Alcohol and older people

6.1 Total mortality

More than 60 conditions have been identified as either partially or entirely caused by exposure to alcohol.³⁰ There is no doubt that alcohol is a major preventable cause of premature mortality for all ages. Increased risk of death is not limited to high levels of consumption and in fact low level drinking has been identified as a major cause of excess mortality arising from a range of alcohol-related cancers. Past meta-analyses of the observational studies which dominate this literature have nevertheless shown J-shaped curves for all-cause mortality with beneficial effects at low doses. Interestingly, these studies also suggest that the level at which maximum potential benefit may be achieved is probably less than 10 grams of ethanol per day and for women may be as low as 5 grams a day (e.g. Bagnardi et al 2004³¹). In fact, the latest, and largest, report on this subject, from the Global Burden of Disease project, finds that the safest consumption level is 0.³⁰ What is more, social and economic cost studies based on the supposition that health benefits from alcohol exist, show that the total social and economic costs of harms from alcohol still outweigh the sum of the benefits.^{32, 33}

In recent years, with the emergence of novel research approaches that substantially reduce potential for confounding and bias compared to traditional observational study designs (i.e. Mendelian randomization studies), long-standing doubts about the veracity of J-shaped curves for alcohol have been reinforced.³⁴ Mendelian randomization studies for instance suggest no protective effect for low-dose consumption and coronary heart disease events.²⁹ This is important because the apparent protective effects for low-volume consumption on total mortality are due to observed protective effects on coronary heart disease, so the veracity of the J-shaped curve and protective associations for low-volume consumption for all-cause mortality remains in doubt. However, based on observational study data, across the general population, including for older persons when assessed, the lowest risk of death is generally found among those drinking less than 10 grams of ethanol daily.^{27, 35–39} Furthermore, a Mendelian randomization study of older persons finds reduced all-cause mortality with a genetic variant associated with less alcohol consumption.⁴⁰

6.2 Cardiovascular diseases and diabetes

Coronary heart disease, stroke, hypertension and diabetes are all more prevalent among older people, and so studies of these outcomes among the general population are relevant to those in later life, even if the cohorts used in those studies are sometimes not restricted to older people. Both congestive heart failure (associated with heavy consumption) and atrial fibrillation (associated with any consumption) are for example far more common among older individuals but will not be discussed in detail in this report.

There has been no randomized study of low-volume alcohol consumption and any morbidity or mortality outcome related to coronary heart disease. Excessive alcohol consumption, including heavy average consumption and binge drinking, are risk factors for coronary heart disease, stroke, hypertension and the development of diabetes or to poor diabetes control.

In terms of low-dose or “moderate” consumption, most observational studies find protective effects for coronary heart disease. Furthermore, protective associations appear to be greater for relatively older persons compared with the general population.^{28, 41} However, the concern is that older ages would also tend to magnify sources of bias including the ‘sick quitter’ phenomenon and the ‘healthy survivor’ bias, and protective effects attenuate when attempting to account for these biases.²⁸ In addition, Mendelian randomization studies find no protective effect for low-volume alcohol on coronary heart disease events or coronary calcification. ^{29, 42, 43} 

While randomized trials of the Mediterranean diet find protective effects for the primary and secondary prevention of CHD outcomes^{44, 45}, those trials were not randomized with respect to alcohol consumption. Although these latter findings do not preclude protective effects of low-dose alcohol consumption, they mean that the apparent benefits of a Mediterranean diet (i.e. one in which the moderate use of alcohol is one small part) may be explained by factors other than alcohol consumption.

Strokes may either be ischemic (in which blood flow is blocked or reduced to a portion of the brain) or hemorrhagic (in which a blood vessel bursts, interrupting flow and also sometimes creating pressure on surrounding brain tissue); ischemic are the most prevalent of the two. In observational studies, high-volume consumption is a risk factor for both stroke types. Low volume alcohol consumption appears protective for ischemic (but not hemorrhagic) strokes among both men and women, though at lower levels for men (less than one drink per day) than for women (about one drink per day).^{46, 47} A recent well-designed and large study questions whether even light drinking protects against ischemic stroke³⁹ while a Mendelian randomization study finds no protection from having a gene related to reduced alcohol consumption on all strokes.²⁹

For hypertension, in itself an important risk factor for cardiovascular disease and mortality, most but not all observational studies find a positive relationship between alcohol consumption and both higher blood pressure and incident hypertension across all levels of consumption, particularly among men.⁴⁸ Furthermore, a meta-analysis finds that reductions in alcohol consumption are associated with reductions in blood pressure.⁴⁹ A meta-analysis of Mendelian randomization studies also finds positive relationships between alcohol consumption with increased blood pressure and hypertension among men.⁵⁰

In terms of blood glucose and diabetes, observational studies generally find a protective effect of alcohol consumption (e.g. ^51,  52), with the lowest level of risk among those consuming about one drink daily. However, sex stratified results find significant protective effects only among women⁵³ and even these may be questioned as there are relatively few well designed studies of women. Moreover, in analyses restricted to trials with a never-drinking comparison group (as opposed to all non-current drinkers), alcohol was not associated with any protective effects for either women or men.⁵³ Mendelian randomization studies also find no significant protective effect of alcohol consumption on diabetes.^29, 54, 55 Finally, a meta-analysis of randomized clinical trials finds no protective effect of low-volume alcohol consumption on blood glucose control among those with diabetes.⁵⁶

In experimental studies, alcohol administration raises HDL cholesterol levels. However, a Mendelian randomization study found no cardio-protection for those carrying a gene variant that raises HDL. Furthermore, after controlling for their effects on LDL, randomized studies of statin drugs have not found a significant relationship between changes in HDL and coronary heart disease. Finally, several drugs that effectively raise HDL have not resulted in reduced coronary heart disease events in clinical trials. For these reasons, although HDL is associated with cardiac outcomes it is not likely causal for heart disease.^57–59 Alcohol consumption does not appear to meaningfully affect levels of LDL (bad) cholesterol.⁶⁰

In summary, taking account of studies of biological plausibility and studies with stronger designs, there is considerable evidence that heavy alcohol use is responsible for worse outcomes and increased risks for cardiovascular diseases and diabetes. In particular, studies of hypertension, a major risk factor for cardiovascular disease, find only negative impacts from the consumption of alcohol even at low doses. In better designed studies there is little or no evidence of protective effects for low-dose alcohol, especially for men. Furthermore, genetic studies using Mendelian randomisation have found no evidence of protective effects from low-dose alcohol use on cardiovascular disease or diabetes for either men or women.

6.3 Cancer

Advancing age is the leading risk factor for incident cancer and for cancer mortality, and those over age 65 account for approximately 70% of all cancer deaths. Increased cancer susceptibility among older persons is likely due to a combination of reduced cellular mechanisms (e.g., DNA repair mechanisms) and an accumulation of carcinogenic damage from environmental exposures over the life course (e.g., tobacco, alcohol).⁶¹ In addition, cancer treatment options may be limited or less desirable on the basis of age (e.g., stem cell transplants, aggressive surgical debulking), or available treatments may have greater relative toxicities than among relatively younger persons.

Alcohol has been classified by the World Health Organisation as a Group 1 carcinogen since 1988 when it was concluded that there was sufficient evidence for its causal role in cancers of the oral cavity, pharynx, larynx, esophagus and
liver. Since that time, several hundred more epidemiological studies have reported on the association between the consumption of alcoholic beverages and the risk for cancer at various sites. In 2007, the IARC added cancers of the female breast, colon and rectum to the list of cancers caused by alcohol (IARC, 2007).⁶² World Cancer Research Fund (2016)⁶³ more recently concluded that stomach cancer also was causally related to alcohol consumption. There is also accumulating evidence for causal association between alcohol use and both prostate and pancreatic cancers.^64–66

The increased risk of these cancers is normally found to be increased even for light or moderate drinkers with no safe level of consumption (e.g. ⁶⁴). However, the appearance of a protective effect does occur in some studies, likely due to the same kinds of confounding factors described earlier in relation to the apparent J-shaped curve for alcohol and mortality generally. Most studies of alcohol and cancer outcomes suffer from the same kinds of systematic bias discussed earlier in particular relation to cardiovascular disease. In one meta-analysis of prostate cancer, it was found that when former drinker bias was corrected the estimated risk for moderate drinkers increased from 8% to 22%.⁶⁶ This means that the increased risk of cancers from alcohol consumption are likely underestimated when based on meta-analysis of the existing literature.

In one of our previous reports⁶⁷, we applied the latest methods used internationally to quantify the burden of cancers attributable to alcohol in Sweden.⁶⁸

This is likely to be a substantial underestimate as there are likely more cancers than these causally related to alcohol use and the risk estimates derived from the existing literature are themselves likely to be underestimated.

6.4 Cognitive function

Alcohol is recognized as a powerful neurotoxin, known to cause or contribute to a wide range of neurological disorders including dementia and fetal alcohol syndrome, among many others. Furthermore, the burden of dementia attributable to alcohol is now recognized as much larger than previously thought. Heavy drinking is strongly associated with the development of Alzheimer’s disease. Alcohol use disorder is the strongest modifiable risk factor for dementia onset and is associated with all other independent risk factors for dementia onset.⁶⁹ 

The areas of cognitive decline and dementia are areas of particular concern for older persons, and have been increasing in developed nations worldwide. Although it is clear that heavy drinking can cause alcoholic dementia, the effects of low-volume consumption on dementia, including Alzheimer’s dementia, are more controversial. Summarizing the literature on the effects of low volume drinking on cognition, four lines of research are considered. One of these supports protective effects for cognitive impairment while the other three do not. In the first, the observational studies, a majority of studies find J-shaped curves, where low-volume drinking appears protective of dementia (e.g. ⁷⁰). While most of these studies find a J- or U-shaped curve for alcohol consumption and the risk of dementia, some have investigated the effects of the apolipoprotein e4 allele, where they find that carriers of this allele have an increased risk of dementia with increasing alcohol consumption.⁷¹ Nearly every review of observational (i.e., non-randomized) studies however describes methodological problems of underlying studies, such as under-representation of heavy drinkers in population-based cohorts; inconsistent measurement of alcohol use or dementia, or both; insufficient control of potential confounders; and insufficient consideration of sample attrition in patients with alcohol use disorders.

In the second, the structural studies consistently report increased rates of atrophy of the brain, and especially the hippocampus (an area of the brain associated with memory), at all levels of drinking including relatively low levels.⁷² In the third, the Mendelian randomization studies, the outcomes are mixed, but overall there is no evidence for protective effects from moderate drinking.⁷³ In the fourth, the animal studies, of which some have the advantage of having an experimental design with controls, all levels of alcohol consumption appear to be harmful to the brain and cognitive function.⁷⁴

Some research has shown a relationship between the development of cognitive impairment and lifestyle-related risk factors that are shared with other non-communicable diseases. These risk factors include physical inactivity, obesity, unbalanced diets, tobacco use and harmful use of alcohol as well as diabetes mellitus and mid-life hypertension. Other potentially modifiable risk factors more specific to dementia include mid-life depression, low educational attainment, social isolation and cognitive inactivity.⁷⁵ A randomized study of the Mediterranean diet showed protective effects for the development of dementia; however, the study was not randomized with respect to alcohol consumption and so the results were inconclusive.⁷⁶

The scientific controversy is illustrated by conflicting evidence from the Whitehall II study. On the one hand, an observational study on the Whitehall II cohort, on alcohol consumption and dementia, found a J-shaped risk curve in which low-volume drinking appeared to be protective for dementia.⁷⁷ On the other hand, the Whitehall II imaging study found increased odds of hippocampal atrophy in a dose dependent fashion, with increased risk starting at low levels of consumption.⁷⁸

A recent Swedish study, based on the Swedish twin registry adds to the growing body of research in which all levels of alcohol consumption appear detrimental to cognitive function.⁷⁹ The study used information from a sample of people in the Swedish Twin Registry, who in their midlife (1967) participated in a survey on alcohol intake and 25 years later participated in a longitudinal study on cognitive aging. This study showed that alcohol intake was related to lower cognitive performance in a dose-response manner, starting at low levels.

6.5 Injury

Injuries may result from a wide range of external causes such as road crashes, falls, near drowning, burns, or violence. Here we have considered the research evidence for three major forms of injury, road crashes, falls and physical/psychological abuse. Several meta-analyses find significant associations between alcohol use and risk of injury within six hours of drinking compared to abstainers for all ages.^80–82 For older drinkers specifically, risk of injury is also likely to be higher immediately after alcohol exposure but may also be compounded by the onset of physiological changes that occur with aging including reduced bone density, decline in balance, coordination and reaction time, vision and hearing problems and medication use.⁸³

6.6 Road crashes

It is well established that in relation to motor vehicle injury, alcohol use, even at relatively low levels, is a risk factor for drivers of all ages.^{84, 85} Road crashes are of concern in most industrialized societies and as increased wealth increases the availability of private automobiles for personal transportation, the risk of traffic crashes and resulting injuries and fatalities becomes increasingly serious across the world. Increased wealth also leads to increased resources to traffic safety work which in the longer run decreases the risks for road crashes.⁸⁶ Since safe driving is dependent upon both training and personal driver skills and ability to operate complex tasks, the aging process may affect some older driver abilities and also increase driving risks due to changes in vision and cognitive functioning (ability to reason and remember), as well as physical changes.⁸⁷ In 2002, internationally more than 193,000 traffic deaths traffic were registered among people who were 60 years and older. The mortality rate for such accidents for 100,000 persons in this age population was the highest compared to other age groups based upon a systematic review of scientific papers published between 2003 and 2013.⁸⁸ The U.S. Centers for Disease Control in reviewing national crash records for all of the U.S. found that involvement in fatal crashes per mile travelled begins increasing among drivers ages 70 – 74 years and are highest among drivers aged 85 and older in the U.S. This later trend has been attributed more to an increased susceptibility to injury and medical complications among older drivers rather than an increased risk of crash involvement.⁸⁹

This has been confirmed by one controlled laboratory study of driving simulation in which participants were given alcohol prior to the simulation. The study found in general that older adults performed more poorly with driving precision and impairment as a result of alcohol consumption measured by steering rate and the ability to maintain a constant speed than younger drivers. The study also found that one serving of alcohol was enough to affect seniors’ driving abilities in contrast to younger age groups with the same level of alcohol consumption.⁹⁰ A national study in the U.S. based upon fatal crash records examined BAC levels for drivers and found that among fatal crashes in which drivers had BAC levels less than 0.08% there was a much higher proportion of older drivers compared with the youngest aged driver group, despite the latter also being a sub-population at high crash risk.⁹¹

In short, not only are older drivers of greater risk of injury and death than all younger ages, but this accident risk is increased with the consumption of alcohol, even at very low dosages. Moreover, there is the additional risk of interaction between alcohol and psychoactive medications taken by the older driver. While younger drivers are at greater risk for crashes at a given BAC, the risk of serious outcomes is greater for older drivers at similar BAC.

6.7 Falls

For the general population research evidence on the association between falls and episodic alcohol use generally shows a linear dose-response relationship such that the more alcohol that is consumed the greater the risk of falling.⁸¹ Older people fall more often than their younger counterparts often as a consequence of chronic health conditions, impaired vision, ear problems, muscle weakness or possibly medication use. They may also have an increased fear of falling as the consequences can be more serious, long lasting and occur at times and places where they are alone.⁸³

One of the few systematic reviews of alcohol’s effect on risk of falling among older drinkers specifically (65+ years) was conducted by Ridolfo and Stevenson (1998).⁹² They noted that one of the problems with studies in this area is that they often recruit participants from nursing homes and that both self-reported alcohol use and the nature of falls among those residing in institutional settings will differ to individuals living in unsupervised situations. People living in residential aged care are likely to be more frail, take more medications and be at higher risk of falling but also less likely to consume significant quantities of alcohol. They concluded that for people aged 65 years and older, acute intoxication is a cause of 12% of male and 4% of female falls. These findings matched results from a recent experimental study showing that even low level alcohol use adversely affects postural stability among persons over 65 years, especially among those who already have poor balance.⁹³

However, some recent studies have produced mixed findings with some showing no association at all (e.g. ⁹⁴) and others showing a protective effect for falls at low doses (e.g. ⁹⁵). Research on hip fracture, osteoporosis and low bone mass density also suggests that compared to abstainers, those who regularly drink small amounts of alcohol (between 0.5 to 1 drink per day) have a lower risk of developing these conditions.^96–99 At least two studies suggest a J-shaped curve where older drinkers, particularly women, who report drinking lower amounts of wine have a significantly lower risk of falls and hip fracture compared to non-drinkers^{100, 101} but that risk increases at 2 or more drinks per day.⁹⁶ However, these observational studies are subject to the same kinds of design limitations described in earlier and apparent protective effects at low doses are likely to be due to uncontrolled confounding and bias.

6.8 Physical and psychological abuse of older people

Physical and/or psychological abuse of older people by others (‘elder abuse’) has gained attention as a significant societal problem in more recent years. A large meta-analysis of data collected from 28 countries estimated that about one in six people aged 60 years or older had experienced elder abuse of some form including psychological, financial, neglect, physical and sexual.¹⁰² Elder abuse may be even higher among vulnerable dementia patients with estimates ranging from about 28% to 52%.¹⁰³ As average population age continues to increase worldwide, the prevalence of elder abuse is also likely to grow rapidly.

Alcohol use by either the victim or the abuser or both has been repeatedly identified as a key risk factor in elder abuse. A 2012 Swedish survey of 20,000 men and women aged between 60 and 74 years found that 4.2% of participants (5.8% of women and 2.7% of men) had been a victim of violence in the past year, often from a current partner or family member. Older people who drank at an ‘at-risk level’ (according to the AUDIT) were more than four times more likely to have been a victim of violence compared to low risk drinkers.¹⁰⁴

A recent review by Warmling et al (2017) identified a range of risk factors for intimate partner violence among older people that included alcohol use, depression, low income, functional impairment and previous exposure to violence.¹⁰⁵ A systematic review of types of physical injuries among general elder abuse cases identified alcohol and/or drug abuse by perpetrators as a major factor and suggested that physical elder abuse occurs more commonly in the evenings and on weekends when greater social interaction and increased perpetrator alcohol occurs.¹⁰⁶

Analysts have nevertheless observed that few studies consider alcohol use among abused older persons as a risk factor. It is plausible that this reflects societal stereotypes of older people that make alcohol consumption less likely to be taken into consideration. Ahnlund and colleagues have recommended that workers in the caring professions should routinely inquire into the experiences of violence among older women and men.¹⁰⁴

6.9 Depression and suicide

Heavy alcohol use is linked to the development of depression; in addition, alcohol use can exacerbate depression or complicate the treatment of depression. Depression, in turn, is closely linked with suicide attempts and completed suicides. Acute impairment from alcohol can induce transient dysphoria, a loss of impulse control and a weakening of executive functioning, all of which can increase the likelihood of attempting suicide among persons who may be depressed or despondent. Rates of suicide have risen in many developed countries in recent years.

¹⁰⁸ Among older persons who completed or attempted suicide, 10 - 20% have alcohol in their bloodstream at the time of the episode.^{108, 109}

6.10 Alcohol use disorders

Alcohol use disorders (AUD) are the most important neuropsychiatric conditions caused by alcohol. Prolonged heavy use of alcohol is the leading risk factor for alcohol dependence. Alcohol has adverse effects on all organ systems, including a number of functions of the brain underlying the development of dependence. The prevalence of AUD is reduced with increasing age, most likely as a result of reduced heavy drinking. Among those who continue to drink however, or increase their consumption, the risk of AUD may be increased as older people are more vulnerable to alcohol related harm from a given volume of alcohol than other age groups. This increased vulnerability may be related to the age-related increased stress produced by alcohol.^110–112

6.11 Other conditions

There is a range of other chronic conditions that occur more frequently with age, often diminishing quality of life, that are partially attributable to or exacerbated by alcohol use. Respiratory diseases such as pneumonia are common causes of morbidity and mortality among older people.¹¹³ Due to associated cardiopulmonary disease and impaired immunity, pneumonia in particular, is more likely to be fatal or have severe consequences when contracted by older compared to younger people.¹¹⁴ A systematic review by Samokhvalov (2010) reported a dose-response relationship between chronic alcohol use and risk of community acquired pneumonia.¹¹⁵ This review also reported that people with alcohol use disorder had more than eight times higher risk of contracting pneumonia compared to non-drinkers.

Gout is a relatively common and debilitating condition and prevalence may be rising, particularly in developed countries, as a result of rising obesity and hypertension.^{116, 117} Two reviews identify a positive relationship between regular alcohol use and gout, with particularly large increases in risk at higher levels of use.^{118, 119} People with gout are also more likely to suffer from osteoarthritis.¹²⁰ Osteoarthritis is a health concern common to older people characterized by painful, progressive degeneration of joint cartilage (e.g. knees, hips). Although we were unable to identify any systematic reviews on whether alcohol use increases the risk of developing osteoarthritis or worsening of the condition, there is growing evidence that disease progression is correlated with oxidative stress and inflammatory processes. An experimental study exposed mice to low/moderate doses of alcohol for eight weeks and found a pathological role for chronic alcohol exposure in the development of osteoarthritis in knee joints compared to control mice.¹²¹ Although these results need to be confirmed in human clinical trials they suggest that for osteoarthritis sufferers, avoiding environmental factors, such as alcohol, that potentially promote oxidative stress and inflammation may be helpful.

Parkinson’s Disease is neurodegenerative disease occurring in the brain that is more common among older ages. Parkinson’s has severely debilitating symptoms and often results in reduced life quality and life expectancy.^{122, 123} A review of 16 studies found a weak protective association between alcohol use and Parkinson’s disease but as noted elsewhere in this report, in terms of strength of design this literature is particularly weak and the association may be due to uncontrolled confounding, selection and recall bias.¹²⁴

Sleep disorders are common among the general population and may become more noticeable or bothersome as people age.¹²⁵ Alcohol at all levels reduces sleep onset and may also disrupt sleep-related physiology and exacerbate day time sleepiness. Alcohol consumed before bedtime has an adverse effect on sleep overall by reducing sleep quality, even if it does help some people to fall asleep.^126–12