This question is really two conflated questions:  is ethanol (i.e. the active ingredient in all alcoholic beverages) a risk-factor for obesity, and;  are alcoholic beverages themselves a risk factor for obesity. There’s a third, related question:  does consuming food with either of these (ethanol or beverages) alter the risk for obesity. Let’s start by looking at what the body does with ingested ethanol.
The metabolism of ethanol
Ethanol is readily absorbed and enters the circulation where it can travel throughout the body and cross into the cells of multiple organs or cross the blood brain barrier and enter the brain. Our bodies do not have the means to control or regulate this toxic attack. Instead, ethanol must be physically eliminated from the body (mostly through the lungs and kidneys) or neutralised (mostly by being metabolised for energy). While some ethanol is metabolised in the gastro-intestinal tract (and other organs including the brain), the bulk (~90%) gets broken down by the liver into acetate (acetic acid).
The liver does this is in a two-step process driven by two enzymes. In the first step, ethanol is converted to acetaldehyde by an enzyme abbreviated ADH (alcohol dehydrogenase). Acetaldehyde is more toxic than ethanol and it must be immediately converted to acetate by another enzyme called ALDH (aldehyde dehydrogenase).
The end product (acetate) is a non-toxic molecule that can be burned for energy by cells in the liver, or released into the circulation and taken up by other cells, also for their energy needs. Burning acetate produces CO2 and water as byproducts, and the process of ethanol elimination is complete.
Readers who recall my previous posts on ketones and the ketogenic diet may be interested to know that some acetate can also be converted into ketones, which in turn can be used as fuel or for other signalling purposes.
So, the primary means for eliminating ethanol is to break it down to a fuel and burn that for energy. This is why ethanol is assigned a caloric value – about 7 cal/g (c.f. carbs = 4, protein = 4, fat = 9, all huge approximations).
It is also why ethanol, per se, is not a risk factor for obesity – we don’t store it, we burn it. Indeed, chronic alcoholics are usually anorexic and malnourished. However, the calories we get from ethanol displace the calories from other macronutrients (such as the glucose in carbohydrates), which instead get put into storage as fat. So it is the calories that accompany ethanol, either in the alcoholic beverage, in mixers or in food, and that are not burned for energy, that are the potential problem.
Still, it’s not that simple either. It would be assumed that a regular calorie surplus from say, food and wine consumption, should translate into a long-term risk for obesity. The dilemma is that this has not been shown to be true and, if anything, alcoholic beverages in moderation may be protective against weight-gain.
What happens during and after a meal?
This is studied by asking people to drink either an alcoholic beverage or a non-alcoholic control drink before or during consumption of a meal in which they can chose for themselves how much to eat. Many studies of this kind have been carried out in various ways, and they report consistent results – the participants do not eat less food to compensate for the extra calories in the alcoholic beverage. In fact, they eat more. That’s a double-whammy of extra calories from ethanol and extra calories from food.
Part of the explanation is that we don’t perceive the extra calories in the alcoholic beverage and therefore don’t adjust our eating to compensate. In fact, we don’t perceive calories in caloric drinks – people drinking fruit juice or sodas under similar free-eating conditions do not compensate for the additional calories either. It might be that, for most of our evolution, the only fluid we drank was water (non-caloric) and so we didn’t evolve a mechanism for perceiving calories in fluids, or for finding fluids satiating (just thirst-quenching).
But, we don’t just fail to compensate for the calories in the alcoholic beverage by eating less, instead we eat more. It’s not that we become disinhibited by ethanol, but rather that ethanol increases the perceived pleasantness of food cues (visual, aroma etc), and thus increases appetite by making food appear more appealing. It is also not the case that ethanol disturbs appetite-satiety hormonal balances to favour appetite. Ghrelin is the main hormone that drives appetite and eating, and so theoretically ethanol could have increased the release of this hormone and driven appetite up. Paradoxically, ethanol does the opposite – it significantly decreases the level of ghrelin, and so should drive appetite down. Similarly, ethanol does not inhibit satiety-signalling hormones.
The long-term effects
The idea that alcohol is a risk factor for obesity is poorly supported across studies, and if anything, long-term moderate alcoholic intake appears to reduce the risk of obesity and cardiovascular events. In this context, moderate means 1-2 standard drinks a day (as alarmingly-few as that may seem). Somehow, the excess calories consumed at a meal (with alcohol) are adjusted for at other times of the day. The role of ethanol in suppressing ghrelin might have something to do with that, and could contribute to a long-term reduction in the risk of obesity.
However, there are many confounds in these long-term studies. For example, people who consume alcohol moderately may be more physically or socially active or make other beneficial lifestyle choices. People who abstain may consume sugary drinks instead, such as fruit juice or soda, thereby over-consuming calories (without a ghrelin benefit), or be treating a past history of alcoholism. The way alcohol is consumed also matters – binge drinking may be a risk factor for obesity (and a lot of other things) but it could be part of a more general form of impulsivity that might be the underlying factor. The type of beverage probably matters – beer vs. wine for example. There are very few studies of spirits.
Then there are individual variations. The ADH and ALDH enzymes (that facilitate the breakdown of ethanol) are encoded by multiple genes leading to multiple sub-types of these enzymes that have differing efficacy across individuals and ethnicities. In regard to a predisposition for alcoholism, the strongest predictors are not genetics, but rather socio-economic circumstances: poverty, social conditions, disenfranchisement.
The brain and nervous system
The liver eliminates ethanol at the maximum possible rate, however it cannot increase this rate to match peak demand. This results in a backlog of ethanol in the circulation waiting to be eliminated.
Ethanol is a small water soluble molecule that travels in the circulation, however it is also slightly fat soluble and it can diffuse across the blood-brain barrier (BBB) and enter the brain. While cells in the brain can metabolise the intruder (through a different pathway than the liver uses), it is the liver that has to do the bulk of the work. As the liver brings blood ethanol levels down, ethanol will diffuse back out of the brain and into the peripheral circulation (it is a passive gradient-driven diffusion that depends on relative concentrations on each side of the BBB). Until this happens, ethanol in the brain disrupts communication between neurones by binding to post-synaptic receptors (receptors on the target neurone during neurone-to-neurone signalling). These disruptions, that can happen across multiple neurotransmitter systems, lead to the intoxicating effects of ethanol.
Ethanol-related studies are almost certainly carried out in people who are in a state of glucose-adaptation, since this has been the default condition of western societies for the last 30-40 years (since the Dietary Guidelines for Americans were proclaimed). However, the findings could be different with keto-adapted individuals. The liver breaks down ethanol, but it is also the only source of ketones and it can produce glucose, meaning that there are many possibilities for interactions. It is known that ethanol inhibits the liver’s ability to generate glucose. It also appears that ethanol raises liver ketone production (perhaps because glucose is lowered). Beyond that, little is known because the studies have not been done. Anecdotally, some individuals report that, after they keto-adapt, alcoholic beverages make them intoxicated quicker. However, this is far from universal (and not my experience).
Keto-adapted people do need to be aware of the carbohydrate content of their alcoholic beverage in order to remain in ketosis. Spirits have zero carbohydrates, so it’s their mixer, if used, that needs scrutiny. For example, the sugar content of the following three mixers is much the same (about 9-10g/100ml): tonic water; coca-cola; unsweetened 100% orange juice. So, be cautious with gin and tonic, rum and coke and vodka and orange. Also, mixed drinks are usually larger than 100ml and the bulk is made up with the mixer – a ‘standard’ vodka and orange will contain about 30g of carbohydrates (for example). The sugar in tonic water is there to balance the bitterness of the quinine (and is masked by it). The best mixer for ketosis is plain soda water or ‘on the rocks’. Many cocktails will have an added sweetener, in the form of liqueurs and/or sugar syrups.
Full-strength beer comes in around 3.5 g per 100ml. Perhaps surprisingly, brown ales or even Guinness are not much more than this either. However, it soon adds up – 13 g per can (350 ml). Also, beers are much higher in carbohydrates than wine when adjusted for alcohol content.
Dry wines, thank goodness, are low in carbohydrate. Both red and white dry wines are only about 2.5 g of carbohydrate per 100 ml of wine, and only about 0.5g of that is sugar, the rest is primarily sugar alcohols that are not readily digested. Champagne is lower still (~1g/100ml).
While high chronic consumption of alcoholic beverages increases the risk of serious conditions, moderate consumption may be protective, particularly for cardiovascular disease. The mechanisms for this remain uncertain however, an intriguing possibility is that it is the acetate (the final break down product of ethanol) that is protective. Acetate can suppress insulin signalling (which reduces fat accumulation in adipose tissue and lessen long-term weight gain). Acetate receptors are highly expressed in immune cells, which in turn could favourably regulate immune and inflammatory responses. Acetate can also inhibit a protein responsible for breaking fats down into fatty acids and thus reduce free fatty acids in the circulation. Like ketones, acetate burns more cleanly than glucose (fewer free radicals are produced).
How all this comes together is not certain, however it raises the possibility that consuming acetate might confer the same benefit as consuming ethanol, without the concerns surrounding ethanol. We already consume acetate, otherwise known as acetic acid – it is the acidic ingredient in vinegar (~5% acetic acid). Thus a vinaigrette tossed over a mediterranean salad may be cardio-protective both from the monounsaturated EVOO as well as from the vinegar. Vinegar taken with a meal can reduce the glycemic index of carbohydrates in the meal. This is also true for diabetics, and before modern drugs people with diabetes used vinegar teas to help control their symptoms. Vinegar was an important ingredient in the cordials of an earlier age (known as shrubs).
Why can we metabolise ethanol?
It’s a valid question because we have only been consuming alcoholic beverages for a very short period of our evolution. However, we are not one, we are a community. Our gut supports trillions of microorganisms – ten-fold more than there are cells in our body – and there is more DNA in our microbiome than in us (by a factor of about 150). These organisms span thousands of species, are ever changing and we need them not just to be healthy, but to be us (they can regulate the expression of our DNA). All of which challenges the very concept of our identity. Scientific interest in the microbiome has exploded in just the last few years. But, to get back to the point, some of these microorganisms (bacteria and yeasts) ferment glucose and fructose to produce ethanol.
The two main bacterial strains that produce ethanol are the appropriately and delightfully named L. fermentum and W. confusa (respectively). The yeast S. Cerevisiae (brewer’s yeast) does too, and can inhabit our gut. It is possible that this left our arboreal primate ancestors in a mellow state of inebriation, but it all changed when early hominids adopted a terrestrial lifestyle. Now fresh tree fruits were less accessible, whereas fruit on the forest floor was more available. Fallen fruit would be expected to contain more fermenting yeast (and ethanol) than similar fruits hanging on trees. The increase in dietary ethanol, beyond that produced by gut microorganisms, favoured the evolution of a mechanism to metabolise it. The genetic mutation enabling this occurred around 10 million years ago (after orangutans diverged from the our lineage but before gorillas and chimpanzees did). Those with this mutation would have been more capable, mentally and physically, to meet the challenges of terrestrial survival, and thus have an evolutionary advantage.
Since the advent of agriculture (about 12,000 years ago), human societies of every kind around the world learned how to ferment local fruits, sugars and grains to produce ethanol. Alcoholic beverages took on social, religious, and medical significance. The earliest evidence of fermented beverages found so far has been traced to the early neolithic period in a village in Henan province China. There, about 9,000 years ago, an alcoholic beverage was being produced from the mixed fermentation of rice, honey and grape (as determined from chemical analysis of pottery shards). About 5,000 years later, during the Shang dynasty, cereal beverages were being fermented and, remarkably, such beverages have been found preserved as liquids inside sealed bronze vessels.
In Australia, ‘responsible’ drinking is defined as up to 2 standard drinks a day for most of the adult population. A standard drink is defined as one that contains 10g (12.5 ml) of ethanol. Recommendations vary worldwide, which suggests that there are confounds in arriving at this number. So, where did it come from? I will use the Australian guidelines (here).
A scientific committee of the National Health and Medical Research Council is charged with coming up with a recommendation (concluding that the task is impossible with any certainty is not an option for them). The committee uses population modelling to estimate alcohol-attributable lifetime risk of death from chronic conditions for different levels of average consumption. This is done by deciding which chronic conditions might causally be linked to alcohol, and then computing alcohol-attributable fractions (AAFs) for these conditions (i.e. what fraction might alcohol have contributed to the condition). For example, to what extent is liver cirrhosis due to ethanol or to non-alcoholic fatty liver disease (driven by carbohydrate consumption). Various equations and risk-estimates are incorporated. Data from scientific publications is incorporated where available. Assumptions are made to fill in the gaps. The end result of the computational model is an estimate of age-adjusted lifetime risk. The reference group is lifetime-abstainers. The calculation is model-based because it would be unethical to conduct prospective control trials of adverse outcomes.
The final step is to decide what the acceptable level of risk is. Currently, this is set to 1 in 100 (one alcohol-attributable death for every 100 people over their lifetime). To put this in perspective, this is less likely than death from a traffic accident. The lifetime risk of death from heart disease is 1 in 5.
So, that’s where 2 standard drinks a day comes from. A nice round number with an implied exactness. There are many potential pitfalls in determining this number: e.g. the paucity of large-scale scientific studies providing solid data, the recognised heterogeneity of abstainers as a control group, estimation of AAFs and modelling assumptions to name just a few. Furthermore, this is a population-level assessment, and as the committee points out: “due to individual variability, there is no amount of alcohol that can be said to be safe for everyone”.
Unfortunately, these subtleties get lost once a fixed number is determined and gets spread by campaigns.
By way of clarification, science often uses modelling as a tool, and this is often appropriate. For example, climate science is modelled (because we do not have multiple earths to experiment on). A difference here is that there is so little experimental observational data driving dietary models and the models themselves are conceptually simple despite our physiological, genetic and environmental diversity.
The post is long but there is much I have lef out. There is much that I do not understand. There are multiple gaps in our knowledge (and mine). There are paradoxes and there are confounds. No matter… Cheers.