16. No safe level for alcohol!


In January 2016 the recommended weekly limits for adult men in the UK was reduced to a maximum of 14 units which meant that the limits for men and women became equalized. According to this article :




the limits were set at a level that equates to a one percent increased risk of dying from alcohol related causes and the following quote from that article translates that risk factor into something more relevant to daily life experience :


“Prof Sir David Spiegelhalter, an expert in understanding risk from the University of Cambridge, said it was important to put the 1% risk in context. He said an hour of TV watching or a bacon sandwich a couple of time a week was more dangerous.”


Wow! If one hour of watching TV carries the equivalent risk of dying as 14 units of alcohol then alcohol is not the main risk factor in my life! The impression I took away from this article is that consuming 14 units of alcohol still carries risk but that risk is relatively minor or comparable to other daily risk factors that we completely ignore and for which there are no recommended government health limits.


How much do 14 Units of alcohol equate to? 1 Unit is defined to be 8g of pure alcohol which, because alcohol is less dense than water, equates to 10ml of pure alcohol. If we drink a 25ml measure of spirits at 40% ABV (Absolute by Volume) that means we are consuming 25x0.40ml = 10ml of pure alcohol. So the definition of 1 Unit corresponds exactly to a 25ml pub measure of 40% spirit. If the measure differs from 25ml, or the alcoholic strength differs from 40%ABV, then the Units should be scaled pro rata.


What about wine and the notional idea of a “small glass” = 1 Unit? Well that’s just wishful thinking! If you manage to find a bottle of wine at 10% ABV and a wine glass that holds just 100ml then you would have exactly 1 Unit, but as the average wine these days is around 13.5% and a “Medium” pub wine glass is 175ml, that means a Medium glass of wine is actually 2.4 Units, and even a “Small” glass, at 125ml, equates to 1.7 Units. This table shows how many units you can expect from your glass of wine :

When I started my fat loss program I bought some wine and spirit measures to allow me to track how many empty calories I was consuming from alcohol and that’s when I realized to my horror that a “small” wine glass was actually a lot smaller than I thought.


If you are a beer drinker then you might think that the lower alcoholic content would be favourable compared to wine, and it is, but the volume consumed is much greater so a 450ml pint of beer at a typical 4.5% ABV works out at 2.0 Units.


So 2016 saw the weekly limits redefined downwards for alcohol intake but it got much worse in 2018 when, on 13th April, the Lancet published the results of a study of 599,912 drinkers which resulted in the following scary headline :


“Regular excess drinking can take years off your life, study finds” https://www.bbc.co.uk/news/health-43738644)


If you were still clinging to the hope that the odd glass of red wine was actually beneficial for your heart then that headline appears to bury that illusion. However if you read the full paper then the position appears to be not quite so alarmist,  You can find the full paper here :




Using studies from three sources covering 19 countries globally, 10 European countries, as well as England, Scotland, and Wales, the authors measured Hazard Ratios for weekly consumption of pure alcohol: up to 100g; 100g to 200g; 200g to 350g; and more than 350g. They found that :


“The threshold for lowest risk of all-cause mortality was about 100g/week. For cardiovascular disease subtypes other than myocardial infarction, there were no clear risk thresholds below which lower alcohol consumption stopped being associated with lower risk disease. These data support limits for alcohol consumption that are lower than those recommended in most current guidelines.”


There is a graph in the paper that plots the Hazard Ratio against alcohol consumption for “All-cause mortality” and shows very clearly that the Hazard Ratio is non-zero but constant with an alcohol consumption from zero up about 100g/week. However, beyond that threshold the risk increases steeply and from about 150 to 350g/week the line is essentially a linear gradient. From the data they then produced estimates of “Years of Life Lost” against the four levels of alcohol consumption that they defined :

This table applies equally to men and women.


The data in their Hazard Ratio vs All cause Mortality the curve is essentially flat up to about 12.5 units which implies that alcohol has minimal effect up to that level. Between 12.5 and 25 Units per week the graph shows an upward trend that equates to about 0.5 Years Lost to Life at 25 units per week.  The Years Lost to Life then  increase rapidly as consumption rises beyond 25 units a week.


Although this paper cannot be directly correlated to the downward change in UK alcohol limits to a maximum of 14 units per week, the two seem to be saying the same thing : keep to less than 12.5 to 14 units and the risk is low, but be aware that the risk rises rapidly for consumption beyond about 12.5 -14 units per week. Note that they defined drinking more than 100g/12.5 Units of alcohol in one session as “binge drinking” for the purpose of this study, whereas I have seen other sources define “binge drinking” as low as 6 Units in one session. The latter definition essentially classifies 3 small glasses of wine as binge drinking in which case I need to confess to many offences!


A few months later, on 24th August 2018, another large study was published in the Lancet which also confirmed that there was no safe limit for alcohol. That gave rise to the following BBC article :


“No alcohol safe to drink, global study confirms”



This was a huge study covering 195 countries and looked at a wide range of health “outcomes” arising from alcohol consumption, not just mortality rates. Although I don’t think that this paper states anything more serious than the previous paper, the reporting headlines appeared a little more sensational.


It was stated that in a population of 100,000 non-drinkers it would be expected that 914 would develop an alcohol “type” of problem such as cancer or suffer an injury, such as falling over, so the background risk rate of such health outcomes in non-drinkers is 0.914%. If the 100,000 non drinkers consumed 1 “standard drink” per day (1 “standard drink” = 10g pure alcohol = 1.25 Units) then an additional four people would be affected. Really? That means that a consumption of 8.75 Units per week would result in only 4 more people in 100,000 suffering one of these alcohol “type” outcomes. That represents a change in risk from 0.914% to 0.918% which, although not zero, is nevertheless insignificant compared with the background risks of just being alive experienced by drinkers and non drinkers alike! So that while the headline that there is no safe lower limit is absolutely true, it is also very misleading because the increased risk at 1 standard drink per day is insignificant.


If 100,000 non drinkers consumed 2 standard drinks per day (17.5 Units per week) then there would be an additional 63 people affected by the defined health outcomes, so the risk factor would increase from 914 in 100,000 for non drinkers to 977 in 100,000 for those drinkers consuming 17.5 units per week, which in absolute terms is only an increase of risk from 0.914% to 0.977%. Although 63 people is a significant number of people it is still 15 times smaller than the number of non drinkers who would still be affected by background risk factors. Basically the background factors still appear to dwarf the affects of alcohol at 17.5 units per week.


Because this study does not use exactly the same methodology or data sets as the earlier one in April, it is not possible to do an accurate like for like comparison, but it seems to me that both studies show similar findings. They show that while there here is no zero risk threshold for alcohol consumption, the risks at the UK’s limit of 14 Units per week seems to be pretty small compared to the background risks of cardiovascular disease and cancers that we are also exposed to through genetics, environment, diet, and exercise (or lack off).


I started reading this article with a sense of dread that alcohol would have to go but when I looked at the numbers presented they just did not add up to the scary headlines. Once again the BBC obtained a quote from Prof David Spiegelhalter, Winton Professor for the Public Understanding of Risk at the University of Cambridge, to try to quantify the risks in terms of everyday reality, and I have to say that I concur with the professor : "Given the pleasure presumably associated with moderate drinking, claiming there is no 'safe' level does not seem an argument for abstention. There is no safe level of driving, but the government does not recommend that people avoid driving. Come to think of it, there is no safe level of living, but nobody would recommend abstention."


Professor Spiegelhalter’s role is to help communicate the idea of risk to the public in simple terms. He came up with the concept of a unit of risk called a Microlife which represents a reduction in life of 30 minutes. For example, smoking two cigarettes represents a risk of 1 Microlife for a 30 year old man. Put another way, smoking two years will likely reduce the life of a 30 year old man by 30 minutes. That certainly focuses the mind much more than quoting a % hazard ratio. Professor Spiegelhalter has a great blog which explores the subject of risk and microlives :