Skip navigation

Caffeine and health

Drinking coffee, or taking in caffeine from tea or cola, or wherever, always seems to be popping up in the newspapers with one message or another, sometimes with conflicting results. So it is comforting that epidemiologists and others are beginning to pull together information so that we can make sense of it. Three recent studies, two of which are meta-analyses, helps us understand the proper place of caffeine, and when to avoid it. It may help formulate better advice for people as part of evidence-based guidance for healthy living.

In interpreting what follows, where data are often analysed in terms of caffeine consumption in milligrams per day, one cup of coffee is equivalent to about 75 mg caffeine, one cup of tea to about 30 mg and one can of cola about 50 mg. But beware espresso or café solo - there the caffeine content can be double at 150 mg a cup!

Caffeine in pregnancy

A meta-analysis examined the effects of caffeine consumption on spontaneous abortion and low birthweight pregnancies [1]. Spontaneous abortion was defined as expulsion from the uterus of products of conception before about 20 weeks (including foetal loss, foetal death and miscarriage). Low birthweight was defined as less than 2,500 g. The control subjects were women who consumed less than 150 mg caffeine a day (two cups of coffee or less), and exposed women those who consumed more than this. Searching was thorough.


In 42,889 women the rate of spontaneous abortion was 24.4% in exposed women, and 20.0% in controls, with consistency between studies (Figure 1).

The number needed to harm we calculated as 23 (Table 1).

Caffeine intake, spontaneous abortion, and low birthweight infants
  Number (%) of women affected    
  Caffeine >150 mg/day Caffeine <150 mg/day Relative risk (95%CI) NNH (95% CI)
Spontaneous abortion 1994/8181 (24.4) 6941/34708 (20.0) 1.27 (1.22 to 1.33) 23 (19 to 30)
Low birthweight infant 713/9308 (7.7) 3033/55383 (5.5) 1.50 (1.39 to 1.62) 46 (36 to 62)
This means that for every 23 pregnant women who consume more than two cups of coffee or six cups of tea a day, one will have a spontaneous abortion who would not have had they not consumed this much caffeine.

In 64,691 women the rate of low birthweight infants was 7.7% in exposed women, and 5.5% in controls, with consistency between studies (Figure 2).

The number needed to harm we calculated as 46 (Table 1). This means that for every 46 pregnant women who consume more than two cups of coffee or six cups of tea a day, one will have a baby weighing less than 2500 grams who would not have had they not consumed this much caffeine.


This is a useful and thoughtful review, which explains, for instance, how each study deals with confounding issues like maternal age and smoking. That does not exclude unknown confounding factors, but an association between maternal caffeine intake and spontaneous abortion or low birthweight is more likely to be true than not. The review tells us that caffeine is cleared from the body much less rapidly in the second and third trimesters of pregnancy, and advises pregnant women to limit their caffeine intake to less than 150 mg a day. This seems sensible.

Coffee and colorectal cancer

Another meta-analysis comes to a somewhat tentative conclusion that high coffee consumption (perhaps four cups a day or more) reduces the risk of colorectal cancer [2]. Methodologically it is sound, and it does a nice job of discussing the subject, but it gives us relative risks for high versus low coffee consumption in people with colorectal cancer.

So the overall figure for all 17 studies with 6,192 cases was that colorectal cancer was 24% less likely in people who drink four or more cups of coffee a day - relative risk 0.76 (95% confidence interval 0.66 to 0.89).

The finding was consistent across geographically distinct areas, and in most studies except five small cohort studies with fewer than 15% of the patients studied. In three studies with 883 cases with adenomas, the relative risk was 0.57 (95% confidence interval 0.44 to 0.72).

Coffee and stroke in hypertensive non-smokers

Coffee consumption in hypertensive men in older middle-age and the risk of stroke has been examined in a long-term study in Hawaii. Coffee intake (inter alia) was measured in a large cohort of men in the late 1960s, and the incidence of stroke observed over the next 25 years in almost all the men in the study. These were men aged 55 to 68 years who were nonsmokers with hypertension (defined as systolic or diastolic blood pressure above 140 or 90 mmHg respectively).

There were 499 men, and 76 developed a stroke, 55 of which were thromboembolic. After adjusting for age, the risk of thromboembolic stroke, but not haemorrhagic stroke, was significantly related to the amount of coffee consumed (Figure 3).

For non-drinkers the five-year incidence was 2%, compared with 4% in those drinking more than three cups a day.

The risk of thromboembolic stroke was more than doubled (relative risk 2.1, 95% confidence interval 1.2 to 3.7) in those who consumed three cups of coffee a day as compared with non-drinkers.


This is probably the first study to show any relationship between coffee consumption and stroke. It is well done, and long-term, and discusses elegantly the methodological issues, especially the difficulty of disentangling coffee and cigarette smoking, as both are highly positively related one to the other.

Overall comment

These three studies show us how careful we need to be in giving advice about lifestyle. It's a bit like Heisenberg's uncertainty principle - if you see it, it's not there any more. With lifestyle advice, any change for the good in one direction may be a change for the bad in another.

For coffee and men with hypertension, stopping drinking it may reduce the chance of a stroke, but increase the risk of colorectal cancer. Difficult, isn't it? But for pregnant women the evidence seems strong - limiting caffeine consumption will reduce the chance of a spontaneous abortion and a low birthweight infant.


  1. O Fernandes, M Sabharwal, T Smiley et al. Moderate to heavy caffeine consumption during pregnancy and relationship to spontaneous abortion and abnormal fetal growth: a meta-analysis. Reproductive Toxicology 1998 12: 435-44.
  2. E Giovannucci. Meta-analysis of coffee consumption and risk of colorectal cancer. American Journal of Epidemiology 1998 147: 1043-52.
  3. AA Hakim, GW Ross, JD Curb et al. Coffee consumption in hypertensive men in older middle-age and the risk of stroke: the Honolulu heart program. Journal of Clinical Epidemiology 1998 51: 487-94.

previous or next story in this issue