Table 1).

Similarly, current smokers were least in HCC group compared with the other groups (10.9% in HCC group vs. 24.0% in HCE group vs. 20.2% in CLD group, P < 0.001). However, heavy daily (>1 packs/day) or lifetime (>15 pack*years) smokers were most common in the HCC group (daily smoking amount >1 packs, 10.9% in HCC group vs. 7.9% in HCE group vs. 6.5% in CLD group, P < 0.001; lifetime smoking amount >15 pack*years, 44.6% in HCC group vs. 19.0% in HCE group vs. 23.6% in CLD group, P < 0.001 respectively; Table 1).

Independent Protective Effect of High Coffee Consumption on the Risk of Hepatocellular Carcinoma

To find independent predictors of HCC development, multivariable logistic regression analysis was performed. Because HCC patients tended to quit coffee, alcohol drinking and smoking after diagnosis of the cancer, we used lifetime coffee consumption amounts to evaluate the effect of high coffee consumption on the HCC risk. As a result, age, male gender and significant lifetime smoking were significantly frequent in the disease group relative to the two control groups (Table 2). As shown in Table 2, those who smoked more than 15 pack*year had 1.84 times (95% CI 1.14–2.99) and 1.93 times (95% CI 1.28–2.91) higher HCC risk in HCE and in CLD controls respectively. A heavy lifetime alcohol drinking (>150 × 103 g) raised HCC risk by 2.36 times (95% CI 1.55–3.60) in the CLD group; however, it was not an independent risk factor of HCC in HCE group. Besides, CLD patients infected with the HBV had a significantly higher HCC risk compared with those without HBV infection (OR 4.67, 95% CI 2.90–7.54, Table 2). Especially, the high lifetime coffee consumption (>20 000 cups) lowered the HCC risk by 44% (OR 0.56, 95% CI 0.33–0.95) in HCE group and by 45% (OR 0.55, 95% CI 0.36–0.85) in CLD group, after adjusting for the above-mentioned risk factors (Table 2).

Effect of High Coffee Consumption on the Risk of Hepatocellular Carcinoma According to the HBV Infection Status

We performed subgroup analyses in 884 patients with CLD or HCC to adjust for the different viral aetiology of liver diseases. Of 365 patients (83 HCC and 282 CLD patients) who were not infected with HBV, 147 (40.3%) were chronic hepatitis C patients and the remainder included 124 (34.0%) alcoholic and 94 (25.7%) non-alcoholic, non-viral liver disease patients. Among them, high lifetime coffee consumption (>20 000 cups) lowered the HCC risk by 53% (OR 0.47, 95% CI 0.23–0.94) independently after adjusting for age, gender, obesity and lifetime alcohol drinking/smoking (Table 3, left column). Coffee drinking was not, however, an independent protective factor against HCC in CHB patients. According to the results from multivariable analysis carried out on 519 patients infected with HBV (175 HCC and 344 CLD patients), high lifetime coffee consumption tended to lower HCC risk but lost statistical significance after adjustment for age, gender, obesity, heavy lifetime alcohol drinking/smoking, HBeAg status, serum HBV DNA level and antiviral therapy against HBV. Instead, viral factors significantly affected HCC risk among CHB patients. Those without HBeAg had 2.66 times higher risk for HCC compared to those with HBeAg (95% CI 1.52–4.64, Table 3). Moreover, the high serum HBV DNA level (>105 copies/ml) increased the risk of developing HCC by 102% (OR 2.02, 95% CI 1.24–3.31, Table 3). Antiviral therapy affected HCC risk among these CHB patients independently (OR 0.58, 95% CI 0.36–0.94, Table 3).