Cigarette smoking, alcohol drinking and the risk of gallbladder cancer death: A prospective cohort study in Japan
Gallbladder cancer is a rare cancer with a poor prognosis, and few risk factors have been identified to date. This prospective study was conducted to evaluate the association of cigarette smoking and alcohol consumption with the risk of gallbladder cancer death. A baseline survey in 45 areas throughout Japan was conducted from 1988 to 1990 using a self-administered questionnaire, and a total of 113,496 participants (65,740 women) aged 40–89 years at entry were followed for 15 years. During the follow-up period, 165 gallbladder cancer deaths (95 women) were observed. Among women, the hazard ratio (HR) [95 percent confidence interval: 95% CI] of current smoker was 2.00 [0.91–4.42], when adjusted for age and drinking. There was no clear association between alcohol consumption and the risk. Among men, HR of current smoker was 2.27 [1.05–4.90]. HRs of those who smoked 21 cigarettes or more per day and those with 801–1,000 cigarette-years were 3.18 [1.18–8.53] and 3.44 [1.40–8.45], respectively, and positive linear associations were observed between that risk and the number of cigarettes per day (p for trend = 0.007) or “cigarette-years” (p for trend = 0.012). The alcohol dose was linearly associated with risk (p for trend = 0.004), where the HR among those who consumed 72.0 g or more of alcohol per day was 3.60 [1.29–9.85]. Among both men and women, cigarette smoking may elevate the risk of death from gallbladder cancer. Drinking may pose an elevated risk among men, but that seems to be less true among women. © 2007 Wiley-Liss, Inc.
Gallbladder cancer is a rare cancer with a poor prognosis, the incidence of which increases with age.1, 2, 3 Geographic variations and a female preponderance in both incidence and mortality4, 5, 6, 7, 8, 9, 10, 11 have been identified. The case survival after diagnosis is relatively high in Japan,12, 13, 14 and the incidence has shown a steady increase in recent years.15, 16 Because of the rarity of this cancer, most previous studies had to be conducted with limited sample sizes, and therefore lacked sufficient statistical power. Although the etiology of gallbladder cancer is poorly understood, several hypotheses have been suggested, and a history of gallstones or gallbladder diseases are considered to be the major risk factor for gallbladder cancer.17, 18, 19 In several case-control studies, cigarette smoking has been reported as a risk factor,19, 20, 21, 22 while inconsistent results have been shown in WHO reports23 and other studies.18, 24, 25 The relationships between cigarette smoking and the risk of gallbladder cancer have not yet been confirmed, and the dose-response association has been only inconsistently observed. Some epidemiological studies have evaluated the association between alcohol consumption and gallbladder cancer to date,18, 19, 21, 25 but their results were also inconsistent.
Using the dataset from a large-scale prospective study in Japan with a follow-up of 1,104,859 person-years, we earlier reported that constipation and a medical history of hepatic disease elevated the risk of gallbladder cancer death.26 In the current study, we further assessed the association of cigarette smoking and alcohol consumption with that risk using the same dataset with a follow-up of 1,471,330 person-years. This is probably the first cohort study evaluating such an association with stratification by gender.
Material and methods
Study-cohort (JACC study)
The Japan Collaborative Cohort Study for Evaluation of Cancer Risk sponsored by the Ministry of Education, Culture, Science, Sports, and Technology of Japan (JACC Study) is a prospective cohort study conducted to evaluate risk factors for a variety of cancers. Details concerning the design and conduct of the JACC Study have been described elsewhere.27, 28 In brief, the study was initiated between 1988 and 1990, during which period 114,271 healthy individuals aged 40–89 years from 45 areas throughout Japan were enrolled as a basic cohort population, and were followed through December 31, 2003. In the majority of study areas, individuals were enrolled by signing the cover page of baseline questionnaire, while in others, not individually, enrollment was done at the group level with the aim of the study and confidentiality of the data being carefully explained to a leader of the community. During the follow-up period, the vital status of participants was determined annually from their residential registration records. These data were collected at the central office of the study group. Cause of death was confirmed by the official death certificate held at the regional health center, for which permission to access was obtained from the Director-General of the Prime Minister's Office, Ministry of Public Management, Home Affairs, Posts and Telecommunications. Cause of death was classified according to the International Classification of Diseases and Related Health Problems, 10th revision: ICD-10.29
This investigation was approved by the Ethics Board of the National Center for Geriatrics and Gerontology and the Ethics Board of the Aichi Medical University School of Medicine.
The person-time of follow-up for each participant was calculated from the day of enrollment to the day of death from gallbladder cancer or any other cause, or until the time the person moved out of the study area, or to the end of 2003, whichever came first. The end point of this study was death from gallbladder cancer (coded as “C23” for ICD10). Participants who died from other causes or moved out of the study area during the follow-up period have been treated as censored subjects. Verification of vital status was believed to be accurate because of the firmly established population registration system in Japan. However, information on the proportion of gallbladder cancer deaths with histological confirmation at autopsy was not available in our study. We excluded subjects who had a history of digestive cancer (stomach, esophageal, liver, pancreas, colon, and rectum) at baseline (159 males and 597 females). To remove a cancer-related effect, we also excluded subjects who died from gallbladder cancer within 2 years from baseline (7 males and 12 females). Participants finally included in the present analysis numbered 113,496 (47,756 males and 65,740 females), ranging in age from 40 to 89 years at entry.
At baseline, the participants were asked to complete a questionnaire including information on demographic characteristics, medical history (including a specific question on gallstone and/or gallbladder diseases), smoking and drinking habits, and lifestyle factors such as bowel movements.
With regard to smoking habits, participants were asked to denote their smoking status at enrollment as nonsmoker, ex-smoker or current smoker. Current smokers were asked to report both their average number of cigarettes per day and total years of smoking. “Cigarette-years” were calculated by multiplying the number of cigarettes per day by years of smoking. Because distribution of numbers of cigarettes per day and cigarette-year were different for men and women, these variables were categorized in different ways by gender. Among females, the number of cigarettes per day was divided into 2 categories: 1–10 cigarettes or 11 cigarettes or more per day, and cigarette-years were divided into 2 categories: 1–200 or 201 or more. Among males, the number of cigarettes per day was divided into 3 categories: 1–10 cigarettes, 11–20 cigarettes or 21 cigarettes or more per day, and cigarette-years were divided into 5 categories: 1–400, 401–600, 601–800, 801–1,000 or 1,001 or more. As for drinking habits, participants were asked to denote their status at enrollment as nondrinker, ex-drinker or current drinker. Current drinkers were further asked to give additional information about their drinking frequency and the total amount consumed on a single occasion. Average daily consumption of ethanol (g) as a quantity of alcohol was calculated by multiplying the average amount consumed with the frequency of alcohol intake. Again, because of the difference in the quantity of alcohol consumed by men and women, the quantity of alcohol was divided into 4 categories by separate numerical values: 0–5.9 g, 6.0–11.9 g, 12.0–23.9 g, 24.0 g or more per day among females, and 0–23.9 g, 24.0–47.9 g, 48.0–71.9 g, or 72 g or more per day among males, respectively.
Statistical analysis was performed using the SAS Software System.30 Hazard ratios (HR) were estimated using Cox proportional hazard models. All HRs were adjusted for potentially confounding factors such as 10-year-interval age categories at enrollment, smoking habits, drinking habits, frequency of stool, tendency toward diarrhea. Nonsmokers or nondrinkers served as the referent, and all variables were entered as dummy variables. Missing values in each variable were treated as an additional category in the analysis. Tests for trend in the Cox proportional hazard models were performed by treating the score as a continuous variable. All tests of significance were two-sided, with a p value less than 0.05 being considered statistically significant.
The demographic characteristics of study subjects at the initiation of follow-up are shown in Table I. A total of 165 deaths (70 men and 95 women) from gallbladder cancer were identified during the follow-up of 1,471,330 person-years. The mean follow-up period was 13.0 years (standard deviation, 3.6). The crude mortality rate in the study population was estimated to be 11.6 per 100,000 males and 10.9 per 100,000 females. The proportions of current smokers and current drinkers among study subjects were 52.6 and 74.3% in men, and 5.6 and 24.2% in women, respectively.
Table I. Baseline Characteristics of Study Subjects
|No. of subjects||65,740||95||47,756||70|
|Age (years)1||58.6 ± 10.9||65.1 ± 10.4||58.4 ± 11.0||66.8 ± 8.1|
|Body mass index (kg/m2)1||23.3 ± 3.2||24.3 ± 4.1||22.8 ± 2.9||22.0 ± 3.2|
|Medical history of gallstone or cholecystitis (%)||5.8||8.1||4.4||7.6|
|Medical history of hepatic disease (%)||5.9||8.5||8.3||14.6|
|Stool frequency [less than once per 4 days, (%)]||4.2||7.9||1.3||0.0|
|Tendency toward diarrhea [yes, (%)]||9.1||5.3||18.9||13.7|
|Smoking habit|| || || || |
| Nonsmoker (%)||92.6||87.8||20.9||12.1|
| Ex-smoker (%)||1.8||2.7||26.6||31.8|
| Current smoker (%)||5.6||9.5||52.6||56.1|
| Number of cigarettes per day, (numbers)1||13.6 ± 8.7||10.7 ± 7.4||20.8 ± 9.7||20.7 ± 9.2|
| Years of smoking (years)1||25.0 ± 13.2||17.6 ± 10.4||35.1 ± 10.7||42.2 ± 8.0|
| “Cigarette-years”1||356.5 ± 316.6||251.2 ± 159.4||714.4 ± 361.4||848.2 ± 284.0|
|Drinking habit|| || || || |
| Nondrinker (%)||74.1||76.3||19.1||14.1|
| Ex-drinker (%)||1.7||1.3||6.6||7.8|
| Current drinker (%)||24.2||22.5||74.3||78.1|
| Quantity of alcohol per day (g)1||13.1 ± 16.1||10.3 ± 15.4||40.5 ± 26.1||46.8 ± 33.0|
Table II shows the adjusted HRs for gallbladder cancer deaths according to smoking status. Among women, the HRs (95 percent confidence interval (95% CIs)) of both ex- and current smokers were 1.47 (0.36, 6.06) and 2.00 (0.91, 4.42), respectively, but a 2.77-fold elevated risk among current smokers (95% CI: 1.17, 6.57) was the significant result when adjusted for age, drinking habits, frequency of stool and tendency toward diarrhea. Among men, the HR (95% CI) of current smokers was 2.27 (1.05, 4.90) compared with nonsmokers when adjusted for age and drinking habits. The duration of smoking was not associated with the risk of gallbladder cancer (p for trend = 0.08). The HR was 1.83 (95% CI: 0.79, 4.21) for the individuals who had smoked for 41 or more years as compared with nonsmokers. Numbers of cigarettes per day and cigarette-years were dose-dependently associated with the risk (p for trend = 0.007 and 0.012, respectively). Compared with nonsmokers, HRs (95% CIs) of those who smoked 21 cigarettes or more per day and those with 801–1,000 cigarette-years were 3.18 (1.18, 8.53) and 3.44 (1.40, 8.45), respectively. When frequency of stool and tendency toward diarrhea were also adjusted for, the significance of an elevated risk among current male smokers disappeared, and the association of the number of cigarettes per day and cigarette-years with the risk of gallbladder cancer death showed only marginal significance (p for trend = 0.082 and 0.075, respectively).
Table II. Adjusted Hazard Ratios (HR) and 95% Confidence Intervals (95% CI) for Death from Gallbladder Cancer by Smoking Habits
| Nonsmoker||699449.7||65||1.00|| ||1.00|| ||121655.6||8||1.00|| ||1.00|| |
| Current smoker||40392.7||7||2.00||(0.91–4.42)||2.77||(1.17–6.57)||302013.1||37||2.27||(1.05–4.90)||1.66||(0.70–3.90)|
|Number of cigarettes|
| None (nonsmoker)||699449.7||65||1.00|| ||1.00|| || || || || || || |
| 1–10 per day||20415.7||3||1.57||(0.49–5.02)||2.37||(0.73–7.68)|| || || || || || |
| 11- per day||18235.7||3||2.19||(0.68–7.04)||2.46||(0.59–10.26)|| || || || || || |
|p for trend|| || ||0.137|| ||0.080|| || || || || || || |
| None (nonsmoker)|| || || || || || ||121655.6||8||1.00|| ||1.00|| |
| 1–10 per day|| || || || || || ||47583.5||4||1.02||(0.31–3.40)||0.31||(0.04–2.54)|
| 11–20 per day|| || || || || || ||167691.2||23||2.44||(1.08–5.50)||1.97||(0.80–4.83)|
| 21- per day|| || || || || || ||82542.1||9||3.18||(1.18–8.53)||2.01||(0.61–6.64)|
|p for trend|| || || || || || || || ||0.007|| ||0.082|| |
| None (nonsmoker)||699449.7||65||1.00|| ||1.00|| || || || || || || |
| 1–200||13396.8||2||2.16||(0.53–8.88)||1.70||(0.23–12.43)|| || || || || || |
| 201-||23048.2||3||1.41||(0.44–4.51)||2.25||(0.69–7.32)|| || || || || || |
|p for trend|| || ||0.392|| ||0.153|| || || || || || || |
|None (nonsmoker)|| || || || || || ||121655.6||8||1.00|| ||1.00|| |
| 1–400|| || || || || || ||48642.3||2||0.98||(0.21–4.67)||0.55||(0.07–4.50)|
| 401–600|| || || || || || ||75023.7||3||0.87||(0.23–3.31)||0.34||(0.04–2.75)|
| 601–800|| || || || || || ||68925.7||10||2.78||(1.08–7.16)||2.48||(0.88–6.99)|
| 801–1,000|| || || || || || ||48517.7||12||3.44||(1.40–8.45)||2.95||(1.09–7.96)|
| 1,001-|| || || || || || ||43017.9||6||1.87||(0.65–5.39)||1.11||(0.29–4.30)|
|p for trend|| || || || || || || || ||0.012|| ||0.075|| |
Table III shows the adjusted HRs for gallbladder cancer deaths according to drinking status. Among women, no significant result was obtained, and point estimates of HRs were near 1.0 except for drinkers consuming small doses. The HR (95% CI) of current male drinkers was 1.76 (0.86, 3.64) compared with nondrinkers, when adjusted for age and smoking habits. As for quantity of alcohol, the HR (95% CI) of those who drank 72.0 g or more per day was 3.60 (1.29, 9.85), and the risk of gallbladder cancer death substantially increased with the quantity of alcohol consumed (p for trend = 0.004). The addition of several other factors to the model (Model 2) did not affect the results.
Table III. Adjusted Hazard Ratios (HR) and 95% Confidence Intervals (95% CI) for Death from Gallbladder Cancer by Drinking Habits
|Drinking habit|| || || || || || || || || || || || |
| Nondrinker||579554.4||61||1.00|| ||1.00|| ||106475.0||9||1.00|| ||1.00|| |
| Current drinker||195556.0||18||1.01||(0.59–1.72)||0.65||(0.32–1.34)||437781.0||50||1.76||(0.86–3.64)||1.51||(0.62–3.67)|
|Quantity of alcohol|| || || || || || || || || || || || |
| None (nondrinker)||579554.4||61||1.00|| ||1.00|| || || || || || || |
| 0–5.9 g/day||39936.2||7||1.97||(0.90–4.32)||1.49||(0.59–3.74)|| || || || || || |
| 6.0–11.9 g/day||22906.9||3||1.57||(0.49–5.03)||1.14||(0.28–4.70)|| || || || || || |
| 12.0–23.9 g/day||19185.4||0||–||(–)||–||(–)|| || || || || || |
| 24.0- g/day||19032.3||2||1.11||(0.27–4.58)||0.62||(0.09–4.55)|| || || || || || |
|p for trend|| || ||0.979|| ||0.136|| || || || || || || |
| None (nondrinker)|| || || || || || ||106475.0||9||1.00|| ||1.00|| |
| 0–23.9 g/day|| || || || || || ||83705.6||4||0.73||(0.22–2.37)||0.90||(0.25–3.20)|
| 24.0–47.9 g/day|| || || || || || ||110328.2||16||2.06||(0.91–4.67)||1.60||(0.58–4.44)|
| 48.0–71.9 g/day|| || || || || || ||95144.1||11||2.14||(0.87–5.26)||2.13||(0.74–6.15)|
| 72.0- g/day|| || || || || || ||44556.3||7||3.60||(1.29–9.85)||3.07||(0.90–10.44)|
|p for trend|| || || || || || || || ||0.004|| ||0.032|| |
We found similar results for cigarette smoking and alcohol consumption in both women and men, when the estimate was further adjusted for a history of gallstones or cholecystitis (data not shown).
To our knowledge, this is the first large-scale cohort study to have evaluated the effects of cigarette smoking and alcohol consumption on gallbladder cancer (excluding biliary tract cancer of other sites) that analyzed men and women separately. In our investigation, both smoking and drinking elevated the risk of gallbladder cancer death dose-dependently among men, with the maximum point estimates of HR being 3.44 (95% CI: 1.40, 8.45) and 3.60 (95% CI: 1.29, 9.85), respectively. Among women, the HR exceeded 1.0, but its effects were not significant except for current smokers.
Only a few studies have examined the relationship between cigarette smoking and gallbladder cancer, and the findings have been inconsistent. Some case-control studies to date have evaluated the effect of smoking on biliary tract cancers,19, 31 but their findings on gallbladder cancer are inconsistent with some other studies reporting that smoking poses an elevated risk of gallbladder cancer20, 24 or others contending that smoking is unrelated to such a risk.18, 21, 25 Very few prospective studies have examined the relationship between cigarette smoking and death from gallbladder cancer. A cohort study by Chow et al.,32 who evaluated the effect of smoking on biliary tract cancers including sites other than the gallbladder, was the only report limiting the subjects to men. They found that smoking elevated the risk of biliary tract cancer death 1.5-fold. Our results are consistent with several previous studies, while those with inconsistent results were all retrospective studies.
The mechanism by which tobacco smoke affects the gallbladder is not known. Smoking is also a risk factor for liver and pancreatic cancers.33, 34 Tobacco smoke compounds are thought to reach the liver or pancreas through the blood flow after being absorbed from the lungs. In the same way, deleterious compounds of tobacco smoke carried through the blood flow may accumulate in bile, exerting carcinogenic effects on the gallbladder's epithelial cells.
In the current study, alcohol consumption was strongly related to the risk of death from gallbladder cancer in men, while it showed no corresponding risk in women. This may be partly due to female drinkers being far fewer and their alcohol consumption being less often compared with that of male drinkers. However, there may be a gender difference in the effects of alcohol on the risk of gallbladder cancer, since female point estimates varied by nearly 1.0, and were not in accordance with male ones.
Previous case-control studies18, 20, 21 have reported that alcohol consumption was not related to the risk of gallbladder cancer, though there was a report that suggested a positive relationship with alcohol consumption and gallbladder cancer.35 Moerman et al.36 reported that the risk of biliary tract cancer was increased by heavy alcohol consumption, and that, among current alcohol drinkers, long-term consumers had a reduced risk. Although the findings of these retrospective studies to date are inconsistent, it may be concluded that alcohol consumption elevates the risk of gallbladder cancer in men, while its effect in women remains unclear.
Alcohol consumption has been associated with upregulated bile acid excretion, and that upregulated bile acid recirculation increased the production of secondary bile acids,37 which may play a role in elevating the risk of gallbladder cancer.
The strength of the current study is its prospective design with mortality follow-up for ∼15 years. This study has a large number of observed person-years, which permits an evaluation of the relationships between factors and risks of various cancers, even relatively rare ones such as gallbladder cancer. It also enables a review of the gender-specific association of smoking and drinking habits with rare cancers.
The present study has several limitations. First, its primary objective was the study of death from gallbladder cancer. The uncertainty of diagnosis of gallbladder cancer may exist because data on histological confirmation were not available in this study. We relied solely on death certificates, which may have led to some misclassifications. However, incidence and mortality may not be notably different, since it is a cancer with a very poor prognosis. Second, we obtained information about smoking habits and drinking habits at baseline, without considering changes in habitual behavior during the whole follow-up period. Kawado et al.38 reported that the percentage of both current smokers and current drinkers between the baseline and interim surveys had declined, a fact that might have caused us to underestimate the relative risk for cancer. However, the risk of cancer would surely be affected by long-term exposure to smoking or drinking in any case, and it is difficult to believe that it might be significantly impacted by interim changes in habits during the follow-up period.
Despite the fact that a history of gallstones is a major recognized risk factor, the HRs of “history of gallstone or cholecystitis” among women and men were 0.72 (95% CI: 0.22, 2.29) and 1.06 (95% CI: 0.25, 4.43), respectively, and a “history of gallstone or cholecystitis” was not associated with gallbladder cancer in this study. Furthermore, we found similar results in both women and men, when HRs were calculated from another multivariate model which added “history of gallstone or cholecystitis” to the variables. Our results are consistent with Wistuba and Gazdar,39 where one pathway may be involved in explaining the lack of association between gallbladder cancer and history of gallstones in Japan.
In summary, our study suggests that smoking may elevate the risk of death from gallbladder cancer dose-dependently among both men and women. Drinking may pose an elevated risk among men, but that seems to be less true among women.
The authors express sincere thanks to Dr. Kunio Aoki, Professor Emeritus, Nagoya University School of Medicine and the former chairman of the JACC Study Group, and Dr. Haruo Sugano, the former Director of the Cancer Institute of the Japanese Foundation for Cancer Research, who contributed to the initiation of the JACC study.
The present members of the JACC Study and their affiliations are as follows: Dr. Akiko Tamakoshi (present chairman of the study group), National Center for Geriatrics and Gerontology; Dr. Mitsuru Mori, Sapporo Medical University School of Medicine; Dr. Yutaka Motohashi, Akita University School of Medicine; Dr. Ichiro Tsuji, Tohoku University Graduate School of Medicine; Dr. Yoshikazu Nakamura, Jichi Medical School; Dr. Hiroyasu Iso, Osaka University Graduate School of Medicine; Dr. Haruo Mikami, Chiba Cancer Center; Dr. Yutaka Inaba, Juntendo University School of Medicine; Dr. Yoshiharu Hoshiyama, University of Human Arts and Sciences Graduate School; Dr. Hiroshi Suzuki, Niigata University Graduate School of Medical and Dental Sciences; Dr. Hiroyuki Shimizu, Gifu University School of Medicine; Dr. Hideaki Toyoshima, Nagoya University Graduate School of Medicine; Dr. Shinkan Tokudome, Nagoya City University Graduate School of Medicine; Dr. Yoshinori Ito, Fujita Health University School of Health Sciences; Dr. Shuji Hashimoto, Fujita Health University School of Medicine; Dr. Shogo Kikuchi, Aichi Medical University School of Medicine; Dr. Kenji Wakai, Nagoya University Graduate School of Medicine; Dr. Akio Koizumi, Graduate School of Medicine and Faculty of Medicine, Kyoto University; Dr. Takashi Kawamura, Kyoto University Center for Student Health; Dr. Yoshiyuki Watanabe and Dr. Tsuneharu Miki, Kyoto Prefectural University of Medicine Graduate School of Medical Science; Dr. Chigusa Date, Faculty of Human Environmental Sciences, Mukogawa Women's University; Dr. Kiyomi Sakata, Wakayama Medical University; Dr. Takayuki Nose, Tottori University Faculty of Medicine; Dr. Norihiko Hayakawa, Research Institute for Radiation Biology and Medicine, Hiroshima University; Dr. Takesumi Yoshimura, Institute of Industrial Ecological Sciences, University of Occupational and Environmental Health, Japan; Dr. Akira Shibata, Kurume University School of Medicine; Dr. Naoyuki Okamoto, Kanagawa Cancer Center; Dr. Hideo Shio, Moriyama Municipal Hospital; Dr. Yoshiyuki Ohno (former chairman of the study group), Asahi Rosai Hospital; Dr. Tomoyuki Kitagawa, Cancer Institute of the Japanese Foundation for Cancer Research; Dr. Toshio Kuroki, Gifu University; and Dr. Kazuo Tajima, Aichi Cancer Center Research Institute.
The past investigators of the study group were listed in reference 20 except for the following 8 members (affiliations are those at the time they participated in the study): Dr. Takashi Shimamoto, Institute of Community Medicine, University of Tsukuba; Dr. Heizo Tanaka, Medical Research Institute, Tokyo Medical and Dental University; Dr. Shigeru Hisamichi, Tohoku University Graduate School of Medicine; Dr. Masahiro Nakao, Kyoto Prefectural University of Medicine; Dr. Takaichiro Suzuki, Research Institute, Osaka Medical Center for Cancer and Cardiovascular Diseases; Dr. Tsutomu Hashimoto, Wakayama Medical University; Dr. Teruo Ishibashi, Asama General Hospital; and Dr. Katsuhiro Fukuda, Kurume University School of Medicine.