Coffee Consumption and Gallbladder Disease Ruhl and Everhart Association of Coffee Consumption with Gallbladder Disease

Transcription

1 American Journal of Epidemiology Copyright 2000 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved Vol. 152, No. 11 Printed in U.S.A. Coffee Consumption and Gallbladder Disease Ruhl and Everhart Association of Coffee Consumption with Gallbladder Disease Constance E. Ruhl 1 and James E. Everhart 2 Coffee consumption was recently shown to protect against symptomatic gallbladder disease in men. The authors examined the relation of ultrasound-documented gallbladder disease with coffee drinking in 13,938 adult participants in the Third National Health and Nutrition Examination Survey, The of total gallbladder disease was unrelated to coffee consumption in either men or women. However, among women a decreased of previously diagnosed gallbladder disease was found with increasing coffee drinking (p = 0.027). These findings do not support a protective effect of coffee consumption on total gallbladder disease, although coffee may decrease the risk of symptomatic gallstones in women. Am J Epidemiol 2000;152: cholelithiasis; coffee; gallbladder diseases; nutrition surveys; risk factors Gallbladder disease is a common condition affecting more than 20 million adults in the United States (1) and a cause of considerable morbidity and health care expense. Despite its common occurrence and impact, risk factors for gallbladder disease are incompletely understood. This is especially true of modifiable lifestyle factors. Coffee consumption was recently found to be protective for symptomatic gallbladder disease in men (2). Men who drank at least 2 cups (473 ml) of regular coffee per day had a risk of symptomatic gallbladder disease that was 60 percent or less that of men who did not drink coffee. Although this was a large, prospective study, it was restricted to the analysis of diagnosed gallstones, which account for less than a third of all men with gallstones (1). We examined the relation of both diagnosed and previously undiagnosed gallbladder disease (total gallbladder disease) with coffee consumption in a large, national, population-based study of men and women. MATERIALS AND METHODS The Third National Health and Nutrition Examination Survey (NHANES III) was conducted in the United States from 1988 through 1994 by the National Center for Health Statistics of the Centers for Disease Control and Prevention. It consisted of interview, examination, and laboratory data collected from a complex multistage, stratified, clustered Received for publication September 21, 1999, and accepted for publication February 22, Abbreviation: NHANES III, Third National Health and Nutrition Examination Survey. 1 Social and Scientific Systems, Inc., Bethesda, MD. 2 National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD. Reprint requests to Dr. Constance E. Ruhl, Social and Scientific Systems, Inc., 7101 Wisconsin Ave., Suite 1300, Bethesda, MD ( CER@s-3.com). probability sample of the civilian, noninstitutionalized population, with oversampling of the elderly, non-hispanic Blacks, and Mexican Americans (3). The sample for this study consisted of 13,938 participants aged years. Excluded were persons who were interviewed but not examined (n 1,318) or were examined at home (n 152), who did not undergo a gallbladder ultrasound (n 351) or whose gallbladder lumen could not be adequately visualized on ultrasound (n 56), who had missing data on coffee consumption (n 24), and pregnant women (n 276). Gallbladder disease was defined as ultrasounddocumented gallstones or evidence of cholecystectomy by standard criteria (1). Based on videotaped recordings of ultrasound examinations, there was excellent agreement on gallbladder disease diagnosis between the ultrasonographer and reviewing radiologist (agreement of 99 percent with a kappa statistic of 0.97). Participants were asked whether or not they had ever been told by a doctor that they had gallstones. Previously diagnosed gallbladder disease (evidence of cholecystectomy or ultrasound-documented gallstones with a doctor diagnosis) and previously undiagnosed ultrasound-documented gallstones were examined individually as outcomes. Participants were asked how often they drank regular coffee with caffeine in the past month. Coffee consumption was coded as none, less than 1 cup per day,, and more than 2 cups per day. No information was available on decaffeinated coffee consumption. Data were collected on known or possible gallbladder disease risk factors, including age (years); sex; race-ethnicity (non-hispanic White, non-hispanic Black, Mexican American, other); body mass index (weight (kg)/height (m) 2 ); waist/hip circumference ratio; serum total cholesterol level (mg/dl); cigarette smoking (never, former, less than one pack per day, one or more packs per day); alcohol consumption (never, former, less than one drink per day, 1 2 drinks per day, more than two drinks per day); and for 1034

2 Coffee Consumption and Gallbladder Disease 1035 women, number of live births. A physical activity intensity variable was created by summing the products of activity frequency in the previous month and an intensity rating for nine common activities (3). For statistical analysis, we examined the relation between gallbladder disease and coffee consumption by first calculating unadjusted gallbladder disease estimates for each coffee consumption category. To further examine the association of gallbladder disease with coffee consumption, while controlling for the effects of covariates, we computed adjusted gallbladder disease estimates for each coffee consumption category using binary linear regression analysis (4). gallbladder disease ratios were then calculated by comparing the for each subsequent coffee consumption category with the in persons who never drank coffee. Binary linear regression was used to calculate ratios, because odd ratios overestimate ratios for a common disease such as gallbladder disease. Prevalence was estimated by least-squares means computed using SUDAAN PROC REGRESS (5). The calculation of 95 percent confidence intervals for adjusted ratios is described in the Appendix. Finally, p values for trend in adjusted gallbladder disease ratios for coffee consumption were computed by including the coffee consumption category in the model as an ordinal variable. Because of sex differences in gallbladder disease and risk factors (1), separate analyses were conducted for women and men. All covariates that had previously been found to be associated with gallbladder disease (p < 0.1) in the NHANES III sample were included in the final models. Multivariate analyses excluded persons with missing values for any risk factor included in the model. To examine the association of previously undiagnosed gallbladder disease with coffee consumption, persons with previously diagnosed gallbladder disease were excluded from the analysis. A 95 percent confidence interval that did not include one (p < 0.05) was considered to indicate statistical significance. All analyses incorporated sample weights, stratification, and clustering using SUDAAN software (5). RESULTS Total gallbladder disease was found in 1,415 (19.5 percent) women and 578 (8.7 percent) men. Among women, 770 (10.6 percent) had previously diagnosed gallbladder disease and 645 ( percent) had previously undiagnosed gallbladder disease. Among men, 219 (3.3 percent) had previously diagnosed and 359 (5.4 percent) had previously undiagnosed gallbladder disease. There was no relation among women in univariate analysis between total gallbladder disease and coffee consumption (table 1). In multivariate analysis controlling for covariates (table 1), total gallbladder disease remained unrelated to coffee consumption. In men in univariate analysis (table 2), coffee consumption was not related to total gallbladder disease. In multivariate TABLE 1. Prevalence of gallbladder disease according to coffee consumption category in women, Third National Health and Nutrition Examination Survey, * Gallbladder disease outcome by coffee consumption category Total gallbladder disease Previously diagnosed gallbladder disease No. 2,917 1,024 2,115 1,207 2,917 1,024 2,115 1,207 Unadjusted (%) (%) ratio % confidence interval 0.92, , , , , , 1.02 p value for trend Previously undiagnosed gallbladder disease# 2, ,865 1, , , , 1.33 * Total sample (n = 7,263); with gallbladder disease (n = 1,415); with previously diagnosed gallbladder disease (n = 770); with previously undiagnosed gallbladder disease (n = 645). From binary linear regression analysis adjusted for age, race-ethnicity, body mass index, waist/hip ratio, cigarette smoking, serum total cholesterol level, alcohol consumption, physical activity, and number of live births. Ultrasound-documented gallstones or evidence of cholecystectomy. One cup = 237 ml. Ultrasound evidence of cholecystectomy or ultrasound-documented gallstones with a doctor diagnosis. # Previously undiagnosed ultrasound-documented gallstones. 0.92

3 1036 Ruhl and Everhart TABLE 2. Prevalence of gallbladder disease according to coffee consumption category in men, Third National Health and Nutrition Examination Survey, * No. Unadjusted (%) (%) ratio 95% confidence interval Gallbladder disease outcome by coffee consumption category Total gallbladder disease 2,218 1,115 2,039 1, , , , 1.33 p value for trend 0.70 analysis controlling for covariates (table 2), total gallbladder disease remained unrelated to coffee consumption. To further evaluate the relation of gallbladder disease with coffee consumption, we examined previously diagnosed and previously undiagnosed gallbladder disease independently as outcomes. In women, no relation was found with either outcome in univariate analysis (table 1). In multivariate analysis (table 1), a test for trend suggested lower risk of previously diagnosed gallbladder disease with increasing coffee consumption (p 0.027). The of previously undiagnosed gallbladder disease was unrelated to coffee consumption. In men in univariate analysis, no relation was found with either gallbladder disease outcome (table 2). In multivariate analysis (table 2), although the ratio for previously diagnosed gallbladder disease was less than one for each of the three coffee consumption categories, statistical significance was reached only when men who drank less than 1 cup per day were compared with those who never drank coffee. The of previously undiagnosed gallbladder disease was unrelated to coffee consumption. DISCUSSION Previously diagnosed gallbladder disease Previously undiagnosed gallbladder disease# 2,218 1,115 2,039 1,303 2,158 1,091 1,950 1,257 We found no relation between total gallbladder disease and coffee consumption in either women or men. Previous studies have examined this potential association in either total gallbladder disease diagnosed by ultrasonography or in the , , , , , , 1.53 * Total sample (n = 6,675); with gallbladder disease (n = 578); with previously diagnosed gallbladder disease (n = 219); with previously undiagnosed gallbladder disease (n = 359). From binary linear regression analysis adjusted for age, race-ethnicity, body mass index, waist/hip ratio, serum total cholesterol level, and alcohol consumption. Ultrasound-documented gallstones or evidence of cholecystectomy. One cup = 237 ml. Ultrasound evidence of cholecystectomy or ultrasound-documented gallstones with a doctor diagnosis. # Previously undiagnosed ultrasound-documented gallstones subset of symptomatic gallstones. Ultrasound-documented gallbladder disease was not associated with coffee consumption in cross-sectional studies of a Danish population, Irish prenatal patients, and German blood donors (6 8), though there was a trend toward a positive relation in the latter study. In contrast, coffee drinking was inversely related to ultrasound-documented gallstones in multivariate analysis only (p 0.04) in a population-based, prospective Italian study in which women and men were combined (9). Symptomatic gallbladder disease among men was recently reported to occur less commonly with increasing coffee drinking in a large prospective cohort study of male health professionals (2). On the other hand, coffee consumption was not associated with clinical gallbladder disease in two smaller hospital case-control studies (10, 11) and one large prospective study of male college alumni (12). Although we found no relation of total gallbladder disease and coffee, we did find a protective effect of coffee consumption when we restricted our analysis to previously diagnosed gallbladder disease, a category similar to symptomatic gallbladder disease. One interpretation of an inverse association with symptomatic gallbladder disease could be that coffee prevents symptoms from gallstones. Biliary colic, the most typical gallstone symptom, is believed to result from temporary obstruction of the gallbladder neck or cystic duct (13). Theoretically, the forcefulness of gallbladder contractions and the relative sizes of stones and the cystic duct should determine the occurrence of colic. Coffee (caffeine) may increase

4 Coffee Consumption and Gallbladder Disease 1037 gallbladder contractility (14), but it is uncertain how this or any other potential factor may prevent symptoms. It is also possible that nonphysiologic reasons could have caused this inverse association. For example, there may be unknown factors associated with coffee consumption and greater reluctance to seek medical attention for abdominal pain. Women might also have decreased coffee consumption following cholecystectomy or even for subclinical symptoms. Nevertheless, the various possibilities as to why symptomatic gallbladder disease would be associated with less coffee consumption are unrelated to our main finding that total gallbladder disease was not associated with coffee drinking. ACKNOWLEDGMENTS This work was supported by a contract from the National Institute of Diabetes and Digestive and Kidney Diseases (NO1-DK ). The authors thank Dr. Keith Rust for statistical advice, Danita Byrd-Holt for computer programming assistance, and the following persons for their assistance in planning and executing the NHANES III ultrasound examinations: Dr. Kurt Maurer, National Center for Health Statistics; Dr. Michael Hill, principal study radiologist; and Cynthia Runco and Diane Palmer, ultrasonographers. REFERENCES 1. Everhart J, Khare M, Hill M, et al. Prevalence and ethnic differences in gallbladder disease in the United States. APPENDIX* Gastroenterology 1999;117: Leitzmann MF, Willett WC, Rimm EB, et al. A prospective study of coffee consumption and the risk of symptomatic gallstone disease in men. JAMA 1999;281: Plan and operation of the Third National Health and Nutrition Examination Survey, Series 1: programs and collection procedures. Vital Health Stat ;(32): Feldstein MS. A binary multiple regression method of analyzing factors affecting perinatal mortality and other outcomes of pregnancy. Part 1. J R Stat Soc (A) 1966;129: Research Triangle Institute. SUDAAN user s manual. Release 7.0. Research Triangle Park, NC: Research Triangle Institute, April Jørgensen T. Gallstones in a Danish population. Relation to weight, physical activity, smoking, coffee consumption, and diabetes mellitus. Gut 1989;30: Basso L, McCollum PT, Darling MRN, et al. A descriptive study of pregnant women with gallstones. Relation to dietary and social habits, education, physical activity, height, and weight. Eur J Epidemiol 1992;8: Kratzer W, Kächele V, Mason RA, et al. Gallstone in relation to smoking, alcohol, coffee consumption, and nutrition. Scand J Gastroenterol 1997;32: Misciagna G, Leoci C, Guerra V, et al. Epidemiology of cholelithiasis in southern Italy. Part II: risk factors. Eur J Gastroenterol Hepatol 1996;8: Pastides H, Tzonou A, Trichopoulos D, et al. A case-control study of the relationship between smoking, diet, and gallbladder disease. Arch Intern Med 1990;150: La Vecchia C, Nègri E, D Avanzo B, et al. Risk factors for gallstone disease requiring surgery. Int J Epidemiol 1991;20: Sahi T, Paffenbarger RS Jr, Hsieh C, et al. Body mass index, cigarette smoking, and other characteristics as predictors of self-reported, physician-diagnosed gallbladder disease in male college alumni. Am J Epidemiol 1998;147: Sullivan FJ, Eaton SB Jr, Ferrucci JT Jr. Cholangiographic manifestations of acute biliary colic. N Engl J Med 1973;288: Douglas BR, Jansen JB, Tham RT, et al. Coffee stimulation of cholecystokinin release and gallbladder contraction in humans. Am J Clin Nutr 1990;52: To calculate 95 percent confidence intervals for adjusted gallbladder disease ratios, the following method was used. The log of the adjusted ratio, log(x i /x 0 ), was calculated, where x i represents gallbladder disease at each higher level of a factor and x 0 at the lowest level of the factor. The variance of the log of the adjusted ratio was computed using the following formula: The variance and covariance of the adjusted (least-squares means) used in this equation were calculated from the variance-covariance matrix of the estimated regression coefficients and the vector of least-squares means contrast coefficients (mean values of the covariates included in the multivariate model) generated by binary linear regression analysis as follows: where var3log1x i >x >x i 2 2 var1x i >1x i 21x 0 24cov1x i, x >x var1x 0 26 V1L2 C V 1@2C V1L2 variance-covariance matrix of adjusted C transposed vector of least-squares means contrast coefficients V1@2 variance-covariance matrix of estimated regression coefficients

5 1038 Ruhl and Everhart C vector of least-squares means contrast coefficients SAS IML software, version 6, first edition (SAS Institute, Inc., Cary, North Carolina), was used to perform matrix algebra. Since the standard error (SE) of the log of the adjusted ratio, SE[log(x i /x 0 )], is the square root of the variance of the log of the adjusted ratio, var[log(x i /x 0 )], calculated above, the 95 percent confidence interval of the log of the adjusted ratio was computed as follows: log1x i >x 0 2 > 1.96 SE3log1x i >x 0 24 The 95 percent confidence interval of the adjusted ratio was then obtained by taking the exponential of the upper and lower confidence limits of the log of the adjusted ratio: exp5log1x i >x 0 2 > 1.96 SE3log1x i >x * The Appendix is reproduced with permission from Ruhl CE, Everhart JE. Association of diabetes, serum insulin, and C-peptide with gallbladder disease. Hepatology 2000;31: Copyright 2000, W. B. Saunders Company.