Daily Archives: 09/02/2015

Lifetime-risk of alcohol-attributable mortality WHO report

At the beginning of 2015 the World Health Organisation (WHO) published a report looking at the lifetime risk of alcohol attributable mortality.  This is a report on mortality risk based on alcohol consumption levels in seven European countries, and on implications for low-risk drinking guidelines.

You can read the full report here.  Within this post I will just share the discussion part of the research findings:

If the threshold for alcohol-attributable lifetime mortality risk is chosen to be 1 in 1,000, which seems to be the standard for other voluntary risks in modern high-income societies, then drinking 20g* pure alcohol per day exceeds this threshold for both sexes.  This result is consistent for the different European countries examined in the main analysis and both sensitivity analyses.Even if a considerably more lenient lifetime mortality threshold of 1
per 100 was introduced, the guidelines for women would recommend values lower than
20g of pure alcohol per day and those for men less than 30g per day.  However, there does not seem to be a good justification for such high thresholds, since alcohol use is a voluntary behaviour, i.e., neither necessary as part of diet nor as part of any needs to survive.  The present results corroborate the finding, that overall modern high-income societies accept higher thresholds for mortality risks from alcohol use compared to other voluntary risk factors (19), which is also reflected in current lower risk drinking guidelines in Europe.
As indicated above, two different sensitivity analyses were carried out.  The first was based on sex-specific mortality risks, and the second took into consideration competing risks in a more conservative way.  Both of these scenarios to test the stability of our finding arrived at similar conclusions for the average levels of drinking to reach predetermined acceptable
thresholds, while for higher use levels, the differences between scenarios grew. It should be stressed, that lifetime mortality risk associated with drinking at higher levels (e.g., 50g pure alcohol per day and above) would be similar or exceed the lifetime risk of smoking (115, 116) 11.   So if you are drinking 4 cans of beer/cider or 2/3 of a bottle of wine a day (depending on ABV) or over 6 shots of spirits every day the lifetime risk is similar to or exceeds that of smoking ……
(11 This does not mean that the deaths of current drinking in Europe exceed the deaths of smoking for two reasons: 1) A smaller proportion of the population drinks 50g or above compared to the prevalence of smoking. 2) The analyses here are restricted to 75 years of age. Many of the smoking attributable deaths occur later than this age.)**
It would be an interesting exercise to base low risk drinking guidelines on in voluntary risk.
From the sparse research on mortality and morbidity caused by others’ drinking from Australia, it seems that even the yearly risks of current drinking patterns in this country
would exceed lifetime acceptable risk thresholds for involuntary behaviour.  Thus, for the year 2008, an Australian study (117) found a yearly burden of 367 deaths and almost
14,000 hospitalizations due to drinking by others, indicating a yearly risk of higher than 1 in 100,000 for mortality, and about 0.5 per 1,000 for hospitalizations, clearly much
higher than the usually accepted involuntary risk thresholds stated above (i.e., 1 in 1,000,000 lifetime; see also (17, 89)).
As these mortality and morbidity risks of current drinking on others than the drinker
by far exceed acceptable risk, this could be reflected in devising guidelines for alcohol based on involuntary risk.  In fact, alcohol-attributable mortality to others (i.e., involuntary risk considerations) could be used as a benchmark for national alcohol policies.  Such a benchmark would contribute to the initiation of effective policies to reduce not only the risks to non-drinkers, but to reduce the risk to the drinkers as well.
Given the main causes of involuntary risk to others, the following areas
should be highlighted:
  • Measures to reduce alcohol-attributable injury in traffic.
  • Measures to reduce alcohol-attributable injury at the workplace.
  • Measure to reduce alcohol-attributable violence.
  • Measures to reduce drinking in pregnant women.
In addition to specific measures, some general measures such as increase in taxation or reduction of availability have been shown to be effective in reducing alcohol- attributable injury including violence (118, 119).
In assessing mortality risk and health burden due to alcohol consumption, one should not overlook that the burden of alcohol goes well beyond the health field, including such social consequences to those around the drinker and to the wider society as crime, lost productivity, family problems, child neglect or abuse, and social marginalisation (120).
An Australian study found that the reported tangible costs from out-of- pocket expenses and time lost because of others’ drinking were of much the same magnitude as the costs to health, social and legal systems of dealing with problems from drinking (117).  While it may prove hard to integrate the metrics of social burden with those of health burden, they underline the necessity to change our negligent attitude towards alcohol use and its risks.
Overall, while we found that societies accept much higher lifetime mortality risk for alcohol use compared to other risk factors, both for voluntary risk to the drinker and for involuntary risk to others, the reasons for this acceptance are not fully understood (19).
It may be related to lack of knowledge about the true risks of alcohol, especially for cancer, or to historical vagaries, as alcohol use has neither been integrated into food regulations as an ordinary food item, nor into international conventions which exist for all other psychoactive substances, nor has there been a public health action similar to tobacco
after the repeal of prohibition (for more details see (19)).
However, as current evidence clearly indicates an exceptional role for alcohol use, with higher mortality risks being accepted than for other behavioural and non- behavioural risk factors, we may well rethink acceptable risk for alcohol, and this rethinking could be reflected in new low-risk guidelines.
*In the UK one unit of alcohol contains 7.9 grams, in Australia one standard drink contains 10 grams whereas in the US one standard drink contains 14 grams.  That means 2 units should be the limit in the UK to fit the WHO guidance although they actually state that this exceeds thresholds for both sexes.
** I’m not sure I agree with this for the UK.  There are 10 million smokers, so 17% of the population, and from the Health Survey for England findings for alcohol in 2013:

  • 18% of men drank at an increased risk of harm (between 21 and 50 units per week), and 5% drank at higher risk levels (more than 50 units per week).
  • 13% of women drank at increased risk levels (between 14 and 35 units per week), and 3% drank at higher risk levels (more than 35 units per week).

So I would say the risk for smoking and drinking IS the same.

I hope the Chief Medical Officer is consulting this WHO report as she reviews our current alcohol guidelines and recommendations ….