American Journal of Therapeutics 23, e232–e237 (2016)
The Effect of Coffee and Quantity of Consumption on Specific Cardiovascular and All-Cause Mortality: Coffee Consumption Does Not Affect Mortality Rohit S. Loomba, MD,1* Saurabh Aggarwal, MD,2 and Rohit R. Arora, MD, FACC, FAHA2
Previous studies have examined whether or not an association exists between the consumption of caffeinated coffee to all-cause and cardiovascular mortality. This study aimed to delineate this association using population representative data from the National Health and Nutrition Examination Survey III. Patients were included in the study if all the following criteria were met: (1) follow-up mortality data were available, (2) age of at least 45 years, and (3) reported amount of average coffee consumption. A total of 8608 patients were included, with patients stratified into the following groups of average daily coffee consumption: (1) no coffee consumption, (2) less than 1 cup, (3) 1 cup a day, (4) 2–3 cups, (5) 4–5 cups, (6) more than 6 cups a day. Odds ratios, 95% confidence intervals, and P values were calculated for univariate analysis to compare the prevalence of all-cause mortality, ischemia-related mortality, congestive heart failure–related mortality, and stroke-related mortality, using the no coffee consumption group as reference. These were then adjusted for confounding factors for a multivariate analysis. P , 0.05 were considered statistically significant. Univariate analysis demonstrated an association between coffee consumption and mortality, although this became insignificant on multivariate analysis. Coffee consumption, thus, does not seem to impact all-cause mortality or specific cardiovascular mortality. These findings do differ from those of recently published studies. Coffee consumption of any quantity seems to be safe without any increased mortality risk. There may be some protective effects but additional data are needed to further delineate this. Keywords: coffee, caffeine, mortality, myocardial infarction, stroke, congestive heart failure
INTRODUCTION Coffee is increasingly consumed in the United States with nearly 70% of American adults reporting daily intake of coffee.1 With the frequency of coffee consumption being so high, it is important to characterize the effects of coffee
1
Department of Cardiology, Children’s Hospital of Wisconsin/ Medical College of Wisconsin Affiliated Hospitals, Wauwatosa, WI; and 2Department of Cardiology, Chicago Medical School, North Chicago, IL. The authors have no conflicts of interest to declare. *Address for correspondence: Children’s Hospital of Wisconsin/ Medical College of Wisconsin Affiliated Hospitals, 9000 Wisconsin Avenue, Wauwatosa, WI 53226. E-mail: [email protected]
consumption on various health factors and mortality risk. Previous studies have studied the effect of coffee consumption on various cardiovascular outcomes and mortality associated with these specific outcomes. The purpose of this study was to use data from the nationally representative National Health and Nutrition Examination Survey (NHANES) to help characterize the risk of specific cardiovascular mortality and all-cause mortality based on quantity of coffee consumption.
METHODS The NHANES is a national effort aimed at assessing the health and nutrition status of children and adults in the United States, with the patient population being representative of the US population. The NHANES
1075–2765 Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
www.americantherapeutics.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Coffee Consumption Does Not Affect Mortality
database consists of demographic information and information drawn from questionnaires, physical examination, and laboratory testing. There are several iterations of NHANES, which have been assembled during various years. For this study, data from NHANES III was used, which consists of data regarding 33,994 patients collected from 1988 to 1994. This specific iteration of NHANES was selected, as it was the most recent data set that also included linked mortality data. Mortality data were collected on patients included in NHANES III through December 31, 2006 and includes mortality status and cause of mortality among other parameters. Specifics regarding the design and implementation of the questionnaires used to collect the data in NHANES can be found in previous publications that have described the logistical aspects of the NHANES database. The NHANES database was reviewed and approved from the institutional review board and did obtain written consent from participants. Patients older than the 45 years were identified in the NHANES III database. Only patients with a response to questionnaire items regarding caffeinated coffee consumption were eligible for inclusion. Those with no mortality data available were excluded. Included patients were stratified into the following groups of average daily caffeinated coffee consumption; (1) no coffee consumption, (2) less than 1 cup, (3) 1 cup a day, (4) 2–3 cups, (5) 4–5 cups, (6) 6 or more cups a day. Baseline characteristics were then compared for included patients between the various coffee consumption groups using x2 analysis, t tests, and Mann–Whitney U tests as appropriate. Next, odds ratios with 95% confidence intervals (CI) and P values were calculated to compare the prevalence of all-cause mortality, ischemia-related mortality, congestive heart failure– related mortality, and stroke-related mortality. After this univariate analysis, a multivariate analysis was done to adjust for the following covariates: race, gender, total cholesterol, low-density lipoprotein levels, and highdensity lipoprotein levels. These specific covariates were selected as per the Framingham criteria. The group reporting no daily caffeinated coffee intake a day was used as the reference group for the univariate and multivariate analysis. All statistical analysis was performed using SPSS statistical software version 20.0 (SPSS, Chicago, IL). P , 0.05 was considered statistically significant. A power analysis was also conducted.
RESULTS In this study, 8608 patients were ultimately included to analyze the effect of daily coffee intake on mortality. www.americantherapeutics.com
e233
There were 2602 patients who reported no average coffee intake a day, 845 who reported intake of less than 1 cup a day, 3266 who reported intake of 1 cup a day, 1391 who reported intake of 2 or 3 cups a day, 290 who reported intake of 4 or 5 cups a day, and 214 who reported intake of 6 or more cups a day. Baseline characteristics demonstrated statistically significant differences in age between the groups and the length of follow-up in the groups. No other statistically significant differences were noted between the different free water intake groups (Table 1). Univariate analysis for all-cause mortality demonstrated a significant decrease in mortality risk for those who reported coffee intake of less than 1 cup a day (odds ratio, 0.698; 95% CI, 0.597–0.816; P , 0.001), 2 or 3 cups of coffee a day (odds ratio, 0.799; 95% CI, 0.701– 0.910; P 5 0.001), 4 or 5 cups a day (odds ratio, 0.578; 95% CI, 0.450–0.741; P , 0.001), and 6 or more cups a day (odds ratio, 0.590; 95% CI, 0.444–0.786; P , 0.001). None of these associations remained significant after multivariate analysis (Table 2). Univariate analysis for cardiac ischemia-related mortality demonstrated a significant decrease in mortality risk for those who reported coffee intake of less than 1 cup a day (odds ratio, 0.735; 95% CI, 0.577– 0.935; P 5 0.012), 2 or 3 cups of coffee a day (odds ratio, 0.799; 95% CI, 0.657–0.972; P 5 0.025), and 6 or more cups a day (odds ratio, 0.513; 95% CI, 0.309– 0.851; P 5 0.010). After multivariate analysis only those with an intake of 1 cup of coffee a day continued to demonstrate a decrease in mortality risk (odds ratio, 0.606; 95% CI, 0.399–0.918; P 5 0.018) (Table 3). Univariate and multivariate analysis for congestive heart failure–related mortality demonstrated no significant association between coffee consumption and mortality risk (Table 4). Univariate analysis for stroke-related mortality demonstrated a significant decrease in mortality risk for those who reported coffee intake of less than 1 cup a day (odds ratio, 0.633; 95% CI, 0.404–0.988; P 5 0.044) and 4 or 5 cups a day (odds ratio, 0.302; 95% CI, 0.111–0.826; P5 0.020). None of these associations remained significant after multivariate analysis (Table 5). A priori power analysis was conducted to determine the power of analysis between the defined groups and the reference group for multivariate analysis. An anticipated effect size of 0.15 was used with a desired statistical power level of 90% and an alpha of 0.05. Power analysis was done based on multivariate analysis consisting of 5 confounders to be potentially adjusted for. The minimum required sample size to meet these requirements was found to be 116 and in this study, the smallest group (those who reported a daily average of 6 or more cups of coffee a day) was 214. American Journal of Therapeutics (2016) 23(1)
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Loomba et al
e234 Table 1. Baseline characteristics between coffee consumption groups. 0 cups a day
Less than 1 cup a day
1 cup a day
2–3 cups a day
4–5 cups a day
6 or more cups a day
2439
845
3266
1391
290
214
N
P
Age (months)
820.38 6 430.68 778.32 6 350.39 808.12 6 314.08 787.23 6 143.64 741.23 6 131.29 716.41 6 125.33
Total cholesterol
219.47 6 45.94
218.61 6 45.11
220.93 6 45.58
222.15 6 43.25
221.46 6 41.20
216.60 6 41.52
,0.001 0.273
LDL HDL
135.74 6 38.35 51.14 6 16.35
140.83 6 41.98 51.17 6 16.18
139.78 6 39.62 51.29 6 16.54
138.08 6 37.67 51.61 6 16.83
143.30 6 41.26 51.87 6 16.53
140.18 6 33.04 48.80 6 14.59
0.086 0.338
Triglycerides
166.89 6 122.20 164.27 6 125.04 165.13 6 134.26 158.11 6 106.49 152.15 6 106.94 160.48 6 119.12
0.202
27.31 6 5.67
27.40 6 5.36
27.35 6 5.20
27.07 6 5.25
26.70 6 6.28
27.45 6 5.40
0.757
Systolic blood pressure
129.49 6 40.56
128.95 6 33.88
131.55 6 37.09
128.33 6 35.83
124.94 6 33.02
123.74 6 39.35
0.196
Months of follow-up
132.60 6 61.04
142.02 6 57.60
135.77 6 60.72
140.61 6 58.72
149.29 6 50.71
149.34 6 54.62
, 0.001
BMI
DISCUSSION Coffee is increasingly consumed around the world, particularly in the United States where 83% of Americans questioned reported as drinking coffee in the past year, amounting to more than 400 million cups of coffee being consumed in a single day.1 With such large consumption, it becomes necessary to assess the impacts of coffee on overall health, particularly cardiovascular health. Although much attention is paid to caffeine, it is important to keep in mind that there are several other components to coffee, which can also exert effects. In fact, the actual caffeine content of coffee is quite variable with an 8-ounce cup of coffee consisting of approximately 100–200 mg of caffeine. Even decaffeinated coffee contains some caffeine with an 8-ounce cup consisting of approximately 5–15 mg of caffeine.2,3 With respect to stroke, a contemporary meta-analysis found that coffee may afford a protective effect. This analysis found a relative risk of 0.82 for stroke in those
who consumed 1–3 cups of coffee a day and a relative risk of 0.79 in those who consumed 3–6 cups of coffee a day. There was no statistical difference in the relative risk of stroke in those who reported drinking more than 6 cups of coffee a day.4 Similar findings have been found in other studies as well including the Swedish Mammography Cohort and the Nurses’ Health Study. In the Swedish Mammography Cohort, the relative risk of stroke was 0.78 for those who consumed 1–2 cups of coffee a day, 0.75 for those who consumed 3–4 cups a day, and 0.77 for those who consumed 5 or more cups a day.5 Our study did find risk reduction of mortality from stroke with certain amounts of coffee consumption in univariate analysis, although significance was not maintained with multivariate analysis. With respect to congestive heart failure, a recent meta-analysis of 5 prospective studies was conducted in which a total of 140,220 patients were included. The meta-analysis demonstrated a U-shaped relationship between coffee consumption and congestive heart
Table 2. Results of univariate and multivariate analysis of all-cause mortality. No coffee a day N
2602
Less than 1 cup of coffee a day
One cup of coffee a day
Two or 3 cups of coffee a day
Four or 5 cups of coffee a day
More than 6 cups of coffee a day
845
3266
1391
290
214
Unadjusted odds ratio and 95% CI Unadjusted P
0.698 (0.597–0.816)* 0.951 (0.858–1.055) 0.799 (0.701–0.910)* 0.578 (0.450–0.741)* 0.590 (0.444–0.786)*
Adjusted odds ratio and 95% CI
0.856 (0.549–1.333)
Adjusted P
,0.001*
0.491
0.342
0.001*
0.955 (0.710–1.283) 0.861 (0.555–1.337)
0.758
0.505
,0.001*
,0.001*
0.962 (0.429–2.155)
0.954 (0.344–2.653)
0.924
0.929
*Indicates data are statistically significant.
American Journal of Therapeutics (2016) 23(1)
www.americantherapeutics.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Coffee Consumption Does Not Affect Mortality
e235
Table 3. Results of univariate and multivariate analysis of ischemia-related mortality. No coffee a day N Unadjusted odds ratio and 95% CI
Less than 1 cup of coffee a day
One cup of coffee a day
845
3266
2602
0.735 (0.577–0.935)* 0.934 (0.806–1.083)
Unadjusted P Adjusted odds ratio and 95% CI
0.012* 0.657 (0.342–1.256)
Adjusted P
Two or 3 cups of coffee a day
Four or 5 cups of coffee a day
More than 6 cups of coffee a day
1391
290
214
0.799 (0.657–0.972)* 0.686 (0.462–1.015) 0.513 (0.309–0.851)*
0.369
0.025*
0.060
0.606 (0.399–0.918)* 0.677 (0.874–2.494)
0.206
0.018*
0.145
0.010*
0.919 (0.385–2.189) 0.411 (0.092–1.824)
0.848
0.242
*Indicates data are statistically significant.
failure with a statistically significant reduction in congestive heart failure events in those who consumed 4 cups of coffee a day. Those who consumed less than or more than 4 cups of coffee a day had an increased risk of congestive heart failure.6 This is in contrast to the findings of our study in which no significant association between coffee consumption and mortality from congestive heart failure was noted. As regards coronary heart disease, a meta-analysis of 21 prospective studies from 1966 to 2008 was conducted and included a total of 407,806 patients. This study found that moderate coffee consumption may decrease the risk of coronary heart disease, with a relative risk of 0.87 in those who had reported moderate coffee consumption.7 It is interesting to note that earlier studies from the 1970s often demonstrated a positive correlation between coffee consumption and risk of coronary heart disease.8,9 This was likely because of the lack of adjustment for confounding factors. When adjustments were made for such factors in contemporary studies, coffee consumption and coronary heart disease demonstrate no correlation or a negative correlation.10,11 In our analysis, we noted a decrease in the relative risk of mortality
from ischemia across all coffee consumption in the univariate analysis and in those who reported consuming 1 cup of coffee a day after multivariate analysis. With respect to all-cause mortality, the National Institutes of Health—AARP Diet and Health Study consisting of 402,260 patients, a 10% and 15% risk reduction in all-cause mortality was noted in men and women, respectively, who consumed 6 or more cups of coffee a day.11 In our study, a negative correlation was found between all-cause mortality and increasing coffee consumption but this was no longer statistically significant after multivariate analysis. The means by which coffee exerts systemic effects is unclear although its effects have been well characterized. In those who are not habituated to caffeine, coffee consumption does acutely increase blood pressure. There appears to be no long-term effect on blood pressure in this population and there is no long-term effect noted in those who are habitual coffee consumers as tolerance to the hemodynamic effects of coffee has been demonstrated to develop over time.12–15 This holds true even at greater levels of coffee consumption.16 The findings of our study are consistent with these findings as
Table 4. Results of univariate and multivariate analysis of congestive heart failure–related mortality. No coffee a day N Unadjusted odds ratio and 95% CI
Less than 1 cup of coffee a day
One cup of coffee a day
Two or 3 cups of coffee a day
Four or 5 cups of coffee a day
More than 6 cups of coffee a day
845
3266
1391
290
214
2602
0.936 (0.512–1.712) 1.005 (0.680–1.484) 0.893 (0.535–1.490) 0.386 (0.093–1.597) 0.524 (0.126–2.174)
Unadjusted P Adjusted odds ratio and 95% CI
0.830
0.982
0.665
0.189
0.374
0.277 (0.063–1.212) 1.233 (0.336–4.529) 0.615 (0.128–2.941) 0.518 (0.046–5.848) 0.548 (0.214–1.476)
Adjusted P
www.americantherapeutics.com
0.088
0.753
0.543
0.594
0.754
American Journal of Therapeutics (2016) 23(1)
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Loomba et al
e236 Table 5. Results of univariate and multivariate analysis of stroke-related mortality. No coffee a day N Unadjusted odds ratio and 95% CI Unadjusted P Adjusted odds ratio and 95% CI Adjusted P
2602
Less than cup of coffee a day
One cup of coffee a day
Two or 3 cups of coffee a day
Four or 5 cups of coffee a day
More than 6 cups of coffee a day
845
3266
1391
290
214
0.633 (0.404–0.988)* 0.939 (0.729–1.211) 0.739 (0.522–1.048) 0.302 (0.111–0.826)* 0.624 (0.271–1.435) 0.044* 0.456 (0.101–2.062) 0.308
0.631
0.090
0.020*
1.012 (0.498–2.058) 0.801 (0.347–1.845) 0.568 (0.071–4.516) 0.973
0.602
0.593
0.267 0.618 (0.373–1.214) 0.477
*Indicates data are statistically significant.
no differences were noted in blood pressure among the various coffee consumption groups. Serum lipids can also be affected by coffee. Multiple studies have demonstrated that consuming boiled coffee can increase both total cholesterol and low-density lipoprotein levels, whereas consuming filtered coffee has no impact on serum lipid levels.17–19 The findings of our study demonstrated no difference in serum lipids among the various coffee consumption groups. Whether or not the coffee consumed by those in our study was boiled or filtered, however, is not known. Caffeine may also impact insulin sensitivity and has been demonstrated to improve glucose metabolism, improve insulin sensitivity, and decrease insulin resistance.20–23 These effects may even translate into a decrease in the risk of the development of type 2 diabetes mellitus, even in those with high coffee consumption.24,25 These findings held true for both caffeinated and decaffeinated coffee. In our study, there was no difference noted in glucose levels among coffee consumption groups. Strengths of the current study include the use of a nationally representative sample with a large sample size. The ability to adjust for multiple confounders and perform a robust multivariate analysis also helps strengthen the analysis in this study. Limitations of this study include lack of data regarding the preparation of the coffee consumed and the longevity of coffee consumption at the currently reported level. With the nature of questionnaire-based collection of data, there is always the risk of recall bias, which cannot be accounted for in this study. Additionally, the retrospective nature of this study does not allow for any firm conclusions about causality to be made, rather simply association. Randomized studies with large sample size with controlled coffee preparation and regimented coffee consumption data collection can help shed further insight into the effects of coffee consumption. Such American Journal of Therapeutics (2016) 23(1)
studies can be difficult to conduct, however, and registry-based studies may be the most feasible.
CONCLUSIONS Coffee is consumed in great quantities, which have been increasing over the last few years, increasing the need for better delineation of its systemic effects. This study, based on data from the NHANES, demonstrates no increased risk of ischemia-related mortality, congestive heart failure–related mortality, strokerelated mortality, or all-cause mortality. After multivariate analysis, significant risk reduction was noted in ischemia-related mortality. Thus, coffee consumption appears to be safe among all consumption quantities.
REFERENCES 1. National Coffee Drinking Trends 2012, National Coffee Association. New York, NY: NCA; 2012. 2. McCusker RR, Fuehrlein B, Goldberger BA, et al. Caffeine content of decaffeinated coffee. J Anal Toxicol. 2006;30:611–613. 3. McCusker RR, Goldberger BA, Cone EJ. Caffeine content of specialty coffees. J Anal Toxicol. 2003;27:520–522. 4. D’Elia LCG, Garbagnati F, Scalfi L, et al. Moderate coffee consumption is associated with lower risk of stroke: meta-analysis of prospective studies. Journal of hypertension. 2012;e107. 5. Larsson SC, Virtamo J, Wolk A. Coffee consumption and risk of stroke in women. Stroke. 2011;42:908–912. 6. Mostofsky E, Rice MS, Levitan EB, et al. Habitual coffee consumption and risk of heart failure: a dose-response meta-analysis. Circ Heart Fail. 2012;5:401–405. 7. Shechter M, Shalmon G, Scheinowitz M, et al. Impact of acute caffeine ingestion on endothelial function in subjects with and without coronary artery disease. Am J Cardiol. 2011;107:1255–1261. www.americantherapeutics.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Coffee Consumption Does Not Affect Mortality 8. Coffee drinking and acute myocardial infarction. Report from the Boston Collaborative Drug Surveillance Program. Lancet. 1972;2:1278–1281. 9. Jick H, Miettinen OS, Neff RK, et al. Coffee and myocardial infarction. N Engl Journal Med. 1973;289:63–67. 10. Kleemola P, Jousilahti P, Pietinen P, et al. Coffee consumption and the risk of coronary heart disease and death. Arch Intern Med. 2000;160:3393–3400. 11. Freedman ND, Park Y, Abnet CC, et al. Association of coffee drinking with total and cause-specific mortality. N Engl J Med. 2012;366:1891–1904. 12. Mesas AE, Leon-Munoz LM, Rodriguez-Artalejo F, et al. The effect of coffee on blood pressure and cardiovascular disease in hypertensive individuals: a systematic review and meta-analysis. Am J Clin Nutr. 2011;94: 1113–1126. 13. Robertson D, Wade D, Workman R, et al. Tolerance to the humoral and hemodynamic effects of caffeine in man. J Clin Invest. 1981;67:1111–1117. 14. Steffen M, Kuhle C, Hensrud D, et al. The effect of coffee consumption on blood pressure and the development of hypertension: a systematic review and meta-analysis. J Hyperts. 2012;30:2245–2254. 15. Myers MG. Effects of caffeine on blood pressure. Arch Intern Med. 1988;148:1189–1193. 16. Winkelmayer WC, Willett WC, Cirhan GC. Habitual caffeine intake and the risk of hypertension in women. JAMA. 2005;30:2245–2254. 17. Bak AA, Grobbee DE. The effect on serum cholesterol levels of coffee brewed by filtering or boiling. N Engl J Med. 1989;321:1432–1437.
www.americantherapeutics.com
e237 18. Jee SH, He J, Appel LJ, et al. Coffee consumption and serum lipids: a meta-analysis of randomized controlled clinical trials. Am J Epidemiol. 2001;153:353–362. 19. Lopez-Garcia E, van Dam RM, Willett WC, et al. Coffee consumption and coronary heart disease in men and women: a prospective cohort study. Circulation. 2006; 113:2045–2053. 20. van Dam RM. Coffee and type 2 diabetes: from beans to beta-cells. Nutrition, metabolism, and cardiovascular diseases. Nutr Metab Cardiovasc Dis. 2006;16:69–77. 21. Wedick NM, Brennan AM, Sun Q, et al. Effects of caffeinated and decaffeinated coffee on biological risk factors for type 2 diabetes: a randomized controlled trial. Nutr J. 2011;10:93. 22. Egawa T, Hamada T, Kameda N, et al. Caffeine acutely activates 59adenosine monophosphate-activated protein kinase and increases insulin-independent glucose transport in rat skeletal muscles. Metab Clinical Experimental. 2009;58:1609–1617. 23. Park S, Jang JS, Hong SM. Long-term consumption of caffeine improves glucose homeostasis by enhancing insulinotropic action through islet insulin/insulin-like growth factor 1 signaling in diabetic rats. Metab Clinical Experimental. 2007;56:599–607. 24. Huxley R, Lee CM, Barzi F, et al. Coffee, decaffeinated coffee, and tea consumption in relation to incident type 2 diabetes mellitus: a systematic review with meta-analysis. Arch Intern Med. 2009;169:2053–2063. 25. van Dam RM, Hu FB. Coffee consumption and risk of type 2 diabetes: a systematic review. JAMA. 2005; 294:97–104.
American Journal of Therapeutics (2016) 23(1)
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
The Effect of Coffee and Quantity of Consumption on Specific Cardiovascular and All-Cause Mortality: Coffee Consumption Does Not Affect Mortality Rohit S. Loomba, MD,1* Saurabh Aggarwal, MD,2 and Rohit R. Arora, MD, FACC, FAHA2
Previous studies have examined whether or not an association exists between the consumption of caffeinated coffee to all-cause and cardiovascular mortality. This study aimed to delineate this association using population representative data from the National Health and Nutrition Examination Survey III. Patients were included in the study if all the following criteria were met: (1) follow-up mortality data were available, (2) age of at least 45 years, and (3) reported amount of average coffee consumption. A total of 8608 patients were included, with patients stratified into the following groups of average daily coffee consumption: (1) no coffee consumption, (2) less than 1 cup, (3) 1 cup a day, (4) 2–3 cups, (5) 4–5 cups, (6) more than 6 cups a day. Odds ratios, 95% confidence intervals, and P values were calculated for univariate analysis to compare the prevalence of all-cause mortality, ischemia-related mortality, congestive heart failure–related mortality, and stroke-related mortality, using the no coffee consumption group as reference. These were then adjusted for confounding factors for a multivariate analysis. P , 0.05 were considered statistically significant. Univariate analysis demonstrated an association between coffee consumption and mortality, although this became insignificant on multivariate analysis. Coffee consumption, thus, does not seem to impact all-cause mortality or specific cardiovascular mortality. These findings do differ from those of recently published studies. Coffee consumption of any quantity seems to be safe without any increased mortality risk. There may be some protective effects but additional data are needed to further delineate this. Keywords: coffee, caffeine, mortality, myocardial infarction, stroke, congestive heart failure
INTRODUCTION Coffee is increasingly consumed in the United States with nearly 70% of American adults reporting daily intake of coffee.1 With the frequency of coffee consumption being so high, it is important to characterize the effects of coffee
1
Department of Cardiology, Children’s Hospital of Wisconsin/ Medical College of Wisconsin Affiliated Hospitals, Wauwatosa, WI; and 2Department of Cardiology, Chicago Medical School, North Chicago, IL. The authors have no conflicts of interest to declare. *Address for correspondence: Children’s Hospital of Wisconsin/ Medical College of Wisconsin Affiliated Hospitals, 9000 Wisconsin Avenue, Wauwatosa, WI 53226. E-mail: [email protected]
consumption on various health factors and mortality risk. Previous studies have studied the effect of coffee consumption on various cardiovascular outcomes and mortality associated with these specific outcomes. The purpose of this study was to use data from the nationally representative National Health and Nutrition Examination Survey (NHANES) to help characterize the risk of specific cardiovascular mortality and all-cause mortality based on quantity of coffee consumption.
METHODS The NHANES is a national effort aimed at assessing the health and nutrition status of children and adults in the United States, with the patient population being representative of the US population. The NHANES
1075–2765 Copyright © 2014 Wolters Kluwer Health, Inc. All rights reserved.
www.americantherapeutics.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Coffee Consumption Does Not Affect Mortality
database consists of demographic information and information drawn from questionnaires, physical examination, and laboratory testing. There are several iterations of NHANES, which have been assembled during various years. For this study, data from NHANES III was used, which consists of data regarding 33,994 patients collected from 1988 to 1994. This specific iteration of NHANES was selected, as it was the most recent data set that also included linked mortality data. Mortality data were collected on patients included in NHANES III through December 31, 2006 and includes mortality status and cause of mortality among other parameters. Specifics regarding the design and implementation of the questionnaires used to collect the data in NHANES can be found in previous publications that have described the logistical aspects of the NHANES database. The NHANES database was reviewed and approved from the institutional review board and did obtain written consent from participants. Patients older than the 45 years were identified in the NHANES III database. Only patients with a response to questionnaire items regarding caffeinated coffee consumption were eligible for inclusion. Those with no mortality data available were excluded. Included patients were stratified into the following groups of average daily caffeinated coffee consumption; (1) no coffee consumption, (2) less than 1 cup, (3) 1 cup a day, (4) 2–3 cups, (5) 4–5 cups, (6) 6 or more cups a day. Baseline characteristics were then compared for included patients between the various coffee consumption groups using x2 analysis, t tests, and Mann–Whitney U tests as appropriate. Next, odds ratios with 95% confidence intervals (CI) and P values were calculated to compare the prevalence of all-cause mortality, ischemia-related mortality, congestive heart failure– related mortality, and stroke-related mortality. After this univariate analysis, a multivariate analysis was done to adjust for the following covariates: race, gender, total cholesterol, low-density lipoprotein levels, and highdensity lipoprotein levels. These specific covariates were selected as per the Framingham criteria. The group reporting no daily caffeinated coffee intake a day was used as the reference group for the univariate and multivariate analysis. All statistical analysis was performed using SPSS statistical software version 20.0 (SPSS, Chicago, IL). P , 0.05 was considered statistically significant. A power analysis was also conducted.
RESULTS In this study, 8608 patients were ultimately included to analyze the effect of daily coffee intake on mortality. www.americantherapeutics.com
e233
There were 2602 patients who reported no average coffee intake a day, 845 who reported intake of less than 1 cup a day, 3266 who reported intake of 1 cup a day, 1391 who reported intake of 2 or 3 cups a day, 290 who reported intake of 4 or 5 cups a day, and 214 who reported intake of 6 or more cups a day. Baseline characteristics demonstrated statistically significant differences in age between the groups and the length of follow-up in the groups. No other statistically significant differences were noted between the different free water intake groups (Table 1). Univariate analysis for all-cause mortality demonstrated a significant decrease in mortality risk for those who reported coffee intake of less than 1 cup a day (odds ratio, 0.698; 95% CI, 0.597–0.816; P , 0.001), 2 or 3 cups of coffee a day (odds ratio, 0.799; 95% CI, 0.701– 0.910; P 5 0.001), 4 or 5 cups a day (odds ratio, 0.578; 95% CI, 0.450–0.741; P , 0.001), and 6 or more cups a day (odds ratio, 0.590; 95% CI, 0.444–0.786; P , 0.001). None of these associations remained significant after multivariate analysis (Table 2). Univariate analysis for cardiac ischemia-related mortality demonstrated a significant decrease in mortality risk for those who reported coffee intake of less than 1 cup a day (odds ratio, 0.735; 95% CI, 0.577– 0.935; P 5 0.012), 2 or 3 cups of coffee a day (odds ratio, 0.799; 95% CI, 0.657–0.972; P 5 0.025), and 6 or more cups a day (odds ratio, 0.513; 95% CI, 0.309– 0.851; P 5 0.010). After multivariate analysis only those with an intake of 1 cup of coffee a day continued to demonstrate a decrease in mortality risk (odds ratio, 0.606; 95% CI, 0.399–0.918; P 5 0.018) (Table 3). Univariate and multivariate analysis for congestive heart failure–related mortality demonstrated no significant association between coffee consumption and mortality risk (Table 4). Univariate analysis for stroke-related mortality demonstrated a significant decrease in mortality risk for those who reported coffee intake of less than 1 cup a day (odds ratio, 0.633; 95% CI, 0.404–0.988; P 5 0.044) and 4 or 5 cups a day (odds ratio, 0.302; 95% CI, 0.111–0.826; P5 0.020). None of these associations remained significant after multivariate analysis (Table 5). A priori power analysis was conducted to determine the power of analysis between the defined groups and the reference group for multivariate analysis. An anticipated effect size of 0.15 was used with a desired statistical power level of 90% and an alpha of 0.05. Power analysis was done based on multivariate analysis consisting of 5 confounders to be potentially adjusted for. The minimum required sample size to meet these requirements was found to be 116 and in this study, the smallest group (those who reported a daily average of 6 or more cups of coffee a day) was 214. American Journal of Therapeutics (2016) 23(1)
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Loomba et al
e234 Table 1. Baseline characteristics between coffee consumption groups. 0 cups a day
Less than 1 cup a day
1 cup a day
2–3 cups a day
4–5 cups a day
6 or more cups a day
2439
845
3266
1391
290
214
N
P
Age (months)
820.38 6 430.68 778.32 6 350.39 808.12 6 314.08 787.23 6 143.64 741.23 6 131.29 716.41 6 125.33
Total cholesterol
219.47 6 45.94
218.61 6 45.11
220.93 6 45.58
222.15 6 43.25
221.46 6 41.20
216.60 6 41.52
,0.001 0.273
LDL HDL
135.74 6 38.35 51.14 6 16.35
140.83 6 41.98 51.17 6 16.18
139.78 6 39.62 51.29 6 16.54
138.08 6 37.67 51.61 6 16.83
143.30 6 41.26 51.87 6 16.53
140.18 6 33.04 48.80 6 14.59
0.086 0.338
Triglycerides
166.89 6 122.20 164.27 6 125.04 165.13 6 134.26 158.11 6 106.49 152.15 6 106.94 160.48 6 119.12
0.202
27.31 6 5.67
27.40 6 5.36
27.35 6 5.20
27.07 6 5.25
26.70 6 6.28
27.45 6 5.40
0.757
Systolic blood pressure
129.49 6 40.56
128.95 6 33.88
131.55 6 37.09
128.33 6 35.83
124.94 6 33.02
123.74 6 39.35
0.196
Months of follow-up
132.60 6 61.04
142.02 6 57.60
135.77 6 60.72
140.61 6 58.72
149.29 6 50.71
149.34 6 54.62
, 0.001
BMI
DISCUSSION Coffee is increasingly consumed around the world, particularly in the United States where 83% of Americans questioned reported as drinking coffee in the past year, amounting to more than 400 million cups of coffee being consumed in a single day.1 With such large consumption, it becomes necessary to assess the impacts of coffee on overall health, particularly cardiovascular health. Although much attention is paid to caffeine, it is important to keep in mind that there are several other components to coffee, which can also exert effects. In fact, the actual caffeine content of coffee is quite variable with an 8-ounce cup of coffee consisting of approximately 100–200 mg of caffeine. Even decaffeinated coffee contains some caffeine with an 8-ounce cup consisting of approximately 5–15 mg of caffeine.2,3 With respect to stroke, a contemporary meta-analysis found that coffee may afford a protective effect. This analysis found a relative risk of 0.82 for stroke in those
who consumed 1–3 cups of coffee a day and a relative risk of 0.79 in those who consumed 3–6 cups of coffee a day. There was no statistical difference in the relative risk of stroke in those who reported drinking more than 6 cups of coffee a day.4 Similar findings have been found in other studies as well including the Swedish Mammography Cohort and the Nurses’ Health Study. In the Swedish Mammography Cohort, the relative risk of stroke was 0.78 for those who consumed 1–2 cups of coffee a day, 0.75 for those who consumed 3–4 cups a day, and 0.77 for those who consumed 5 or more cups a day.5 Our study did find risk reduction of mortality from stroke with certain amounts of coffee consumption in univariate analysis, although significance was not maintained with multivariate analysis. With respect to congestive heart failure, a recent meta-analysis of 5 prospective studies was conducted in which a total of 140,220 patients were included. The meta-analysis demonstrated a U-shaped relationship between coffee consumption and congestive heart
Table 2. Results of univariate and multivariate analysis of all-cause mortality. No coffee a day N
2602
Less than 1 cup of coffee a day
One cup of coffee a day
Two or 3 cups of coffee a day
Four or 5 cups of coffee a day
More than 6 cups of coffee a day
845
3266
1391
290
214
Unadjusted odds ratio and 95% CI Unadjusted P
0.698 (0.597–0.816)* 0.951 (0.858–1.055) 0.799 (0.701–0.910)* 0.578 (0.450–0.741)* 0.590 (0.444–0.786)*
Adjusted odds ratio and 95% CI
0.856 (0.549–1.333)
Adjusted P
,0.001*
0.491
0.342
0.001*
0.955 (0.710–1.283) 0.861 (0.555–1.337)
0.758
0.505
,0.001*
,0.001*
0.962 (0.429–2.155)
0.954 (0.344–2.653)
0.924
0.929
*Indicates data are statistically significant.
American Journal of Therapeutics (2016) 23(1)
www.americantherapeutics.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Coffee Consumption Does Not Affect Mortality
e235
Table 3. Results of univariate and multivariate analysis of ischemia-related mortality. No coffee a day N Unadjusted odds ratio and 95% CI
Less than 1 cup of coffee a day
One cup of coffee a day
845
3266
2602
0.735 (0.577–0.935)* 0.934 (0.806–1.083)
Unadjusted P Adjusted odds ratio and 95% CI
0.012* 0.657 (0.342–1.256)
Adjusted P
Two or 3 cups of coffee a day
Four or 5 cups of coffee a day
More than 6 cups of coffee a day
1391
290
214
0.799 (0.657–0.972)* 0.686 (0.462–1.015) 0.513 (0.309–0.851)*
0.369
0.025*
0.060
0.606 (0.399–0.918)* 0.677 (0.874–2.494)
0.206
0.018*
0.145
0.010*
0.919 (0.385–2.189) 0.411 (0.092–1.824)
0.848
0.242
*Indicates data are statistically significant.
failure with a statistically significant reduction in congestive heart failure events in those who consumed 4 cups of coffee a day. Those who consumed less than or more than 4 cups of coffee a day had an increased risk of congestive heart failure.6 This is in contrast to the findings of our study in which no significant association between coffee consumption and mortality from congestive heart failure was noted. As regards coronary heart disease, a meta-analysis of 21 prospective studies from 1966 to 2008 was conducted and included a total of 407,806 patients. This study found that moderate coffee consumption may decrease the risk of coronary heart disease, with a relative risk of 0.87 in those who had reported moderate coffee consumption.7 It is interesting to note that earlier studies from the 1970s often demonstrated a positive correlation between coffee consumption and risk of coronary heart disease.8,9 This was likely because of the lack of adjustment for confounding factors. When adjustments were made for such factors in contemporary studies, coffee consumption and coronary heart disease demonstrate no correlation or a negative correlation.10,11 In our analysis, we noted a decrease in the relative risk of mortality
from ischemia across all coffee consumption in the univariate analysis and in those who reported consuming 1 cup of coffee a day after multivariate analysis. With respect to all-cause mortality, the National Institutes of Health—AARP Diet and Health Study consisting of 402,260 patients, a 10% and 15% risk reduction in all-cause mortality was noted in men and women, respectively, who consumed 6 or more cups of coffee a day.11 In our study, a negative correlation was found between all-cause mortality and increasing coffee consumption but this was no longer statistically significant after multivariate analysis. The means by which coffee exerts systemic effects is unclear although its effects have been well characterized. In those who are not habituated to caffeine, coffee consumption does acutely increase blood pressure. There appears to be no long-term effect on blood pressure in this population and there is no long-term effect noted in those who are habitual coffee consumers as tolerance to the hemodynamic effects of coffee has been demonstrated to develop over time.12–15 This holds true even at greater levels of coffee consumption.16 The findings of our study are consistent with these findings as
Table 4. Results of univariate and multivariate analysis of congestive heart failure–related mortality. No coffee a day N Unadjusted odds ratio and 95% CI
Less than 1 cup of coffee a day
One cup of coffee a day
Two or 3 cups of coffee a day
Four or 5 cups of coffee a day
More than 6 cups of coffee a day
845
3266
1391
290
214
2602
0.936 (0.512–1.712) 1.005 (0.680–1.484) 0.893 (0.535–1.490) 0.386 (0.093–1.597) 0.524 (0.126–2.174)
Unadjusted P Adjusted odds ratio and 95% CI
0.830
0.982
0.665
0.189
0.374
0.277 (0.063–1.212) 1.233 (0.336–4.529) 0.615 (0.128–2.941) 0.518 (0.046–5.848) 0.548 (0.214–1.476)
Adjusted P
www.americantherapeutics.com
0.088
0.753
0.543
0.594
0.754
American Journal of Therapeutics (2016) 23(1)
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Loomba et al
e236 Table 5. Results of univariate and multivariate analysis of stroke-related mortality. No coffee a day N Unadjusted odds ratio and 95% CI Unadjusted P Adjusted odds ratio and 95% CI Adjusted P
2602
Less than cup of coffee a day
One cup of coffee a day
Two or 3 cups of coffee a day
Four or 5 cups of coffee a day
More than 6 cups of coffee a day
845
3266
1391
290
214
0.633 (0.404–0.988)* 0.939 (0.729–1.211) 0.739 (0.522–1.048) 0.302 (0.111–0.826)* 0.624 (0.271–1.435) 0.044* 0.456 (0.101–2.062) 0.308
0.631
0.090
0.020*
1.012 (0.498–2.058) 0.801 (0.347–1.845) 0.568 (0.071–4.516) 0.973
0.602
0.593
0.267 0.618 (0.373–1.214) 0.477
*Indicates data are statistically significant.
no differences were noted in blood pressure among the various coffee consumption groups. Serum lipids can also be affected by coffee. Multiple studies have demonstrated that consuming boiled coffee can increase both total cholesterol and low-density lipoprotein levels, whereas consuming filtered coffee has no impact on serum lipid levels.17–19 The findings of our study demonstrated no difference in serum lipids among the various coffee consumption groups. Whether or not the coffee consumed by those in our study was boiled or filtered, however, is not known. Caffeine may also impact insulin sensitivity and has been demonstrated to improve glucose metabolism, improve insulin sensitivity, and decrease insulin resistance.20–23 These effects may even translate into a decrease in the risk of the development of type 2 diabetes mellitus, even in those with high coffee consumption.24,25 These findings held true for both caffeinated and decaffeinated coffee. In our study, there was no difference noted in glucose levels among coffee consumption groups. Strengths of the current study include the use of a nationally representative sample with a large sample size. The ability to adjust for multiple confounders and perform a robust multivariate analysis also helps strengthen the analysis in this study. Limitations of this study include lack of data regarding the preparation of the coffee consumed and the longevity of coffee consumption at the currently reported level. With the nature of questionnaire-based collection of data, there is always the risk of recall bias, which cannot be accounted for in this study. Additionally, the retrospective nature of this study does not allow for any firm conclusions about causality to be made, rather simply association. Randomized studies with large sample size with controlled coffee preparation and regimented coffee consumption data collection can help shed further insight into the effects of coffee consumption. Such American Journal of Therapeutics (2016) 23(1)
studies can be difficult to conduct, however, and registry-based studies may be the most feasible.
CONCLUSIONS Coffee is consumed in great quantities, which have been increasing over the last few years, increasing the need for better delineation of its systemic effects. This study, based on data from the NHANES, demonstrates no increased risk of ischemia-related mortality, congestive heart failure–related mortality, strokerelated mortality, or all-cause mortality. After multivariate analysis, significant risk reduction was noted in ischemia-related mortality. Thus, coffee consumption appears to be safe among all consumption quantities.
REFERENCES 1. National Coffee Drinking Trends 2012, National Coffee Association. New York, NY: NCA; 2012. 2. McCusker RR, Fuehrlein B, Goldberger BA, et al. Caffeine content of decaffeinated coffee. J Anal Toxicol. 2006;30:611–613. 3. McCusker RR, Goldberger BA, Cone EJ. Caffeine content of specialty coffees. J Anal Toxicol. 2003;27:520–522. 4. D’Elia LCG, Garbagnati F, Scalfi L, et al. Moderate coffee consumption is associated with lower risk of stroke: meta-analysis of prospective studies. Journal of hypertension. 2012;e107. 5. Larsson SC, Virtamo J, Wolk A. Coffee consumption and risk of stroke in women. Stroke. 2011;42:908–912. 6. Mostofsky E, Rice MS, Levitan EB, et al. Habitual coffee consumption and risk of heart failure: a dose-response meta-analysis. Circ Heart Fail. 2012;5:401–405. 7. Shechter M, Shalmon G, Scheinowitz M, et al. Impact of acute caffeine ingestion on endothelial function in subjects with and without coronary artery disease. Am J Cardiol. 2011;107:1255–1261. www.americantherapeutics.com
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Coffee Consumption Does Not Affect Mortality 8. Coffee drinking and acute myocardial infarction. Report from the Boston Collaborative Drug Surveillance Program. Lancet. 1972;2:1278–1281. 9. Jick H, Miettinen OS, Neff RK, et al. Coffee and myocardial infarction. N Engl Journal Med. 1973;289:63–67. 10. Kleemola P, Jousilahti P, Pietinen P, et al. Coffee consumption and the risk of coronary heart disease and death. Arch Intern Med. 2000;160:3393–3400. 11. Freedman ND, Park Y, Abnet CC, et al. Association of coffee drinking with total and cause-specific mortality. N Engl J Med. 2012;366:1891–1904. 12. Mesas AE, Leon-Munoz LM, Rodriguez-Artalejo F, et al. The effect of coffee on blood pressure and cardiovascular disease in hypertensive individuals: a systematic review and meta-analysis. Am J Clin Nutr. 2011;94: 1113–1126. 13. Robertson D, Wade D, Workman R, et al. Tolerance to the humoral and hemodynamic effects of caffeine in man. J Clin Invest. 1981;67:1111–1117. 14. Steffen M, Kuhle C, Hensrud D, et al. The effect of coffee consumption on blood pressure and the development of hypertension: a systematic review and meta-analysis. J Hyperts. 2012;30:2245–2254. 15. Myers MG. Effects of caffeine on blood pressure. Arch Intern Med. 1988;148:1189–1193. 16. Winkelmayer WC, Willett WC, Cirhan GC. Habitual caffeine intake and the risk of hypertension in women. JAMA. 2005;30:2245–2254. 17. Bak AA, Grobbee DE. The effect on serum cholesterol levels of coffee brewed by filtering or boiling. N Engl J Med. 1989;321:1432–1437.
www.americantherapeutics.com
e237 18. Jee SH, He J, Appel LJ, et al. Coffee consumption and serum lipids: a meta-analysis of randomized controlled clinical trials. Am J Epidemiol. 2001;153:353–362. 19. Lopez-Garcia E, van Dam RM, Willett WC, et al. Coffee consumption and coronary heart disease in men and women: a prospective cohort study. Circulation. 2006; 113:2045–2053. 20. van Dam RM. Coffee and type 2 diabetes: from beans to beta-cells. Nutrition, metabolism, and cardiovascular diseases. Nutr Metab Cardiovasc Dis. 2006;16:69–77. 21. Wedick NM, Brennan AM, Sun Q, et al. Effects of caffeinated and decaffeinated coffee on biological risk factors for type 2 diabetes: a randomized controlled trial. Nutr J. 2011;10:93. 22. Egawa T, Hamada T, Kameda N, et al. Caffeine acutely activates 59adenosine monophosphate-activated protein kinase and increases insulin-independent glucose transport in rat skeletal muscles. Metab Clinical Experimental. 2009;58:1609–1617. 23. Park S, Jang JS, Hong SM. Long-term consumption of caffeine improves glucose homeostasis by enhancing insulinotropic action through islet insulin/insulin-like growth factor 1 signaling in diabetic rats. Metab Clinical Experimental. 2007;56:599–607. 24. Huxley R, Lee CM, Barzi F, et al. Coffee, decaffeinated coffee, and tea consumption in relation to incident type 2 diabetes mellitus: a systematic review with meta-analysis. Arch Intern Med. 2009;169:2053–2063. 25. van Dam RM, Hu FB. Coffee consumption and risk of type 2 diabetes: a systematic review. JAMA. 2005; 294:97–104.
American Journal of Therapeutics (2016) 23(1)
Copyright © 2015 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
