Coffee and human health

By Dr John Stanley, lecturer in biochemistry at Trinity College, Oxford.

Coffee is one of the most widely consumed beverages in the world and caffeine is the most widely consumed pharmacologically active compound. It is therefore entirely appropriate that attempts should be made to assess the effects of coffee and caffeine on human health. However, as discussed in the next section these attempts are hampered by the weaknesses of the scientific approaches available.

The health effects of coffee should not be equated with those of caffeine. Coffee contains hundreds of other compounds e.g. the antioxidant chlorogenic acid which may reinforce or counteract the effects of caffeine on human health and some examples of such interactions have emerged. Hence effects of caffeine supplements and caffeine in coffee may not be the same. In addition, the effects of compounds like chlorogenic acid on human health remain to be determined.

This article reviews the most recent scientific evidence on effects of coffee consumption on the risk of developing the major diseases affecting the United Kingdom population including cardiovascular disease, cancer and type 2 diabetes. It also considers the evidence for the reputation of coffee as a diuretic.        

The nature of the scientific evidence
Evidence for a relationship between diet and health in general and coffee or caffeine and health in particular comes from two kinds of scientific study. Observational epidemiological studies look for associations between coffee and health whereas randomised controlled trials (RCTs) look for cause effect relationships. Both types of study have their strengths and weaknesses and these need to be born in mind when interpreting the data they generate.

There are several types of observational epidemiological study. The ecological study has the weakest design as coffee intake and disease risk are measured in different individuals because the two kinds of data are usually derived from two databases collected for very different purposes. The cross-sectional study is a snapshot with the disadvantage that when associations are found it is not always clear which is the cause and which is the effect. The case control study in which disease cases are matched with controls derived either from the same hospital or from the same community suffers from two types of bias. Recall bias arises because the study subjects are asked to recall their coffee drinking habits retrospectively, sometimes from years ago, whereas selection bias arises through an inappropriate choice of controls. Community based controls are preferable to hospital based controls. The cohort study has the strongest design as large numbers of healthy subjects are recruited, their coffee consumption assessed prospectively and the study population followed for lengthy periods of time to allow disease to develop. Data from cohort studies are usually corrected for confounders although residual confounding remains a problem.

RCTs compare the effects of coffee with the effects of a placebo. Control of coffee intake is essential and tests of compliance are vitally important to ensure this. Unlike cohort studies relatively small numbers of subjects are followed for relatively short periods of time and the endpoints are almost always disease risk factors rather than disease itself.

The most reliable data are generated by the cohort studies and the RCTs. Data from different cohort studies or different RCTs, carried out at different times and in different places, can be combined and subjected to a meta-analysis. This approach increases statistical power and hence the ability to detect associations or cause effect relationships. The results of such meta-analyses are likely to generate the most reliable data on the relationship between coffee consumption and health and will be cited in this article whenever they exist.  

Cardiovascular Disease (CVD)
Epidemiological studies have looked at associations between the consumption of coffee and the risk of developing CVD. A meta-analysis of 21 independent prospective cohort studies on the relationship between coffee consumption and CVD risk which together comprised a study population of 407,806 including 15,599 cases of CVD was published recently1. This analysis showed that moderate (1-3 or 3-4 cups/d) coffee consumption significantly lowered CVD risk in men and women by comparison with light (<1 cup/d) coffee drinking. Two cohort studies have been published since this meta-analysis and consequently were not included in it. These two studies confirmed the conclusions of the meta-analysis and showed that habitual coffee consumption was not associated with increased risk of CVD in diabetic men2 or diabetic women3. There are too few studies of associations between risk of stroke and coffee consumption for a meta-analysis. However, it has recently been shown that coffee consumption is inversely and significantly associated with a reduced risk of stroke in men4 and women5. The possibility that coffee consumption might protect against stroke deserves further attention.

While coffee consumption is not associated with CVD risk in prospective cohort studies, some RCTs have found effects of coffee on CVD risk factors. These risk factors include the plasma low density lipoprotein (LDL) cholesterol level, the blood homocysteine level and blood pressure. However, it should be born in mind that the contribution of a single risk factor to the total risk of developing CVD can be quite small and is often unknown.

A meta-analysis of 18 RCTs showed that drinking six cups of coffee per day significantly increased LDL-cholesterol levels.6 However, this increase was greater in response to boiled coffee and disappeared when the coffee was filtered. The effect of coffee on LDL-cholesterol is due to two diterpenes, cafestol and kahweol, which are released by boiling and removed by filtering. Although consumption of boiled coffee in the United Kingdom is uncommon this effect of coffee can be avoided if boiled coffee is filtered.     

It has been shown that caffeine and filtered coffee increase plasma homocysteine levels and that both caffeine and chlorogenic acid contribute to this effect7. Although epidemiological studies have shown that high plasma homocysteine levels are associated with a higher risk of developing CVD, RCTs which lower homocysteine levels have been unable to demonstrate any effect on cardiovascular health8. Accordingly a high plasma homocysteine level may be an effect of CVD rather than a cause of it.

Caffeine increases blood pressure in people who do not habitually consume it but tolerance develops within several days9. A meta-analysis of 16 RCTs, with at least seven days of intervention, showed that effects of caffeine in coffee on systolic and diastolic blood pressure are four to five times smaller than effects of caffeine supplements10. This suggests that other components in coffee partially counteract the effects of caffeine on blood pressure. When studies of caffeine in coffee were considered separately, systolic blood pressure increased by 1.2 mmHg and diastolic blood pressure by 0.5mmHg. This is an effect on blood pressure comparable to walking upstairs.

The evidence that coffee causes cardiac arrhythmias is "anecdotal and tenuous"11 and at least four RCTs have failed to find any effect. The most recently published of these showed that there were no effects of caffeine restriction on either visual analogue palpitation scores or ventricular premature beats in a study population at high risk of developing cardiac arrhythmias11.

Early studies of the relationship between coffee consumption and the risk of developing cancer are a striking example of the way that case control studies and cohort studies can give different answers. A large case control study published in 1981 showed an association between coffee consumption and the risk of developing pancreatic cancer12 but these findings have not been confirmed by subsequent cohort studies13. There is a particular reason for distrusting the results of case control studies on pancreatic cancer as poor survival leads to reduced participation rates by cases in interviews and consequently more interviews with surrogates13. By contrast, a meta-analysis of 12 case control studies on coffee consumption and risk of developing colorectal cancer found an inverse association i.e. protective effect14 but a meta-analysis of 12 cohort studies found no association15. Given the biases inherent in case control studies (see above) it would be wiser to accept the results of the cohort studies for cancer at both these sites.

However, liver cancer may be another matter. In a meta-analysis of four cohort studies and five case control studies an increase in coffee consumption of two cups per day was associated with a 31% fall in risk of developing liver cancer in people with no history of liver disease and a 44% fall in risk in people with a history of disease16. In a second meta-analysis of the same data plus a sixth case control study it was found that an increase in coffee consumption of one cup per day was associated with a 23% reduction in risk of developing liver cancer17. Crucially, all ten studies analysed, individually reported this inverse association. This consistency of data across study type argues in favour of a protective effect of coffee against developing liver cancer.  

Alzheimer's and Parkinson's disease
The association between coffee consumption and the risk of developing Alzheimer's disease has been examined in two case control studies and two cohort studies.18 Although coffee consumption was associated with a 27% reduction in the risk of developing Alzheimer's disease the four studies did not comprise a homogeneous group so it was not appropriate to combine their results. A number of cohort studies have examined associations between coffee consumption and the risk of developing Parkinson's disease and again an inverse association has been observed, although women appear worse protected than men 19. More cohort studies are needed to assess the impact of coffee consumption on these two diseases.

Type 2 diabetes
The association of coffee consumption with the risk of developing type 2 diabetes has been examined in 17 prospective cohort studies and all but three have found an inverse association20. The consistency of this association across different study populations supports the idea that the consumption of coffee protects against the development of type 2 diabetes. In a systematic review of nine of these cohort studies, it was concluded that consuming between 6 and 7 cups of coffee a day reduced the risk of developing type 2 diabetes by 35% while consuming between 4 and 6 cups reduced the risk by 28%.21

In six of these cohort studies associations with the consumption of decaffeinated coffee were studied and four of them found a significantly lower risk of developing type 2 diabetes20. This suggests that components of coffee other than caffeine are responsible for the protective effect. While these components remain unidentified, the Maillard reactions involved in roasting coffee are known to transform chlorogenic acid into quinides which are compounds known to affect blood glucose levels22.   

Caffeine as a diuretic
It is not hard to understand where the reputation of caffeine as a diuretic originally came from although it is hard to understand why this reputation persists in the light of recent scientific evidence23. Two individual RCTs can be cited to illustrate this point. In a study often cited in medical textbooks in support of the idea that caffeine containing beverages promote dehydration, the effects of a single 250 mg dose of caffeine on 3-hour urine volume were compared with a placebo24. Urinary volume increased in response to caffeine. However, when the effects of a single dose of either 114 mg or 253 mg caffeine on 24-hour urine volume were compared with water urinary volume was unchanged25. Hence caffeine has no effect on 24-hour urine production but rather shifts the pattern of urine production to earlier in the day. This is hardly evidence for a diuretic action of caffeine over and above that of water.

There is no evidence for effects of coffee consumption on the risk of developing cardiovascular disease although the practice of boiling coffee without filtering it afterwards would be unwise. Effects of coffee consumption on liver cancer are intriguing although a lot more research is needed before a protective effect can be accepted. There is little evidence for protective or detrimental effects of coffee on the risk of cancer at other sites. Research on effects of coffee consumption on Alzeimer's disease and Parkinson's disease is at a very early stage but clearly both deserve further attention. There is a consistency about the evidence for a protective effect of coffee consumption on the risk of developing type 2 diabetes but a convincing mechanism has yet to emerge and this is where future research effort should be concentrated. There is little evidence for the idea that caffeine is a diuretic and this widely held belief should be abandoned.

However, it should be remembered that effects of coffee consumption are dependent on how much coffee is consumed. The evidence discussed above is consistent with the idea that a moderate consumption of 4-5 cups of coffee per day is perfectly safe. However, pregnant women should moderate their intake of caffeine to 200mg per day as advised by the Food Standards Agency.

For further information please visit


1. Wu J-n, Ho SC, Zhou C, Ling W-h, Chen W-q, Wang C-l, Chen Y-m. Coffee consumption and risk of coronary heart diseases: A meta-analysis of 21 prospective cohort studies. Int. J. Cardiol. 2008; 1-10.

2. Zhang W, Lopez-Garcia E, Li TY, Hu FB, van Dam RM. Coffee consumption and risk of cardiovascular diseases and all-cause mortality among men with type 2 diabetes. Diabetes Care 2009; epub.

3. Zhang WL, Lopez-Garcia E, Li TY, Hu FB, van Dam RM. Coffee consumption and risk of cardiovascular events and all-cause mortality among women with type 2 diabetes. Diabetologia 2009; 52: 810-817.

4. Larsson SC, Mannisto S, Virtanen MJ, Kontto J, Albanes D, Virtamo J. Coffee and tea consumption and risk of stroke subtypes in male smokers. Stroke 2008; 39: 1681-1687.

5. Lopez-Garcia E, Rodriguez-Artalejo F, Rexrode KM, Logroscino G, Hu FB, van Dam RM. Coffee consumption and risk of stroke in women. Circulation 2009; 119: 1116-1123.

6. Jee SH, He J, Appel LJ, Whelton PK, Suh I, Klag MJ. Coffee consumption and serum lipids: a meta-analysis of randomized controlled trials. Am. J. Epidemiol. 2001; 153: 353-362.

7. Verhoef P, Pasman WJ, Van Vliet T, Urgert R, Katan MB. Contribution of caffeine to the homocysteine raising effect of coffee: a randomized controlled trial in humans. Am. J. Clin. Nutr. 2002; 76: 1244-1248.

8. Wierzbicki AS. Homocysteine and cardiovascular disease: a review of the evidence. Diab. Vasc. Dis. Res. 2007; 4: 143-150.

9. Robertson D, Wade D, Workman R, Woosley RL, Oates JA. Tolerance to the humoral and hemodynamic effects of caffeine in man. J. Clin. Invest. 1981; 67: 1111-1117.

10. Noordzij M, Uiterwaal CS, Arends LR, Kok FJ, Grobbee DE, Geleijnse JM. Blood pressure response to chronic intake of coffee and caffeine: a meta-analysis of randomized controlled trials. J. Hypertens. 2005; 23: 921-928.

11. Newby DE, Neilson JMM, Jarvie DR, Boon NA. Caffeine restriction has no role in the management of patients with symptomatic idiopathic ventricular premature beats. Heart 1996; 76: 355-357.

12. MacMahon, B, Yen S, Trichopoulos D, Warren K, Nardi G. Coffee and cancer of the pancreas.  N. Engl. J. Med. 1981; 304: 630-633.

13. Michaud DS, Giovannucci E, Willett WC, Colditz GA, Fuchs CS. Coffee and alcohol consumption and the risk of pancreatic cancer in two prospective United States cohorts. Cancer Epidemiol. Biomarkers Prev. 2001; 10:429-437.

14. Giovannucci E. Meta-analysis of coffee consumption and risk of colorectal cancer. Am. J. Epidemiol. 1998; 147: 1043-1052.

15. Je Y, Liu W, Giovannucci E. Coffee consumption and risk of colorectal cancer: a systematic review and meta-analysis of prospective cohort studies. Int. J. Cancer. 2009; 124: 1662-1668.

16. Larsson SC & Wolk A. Coffee consumption and risk of liver cancer: a meta-analysis. Gastroenterology 2007; 132: 1740-1745.

17. Bravi F, Bosetti C, Tavani A, Bagnardi V, Gallus S, Negri E, Franceschi S, La Vecchia C. Coffee drinking and hepatocellular carcinoma risk: a meta-analysis. Hepatology 2007; 46: 430-435.

18. Barranco Quintana JL, Allam MF, Serrano Del Castillo A, Fernandez-Crehuet Navaja R. Alzheimer's disease and coffee: a quantitative review. Neurol. Res. 2007; 29: 91-95.

19. Hu G, Bidel S, Jousilahti P, Antikainen R, Tuomilehto J. Coffee and tea consumption and the risk of Parkinson's disease. Movement Disorders 2007; 22: 2242-2248.

20. van Dam RM. Coffee consumption and risk of type 2 diabetes, cardiovascular disease, and cancer. Appl. Physiol. Nutr. Metab. 2008; 33: 1269-1283.

21. van Dam RM & Hu FB. Coffee consumption and risk of type 2 diabetes. JAMA 2005; 294:97-104.

22. Tunnicliffe JM & Shearer J. Coffee, glucose homeostasis, and insulin resistance: physiological mechanisms and mediators. Appl. Physiol. Nutr. Metab. 2008; 1290-1300.

23. Ganio, MS, Casa DJ, Armstrong LE, Maresh CM. Evidence-based approach to lingering hydration questions. Clinics in Sports Medicine 2007; 26: 1-16.

24. Robertson D, Frolich JC, Carr RK (1978) Effects of caffeine on plasma rennin activity, catecholamines and blood pressure. N. Engl. J. Med. 1978; 298: 181-186.

25. Grandjean AC, Reimers KJ, Bannick KE. The effect of caffeinated, non-caffeinated, caloric and non-caloric beverages on hydration. J. Am. Coll. Nutr. 2000; 19: 591-600.

Have you registered with us yet?

Register now to enjoy more articles and free email bulletins


Already registered?

Sign in

Before commenting please read our rules for commenting on articles.

If you see a comment you find offensive, you can flag it as inappropriate. In the top right-hand corner of an individual comment, you will see 'flag as inappropriate'. Clicking this prompts us to review the comment. For further information see our rules for commenting on articles.

comments powered by Disqus