Drink and exercise for a healthy life

“Cutting out alcohol as part of a health drive could increase the risk of early death”, the Daily Express reported. Multiple news sources covered a study of 12,000 people that found that, compared to teetotallers, those who drink moderate amounts of alcohol are at a 30% lower risk of developing heart disease, while those who drink moderate amounts and combine it with regular exercise have up to a 50% reduced risk.

The Times headline states that “2 ½ bottles of wine a week could save your life”, and claims that a weekly consumption of up to 14 drinks offers the greatest health benefits; while The Daily Telegraph says that “having a pint of beer after going to the gym could be the key to a long healthy life”.

This 20-year study is reported to be one of the first to have looked at the combined effects of both alcohol and exercise on the heart. The benefits of exercise in promoting optimal health are well recognised; however, putting a value on the amount of alcohol that it is safe to consume, or even beneficial to consume, has always been a grey area.

Evidence suggests that a mild to moderate amount of alcohol may offer some benefit to the heart. However, it is important to remember the risks to health, in particular to the liver, of regular heavy alcohol consumption and binge drinking. It is also worth noting that heart disease is not caused or prevented by one factor alone, and other factors such as smoking, blood pressure, diabetes, and cholesterol all have very large impacts on cardiovascular health.

At the current time, it seems sensible to continue to follow standard advice: exercise regularly and observe the maximum number of alcohol units recommended per week.

Where did the story come from?

Jane Østergaard Pedersen and colleagues from the Centre for Alcohol Research, University of Southern Denmark; The Research Unit for Dietary Studies, the Institute of Preventative Medicine; and The Copenhagen City Heart Study, Bispebjerg University Hospital, all in Denmark, carried out the research. The Danish Heart Foundation provided funding. The study was published in the peer-reviewed: European Heart Journal.

What kind of scientific study was this?

In this prospective cohort study - the Copenhagen City Heart Study - the authors investigated the combined effects of exercise and weekly alcohol consumption on death from heart disease and deaths from any cause.

Between 1976 and 1978, 19,329 people over 20 years old, were chosen at random from the Danish civil registration system. They were invited to the Copenhagen University Hospital where they completed a questionnaire of health related questions, including information on physical activity, weekly alcohol intake and general health. Average weekly exercise was graded into four levels ranging from being mostly inactive to vigorous physical activity for more than four hours per week. Weekly alcohol intake was graded as none (less than one drink), moderate (one to 14 drinks per week) and high (15 or more drinks). One drink corresponded to one bottle of beer, one glass of wine or one unit of spirits. All participants also received a general physical examination.

Five years later, a smaller number were contacted and returned for a repeat examination and health questionnaire. After excluding those with pre-existing heart disease, they were left with 11,914 people. These (mostly white) people were followed up until December 2001 to investigate deaths from heart disease, and up to March 2004 to investigate deaths from any cause.

They used identification numbers to trace the members of the study, and only lost about 1% during the follow up. Deaths were identified using the Danish Causes of Death Registry, which uses the international classification of diseases to record causes of death. The researchers calculated the risk of death from heart disease or any cause, taking into account factors that may affect results such as age, sex, smoking, cholesterol, diabetes, and body mass index.

What were the results of the study?

In their total sample, 16% of men and 17% of women were classified as inactive, and 15% of men and 43% of women were classed as non-drinkers. During the total follow up period, there were 1,242 cases of death from heart disease, and 5,901 died from all causes.

Increasing physical activity was linked to a 25-30% lower risk of death from heart disease or any cause than being inactive.

Moderate alcohol intake was linked to significantly reduced risk of death from heart disease in both men and women compared to non-drinking (about 17% reduced risk in men and 24% in women); however there was no difference in risk between non-drinkers and heavy drinkers.

When the researchers also took into account levels of “good cholesterol” (HDL-C), they found that there was no longer a significant difference in risk of death for men from heart disease between moderate drinkers and those who drank no alcohol. Similar effects of alcohol consumption were seen when they looked at all causes of mortality, with non-drinkers and heavy drinkers having a higher risk of death than moderate drinkers.

What interpretations did the researchers draw from these results?

The authors concluded that the combination of physical activity and moderate amounts of alcohol give the greatest health benefits in terms of reduced risk of death from heart disease or any cause.

They suggest that both alcohol and activity may have beneficial effects upon blood circulation, the function of blood vessels and breakdown of fats.

What does the NHS Knowledge Service make of this study?

The benefits of exercise in promoting optimal health are well recognised; however, putting a value on the amount of alcohol that it is safe to consume, or even beneficial to consume, has always been a grey area. There are some important limitations to this study that are worth taking into consideration:

The quantification of alcohol intake may have introduced errors when classifying participants into their alcohol intake group. The measures used - one drink corresponding to one bottle of beer, one glass of wine or one unit of spirits – may mean different things to different people; likewise, it does not account for the type of drink taken and its alcohol content. Similarly, errors may also have been introduced by the exercise groupings.
The researchers only measured average weekly alcohol intake and level of physical activity at the beginning of the study. There is no guarantee that these measures give a reliable representation of an individual’s behaviour from one month, or one year to the next. The snapshot of behaviour at the beginning of the study is unlikely to accurately represent the participants behaviour over the entire 20 years of follow up.
The study relies upon accurate registration of causes of death and misclassification may have led to errors.
The study only considered fatal cases of heart disease; it did not look at how alcohol and exercise relate to the number of cases of diagnosed heart disease, or effects on risk factors such as high blood pressure, among those who survived. It did not assess harm related to alcohol such as alcoholic liver disease.
Other confounding factors may have an effect and which have not been taken into account. Dietary intake for example is not considered; those who exercise more may consume a healthier diet, and it may be this that is causing or at least mediating the benefit.
The researchers report that good cholesterol (HDL-C) “was a mediator” of the effect of alcohol intake on fatal heart disease in both men and women. This is important to note. The paper does not give details on how the researchers measured these cholesterol levels (i.e. whether they physically measured it or whether they asked people about their cholesterol). Inaccuracies here may have underestimated the mediating effects of good cholesterol and therefore overestimated the benefits of alcohol consumption. The same argument would apply to physical activity. However, there was little evidence that levels of HDL-C mediated the benefits of exercise.
This study was mainly conducted in Danish Caucasians, and therefore the results may not be applicable to other cultures or ethnic groups.
Finally, the outcomes of the 30% of the people from the original selected sample who did not participate in further follow up are unknown.