Alcohol and Mortality in the Over 50s

Two studies published this week, conducted in Britain and
Denmark, report on the combined effects of exercise and alcohol on the
risk of death (from any cause). Both studies found that moderate
drinkers (less than 14 units per week) had a lower risk of dying in the
ten or twenty years of the study than non-drinkers and heavier drinkers.
However, neither study asked their non-drinking group whether they had
recently given up drinking (perhaps for health reasons) or whether they
had always been abstainers.

The British study, conducted by Kay-Tee Khaw and colleagues at
Cambridge University, was conducted with the aim of developing and
evaluating a simple measure of health behaviours that could be used by
GPs and the general public. Over 20,000 people aged 45-79 at the start
of the study were surveyed between 1993 and 1997, and again in 2006. The
measure gave one point for each of four health related behaviours,
namely not smoking, being physically active, drinking moderately and
eating five portions of fruit and vegetables per day. The points are
then simply added up to give a score from zero to four. This score was
strongly related to the risk of dying before the end of the eleven year
study, such that people with a score of zero had four times the risk of
dying as people with a score of four.

Because the purpose of the study was to evaluate the new
scale, the report did not include much detail on the effects of
drinking. However, they did compare moderate drinkers (1-14 units per
week) with others (combining non-drinkers with heavier drinkers) on risk
of death from all causes, cardiovascular disease, cancer and other
causes, over the eleven years of the study. Relative to moderate
drinkers, others had a 26% greater risk of dying from any cause. This
increased risk was greater for cardiovascular disease and cancer (29%
and 28% respectively) than other causes (17%).

Considering men and women separately, the increased risk was
higher for men (35% increased risk overall), particularly for cancer
(46% increased risk). For women, there was no overall increased risk,
but for cardiovascular disease, ‘others’ had a 37% higher risk than
moderate drinkers. The different results for men and women may be
explained by the fact that for men ‘others’ were mostly heavier drinkers
(74% of the others drank more than 14 units per week) whereas for women
the ‘others’ were mostly nondrinkers (only 31% drank more than 14 units
per week). This means that for men, the comparison was between moderate
drinkers and (mostly) heavier drinkers whereas for women, the
comparison was between moderate drinkers and (mostly) nondrinkers.

The Danish study, conducted by Jane Ostergaard Pedersen and
colleagues at the National Institute of Public Health, University of
Southern Denmark, investigated the combined effects of drinking and
leisure-time physical activity. Almost 12,000 people aged 20 years or
older took part in the study between 1976 and 1978 and follow up data
were collected in 2001 and 2004. Participants were classified as
inactive (less than 2hrs of light physical activity per week), having a
low level of activity (2-4 hrs of light physical activity per week) or
having a moderate to high level of activity (at least 4 hrs of light
activity or at least 2 hrs of vigorous activity per week). weekly
alcohol intake was classified as none (less than one drink per week),
moderate (1-14 drinks per week) or heavy (more than 14 drinks per week).

Overall, alcohol had no effect on the risk of dying within the
20 years of this study. However, for people with low or moderate to
high levels of physical activity, those with a moderate alcohol intake
reduced the risk of dying by 10% to 13%, relative to nondrinkers (who
were not asked if they had previously been drinkers). For inactive
people, there was no benefit of alcohol consumption. Conversely, any
regular physical activity reduced the risk of death by around 20%,
regardless of alcohol intake.

The study was published in the European Heart Journal and also
specifically investigated heart disease. In this case, both alcohol and
exercise were beneficial, such that inactive moderate drinkers had a
similar risk of dying from heart disease to teetotallers with low levels
of physical activity. Both groups had around 30% lower risk of death
than inactive teetotallers. Relative to the same group, moderate
drinkers who were also somewhat active had around half the risk of dying
from heart disease.

Both studies define moderate drinking as between one and
fourteen units/drinks per week. The UK unit is precisely defined as 8 g
or 10 ml alcohol, though in practice it can mean considerably more, as
it is often equated to a glass of wine, which can be as much as three
times this amount. A ‘standard’ drink in Denmark contains 12 g or 15 ml
alcohol. A small glass of wine (125 ml; the old pub measure) contains 16
ml alcohol. A typical UK wine glass is 250 ml.

The Danish study reported that men classified as moderate
drinkers actually drank an average (median) of seven drinks per week and
women in the same category drank an average of four or five drinks per
week. Therefore, the benefit of moderate drinking found in this study
was gained from a weekly alcohol intake equivalent to two and half
glasses of wine (or pints of beer) for men and one and a half glasses of
wine (or pints of beer) for women.

Neither of the two studies examined different age groups
separately. Whilst the Danish study used an age-stratified sample of
people over the age of 20, the average (median) age of their
participants was 55 for the most active and 60 for the least active, at
the start of their twenty year study. The British study included people
aged between 45 and 79, with an average (mean) age of 58 at the start of
their eleven year study. It is well established that any benefit of
moderate drinking is found only in older people. For men this starts at
the age of about 40 and for women this starts at around 55.

In summary, both studies found a small benefit of having a few
drinks per week, relative to not drinking at all. The majority of
people who took part in both of these studies were in their 50s or older
at the start of the studies.


Khaw et al (2007) Combined impact of health behaviours and
mortality in men and women: The EPIC-Norfolk prospective population
study. PLoS Medicine 5(1) e12

Pedersen et al (2007) The combined influence of leisure-time
physical activity and weekly alcohol intake on fatal ischaemic heart
disease and all-cause mortality. European Heart Journal.