How alcohol intake can change over a lifetime

"Binge drinking peaks at 25 … but by middle age he's drinking daily," the Mail Online reports. In what has been described as the first of its kind, a new study has tried to track the average adult drinking pattern over the course of a lifespan.

Researchers combined information from nine studies, following almost 60,000 people, to model how average alcohol intake changes over a lifetime in men and women in the UK.

It found that, in men, alcohol consumption rose considerably in adolescence and peaked at around 20 units per week (around six pints of higher-strength lager) at age 25, before decreasing. Drinking daily or on most days of the week became more common in mid-life to older age. A similar pattern was seen in women, but they drank less (around seven to eight units a week).

The authors note that they were not able to obtain complete information on drinking patterns, as all the studies collected information in different ways. This means that while the study can tell us about average behaviours, it can’t tell us whether people were binge drinking or not.

While there is more to be learned in this area, the study gives an insight into estimated average alcohol consumption over time in the UK.

Where did the story come from?

The study was carried out by researchers from University College London and other universities in the UK. Two of the researchers were funded by the European Research Council, and the studies they utilised data from were funded by the Medical Research Council, British Heart Foundation, Stroke Association, National Heart Lung and Blood Institute, and National Institute on Aging.

The study was published in the peer-reviewed journal BMC Medicine on an open-access basis, so it is free to read online or download as a PDF.

All of the UK’s news sources refer to "binge drinking" when talking about the results, but the study did not look at binge drinking specifically. It looked at total weekly alcohol consumption and frequency of drinking alcohol, but did not assess how much people drank on a single occasion. While it is plausible that in some cases the 20 units per week average seen in men in their twenties were drunk during a single Friday or Saturday night, it’s not possible to know this for certain, based on the data presented.

In addition, there was a level of alarm in the reporting about the elderly drinking daily, but the average consumption per week in these age groups – particularly in women – was within the recommended UK drinking limits.

What kind of research was this?

This research analysed data from different UK cohort studies, which looked at how people’s alcohol consumption changed over their lifespans.

This is the best way to assess this question, although, ideally, the same people would be followed throughout their lifespan. This study used overlapping data from different people and combined it, to assess overall patterns.

What did the research involve?

The researchers used data from nine prospective cohort studies, each providing at least three measurements of alcohol consumption for each individual. These studies included 59,397 people in total, and had data from people ranging in age from 15 to more than 90. They turned data from these studies into statistical models, to predict patterns of alcohol consumption by age.

The studies covered data collected from 1979 to 2013, in individuals born between 1918 and 1973. The nine cohort studies included:

three nationally representative cohorts recruited at birth
three cohorts born 20 years apart, representative of the West of Scotland
a representative cohort of older people in England
a cohort of civil servants in London aged 35 to 55 years at recruitment
a population-based cohort of men aged 45 to 59 years from South Wales

Average weekly alcohol consumption was extracted from each study and expressed in UK units (one unit = eight grams of ethanol). Frequency of alcohol consumption was also extracted and classed as:

none in past year
monthly/special occasions
weekly – infrequent (not daily or almost daily)
weekly – frequent (daily or almost daily)

Statistical models were generated for men and women separately, and the researchers tested how well their models fit the observed data.

What were the basic results?

The researchers found that average alcohol consumption rose considerably in adolescence and peaked at around 20 units per week at age 25. After this, it decreased, until it started to level off in mid-life, and then finally decreased again from age 60, to around five to 10 units per week.

A similar pattern was seen for women, but they had lower levels of consumption, with a peak of around seven to eight units a week and consumption of two to four units in those aged 70 and over.

Drinking daily or on most days of the week became more common in mid-life to older age, particularly in men, with half of men drinking this frequently at this time of life in one cohort. Drinking this frequently reduced in very old age.

How did the researchers interpret the results?

The researchers concluded that this is the “first attempt to synthesise longitudinal data on alcohol consumption from several overlapping cohorts to represent the entire life course, and illustrates the importance of recognising that this behaviour is dynamic”. They say that this sort of information may help to design better strategies to reduce alcohol consumption. They also say that it suggests studies looking at the effects of alcohol consumption which only assess intake at one time point should be treated with caution.

Conclusion

This research combined information from nine studies, to model how alcohol intake changes over a lifetime in men and women in the UK, and is reportedly the first to do this.

There are some points and potential limitations to note:

The authors note that the different studies captured information on drinking in different ways, and although they tried to standardise this, they were not able to obtain complete information on drinking patterns.
The data on which the models are based come from different periods, when alcohol consumption habits and strength of alcohol available may have differed.
The authors looked at this, and say that the patterns seen in different periods appeared similar, although the rate at which consumption changed differed slightly.
There were also some differences in patterns seen between cohorts which were not due to differences in the period. For example, older women in a Scottish cohort had lower consumption than women of a similar age in the cohort of civil servants in London, despite this data being collected in a similar period. Other factors, such as socioeconomic status, could be contributing to these.
Not all cohorts covered the same age range, so although the total number of people being analysed was large (almost 60,000), each individual age would have a smaller number of people.
It was not clear how the nine cohorts were identified, and whether there were others available that were missed.