By Nancy J. Selfridge, MD
Professor and Chair, Department of Clinical Medicine, Assistant Dean for Clinical Science, Ross University School of Medicine, Commonwealth of Dominica, West Indies
Dr. Selfridge reports no financial relationships relevant to this field of study.
- Brewed coffee contains many biologically active substances with antioxidant and other putative positive effects on human physiology.
- Meta-analyses of pooled data have shown mostly inverse associations of moderate coffee consumption of 2-5 cups/day and risk of all-cause mortality, cancers, cardiovascular disease, and several other diseases.
- Higher levels of coffee consumption appear to be associated with increased fracture risk, increased risk of bladder cancer, and an increased risk of gastric cancer in U.S. consumers.
Worldwide, coffee is the second most consumed beverage after water. Chosen because of its rich complex flavors and pleasant stimulating effects, recent research suggests potential health benefits, including risk reductions for cardiovascular disease, cancers, diabetes, and Parkinson’s disease. The global consumption statistics suggest that the public health effect of coffee consumption could be significant, even if the effect on an individual consumer is small.
Roasted coffee beans and the various brews derived from them contain many bioactive substances with potentially beneficial and detrimental effects on health. A summary of the biologically active classes of compounds and their putative effects is shown in Table 1.1 Varying concentrations of substances in the variety of coffee preparations available, confounded by the fact that most of the substances are highly metabolized after consumption and do not reach significant plasma concentrations, dictate that the mechanisms of any observed beneficial effects remain obscure.
It is clear that describing coffee’s benefits as primarily due to antioxidant effect is an oversimplification.1 Coffee has been touted as an unhealthy beverage in the past based on epidemiologic studies published between 1980 and 2000, suggesting a possible association with increased risk of fractures, hypertension, and some cancers. However, more recent research has not supported these early conclusions. Several analyses of cumulative data form the focus of this review.
Coffee Consumption and all-cause mortality
Numerous epidemiological studies have investigated the effect of coffee consumption on health and disease. Analyzing pooled data often helps clarify or confirm statistical associations when an effect size is small and individual studies are not adequately powered or differ in methodology. Four recent publications have done just this, looking at coffee consumption and all-cause mortality. Malerba et al analyzed data from 23 prospective cohort studies comparing relative risk (RR) of all-cause mortality of the lowest coffee consumption (≤ 1 cup per day) to moderate (1-3 cups/day) and high (> 3 cups/day) consumption, using random-effects models for the heterogeneity in the studies.2 The RR of highest consumption vs. lowest consumption categories was 0.88 (95% confidence interval [CI], 0.84-0.93) for all studies and was 0.87 (95% CI, 0.82-0.93) for the 19 studies that adjusted for smoking.2
Je and Giovanucci included 20 of the same studies in their analysis,3 all controlling for smoking, and as well as two earlier studies not considered in the Malerba analysis. These authors indicated that their pooled studies included 973,904 participants and 129,538 deaths. They similarly defined lowest coffee consumption as < 1 cup/day and moderate intake as 1-2 cups/day, but they stratified high consumption as 2-4 cups per day and 5-9 cups per day based on the variations in this definition in the original research publications. Pooled RR for total mortality comparing high vs. low consumption was 0.86 (95% CI, 0.80-0.92), and this result did not vary at all using the two different definitions of high consumption. Moderate coffee consumption was associated with a weak inverse association, RR 0.92 (95% CI, 0.87-0.98). Interestingly, these authors noted that associations differed according to geographical region, with European studies showing the strongest inverse association, followed by Japan, and finally by the United States.3
Crippa et al conducted a similar analysis of 21 prospective cohort studies, including four studies not included in the analyses above.4 Again, all chosen studies adjusted risk for smoking and assessed risk based on three categories of coffee consumption. But unlike the other analyses, which assessed highest to lowest category relative risks, these authors performed a dose-response meta-analysis and reported risk reductions for total mortality, cardiovascular disease, and cancer deaths. Pooled data included 997,464 participants and 121,915 deaths. The authors reported strong evidence of a nonlinear inverse association between coffee consumption and mortality due to all causes and cardiovascular disease (P < 0.001). In the dose-response analysis, 4 cups/day showed the largest risk reduction for all-cause mortality (16%; 95% CI, 13-18) and 3 cups/day showed the largest risk reduction for cardiovascular disease mortality (21%; 95% CI, 16-26). There was no association between coffee consumption and cancer mortality.4
Ding et al examined associations between coffee consumption and total and cause-specific mortality in pooled data from three large cohort studies: the Nurses’ Health Study, the Nurses’ Health Study II, and the Health Professionals Follow-up Study. Total participants numbered 124,821 and there were 31,956 deaths over 4,690,072 person years of follow-up. Hazard ratios were reported instead of RRs, and these data sets included periodic updates of coffee consumption through food frequency questionnaires. Compared to no coffee consumption, hazard ratios for death were as follows: 0.95 (95% CI, 0.91-0.99) for ≥ 1 cup per day, 0.91 (95% CI, 0.88-0.95) for 1.1-3 cups/day, 0.93 (95% CI, 0.89-0.97) for 3.1-5 cups/day, and 1.02 (95% CI, 0.96-1.07) for > 5 cups/day. This non-linear association changed to a linear inverse association when restricting the analysis to never smokers (P < 0.001).5
Recently, several meta-analyses looking at the associations between coffee consumption and specific disease risk have been published. These are summarized below.
Ding at al performed a dose-response meta-analysis on 36 prospective cohort studies (including many of the studies cited in the analyses of all-cause mortality, above) investigating the relationship between coffee consumption and risk of cardiovascular disease, including coronary heart disease, heart failure, stroke, and cardiovascular disease mortality. The pooled data included 1,279,804 participants and 36,352 cases of cardiovascular disease. RR of developing cardiovascular disease compared to the lowest intake category (< 1 cup/day) was 0.95 for 5 cups per day (95% CI, 0.87-1.03), 0.85 for 3.5 cups per day (95% CI, 0.80-0.90), and 0.89 for 1.5 cups/day (95% CI, 0.84-0.94).6 Mostofsky et al looked at coffee consumption and heart failure risk, performing a dose-response meta-analysis on five prospective cohort studies that consisted of 140,220 participants and 6,522 incidents of heart failure. They found an inverse relationship for coffee consumption up to 9-10 servings per day, with the lowest risk for 3-4 servings per day, RR 0.89 (95% CI, 0.81-0.99). More than 10 servings per day was associated with RR > 1.0.7 Thus, even high coffee intake of up to 10 servings per day appears to have a protective effect for cardiovascular disease, with 3-4 cups per day associated with the lowest risk.
Larsson and Orsini analyzed 11 prospective studies to assess the association between coffee dose and stroke risk. Pooled data included 479,689 participants and 10,003 cases of stroke. Risk for each category of consumption was compared to no consumption. RR was 0.86 (95% CI, 0.78-0.94) for 2 cups/day, 0.83 (95% CI, 0.74-0.92) for 3-4 cups/day, 0.87 (95% CI, 0.77-0.97) for 6 cups/day, and 0.93 (95% CI, 0.79-1.08) for 8 cups/day.8 Kim at al arrived at similar risks comparing only highest to lowest intake categories for nine studies, six of which were included in Larsson’s analysis above: RR 0.83 (95% CI, 0.76-0.91).9
Lee et al performed a dose-response meta-analysis to assess coffee consumption and risk of fracture. Nine prospective cohort studies and six case-control studies were included with pooled data from 253,514 participants and 12,939 fracture cases. They found that the RR of fracture comparing highest to lowest categories of consumption was 1.14 (95% CI, 1.05-1.24) for women and 0.76 (95% CI, 0.62-0.94) for men. RR for women increased from 1.02 (95% CI, 1.01-1.04) to 1.54 (95% CI, 1.19-1.99) for 2 cups/day and 8 cups/day, respectively.10 Li et al assessed 10 prospective cohort studies (five in common with Lee, above) for coffee and hip fracture risk. Highest coffee consumption (2-9 cups/day) compared to lowest (never to seldom) was associated with an RR of 1.13 (95% CI, 0.86-1.48).11 Both of these analyses included studies involving participants who were middle-aged and elderly, so menopausal women, a cohort with significant osteoporosis and fracture risk, were included. Although coffee consumption is associated with an increase in urinary calcium excretion, it appears that there is no increased risk of fracture for men, and a negligible risk for women at around 2 cups/day, which may increase with higher daily doses.
Since cancers are much less common than cardiovascular events, the large participant and event numbers from pooled data have been especially helpful in determining how coffee consumption may be associated with risks.
Gastric cancer: Five recent meta-analyses of coffee consumption and gastric cancer risk concluded that there was no association or a low positive association.12,13,14,15,16 All of these authors included most of the same prospective cohort studies in their analyses. Results by Li et al were representative: comparing to lowest coffee consumption (< 1 cup/day), RR was 1.13 (95% CI, 0.94-1.35).16 Two of these papers included subgroup analyses indicating that RR for participants from the United States was higher, 1.36 (95% CI, 1.06-1.75), despite the known fact that coffee consumption is much higher in European countries.12,13 Authors believed that these differences might be due to genetic factors influencing gastric cancer risk or to artifact due to limited statistical power for the sub-analyses and that further studies were warranted. Coffee consumption may be associated with an increased risk of gastric cancer in the United States, but not in other countries included in these studies.
Breast cancer: Li et al analyzed pooled data from 26 published articles including 863,096 participants and 49,497 breast cancer cases from both prospective cohort and case-control studies. Results of this and previous meta-analyses were similar. RR of breast cancer comparing lowest and highest consumption of coffee was 0.96 (95% CI, 0.93-1.00), an insignificant association.17 They noted that RR was slightly higher for premenopausal cancers (1.00) and slightly lower for postmenopausal cancers (0.92). A more recent meta-analysis of 37 studies yielded similar results, but the authors noted that an inverse association of coffee consumption with breast cancer risk was found for BRCA1 mutation carriers (RR = 0.69; P < 0.01).18
Prostate cancer: Cao et al included 10 prospective cohort studies with 206,096 participants and 8,973 cases of prostate cancer in their recent meta-analysis.19 An inverse association between coffee consumption and prostate cancer was noted with pooled RR, compared to non- or seldom drinkers, of 0.88 (95% CI, 0.82-0.95). Thus, coffee consumption appears to slightly reduce the risk of prostate cancer.
Pancreatic cancer: A recent meta-analysis of 20 cohort studies by Ran et al reported a RR for highest vs. lowest coffee consumption and pancreatic cancer of 0.75 (95% CI, 0.63-0.86).20 This analysis included three more studies than a previous publication by Turati et al, in which the pooled RR for cohort studies was 1.04 (95% CI, 0.80-1.36).21 Despite a sentinel case-control study published in 1981 showing a positive association between coffee consumption and pancreatic cancer, cumulative data consistently demonstrate no increased risk and possibly a protective effect.
Colorectal cancer: Gan et al analyzed 19 prospective cohort studies (2,046,575 participants and 22,629 cases of colorectal cancer) to assess risk associated with coffee consumption. Pooled RR for highest vs. lowest consumption categories was 0.98 (95% CI, 0.90-1.06). However, dose-response analysis demonstrated that higher levels of consumption were associated with lower risk: RR was 0.93 (95% CI, 0.89-0.99) for 6 cups/day, 0.90 (95% CI, 0.85-0.97) for 7 cups/day, and 0.87 (95% CI, 0.80-0.95) for 8 cups/day.22
Other cancers: Liu et al reported a pooled RR of 0.81 (95% CI, 0.68-0.97) in an analysis of five cohort studies and two case-control studies assessing coffee and melanoma risk. This inverse association existed only for caffeinated coffee.23 Oral cancer RR was reported to be 0.69 (95% CI, 0.54-0.88) in an analysis of 12 prospective cohort and case-control studies.24 Sang et al assessed 16 case-control and prospective cohort studies for liver cancer risk. Pooled RR for highest vs. lowest coffee consumption was 0.50 (95% CI, 0.42-0.59). Of interest is that the RR was 0.39 (95% CI 0.28-0.54) with no adjustment for a history of liver disease and 0.54 (95% CI, 0.46-0.66) after adjusting for a history of liver disease, indicating a potentially protective effect even for those patients at higher risk of developing liver cancer.25 Wu et al analyzed 34 case-control and six cohort studies to assess risk of bladder cancer. Odds ratios were reported due to the large number of case-control studies. Pooled OR for highest vs. lowest consumption was 1.33 (95% CI, 1.19-1.48). In this analysis, odds ratios were actually higher in non-smokers, 1.72 (95% CI, 1.25-2.35), vs. smokers, 1.24 (95% CI, 0.91-1.70). The authors hypothesized that this might be due to smoking effects on caffeine metabolism.26 Zhou et al explored coffee consumption and risk of endometrial cancer in a meta-analysis. Pooled RR from 13 studies including 1,534,039 participants and 10,100 cases of endometrial cancer comparing highest and lowest consumption categories was 0.66 (95% CI, 0.52-0.84) for caffeinated coffee and 0.77 (95% CI, 0.63-0.94) for decaffeinated coffee.27
Other diseases: An updated meta-analysis of 29 prospective studies, including 1,109,272 participants and 45,335 cases of type 2 diabetes, demonstrated an inverse association that was dose related. RR was reported at 0.92 (95% CI, 0.90-0.94) for 1 cup/day and 0.67 (95% CI, 0.61-0.74) for 6 cups/day. These findings were consistent with earlier meta-analyses.28 Coffee consumption showed an inverse association with urolithiasis in a meta-analysis by Wang et al: odds ratio 0.70 (95% CI, 0.60-0.82).29 An inverse association was also noted for Parkinson’s disease: RR 0.72 (95% CI, 0.6-0.81).30 Pooled RR for coffee and depression in a meta-analysis by Wang et al was 0.75 (95% CI, 0.62-0.91).31 Habitual coffee consumption and dementia risk was reported by Liu et al, who noted a significant inverse relationship between highest levels of coffee consumption and Alzheimer’s disease, RR 0.73 (95% CI, 0.55-0.97).32
The associations between coffee consumption and various conditions presented earlier is summarized in Table 2.
A plethora of meta-analyses have been published recently assessing the association of coffee consumption with risk of various disease and mortality. An inverse association (potentially protective effect) has been noted for all-cause mortality, cardiovascular disease including stroke, breast cancer in BRCA1 mutation carrying women, prostate cancer, colorectal cancer, pancreatic cancer, oral cancer, liver cancer, melanoma, endometrial cancer, type 2 diabetes, Parkinson’s disease, urolithiasis, depression, and Alzheimer’s disease. No significant association has been noted for breast cancer in pre- and postmenopausal women and gastric cancer in non-U.S. consumers. A small increased risk of fracture in women is associated with coffee intake and it appears to increase at higher levels of consumption. Significant positive associations have been noted for gastric cancer in U.S. consumers and for bladder cancer.
Several meta-analyses have investigated the associations between coffee consumption and pregnancy outcomes, but these studies are beyond the scope of this present review. Omitting conclusions relevant to pregnant women, in general, coffee consumption is not associated with significant risk of adverse health outcomes or death, with the exceptions noted above, and it may have some protective effect against certain cancers and diseases. The reported studies all had similar limitations. Associations in observational studies tell us nothing about causation, and we can only hypothesize about reasons and mechanisms.
Moderate heterogeneity, which can influence the reliability of the analyses, frequently was reported in the studies. Adjustment for all possible confounders in the original studies could hamper results, although many did adjust for the significant confounder of smoking status, which has been strongly associated with high coffee consumption. In addition, patients with chronic diseases may reduce coffee consumption voluntarily, yet baseline health status would affect outcomes independent of coffee consumption. All studies used Food Frequency Questionnaires and self-reporting for stratifying coffee consumption, which was generally given in cups per day or cups per week, both contributing to potential measurement errors. Coffee type (Arabica vs. Robusta) and method of preparation (boiled, steeped, filtered, percolated) were not specified; these affect the quantities and proportions of biologically active substances in coffee and could affect outcomes. Finally, coffee consumption may have other confounding aspects that could affect health outcomes. Many people drink coffee in social settings or while “taking a coffee break” and experience true pleasure associated with the coffee drinking experience. There may be psychosocial aspects intrinsic to these contexts that contribute to a beneficial effect of coffee on health outcomes.
Based on a significant amount of data, we can safely say that coffee consumption poses no significant health hazards for a given individual who is otherwise healthy, and where there is no concern for development of gastric cancer or bladder cancer. Women at risk for fracture, especially those with osteoporosis or osteopenia, should be informed of the increased risk of fracture associated with coffee consumption of > 2 cups/day. Statistics and what is known about the putative effects of the biologically active chemicals in coffee suggest that it may be a health-promoting beverage. People who enjoy their daily coffee consumption without suffering adverse effects, such as jitteriness, insomnia, dyspepsia or palpitations, can be reassured that they don’t need to give this up, though moderate intake of 3-4 cups/day appears a most wise limit. However, it is premature to suggest coffee as a preventive or therapeutic intervention until more is understood about the mechanisms of its impact on health and disease.
- Ludwig IA, et al. Coffee: Biochemistry and potential impact on health. Food Funct 2014;5:1633-1978.
- Malerba S, et al. A meta-analysis of prospective studies of coffee consumption and mortality for all causes, cancers and cardiovascular disease. Eur J Epidemiol 2013;28:527-539.
- Je Y, et al. Coffee consumption and total mortality: A meta-analysis of twenty prospective cohort studies. Br J Nutr 2014;111:1162-1173.
- Crippa A, et al. Coffee consumption and mortality from all causes, cardiovascular disease and cancer: A dose-response meta-analysis. Am J Epidemiol 2014;180:763-775.
- Ding M, et al. Association of coffee consumption with total and cause-specific mortality in 3 large prospective cohorts. Circulation 2015;132:2305-2315.
- Ding M, et al. Long-term coffee consumption and risk of cardiovascular disease: A systematic review and dose-response meta-analysis of prospective cohort studies. Circulation 2014;129:643-659.
- Mostofsky E, et al. Habitual coffee consumption and risk of heart failure: A dose-response meta-analysis. Circ Heart Fail 2012;5:401-405.
- Larsson SC, et al. Coffee consumption and risk of stroke: A dose-response meta-analysis of prospective studies. Am J Epidemiol 2011;174:993-1001.
- Kim B, et al. Coffee consumption and stroke risk: A meta-analysis of epidemiologic studies. Korean J Fam Med 2012;33:356-365.
- Lee DR, et al. Coffee consumption and risk of fractures: A systematic review and dose-response meta-analysis. Bone 2014;63:20-28.
- Li S, et al. Effect of coffee intake on hip fracture: A meta-analysis of prospective cohort studies. Nutr J 2015;14:38-44.
- Zeng SB, et al. Long-term coffee consumption and risk of gastric cancer. A PRISMA-compliant dose-response meta-analysis of prospective cohort studies. Medicine 2015;94:e1640.
- Xie F, et al. Coffee consumption and risk of gastric cancer: A large updated meta-analysis of prospective cohort studies. Nutrients 2014;6:3734-3746.
- Liu H, et al. Effect of coffee consumption on the risk of gastric cancer: A systematic review and meta-analysis of prospective cohort studies. PLoS ONE 2015;5:e0128501.
- Shen Z, et al. Coffee consumption and risk of gastric cancer: An updated meta-analysis. Clin Res Hepatol Gastroenterol 2015;39:245-53.
- Li L, et al. Coffee consumption and the risk of gastric cancer: A meta-analysis of prospective cohort studies. BMC Cancer 2015;15:733.
- Li X J, et al. Coffee consumption and risk of breast cancer: An up-to-date meta-analysis. PLoS One 2013;8:e52681.
- Jiang W, et al. Coffee and caffeine intake and breast cancer risk: An updated dose-response meta-analysis of 37 published studies. Gynecol Oncol 2013;129:620-629.
- Cao S, et al. Coffee consumption and risk of prostate cancer: A meta-analysis of prospective cohort studies. Carcinogenesis 2014;35:256-261.
- Ran HQ, et al. Coffee consumption and pancreatic cancer risk: An update meta-analysis of cohort studies. Pak J Med Sci 2016;32:253-259.
- Turati F, et al. A meta-analysis of coffee consumption and pancreatic cancer. Ann Oncol 2012;23:311-318.
- Gan Y, et al. Association of coffee consumption with risk of colorectal cancer: A meta-analysis of prospective cohort studies. Oncotarget 2016; doi: 10.18632/oncotarget.8627. [Epub ahead of print]
- Liu J, et al. Higher caffeinated coffee intake is associated with reduced malignant melanoma risk: A meta-analysis study. PLoS One 2016;11:e0147056.
- Zhang Y, et al. Association between coffee consumption and the risk of oral cancer: A meta-analysis of observational studies. Int J Clin Exp Med 2015;8:11657-11665.
- Sang LX, et al. Consumption of coffee associated with reduced risk of liver cancer: A meta-analysis. BMC Gastroenterol 2013;13:33-46.
- Wu W, et al. Coffee Consumption and bladder cancer: A meta-analysis of observational studies. Sci Rep 2015;5:9051-9059.
- Xhou Q, et al. Coffee consumption and risk of endometrial cancer: A dose response meta-analysis of prospective cohort studies. Sci Rep 2015;5:13410-13422.
- Ding M, et al. Caffeinated and decaffeinated coffee consumption and risk of type 2 diabetes: A systematic review and dose-response meta-analysis. Diabetes Care 2014;37:569-586.
- Wang S, et al. A meta-analysis of coffee intake and risk of urolithiasis. Urol Int 2014;93:220-228.
- Qi H, et al. Dose-response meta-analysis on coffee, tea and caffeine consumption with risk of Parkinson’s disease. Geriatr Gerontol Int 2014;14:430-439.
- Wang L, et al. Coffee and caffeine consumption and depression: A meta-analysis of observational studies. Aust N Z J Psychiatry 2016;50:228-242.
- Liu QP, et al. Habitual coffee consumption and risk of cognitive decline/dementia: A systematic review and meta-analysis of prospective cohort studies. Nutrition 2015;15:doi: 10.1016/j.nut.2015.11.015. [Epub ahead of print]