From Liver International
The Effect of Caffeine and Alcohol Consumption on Liver Fibrosis
A Study of 1045 Asian Hepatitis B Patients Using Transient Elastography
Arlinking Ong; Vincent Wai-SunWong; Grace Lai-Hung Wong; Henry Lik-Yuen Chan
Posted: 01/25/2012; Liver International. 2011;31(7):1047-1053. © 2011 Blackwell Publishing
Abstract and IntroductionBackground:
Role of caffeine consumption in chronic hepatitis B virus (HBV)-infected patients and the interaction with alcohol consumption is unclear.
This study aimed to investigate the relationship between caffeine and alcohol consumption and liver stiffness in chronic HBV-infected patients.
Chronic HBV-infected patients who underwent transient elastography examination in 2006–2008 were studied. Advanced fibrosis was defined as liver stiffness >9 kPa for patients with normal alanine aminotransferase (ALT) or >12 kPa for those with elevated ALT according to previous validation study. Caffeine and alcohol consumption was recorded using a standardized questionnaire. Excessive alcohol intake was defined as 30 g/day in men and 20 g/day in women.
The liver stiffness of 1045 patients who completed the questionnaire was 8.3 ± 6.2 kPa. Two hundred and sixteen (20.7%) patients had advanced fibrosis. Ninety-five (19.0%) patients who drank ≥1 cup of coffee had advanced fibrosis, compared with 121 (22.2%) patients who drank <1 cup (P=0.21). The amount of caffeine intake had positive correlation with the amount of alcohol intake (rs=0.167, P<0.001). Although 231 (22.1%) patients reported alcohol consumption, only 11 (1%) had excessive alcohol intake. The prevalence of advanced fibrosis among patients with mild to moderate alcohol intake (26, 18.8%) was comparable to that among non-drinkers (190, 21.0%) (P=0.57).
Conclusion: Caffeine intake does not affect liver stiffness in chronic HBV-infected patients. Patients who drink coffee regularly tend to drink alcohol. Most chronic HBV-infected patients do not have excessive alcohol consumption. The prevalence of advanced fibrosis among mild to moderate alcohol drinkers was low in this population.
Chronic hepatitis B virus (HBV) infection is a global health problem affecting over 350 million people. Persistent hepatic inflammation because of chronic HBV infection will result in progressive liver fibrosis, which will eventually result in cirrhosis, hepatic decompensation and hepatocellular carcinoma (HCC).
Alcoholism is a primary chronic disease with genetic, psychosocial and environmental factors influencing its development and manifestations. Heavy alcohol consumption commonly causes alcoholic liver disease, cirrhosis and even HCC. However, it is unclear whether consumption of a smaller amount of alcohol is safe in chronic hepatitis B patients.
On the other hand, consumption of coffee or caffeine may reduce liver injury. Greater coffee, and especially caffeine, intake was associated with a lower prevalence of abnormal alanine aminotransferase (ALT) activity in USA and Japan. Plasma glutathione was increased by 16% on coffee consumption in an Italian population. In a prospective study among advanced Hepatitis C-related liver disease, regular coffee consumption was associated with lower rates of disease progression. In human hepatoma cell lines, coffee and its major components (caffeine, cafestol and kahweol) alter expression and activity of enzymes involved in xenobiotic mechanisms. Mice pretreated with cafestol and kahweol were protected from carbon tetrachloride toxicity by inhibiting cytochrome CYP 2E1, an enzyme responsible for carbon tetrachloride bioactivation while caffeine specifically inhibited expression of connective tissue growth factor by interfering with transforming growth factor-β signaling through the SMAD pathway and upregulate peroxisome proliferator activated receptor γ levels. These in vitro and in vivo data suggested that caffeine has antifibrotic effects.
A recent case control study among chronic hepatitis B carriers showed that coffee consumption reduced the risk of HCC by half with a significant dose-response effect. It is uncertain if the beneficial effect is mediated through prevention of liver fibrosis and cirrhosis. In the past, large-scaled studies on risk factors of liver fibrosis and cirrhosis were difficult to conduct because the assessment required liver biopsy, which was invasive and not acceptable by all patients. In this study, we aimed to determine the effect of alcohol and caffeine consumption on the prevalence of advanced liver fibrosis among chronic HBV-infected patients using transient elastography.
Study Population We prospectively recruited chronic HBV-infected patients regardless of the disease activity for transient elastography. We received referrals from all primary care and hospital clinics in Hong Kong from July 2006 to February 2008. Chronic HBV infection was diagnosed by positive serology tests for serum HBsAg for at least 6 months. We excluded patients with evidence of other chronic liver disease by screening with antibody against hepatitis C virus, antinuclear antibody, antismooth muscle antibody, antimitochondrial antibody, serum ceruloplasmin, transferrin saturation and ferritin.
From April to May 2008, patients from the original cohort were phone-interviewed with a verbal consent for this study. Questions concerning their alcohol and caffeine consumption (days per week of consumption and amount per day) in the previous year were asked from a standardized questionnaire. Patients who refused to give verbal consent for the phone-interview were not included in this study.
All patients received comprehensive clinical and laboratory assessment at the time of transient elastography. Serum HBV DNA levels were measured by the TaqMan real-time polymerase chain reaction assay with a range of detection from 100 to 109 copies/ml. Anthropometric parameters including body weight, body height, hip circumference and waist circumference were measured. Body mass index (BMI) was calculated as weight (kg) divided by height (m) squared. Overweight was defined as BMI≥23 kg/m2, and obesity as BMI≥25 kg/m2 according to the Asian and Chinese criteria. We defined moderate and severe obesity as BMI≥28 kg/m2 M and ≥30 kg/m2 respectively. Excessive alcohol intake was defined as >30 g/day in men and >20 g/day in women.
Liver Stiffness Measurement by Transient Elastography
Transient elastography (FibroScan®, Echosens, Paris, France) is one of the new noninvasive modality to evaluate liver fibrosis by measuring liver stiffness. It uses an ultrasound-based technique to measure the speed of propagation of the shear wave through the liver. It can assess approximately 1/500 of the liver's total mass thus reducing the sampling error. Transient elastography has been shown to accurately predict histological advanced fibrosis in different liver diseases.[16–21]
Liver stiffness measurement (LSM) was performed using transient elastography according to the instructions of the manufacturer. Details of the technical background and examination procedure was described previously. Officially trained operators who had performed at least 50 measurements prior this study were responsible to perform the LSM for the patients who had kept fast for at least 8 h. The LSM was considered reliable only if 10 successful acquisitions were obtained, an interquartile range (IQR)/LSM of ≤30% and the success rate was ≥60%. Liver stiffness was expressed in kPa. We defined advanced fibrosis as liver stiffness >9 kPa for those with normal ALT or >12 kPa for those with elevated ALT (F3 fibrosis on histology).[23,24]
Hui's index was a non-invasive model comprising of BMI, platelet count, serum albumin and total bilirubin [predictive probabilities=exp(3.148+0.167 × BMI+0.088 × bilirubin [μM]−0.151 × albumin [g/L]−0.019 × platelet [109/L])/(1+exp(3.148+0.167 × BMI+0.088 × bilirubin [μM]−0.151 × albumin [g/L]−0.019 × platelet [109/L]))] to predict significant fibrosis (F2 fibrosis on histology). Among 235 chronic hepatitis B patients with liver biopsy, at a cutoff value of 0.15, the sensitivity, specificity, positive and negative predictive values for significant fibrosis were 93, 49, 41 and 95% respectively.
Phone-interview and Questionnaire
A structured questionnaire interview was conducted over the phone to collect data on alcohol and caffeine intake. Participants were specifically asked about the average quantity and frequency of beer, wine, or liquor consumption to estimate the amount of alcohol consumption. The average quantity and frequency of consumption of coffee (regular, instant, ground and decaffeinated), tea (any kind), cola-type soda (regular or decaffeinated) and chocolate (including chocolate milk) was also enquired to estimate the amount of caffeine consumption. Effect of alcohol and caffeine consumption was analyzed with respect to the severity of liver fibrosis as assessed by transient elastography, Hui's Index and other clinical and biochemical parameters
Total caffeine consumption (mg/day) was estimated by summing caffeine from regular coffee (137 mg per cup), regular tea (47 mg per cup), regular and diet cola-type soda (46 mg per bottle or can) and chocolate (7 mg per serving). As for the coffee cup equivalent, the total caffeine consumption per day was divided by 137 (Table 1).
Table 1. Caffeine content in beverages or food
|Beverage or food||Caffeine content (mg per standard unit of consumption)||Units consumed per day (mean ± SD)||Proportion of caffeine consumed in the cohort (%)|
|Coffee (8 oz)||137||0.30 ± 0.62||20.7|
|Decaffeinated coffee (8 oz)||3||0.01 ± 0.12||0|
|Black tea (8 oz)||47||3.21 ± 5.35||75|
|Cola (12 oz)||46||0.12 ± 0.24||2.8|
|Caffeine free cola (12 oz)||0||0.00 ± 0.06||0|
|Chocolate bar (1 oz)||7||0.38 ± 0.71||1.3|
|Cocoa (8 oz)||6||0.05 ± 0.16||0.2|
Continuous variables were expressed in mean ± standard deviation or median (IQR) as appropriate. Qualitative and quantitative differences between subgroups were analyzed using χ2 or Fisher's exact test for categorical parameters as appropriate, and Student's t-test or Mann–Whitney test for continuous parameters as appropriate. Logistic regression analyses were performed to determine if caffeine and alcohol consumption was associated with liver fibrosis. The multivariate adjusted logistic regression model included covariates that were included in univariate analysis. Spearman's rank correlation was used to correlate caffeine and alcohol intake. All statistical tests were two-sided. Statistical significance was taken as P<0.05. Statistical analysis was performed by Statistical Package for Social Science (SPSS version 17.0, Chicago, IL, USA).
Patient CharacteristicsOverall, 22.7% of 1532 patients in this study came from primary care clinics and 77.3% were from hospital clinics. Six patients were excluded because of anti-HCV positivity, 60 patients because of unreliable LSM results and 421 as we failed to contact for phone-interview because of various reasons (e.g. changed telephone numbers or no answer despite multiple attempts). For patients we able to contact via phone, none of them refused to give verbal consent for the phone interview. A total of 1045 chronic hepatitis B patients were thus included in this study. Majority of patients were males (652, 62.4%), of which 175 (26.8%) were overweight and 210 (32.2%) were obese. Two hundred forty seven (37.9%) male patients had elevated ALT with 71% having HBV DNA of more than 4 logs copies/ml. Four hundred and ninety-nine (47.8%) patients consumed one cup of coffee or more per day, while 231 (22.1%) patients drank alcohol. Only 26 (2.5%) patients reported no intake of both coffee and alcohol. The median (IQR) daily consumption of caffeine from food and beverages was 124 (51.6–275.6) mg/day or 1 (0.4–2.0) cup of coffee/day while alcohol consumption of >20 g/day was 1.9% (Table 2). Two hundred and sixteen (20.7%) patients had LSM suggestive of advanced fibrosis.
Table 2. Baseline characteristics of patients (n=1045)
|Age (years, mean ± SD)||46.5 ± 12.2|
|BMI (kg/m2, mean ± SD)||23.1 ± 3.3|
|Waist circumference (cm, mean ± SD)||84.5 ± 10|
|Platelet [ × 109/L, median (IQR)]||206 (167–250)|
|ALT [IU/L, median (IQR)]||41 (26–68)|
|HBeAg positive (%)||224 (21.4%)|
|HBV DNA (log copies/ml, mean ± SD)||7.82 ± 8.51|
|LSM [kPa, median (IQR)]||6.4 (4.9–9.1)|
|Caffeine per day|
|Overall [mg, median (IQR)]||124.0 (51.6–275.6)|
|Coffee-cup equivalent [median (IQR)]||1 (0.4–2.0)|
|Black-tea-cup equivalent [median (IQR)]||3 (1.2–5.8)|
|Alcohol consumption per day [g, median (IQR)]||0 (0–0)|
|0 g (%)||810 (77.5%)|
|>0–20 g (%)||215 (20.6%)|
|>20 g(%)||20 (1.9%)|
ALT, alanine aminotransferase; BMI, body mass index; IQR, interquartile range; kPa, kilopascal; LSM, liver stiffness measurement; SD, standard deviation.
Caffeine Intake and Liver Fibrosis
Table 3 shows the median total caffeine consumption in mg/day and the coffee cup equivalent according to different patient characteristics where there was no significant difference in caffeine consumption among those with no fibrosis and advanced fibrosis.
Table 3. Daily caffeine consumption according to patient characteristics (n=1045)
|Characteristics||N||Median caffeine intake (mg/day)||Coffee-cup equivalent (caffeine in mg divided by 137)||Black-tea-cup equivalent (caffeine in mg divided by 47)||P-value*|
|Male||652||131.5 (53.7–281.7)||1.0 (0.4–2.1)||2.9 (1.2–6.1)||0.129|
|Female||393||111.9 (48.0–259.4)||0.8 (0.4–1.9)||2.3 (1.2–5.5)|
|≤45||483||113.9 (47.0–263.1)||0.8 (0.3–1.9)||2.3 (0.9–5.5)||0.163|
|>45||562||128.9 (56.9–282.0)||0.9 (0.4–2.1)||2.6 (1.2–6.1)|
|Body mass index|
|Normal||515||123.0 (48.9–247.7)||0.9 (0.4–1.8)||2.6 (1.2–5.2)||0.315|
|Overweight||253||141.9 (53.6–284.4)||1.0 (0.4–2.1)||2.9 (1.2–6.1)|
|Obese||277||116.7 (53.1–282.0)||0.9 (0.4–2.1)||2.6 (1.2–6.1)|
|≤58 IU/L||715||113.7 (47.4–262.1)||0.8 (0.3–1.9)||2.3 (0.9–5.5)||0.074|
|>58 IU/L||330||144.1 (59.7–282.0)||1.1 (0.4–2.1)||3.2 (1.2–6.1)|
|Positive||224||137.3 (48.9–280.8)||1.0 (0.4–2.0)||2.9 (1.2–5.8)||0.709|
|Negative||821||121.2 (53.1–274.7)||0.9 (0.4–2.0)||2.6 (1.2–5.8)|
|No||814||113.0 (46.7–243.8)||0.8 (0.3–1.8)||2.3 (0.9–5.2)||<0.001|
|Yes||231||180.1 (83.0–313.1)||1.3 (0.6–2.3)||3.8 (1.7–6.7)|
|Liver stiffness measurement|
|No fibrosis||829||127.6 (48.1–282.0)||0.9 (0.4–2.1)||2.6 (1.2–6.1)||0.355|
|Advanced fibrosis and cirrhosis||216||111.3 (54.1–239.0)||0.8 (0.4–1.7)||2.3 (1.2–4.9)|
All caffeine intakes were presented as median (interquartile range).
*P-values refer to the comparison of caffeine intake by Kruskal–Wallis's test between groups under the same characteristic (e.g. male vs. female).
There was no significant difference in LSM (Table 4). In addition, caffeine intake had no significant correlation with liver stiffness (rs=0.011, P=0.722) or the Hui's index (rs=−0.002, P=0.948).
Table 4. Characteristics of patients by category of caffeine consumption
|Coffee-cup equivalent of caffeine intake||P-value*|
|Number in cohort||27||519||237||262|
|Age (years, mean ± SD)||49 ± 13||46 ± 12||46 ± 13||47 ± 12||0.55|
|≤45||11 (40.7%)||246 (47.4%)||110 (46.4%)||116 (44.3%)||0.80|
|>45||16 (59.3%)||273 (52.6%)||127 (53.6%)||146 (55.7%)|
|Female||13 (3.3%)||204 (51.9%)||85 (21.6%)||91 (23.2%)||0.37|
|Male||14 (2.2%)||315 (48.3%)||152 (23.3%)||171 (26.2%)|
|Body mass index (kg/m2, mean ± SD)||23.9 ± 3.3||23.1 ± 3.2||22.3 ± 8.7||24.6 ± 21.4||0.16|
|Body mass index|
|Normal||12 (44.4%)||262 (50.5%)||124 (52.3%)||117 (44.7%)||0.36|
|Overweight||5 (18.5%)||117 (22.5%)||60 (25.3%)||71 (27.1%)|
|Obese||10 (37.1%)||140 (27.0%)||53 (22.4%)||74 (28.2%)|
|Waist circumference (cm, mean ± SD)||86.9 ± 11.1||84.5 ± 10.1||82.8 ± 15.2||84.9 ± 9.8||0.09|
|Alanine aminotransferase (IU/L)||38 (21–64)||41 (26–64)||42 (25–72.5)||41 (27–72)||0.37|
|≤58 IU/L||19 (2.7%)||371 (51.9%)||154 (21.5%)||171 (23.9%)||0.19|
|>58 IU/L||8 (2.4%)||148 (44.8%)||83 (25.2%)||91 (27.6%)|
|Albumin (g/L, mean ± SD)||44 ± 4||44 ± 4||44 ± 3||44 ± 3||0.68|
|Total bilirubin [μmol/L, median (IQR)]||13 (10–19)||13 (9–17)||13 (9–18)||13 (9–18)||0.58|
|Positive||4 (1.8%)||108 (48.2%)||55 (24.6%)||57 (25.4%)||0.73|
|Negative||23 (2.8%)||411 (50.1%)||182 (22.2%)||205 (24.9%)|
|HBV DNA (log copies/ml, mean ± SD)||7.28 ± 7.98||7.91 ± 8.57||7.81 ± 8.51||7.62 ± 8.23||0.34|
|Liver stiffness measurement (kPa, mean ± SD)||8.0 ± 6.5||8.4 ± 6.6||8.4 ± 6.1||8.2 ± 5.5||0.95|
|Liver stiffness measurement|
|No fibrosis||24 (88.9%)||401 (77.3%)||189 (79.7%)||215 (82.1%)||0.25|
|Advanced fibrosis and cirrhosis||3 (11.1%)||118 (22.7%)||48 (20.3%)||47 (17.9%)|
|Unlikely to have fibrosis||20 (74.1%)||351 (67.6%)||164 (69.2%)||179 (68.3%)||0.89|
|With fibrosis||7 (25.9%)||168 (32.4%)||73 (30.8%)||83 (31.7%)|
|No||26 (96.3%)||422 (81.3%)||181 (76.4%)||185 (70.6%)||0.001|
|Yes||1 (3.7%)||97 (18.7%)||56 (23.6%)||77 (29.4%)|
|Daily alcohol intake [g, median (IQR)]||0 (0–0)||0 (0–0)||0 (0–0)||0 (0–1.1)||0.001|
HBV, hepatitis B virus; IQR, interquartile range; SD, standard deviation.
Based on LSM and the Hui's index, caffeine intake by univariate and multivariate analysis was not associated with the risk of advanced liver fibrosis (Table 5).
Table 5. Relationship of caffeine intake, advanced fibrosis and alcohol consumption
Alcohol Intake and Liver Fibrosis
Majority of patients (77.9%) did not consume alcohol while only 11 out of 231 patients with alcohol consumption had excessive alcohol drinking. The highest alcohol consumption per day was 88 g. Advanced fibrosis was only prevalent in 26 (18.8%) of those with mild to moderate alcohol consumption compared with 190 (21.0%) of those non-alcoholic drinkers (P=0.57).
Those who drink alcohol have significantly higher caffeine consumption than those who do not drink alcohol (Table 3). Caffeine intake had positive correlation with alcohol intake (rs=0.167, P<0.001).
DiscussionIn this large territory-wide observational study with prospective recruitment, we assessed the relationship between caffeine and alcohol consumption and the risk of advanced liver fibrosis in chronic hepatitis B patients with different disease severity. Caffeine intake was not associated with advanced fibrosis. Patients who drank coffee regularly were more likely to consume alcohol. On the other hand, few chronic hepatitis B patients in Hong Kong had excessive alcohol consumption. Mild to moderate alcohol consumption did not increase the risk of advanced fibrosis.
Coffee is a rich source of a number of phenol compounds with antioxidant effects in vitro, with main polyphenols are phenolic acids such as chlorogenic and caffeic acid. Caffeine and its metabolites, 1-methylxanthine and 1-methyluric acid, have also been shown to have antioxidant properties. Caffeine has also been reported to inhibit chemical carcinogenesis and ultraviolet B light induced carcinogenesis in mice.[29,30] Coffee is also found to decrease the progression of liver disease among those with advanced hepatic fibrosis and even reduce the risk of HCC[31,32] One study concluded that approximately 2 coffee cup equivalents/day was associated with less severe hepatic fibrosis, but the beneficial effect was only shown in patients with chronic hepatitis C rather than patients with other liver diseases. A case control study in Italy showed that the protective effect of coffee on HCC was mainly in people who are not chronically infected with HBV. Overall, previous studies included mostly chronic hepatitis C patients and HBV patients were underrepresented. Based on our results, it appears that liver fibrosis in chronic hepatitis B patients is determined mainly by virological and genetic factors and less affected by caffeine intake.
In previous reports, men who have daily alcohol intake of 30–39 g and women who consumed 20–29 g of alcohol daily would have an increased risk of all cause mortality among the general population. Seventy-eight percent of our cohort did not drink alcohol at all while only 1% of patient had history of excessive alcohol consumption. Nonetheless, we could not find any deleterious effect of mild to moderate alcohol consumption to liver fibrosis. As alcohol drinkers were mostly coffee drinkers as well, the concomitant caffeine intake might but one of the confounders leading to the absence of increased risk of advanced fibrosis in alcohol drinkers. Nonetheless only a minority (1% of the study population has excessive alcohol intake, we believe that the major reason of the absence of increased risk of advanced fibrosis in alcohol drinkers was the modest instead of excess amount of alcohol use. Because most of our patients had no advanced liver disease as compared with other studies (33, 35–38), our results could not be extrapolated to cirrhotic patients. Less than 20 g of alcohol intake tends not to increase risk of advanced fibrosis in our present study. Controlled prospective studies may be done in the future to verify this.
One of the limitations of our study was the lack of liver biopsy. As liver biopsy is an invasive procedure, studies using histological fibrosis as an endpoint might suffer from the problem of small sample size and selection bias towards patients with active or advanced liver disease. On the other hand, a non-invasive test such as transient elastography could be applied to a large number of patients with different disease severity. In fact, transient elastography has been shown to be highly accurate in detecting histological advanced fibrosis and cirrhosis in chronic hepatitis B patients,[16,23] though the accuracy might be limited in settings of grossly elevated transaminase levels, but not in the presence of steatosis. Secondly, recall bias might occur during the questionnaire survey and affect the accuracy of the measurement of the caffeine intake measured. We have tried to minimize this bias by using a quantity-frequency questionnaire, which has been previously adapted in a study demonstrating the protective effect of coffee consumption from HCC. There were 159 (15.2%) participants drinking >2–3 coffee-cup equivalents/day and 103 (9.9%) drinking >3 coffee-cup equivalents/day. Because the above numbers for these groups were low and previous finding showing 2 coffee-cup equivalents/day was associated to less severe liver fibrosis, we combined these groups for analysis. Because this is a cross sectional study, possible unmeasured as well as poorly measured confounding factors that may affect the interpretation of our data such as changes in the drinking habit of the patients, variability of caffeine intake over time, socioeconomic status, educational level were not considered in our analysis.
In conclusion, cross-sectional caffeine intake does not affect liver stiffness in chronic HBV-infected patients. The protective effect of caffeine on HCC demonstrated in previous studies is probably via the pathway other than reducing liver fibrosis. The prevalence of advanced liver fibrosis is low (20%) in chronic hepatitis B patients with daily alcohol consumption below 20 g.