Coffee and Tea Consumption Are Associated With a Lower Incidence of Chronic Liver Disease in the United States
2005; Elsevier BV; Volume: 129; Issue: 6 Linguagem: Inglês
10.1053/j.gastro.2005.08.056
ISSN1528-0012
AutoresConstance E. Ruhl, James E. Everhart,
Tópico(s)Nutrition, Health and Food Behavior
ResumoBackground & Aims: Coffee drinking has been suggested to protect against liver injury, but it is uncertain whether this is of clinical significance. We examined the relationship of coffee and tea consumption with the incidence of hospitalization or death from chronic liver disease (CLD). Methods: Participants in the population-based, first National Health and Nutrition Examination Survey, 1971–1975, were asked about coffee and tea consumption, which was categorized as 2 cups per day (mean, 4.0 cups). A second analysis included persons who, in 1982–1984, were asked more detailed questions on coffee and tea drinking. Participants were followed through 1992–1993 for a hospital or death certificate diagnosis of CLD or cirrhosis (ICD-9-CM 571). Hazard rate ratios for CLD according to coffee and tea intake were calculated using Cox proportional hazards analysis. Results: Among 9849 persons followed for a median of 19.0 years (range, 0.02–22.1), the cumulative incidence of CLD was 1.4%. In multivariate analysis, participants who drank >2 cups per day had less than half the rate of CLD as those who drank <1 cup per day (hazard ratio, 0.43, 95% confidence interval: 0.24–0.78). Protection by coffee and tea was limited to persons at higher risk for liver diseases from heavier alcohol intake, overweight, diabetes, or high iron saturation. Among 9650 participants who provided detailed drink information in 1982–1984, intake of regular ground coffee and of caffeine was associated with lower incidence of CLD. Conclusions: Coffee and tea drinking decreases the risk of clinically significant CLD. Background & Aims: Coffee drinking has been suggested to protect against liver injury, but it is uncertain whether this is of clinical significance. We examined the relationship of coffee and tea consumption with the incidence of hospitalization or death from chronic liver disease (CLD). Methods: Participants in the population-based, first National Health and Nutrition Examination Survey, 1971–1975, were asked about coffee and tea consumption, which was categorized as 2 cups per day (mean, 4.0 cups). A second analysis included persons who, in 1982–1984, were asked more detailed questions on coffee and tea drinking. Participants were followed through 1992–1993 for a hospital or death certificate diagnosis of CLD or cirrhosis (ICD-9-CM 571). Hazard rate ratios for CLD according to coffee and tea intake were calculated using Cox proportional hazards analysis. Results: Among 9849 persons followed for a median of 19.0 years (range, 0.02–22.1), the cumulative incidence of CLD was 1.4%. In multivariate analysis, participants who drank >2 cups per day had less than half the rate of CLD as those who drank <1 cup per day (hazard ratio, 0.43, 95% confidence interval: 0.24–0.78). Protection by coffee and tea was limited to persons at higher risk for liver diseases from heavier alcohol intake, overweight, diabetes, or high iron saturation. Among 9650 participants who provided detailed drink information in 1982–1984, intake of regular ground coffee and of caffeine was associated with lower incidence of CLD. Conclusions: Coffee and tea drinking decreases the risk of clinically significant CLD. Consumption of coffee or caffeine or both may reduce liver injury, as suggested by studies of liver enzymes. Greater coffee, and especially caffeine, intake was associated with a lower prevalence of abnormal alanine aminotransferase (ALT) activity in the United States population.1Ruhl C.E. Everhart J.E. Coffee and caffeine consumption reduce the risk of elevated serum alanine aminotransferase activity in the United States.Gastroenterology. 2005; 128: 24-32Abstract Full Text Full Text PDF PubMed Scopus (201) Google Scholar Studies in Europe and Japan also found inverse relationships of coffee with serum levels of γ-glutamyltransferase2Arnesen E. Huseby N.E. Brenn T. Try K. The Tromso Heart Study distribution of, and determinants for, γ-glutamyltransferase in a free-living population.Scand J Clin Lab Invest. 1986; 46: 63-70Crossref PubMed Scopus (88) Google Scholar, 3Nilssen O. Forde O.H. Brenn T. The Tromso Study. Distribution and population determinants of γ-glutamyltransferase.Am J Epidemiol. 1990; 132: 318-326PubMed Google Scholar, 4Casiglia E. Spolaore P. Ginocchio G. Ambrosio G.B. Unexpected effects of coffee consumption on liver enzymes.Eur J Epidemiol. 1993; 9: 293-297Crossref PubMed Scopus (116) Google Scholar, 5Kono S. Shinchi K. Imanishi K. Todoroki I. Hatsuse K. Coffee and serum γ-glutamyltransferase a study of self-defense officials in Japan.Am J Epidemiol. 1994; 139: 723-727PubMed Google Scholar, 6Pintus F. Mascia P. Distribution and population determinants of γ-glutamyltransferase in a random sample of Sardinian inhabitants. ATS-SARDEGNA Research Group.Eur J Epidemiol. 1996; 12: 71-76Crossref PubMed Scopus (47) Google Scholar, 7Poikolainen K. Vartiainen E. Determinants of γ-glutamyltransferase positive interaction with alcohol and body mass index, negative association with coffee.Am J Epidemiol. 1997; 146: 1019-1024Crossref PubMed Scopus (85) Google Scholar, 8Tanaka K. Tokunaga S. Kono S. Tokudome S. Akamatsu T. Moriyama T. Zakouji H. Coffee consumption and decreased serum γ-glutamyltransferase and aminotransferase activities among male alcohol drinkers.Int J Epidemiol. 1998; 27: 438-443Crossref PubMed Scopus (127) Google Scholar, 9Honjo S. Kono S. Coleman M.P. Shinchi K. Sakurai Y. Todoroki I. Umeda T. Wakabayashi K. Imanishi K. Nishikawa H. Ogawa S. Katsurada M. Nakagawa K. Yoshizawa N. Coffee drinking and serum γ-glutamyltransferase an extended study of Self-Defense Officials of Japan.Ann Epidemiol. 1999; 9: 325-331Abstract Full Text Full Text PDF PubMed Scopus (49) Google Scholar, 10Nakanishi N. Nakamura K. Nakajima K. Suzuki K. Tatara K. Coffee consumption and decreased serum γ-glutamyltransferase a study of middle-aged Japanese men.Eur J Epidemiol. 2000; 16: 419-423Crossref PubMed Scopus (62) Google Scholar, 11Nakanishi N. Nakamura K. Suzuki K. Tatara K. Lifestyle and the development of increased serum γ-glutamyltransferase in middle-aged Japanese men.Scand J Clin Lab Invest. 2000; 60: 429-438Crossref PubMed Scopus (19) Google Scholar, 12Nakanishi N. Nakamura K. Suzuki K. Tatara K. Lifestyle and serum γ-glutamyltransferase a study of middle-aged Japanese men.Occup Med. 2000; 50 (Lond): 115-120Crossref Scopus (22) Google Scholar and aminotransferases.4Casiglia E. Spolaore P. Ginocchio G. Ambrosio G.B. Unexpected effects of coffee consumption on liver enzymes.Eur J Epidemiol. 1993; 9: 293-297Crossref PubMed Scopus (116) Google Scholar, 8Tanaka K. Tokunaga S. Kono S. Tokudome S. Akamatsu T. Moriyama T. Zakouji H. Coffee consumption and decreased serum γ-glutamyltransferase and aminotransferase activities among male alcohol drinkers.Int J Epidemiol. 1998; 27: 438-443Crossref PubMed Scopus (127) Google Scholar, 13Nakanishi N. Nakamura K. Suzuki K. Tatara K. Effects of coffee consumption against the development of liver dysfunction a 4-year follow-up study of middle-aged Japanese male office workers.Ind Health. 2000; 38: 99-102Crossref PubMed Scopus (17) Google Scholar, 14Honjo S. Kono S. Coleman M.P. Shinchi K. Sakurai Y. Todoroki I. Umeda T. Wakabayashi K. Imanishi K. Nishikawa H. Ogawa S. Katsurada M. Nakagawa K. Yoshizawa N. Coffee consumption and serum aminotransferases in middle-aged Japanese men.J Clin Epidemiol. 2001; 54: 823-829Abstract Full Text Full Text PDF PubMed Scopus (89) Google Scholar Fewer studies have investigated the relationship of coffee and caffeine consumption with chronic liver disease and hepatic cirrhosis. In a large prospective study of Kaiser Permanente Medical Care Program members in California, coffee but not tea drinking was independently associated with a lower incidence of alcoholic but not nonalcoholic cirrhosis hospitalization or death.15Klatsky A.L. Armstrong M.A. Alcohol, smoking, coffee, and cirrhosis.Am J Epidemiol. 1992; 136: 1248-1257PubMed Google Scholar Three Italian hospital-based case-control studies found inverse relationships of coffee consumption with alcoholic16Corrao G. Lepore A.R. Torchio P. Valenti M. Galatola G. D’Amicis A. Arico S. di Orio F. Provincial Group for the Study of Chronic Liver DiseaseThe effect of drinking coffee and smoking cigarettes on the risk of cirrhosis associated with alcohol consumption. A case-control study.Eur J Epidemiol. 1994; 10: 657-664Crossref PubMed Scopus (120) Google Scholar or unspecified cirrhosis.17Corrao G. Zambon A. Bagnardi V. D’Amicis A. Klatsky A. Coffee, caffeine, and the risk of liver cirrhosis.Ann Epidemiol. 2001; 11: 458-465Abstract Full Text Full Text PDF PubMed Scopus (126) Google Scholar, 18Gallus S. Tavani A. Negri E. La Vecchia C. Does coffee protect against liver cirrhosis?.Ann Epidemiol. 2002; 12: 202-205Abstract Full Text Full Text PDF PubMed Scopus (84) Google Scholar Most recently, a large Norwegian population-based prospective study found a lower risk of both alcoholic and all-cirrhosis mortality among coffee drinkers.19Tverdal A. Skurtveit S. Coffee intake and mortality from liver cirrhosis.Ann Epidemiol. 2003; 13: 419-423Abstract Full Text Full Text PDF PubMed Scopus (122) Google Scholar In addition, hepatocellular carcinoma, for which cirrhosis is the primary risk factor, has been associated with lower coffee intake in some studies,20Gelatti U. Covolo L. Franceschini M. Pirali F. Tagger A. Ribero M.L. Trevisi P. Martelli C. Nardi G. Donato F. Brescia HCC Study GroupCoffee consumption reduces the risk of hepatocellular carcinoma independently of its aetiology a case-control study.J Hepatol. 2005; 42: 528-534Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar, 21Shimazu T. Tsubono Y. Kuriyama S. Ohmori K. Koizumi Y. Nishino Y. Shibuya D. Tsuji I. Coffee consumption and the risk of primary liver cancer pooled analysis of two prospective studies in Japan.Int J Cancer. 2005; 116: 150-154Crossref PubMed Scopus (117) Google Scholar, 22Inoue M. Yoshimi I. Sobue T. Tsugane S. Influence of coffee drinking on subsequent risk of hepatocellular carcinoma a prospective study in Japan.J Natl Cancer Inst. 2005; 97: 293-300Crossref PubMed Scopus (164) Google Scholar, 23Gallus S. Bertuzzi M. Tavani A. Bosetti C. Negri E. La Vecchia C. Lagiou P. Trichopoulos D. Does coffee protect against hepatocellular carcinoma?.Br J Cancer. 2002; 87: 956-959Crossref PubMed Scopus (80) Google Scholar although not in all.24La Vecchia C. Ferraroni M. Negri E. D’Avanzo B. Decarli A. Levi F. Franceschi S. Coffee consumption and digestive tract cancers.Cancer Res. 1989; 49: 1049-1051PubMed Google Scholar, 25Kuper H. Tzonou A. Kaklamani E. Hsieh C.C. Lagiou P. Adami H.O. Trichopoulos D. Stuver S.O. Tobacco smoking, alcohol consumption and their interaction in the causation of hepatocellular carcinoma.Int J Cancer. 2000; 85: 498-502Crossref PubMed Scopus (302) Google Scholar To our knowledge, no studies have prospectively investigated the relationship of coffee or tea consumption with chronic liver disease in the general US population. Therefore, we examined the relationship of coffee and tea drinking with incident chronic liver disease in the first National Health and Nutrition Examination Survey Epidemiologic Follow-Up Study (NHEFS), a large US national, population-based study. Consistent findings based on beverage consumption at 2 time points approximately 10 years apart established more conclusive evidence of a protective effect. Conducted between 1971 and 1975, the first National Health and Nutrition Examination Survey (NHANES I) included interview, examination, and laboratory data collected from a national probability sample of the civilian, noninstitutionalized population.26Miller H.W. Plan and operation of the health and nutrition examination survey, United States 1971-1973, part A—development, plan, and operation. National Center for Health Statistics, Hyattsville, MD1973Google Scholar, 27Engel A. Murphy R.S. Maurer K. Collins E. Plan and operation of the NHANES I Augmentation Survey of adults 25-74 years, United States 1974-1975. National Center for Health Statistics, Hyattsville, MD1978Google Scholar At the NHANES I interview, participants were asked about coffee or tea consumption, which was categorized as 0 to 2 cups per day for this analysis. Participants were also asked about frequency and quantity of beer, wine, or liquor consumption, and alcohol use (drinks per day) was summarized as none, 2. Doctor-diagnosed diabetes mellitus was determined by interview. Body mass index (BMI) [weight in kg/(height in m)2] and subscapular and triceps skin folds were measured at examination. Serum iron and total iron-binding capacity were measured from blood samples by a modification of the automated Technicon AAII-25 Method,28NHANES I hematology and clinical chemistry procedures developed or utilized by the Center for Disease Control, Bureau of Laboratories, 1971-1975, instruction manual, data collection, part 16. US Government Printing Office, Washington, DC1979Google Scholar and transferrin saturation (percentage) was calculated. Demographic variables were age (years; 25–39, 40–59, 60+), sex, ethnicity (white, black, other), and education (less than high school graduate, high school graduate). The NHANES I Epidemiologic Follow-Up Study (NHEFS) was a longitudinal study of 14,407 NHANES I participants aged 25–74 years conducted in 4 waves: 1982–1984, 1986, 1987, and 1992–1993.29Cohen B.B. Barbano H.E. Cox C.S. Feldman J.J. Finucane F.F. Kleinman J.C. Madans J.H. Plan and operation of the NHANES I Epidemiologic Follow-up Study, 1982-84. National Center for Health Statistics, Hyattsville, MD1987Google Scholar, 30Finucane F.F. Freid V.M. Madans J.H. Cox C.S. Kleinman J.C. Rothwell S.T. Barbano H.E. Feldman J.J. Plan and operation of the NHANES I Epidemiologic Follow-up Study, 1986. National Center for Health Statistics, Hyattsville, MD1987Google Scholar, 31Cox C.S. Rothwell S.T. Madans J.H. Finucane F.F. Freid V.M. Kleinman J.C. Barbano H.E. Feldman J.J. Plan and operation of the NHANES I Epidemiologic Follow-up Study, 1987. National Center for Health Statistics, Hyattsville, MD1992Google Scholar, 32Cox C.S. Mussolino M.E. Rothwell S.T. Lane M.A. Golden C.D. Madans J.H. Feldman J.J. Plan and operation of the NHANES I Epidemiologic Follow-up Study, 1992. National Center for Health Statistics, Hyattsville, MD1997Google Scholar Follow-up included measured weight in 1982–1984 and interviews and collection of hospital and nursing home records and death certificates at each wave. Excluded from the current analysis were 546 survey participants who could not be traced; 1138 who at baseline reported ever having yellow jaundice, had hepatomegaly or splenomegaly on abdominal examination, or had a serum albumin level <3.0 gm/dL; and 2874 who had missing data on coffee and tea consumption. Remaining for analysis were 9849 persons with a median follow-up time of 19.0 years (range, 0.02–22.1 years). Records were obtained for overnight medical facility stays occurring since the participant’s NHANES I examination. Hospital discharge and nursing home admission diagnoses were recoded by trained medical coders using the Ninth Revision of the International Classification of Diseases with Clinical Modification (ICD-9-CM).33International classification of diseases. 4th ed. US Government Printing Office, Washington, DC1991Google Scholar We defined chronic liver disease as a facility stay with an ICD-9-CM diagnosis of a 3-, 4-, or 5-digit code, for which the first 3 digits were 571, chronic liver disease and cirrhosis. Of 94 first facility stays with a chronic liver disease diagnosis, 97% were hospitalizations, and 3% were nursing home admissions. An additional 14 participants died, with chronic liver disease listed as underlying cause of death without a prior facility stay diagnosis. The date of diagnosis was considered to be the admission date of the first facility stay with a chronic liver disease diagnosis or date of death. For each level of coffee and tea consumption, we calculated the cumulative percentage of participants with chronic liver disease during the 20 years of follow-up. We estimated curves for percentage of participants with chronic liver disease (reciprocal of the survival curve) by coffee and tea intake categories using Kaplan–Meier analysis. We then calculated hazard ratio estimates (relative risk) using Cox proportional hazard regression models (SAS, PROC PHREG; SAS OnlineDoc, version 8, 1990; SAS Institute Inc, Cary, NC) to take into consideration varying lengths of follow-up. Time at risk was from the date of the NHANES I examination to the date of chronic liver disease diagnosis for cases or date of last contact or death for noncases. All factors met the proportional hazard assumption of a relatively constant risk ratio through examination of −log (−log) plots of survival versus time by categories.34Kleinbaum D.G. Survival analysis. Springer, New York, NY1996Crossref Google Scholar The relationship of incident chronic liver disease with coffee and tea drinking was examined further in multivariate proportional hazards analysis while controlling for effects of other potential risk factors. Multivariate analyses excluded persons with missing values for any risk factor included in the model. The trend in the relative risk of chronic liver disease across coffee and tea intake categories was computed by including this factor in analyses as an ordinal variable with 3 levels. Stratified analyses were performed to examine the consistency of relationships of chronic liver disease with coffee and tea drinking among persons at low and higher risk for liver diseases. Multiplicative interaction between coffee and tea intake and liver injury risk level was also evaluated. Consistent with previous analyses utilizing the NHEFS, our main analysis was conducted without weights or design effects. To assess the effect of the complex sampling design on the results, a supplemental multivariate Cox proportional hazard regression analysis was performed using SUDAAN (SUDAAN 7.0., 1996; Research Triangle Institute, Research Triangle Institute Park, NC) to incorporate sample weights, stratification, and clustering into the analysis. A new set of sample weights was calculated for use with the entire NHEFS sample.35Ingram D.D. Makuc D.M. Statistical issues in analyzing the NHANES I Epidemiologic Follow-up Study. National Center for Health Statistics, Hyattsville, MD1994Google Scholar A second analysis was performed using a follow-up of study participants who were interviewed in 1982–1984 when more detailed information on coffee and tea drinking was obtained. As part of an expanded food frequency inventory, participants were asked about the average quantity and frequency of consumption of coffee (instant decaffeinated, other instant, ground decaffeinated, other ground), tea (herb, regular, including ice tea), regular or diet cola-type soda, and chocolate. Intake of total coffee and tea, ground regular coffee, decaffeinated coffee, instant coffee, regular tea, and cola was calculated. Beverage intake was coded as 0 to 2 cups per day. 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).36Michels K.B. Willett W.C. Fuchs C.S. Giovannucci E. Coffee, tea, and caffeine consumption and incidence of colon and rectal cancer.J Natl Cancer Inst. 2005; 97: 282-292Crossref PubMed Scopus (124) Google Scholar Caffeine intake was expressed as quintiles. Participants were also asked about alcohol use, which was coded as for NHANES I, and about doctor-diagnosed diabetes mellitus. Information on cigarette smoking was obtained from all participants and categorized as never, former, <1 pack per day, and ≥1 pack per day. BMI was calculated from measured 1982–1984 weight and NHANES I height. We excluded 3908 participants who were not interviewed in 1982–1984 or had no further follow-up; 786 who had a history of yellow jaundice, hepatomegaly or splenomegaly on abdominal examination, or a serum albumin level <3.0 gm/dL at NHANES I; 41 who had a facility stay with a chronic liver disease diagnosis between NHANES I and the 1982–1984 examination; and 22 who reported in 1982–1984 ever having doctor-diagnosed cirrhosis. Remaining for analysis were 9650 persons with a median follow-up time of 9.1 years (range, 0.01–11.1 years). Incident chronic liver disease was coded and analyzed as in the entire follow-up. Forty-six persons had a facility stay with chronic liver disease, and an additional 7 died with chronic liver disease, without a prior facility stay diagnosis. The trend in the relative risk of chronic liver disease across caffeine quintiles was computed by including this factor in analyses as an ordinal variable with 5 levels. The baseline characteristics of the 9849 participants who were followed for chronic liver disease are shown in Table 1 according to coffee and tea consumption. Coffee and tea intake ranged from zero to 16 cups per day with a median of 2 cups per day. Mean consumption by category was 0.2 for less than 1 cup per day, 1.5 for 1 to 2 cups per day, and 4.0 for more than 2 cups per day. Of the persons who reported less than 1 cup per day, 35% drank no coffee or tea at all. Higher coffee and tea consumption was associated with middle age, male sex, white race, greater education, lower BMI, and less central fat as determined by subscapular-to-triceps skin-fold ratio. Moderate alcohol consumption (1–2 drinks per day) was associated with greater coffee and tea drinking, but higher consumption was not.Table 1Coffee and Tea Drinking Distribution of Participants by Categories of Baseline CharacteristicsCharacteristicNumber of participantsCoffee and tea (cups/day) (%)P valueaFrom χ2 test. 2Total sample9849144541Age (y)<.001 25–393311204238 40–593055114247 60 +3483125137Gender<.001 Men3854144343 Women5995154639Race<.001 White8106114346 Black1633295714 Other110124345Education<.001 <12th grade4707144838 ≥12th grade5059144343Alcohol drinking (drinks/day)<.001 0 (nondrinker)4315164737 >0 to 2501164737Serum transferrin saturation (%).13 ≤508918154639 >50357144145BMI (kg/m2)<.001 <254933144343 25 to <303251134740 ≥301665184834Subscapular-to-triceps skin-fold ratiobCut points were tertiles (thirds).<.001 <0.83292134344 0.8 to <1.23285144640 ≥1.23272164638Diabetes.87 No9426144541 Yes383154441a From χ2 test.b Cut points were tertiles (thirds). Open table in a new tab The overall risk of hospitalization or death with chronic liver disease was 1.4% at 20 years. Unadjusted cumulative incidence of chronic liver disease over the 20 years of follow-up according to coffee and tea consumption was 1.8% for less than 1 cup per day, 1.6% for 1 to 2 cups per day, and 1.1% for more than 2 cups per day (Table 2). A lower incidence of chronic liver disease was associated with higher consumption of coffee and tea (Figure 1), with a significant test for trend (P = .002). Other factors associated with increased risk of chronic liver disease were age greater than 40 years, male sex, less than a high school education, 1 or more alcoholic drinks per day, elevated transferrin saturation, a subscapular-to-triceps skin-fold ratio in the upper 2 tertiles, and diabetes. Blacks and overweight and obese persons (BMI ≥25) had nonsignificantly higher rates of chronic liver disease.Table 2Cumulative Unadjusted Probability Over 20 Years and Unadjusted Relative Risk and 95% Confidence Interval for Chronic Liver DiseaseCharacteristicNumber with chronic liver diseaseCumulative chronic liver disease (%)Unadjusted hazard ratioaEstimated using Cox proportional hazard regression analysis.95% Confidence intervalCoffee and tea (cups/day) 2291.10.460.26–0.80Age (y) 25–39220.741.0 40–59502.12.61.6–4.4 60+361.42.41.4–4.0Gender Men672.61.0 Women410.830.340.23–0.50Race White841.41.0 Black242.01.50.96–2.4 Other0———Education 0 to 2286.58.65.1–14.5Serum transferrin saturation (%) ≤50921.41.0 >50155.04.32.5–7.4BMI (kg/m2) <25441.21.0 25 to <30431.61.50.99–2.3 ≥30211.81.50.87–2.5Subscapular-to-triceps skin-fold ratiobCut points were tertiles (thirds). <0.8170.631.0 0.8 to 2 alcoholic drinks per day, transferrin saturation >50%, diagnosed diabetes, BMI ≥30, and subscapular-to-triceps skin-fold ratio in the highest third. Persons in the higher risk subgroup had 2.7 times the risk of chronic liver disease (95% confidence interval [CI]: 1.6–4.5) compared with those in the low-risk group, independent of demographics and education. The cumulative incidence of chronic liver disease differed by coffee consumption among persons at higher risk for liver injury but not among those at low risk (Figure 2). These findings were substantiated in multivariate analysis (Table 3), in which higher risk participants who consumed more than 2 cups per day of coffee and tea had only 40% of the risk of those drinking less than 2 cups per day. A nonstatistically significant trend for interaction of coffee and tea consumption with risk group for chronic liver disease was found in unadjusted (P = .12) and multivariate-adjusted analysis (P = .11).Table 3Coffee and Tea Consumption and Multivariate-Adjusted Relative Risks and 95% Confidence Intervals for Chronic Liver Disease Among All Participants and Persons at Low and High Risk for Liver DiseaseCoffee and tea (cups/day)Number of participantsNumber with chronic liver diseaseHazard ratioaEstimated using Cox proportional hazard regression analysis. Hazard ratios for all participants adjusted for age, sex, race, education, alcohol drinking, serum transferrin saturation, BMI deciles, subscapular-to-triceps skin-fold ratio deciles, and diabetes. Hazard ratios for liver disease risk subgroups adjusted for age, sex, race, and education.95% Confidence intervalP valueAll participants.003 23622280.430.24–0.78Persons at low risk for liver diseasebHigh risk defined as one or more of the following: >2 alcoholic drinks per day, transferrin saturation >50%, diagnosed diabetes, BMI ≥30, subscapular-to-triceps skin-fold ratio ≥1.2 (highest third)..63 ≤22660131.0— >21995101.20.52–3.0Persons at high risk for liver disease<.001 ≤22844651.0— >21627180.400.23–0.68a Estimated using Cox proportional hazard regression analysis. Hazard ratios for all participants adjusted for age, sex, race, education, alcohol drinking, serum transferrin saturation, BMI deciles, subscapular-to-triceps skin-fold ratio deciles, and diabetes. Hazard ratios for liver disease risk subgroups adjusted for age, sex, race, and education.b High risk defined as one or more of the following: >2 alcoholic drinks per day, transferrin saturation >50%, diagnosed diabetes, BMI ≥30, subscapular-to-triceps skin-fold ratio ≥1.2 (highest third). Open table in a new tab To limit a theoretic effect of occult liver disease on coffee and tea consumption, we performed an analysis that excluded participants who died or had a chronic liver disease diagnosis within 5 years of baseline. This restriction had little effect on the results. Relative to consumption of less than 1 cup per day, the multivariate-adjusted hazard ratio was 0.90 (95% CI: 0.50–1.6) for 1 to 2 cups per day and 0.42 (95% CI: 0.21–0.86) for more than 2 cups per day (P value for trend = .008). Because of the small number of coffee and tea abstainers (5.0%), we used an intake of less than 1 cup per day as the reference group. If non-coffee and non-tea drinkers were used instead as the comparison group, a similar but less precise inverse relationship was found. The multivariate-adjusted hazard ratio was 0.95 (95% CI: 0.39–2.3) for more than zero but less than 1 cup per day, 0.76 (95% CI: 0.36–1.6) for 1 to 2 cups per day, and 0.42 (95% CI: 0.18–0.95) for more than 2 cups per day (P value for trend = .005). An additional analysis was performed in the entire population using sample weights and design effects. The results were similar to those of the unweighted analysis. The multivariate-adjusted hazard ratio for 1 to 2 cups of coffee or tea per day compared with less than 1 cup per day was 0.82 (95% CI: 0.42–1.6) and, for more than 2 cups of coffee or tea per day, was 0.43 (95% CI: 0.19–0.95) (P value for trend = .013). The consistency of coffee and tea drinking over time was invest
Referência(s)