Drinking patterns and the risk of serious liver disease

In the UK most adults drink some alcohol and, as a population, we drink twice as much as we did 50 years ago [1]. The current recommendations are that men should not drink more than 3–4 units of alcohol a day and women not more than 2–3 units [2]; however, in 2007, 41% of men and 34% of women questioned had exceeded these recommendations at least once in the previous week [2]. There is a clear relationship between rates of liver cirrhosis and alcohol consumption at a population level [3]; liver cirrhosis mortality rates in the UK have increased dramatically as alcohol has become relatively more affordable. Of the 6541 deaths in England directly due to alcohol in 2007, 65% were due to liver disease and trends in liver mortality are an excellent marker of alcohol harm [2].

Most chronic, heavy drinkers cause some fatty damage to their liver, but only approximately 20% will develop cirrhosis [4,5]. Individual susceptibility to alcoholic liver disease (ALD) remains incompletely understood but undoubtedly represents a balance between genetic and environmental factors. Some factors may offer protection and explain why some drinkers never develop symptomatic liver disease. Others may be detrimental and lead to liver damage at ‘lower risk’ levels of consumption.

Large population-based studies have investigated how much an individual needs to drink to develop ALD. Data from 13,000 people in Copenhagen, Denmark, showed that the relative risk of clinically apparent liver disease was detectable above 21 units (cl) per week, with an odds ratio of 3.7 in men and 7.3 in women with intake above 42 units (cl), rising 14-fold in men drinking more than 100 units (cl) per week [6]. In an Italian study of almost 7000 people (largely men) the risk of any liver disease became significant in individuals drinking more than 26 units (cl) per week calculated on the basis of lifetime consumption [7]; at this level the odds ratio for liver disease was 7.5, increasing to 35.8 at an intake of 100 units (cl) per week. The data in each case are heterogeneous, but we can be relatively confident that the risk of liver disease becomes significant at approximately 20 units (cl) per week, but then increases dramatically. The average intake in our patients with cirrhosis is approximately 80 units (cl) per week [8].

Genetics determine which versions of the enzymes involved in alcohol metabolism are inherited. These can both increase or reduce vulnerability to ALD [4,9]. Female gender also appears to increase susceptibility, owing to differences in alcohol metabolism [4].

Pre-existing liver disease plus drinking alcohol increases susceptibility to liver damage. Fatty liver disease is increasingly common; its prevalence is fuelled by rising levels of obesity and Type 2 diabetes [10]. These conditions may have a synergistic effect with alcohol and it has been suggested that drinking guidelines should be related to BMI [11]. In chronic hepatitis C virus infection (another common liver disease), alcohol accelerates disease progression and reduces treatment response [12,13]. These negative effects have been reported in those drinking over 6 units (cl) of alcohol per day, but other research proposes increased risk at lower levels [14,15]. Socioeconomic status is relevant. In a UK study, unskilled, male manual workers aged 25–39 years were 10–20-times more likely to die owing to alcohol than their professional counterparts [16]. Similar conclusions were reached by a Finnish study, which showed that deprivation doubled the risk of alcohol-related mortality or hospital admission [17].

Alcohol damages other organs besides the liver. The WHO suggests that alcohol is implicated in over 60 types of disease and injury [18]. A Russian study of almost 50,000 people showed that accidents, violence, alcohol poisoning and acute ischemic heart disease accounted for most alcohol-associated deaths [19]. Notably, it demonstrated a heavy impact on younger people; alcohol was responsible for over 50% of deaths and 90% of excess mortality in those aged 15–54 years [19]. UK data also highlight the fact that drinking has a greater impact on younger people. Peak ALD mortality has been in those in their fifties, but deaths due to ALD of those in their thirties and forties (or even their twenties) are now increasingly common [20].

Alcohol is strongly implicated in a variety of cancers, especially in oral, breast, bowel and liver cancers. The risk of cancer is slightly increased with any level of alcohol consumption, and then rises in a linear fashion. The more common a cancer is, the greater an impact alcohol will have on individual risk.

Drinking and smoking are the major causes of oral cancers, with a relative risk of up to 38 in heavy-drinking smokers [21]. In line with increased levels of population drinking in the UK, the incidence of oral cancer in men has doubled in the last two decades [22]. Fortunately, oral cancers are still relatively uncommon, with nine cases per 100,000 each year [22].

On the other hand, breast cancer is very common – 110 women in every 1000 will get breast cancer at some stage in their lives – and the absolute risk from alcohol is substantial. For every 1000 women drinking a bottle of wine each week throughout their lives, approximately ten (i.e., 1%) will develop breast cancer as a direct result of the alcohol; for two bottles a week it is 1.5–2%. These risks compare with the risk of dying from a single BASE jump (jumping off a building or cliff with a parachute) of one in 2000 (i.e., 0.05%) [23].

In 1995, UK drinking guidelines changed from weekly limits (21 units for men and 14 for women) to daily benchmarks in response to a Department of Health report [24]. A move perhaps influenced by the perceived health benefits of low-level daily drinking (1–2 units per day). There is some evidence to suggest that light drinking may be protective against coronary heart disease, ischemic stroke and dementia [25,26]; although this remains a subject of debate. Conversely, heavier drinking increases the risk.

The ‘J’ shaped curve has been popular in describing this relationship, whereby light-to-moderate drinkers have reduced cardiovascular risk compared with abstainers, while heavy drinkers are at greatest risk [26,25]. If there is a benefit to health it has nothing whatsoever to do with ‘daily drinking’; the data are based on studies of weekly intake.

Puzzled by increases in liver-related deaths that appeared to outweigh changes in overall consumption, in 2001 the Chief Medical Officer suggested that perhaps binge drinking (drinking a week’s units at once) could be responsible [27]. However, evidence to support this hypothesis has not been forthcoming so far. A large North American study of over 22,000 also showed that daily drinking carried more than twice the risk of liver damage compared with intermittent drinking once or twice per week [28]. Other studies have drawn similar conclusions [7,9,29–35]; and our own study in Southampton, UK, showed that from 80 patients with severe ALD only one reported a consistent pattern of binge drinking [8], the rest were daily or near-daily drinkers. The study also highlighted that risky drinking patterns reported by ALD patients may be established in their early twenties – patients with alcohol-related liver disease were found to have drunk more heavily and more frequently when in their early twenties.

There are well-documented risks from binge drinking but these are to some extent separate from the risks of regular heavy drinking. They involve younger people and include: aggression and violence, self harm including suicide, road traffic accidents, mental and behavioral disorders, sudden cardiac death, stroke, various social problems and impaired performance at work [36,37].

There is little in the way of evidence to suggest that guidelines make a clear difference to levels of alcohol-related harm. However, they are taken very seriously by people and there are legitimate concerns over the shift to daily limits that took place in 1995. The suggestion that daily drinking is in the ‘lowest risk’ category runs counter to the body of evidence showing that frequency of drinking is a significant risk factor. Question one in the gold standard ten-item WHO Alcohol Use Disorders Identification Test (AUDIT) [38] is: “How often do you have a drink containing alcohol?”, with the highest score being more than four-times per week. The question is retained in the four-item AUDIT-C test [39], which features the four most discriminatory questions differentiating hazardous from safe drinking.

Differences between the Greco–Roman pattern of regular drinking with meals and the Celtic–Saxon patterns of feast drinking were noted by Tacitus in the 2nd Century, St Caesarius of Arles in the 6th Century and St Boniface in the 8th Century. The UK habit of weekend binging to celebrate the battles won or lost in the course of the week dies hard, and tends to remain superimposed on top of that daily glass or three of wine.

Furthermore, the natural tolerance induced by regular daily drinking can result in the so-called ‘safe’ third of a bottle of wine on Monday night increasing to a half bottle, and then to a whole one.

Finally, 3–4 units still translates in many people’s minds to three-to-four glasses, which is more likely to represent 7–12 units – the equivalent of more than two bottles of vodka a week, and more than enough to cause liver cirrhosis.

This last point is perhaps the most important in communicating messages to the drinking public at risk. There remains widespread misunderstanding of exactly what comprises a unit, not helped by the fact that a standard ‘drink’ or ‘unit’ varies within Europe from 6 g in Austria and 8 g in the UK, to 10 g in France, 12 g in Sweden and 17 g in Hungary.

Some members of the European public health community would prefer a move to standardize alcohol in grams, on the basis that this is the definition preferred in the scientific literature. Given that alcohol is a liquid and few people go to the shops to purchase an alcoholic beverage by the gram, the use of a dry measure seems counterintuitive. The obvious solution would be to dump the concept of units all together, and refer only to centiliters of pure alcohol. This has the twofold advantage that the percentage alcohol by volume equals the number of centiliters of pure alcohol in a liter of any beverage, and conveniently for the UK consumer 1 cl equals 1 UK unit already.

As for the guidelines themselves, at the very least authorities should consider recommending 3 alcohol-free days each week, together with the concept that the risks associated with drinking are not subject to an on–off threshold but are continuous and in some cases can escalate exponentially.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.