Critique 251: Does moderate drinking increase, or decrease, the risk of falls in middle-age to elderly adults? 1 November 2021

Tan GJ, Tan MP, Luben RN, Wareham NJ, Khaw K-T, Myint PK. The relationship between alcohol intake and falls hospitalization: Results from the EPIC-Norfolk. Geriatr Gerontol Int 2021;21:657-663.

Authors’ Abstract

Aim: To evaluate the relationship between habitual alcohol consumption and the risk of falls hospitalization.

Methods: The EPIC-Norfolk is a prospective population-based cohort study in Norfolk, UK. In total, 25,637 community dwelling adults aged 40–79 years were recruited. Units of alcohol consumed per week were measured using a validated Food Frequency Questionnaire. The main outcome was the first hospital admission following a fall.

Results: Over a median follow-up period of 11.5 years (299 211 total person years), the cumulative incidence function (95% confidence interval) of hospitalized falls at 121–180 months or non-users, light (>0 to ≤7 units/week), moderate (>7 to ≤28 units/week) and heavy (>28 units/week) were 11.08 (9.94–12.35), 7.53 (7.02–8.08), 5.91 (5.29–6.59) and 8.20 (6.35–10.56), respectively. Moderate alcohol consumption was independently associated with a reduced risk of falls hospitalization after adjustment for most major confounders (hazard ratio = 0.88; 95% confidence interval 0.79–0.99). The relationship between light alcohol consumption and falls hospitalization was attenuated by gender differences. Alcohol intake higher than the recommended threshold of 28 units/week was associated with an increased risk of falls hospitalization (hazard ratio 1.40 [1.14–1.73]).

Conclusions: Moderate alcohol consumption appears to be associated with a reduced risk of falls hospitalization, and intake above the recommended limit is associated with an increased risk. This provides incentive to limit alcohol consumption within the recommended range and has important implications for public health policies for aging populations.

Forum Comments

Forum members consider that this is a well-done analysis to evaluate whether or not people who consume alcohol are more or less likely than non-drinkers to have falls requiring hospitalization. Data used in the analysis are from the EPIC-Norfolk study of 25,000 subjects, initially 40-79 years of age, followed for 11.5 years. These analyses are based on a “unit of alcohol” being 10 ml (8 gm of alcohol) (versus the typical value of a drink averaging 12-14 grams of alcohol, frequently used in American studies). The categories of alcohol consumption were none, low amounts (>0 but ≤7 units/week), moderate (> 7 to 28 units/week, equal to 56 – 228 g of alcohol/week), or heavy (> 28 units/week). There were adequate numbers in each of the alcohol categories, with approximately 3,600 abstainers, 13,000 low drinkers, 7,300 moderate, and 959 heavy drinkers.

Estimates of the relation of alcohol intake at baseline were related to the subsequent risk of hospitalization for a fall. Estimates of effect were adjusted for a long list of potential confounders (including age, gender, smoking, educational level, level of social deprivation, antidepressant use, level of physical activity, BMI, history of stroke or diabetes, and dietary factors).

The key results of the study were that 19.2 % of non-drinkers were hospitalized for a fall during follow up, compared with 13.7% of low drinkers, 10.9 % of moderate drinkers, and 12.3% of heavy drinkers.   In addition, the cumulative incident rates at 121-180 months of follow up were 11.08% for teetotalers, 7.53% for low drinkers, 5.91% for moderate drinkers, and 8.20 % for heavy drinkers. With the final analytic model containing a large number of potential confounders, there were significantly fewer falls for moderate drinkers, and increased risk for heavy drinkers.

Comments from individual Forum members

Reviewer Skovenborg wrote: “The study is as good as an observational study can get with a very robust control for potential confounders (including physical activity, dietary factors, indices of obesity and socio-economic status) and a robust outcome measure (falls as a reason for hospital admission identified by ICD-10 codes W00-W19). The threshold for heavy alcohol consumption (>224 g of alcohol per week) is rather low for men but high for women, as a consequence of the very strange decision of the UK Chief Medical Officers’ ‘low risk drinking guidelines’ that established 14 units a week as low-level drinking for both women and men.

“I agree with other Forum members that this is an important study with important results. One of the reasons is that old age sensitivity to alcohol and an increased risk of falls are used as part of the evidence base for the US drinking guidelines to recommend that persons older than 65 consume no more than one drink per day. That recommendation was based primarily on papers by Dufour et al and McKim & Mishara, which were quoting results from a study by Forbes & Reina.

“For the limitations on alcohol intake after age 65, a loss of total body water is one of the key factors used as support for the recommendation. Evidence for a considerable reduction of total body water with age (according to Total Body Water reference values and prediction equations for adults from Chumley, et al) was that the average man will lose around 3.1 litres of body water going from 20 to 80 years of age, or about 6.9 % of total body weight (TBW) and the average woman loses 1.8 litres of body water (5.6 % of TBW). I consider the evidence is very weak for these assertions, as the basic studies used to reach these estimates involved only 4 individuals! Yet, these recommendations are used for all adults after the age of 65 years. We should instead focus on much better evidence, such as those from the EPIC-Norfolk study or other large cohort studies, to judge the validity of drinking guidelines from the UK and the USA and elsewhere.”

Forum member de Gaetano wrote: “I agree that this is an important study with important results. I noticed, however, that non-drinkers had a greater incidence of events than heavy drinkers.  This should be considered in light of the suggestion made by some ‘anti-alcohol’ investigators that non-drinkers should not be included in alcohol studies, as they may be quite different by many aspects from drinkers (at any dose). I disagree with the latter statement, but I would enjoy reading comments from other Forum members.”

Skovenborg responded: “I have 3 suggestions for a possible explanation of why the heavier drinkers in this study did not have more falls:

1. The EPIC-Norfolk has been up and running for about 25 years and there is nothing to suggest that the 3638 non-drinkers in the cohort are “a basket of deplorables”; I note however, that only about one-third of the non-drinkers were moderately active or active. That might be a sort of an explanation even though Tan et al adjusted for physical activity in their Model 8 adjustment (Model 8 = adjusted for age, physical activity, stroke, diabetes, asthma, antidepressant use, Townsend index, fish intake and fruit & vegetable consumption, gender, aspirin usage, anti-hypertensive medication use, vitamin D supplementation, systolic and diastolic blood pressure, body mass index, waist-hip ratio, and prevalent cancer). There might still be some residual confounding here.
2. The threshold for heavy drinking in men (>224 g of alcohol per week) is below the upper limit for sensible drinking in Denmark (21 units/week; 1 unit = 12 g alcohol; threshold = 252 g of alcohol per week) that existed until the sensible drinking limit was reduced to 14 units per week for men (a reduction based, somewhat surprisingly, on the alcohol-breast cancer association as argument). [As suggested by reviewer Ellison, these Norwich ‘heavy drinking men’ might not have been drinking that heavily after all.]
3. In our focus on nature and nurture we should never forget chance as a possibility.”

Reviewer Finkel added: “I, too, was intuitively surprised that this paper reports, with many of the confounders included in the analysis, that heavy drinkers were not more prone to fall more. But intuition is not science — another of our stereotypes has fallen!”

Reviewer Mattivi wrote: “I agree that the outcome for those classified as heavy drinkers is surprising and in contrast with the behavioural risk factors for falls in elder age in the current WHO Guidelines global report. In that document from WHO, where the negative factors are excessive alcohol, sedentary behaviour and multiple medications, the inclusion of alcohol was based on an unpublished paper by Scott et al.

“The results of the EPIC-Norfolk study are instead in agreement with the WHO guidelines on the conclusion that moderate alcohol consumption is a protective factor. This statement in the WHO document is based on a study by Peel, et al, who included moderate alcohol consumption in mid- and older age among the behavioural factors which had a significant independent protective effect on the risk of hip fracture (AOR: 0.40, CI 0.25-0.95). In that study moderate drinkers were those drinking no more than two standard drinks per day in older age and no more than four drinks per day in middle and younger ages for males, and half these measures for females.”

Forum member Van Velden considered that this was an interesting paper, but pointed out that moderate drinkers usually have a responsible lifestyle, with many other favorable behaviors, that may contribute to health-related issues, including the risk of falls requiring hospitalization.

Forum member Goldfinger wrote: “I concur; this is a very interesting paper with somewhat surprising results. Having said that, I agree with Van Veldens’ assessment that moderate drinkers usually have a responsible lifestyle, with many other favourable behaviours, that may contribute to a benefit with respect to health-related issues, including risk of falls requiring hospitalization.”

Reviewer Ellison added: “As stated by Skovenborg, there is always the risk of residual confounding despite the long and well-chosen list of potentially confounding factors included in this analysis. Nevertheless, this analysis certainly does not allow us to include a tendency to fall as a factor to advise the elderly to avoid any consumption of alcohol. Rather, the moderate drinkers in this study had the lowest risk of having a fall requiring hospitalization.”

References from Forum critique

Chumlea WC, Guo SS, Zeller CM, et al. Total body water reference values and prediction equations for adults. Kidney International 2001;59:2205-2258.

Dufour MC, Archer L, Gordis E. Alcohol and the elderly. Clin Geriatr Med 1992;8:127-141.

Forbes GB, Reina JC. Adult lean body mass declines with age some longitudinal observations. Metabolism 1970;19:653-663.

McKim WA, Mishara BL: Drugs and Aging. Toronto, Butterworths, 1987, p 63 (here we find the actual reference to a study by Forbes & Reina).

Peel NM, McClure RJ, Hendrikz JK. Health-protective behaviours and risk of fall-related hip fractures: a population-based case-control study. Age Ageing 2006;35:491-497.   doi: 10.1093/ageing/afl056

Scott VJ, et al. Canadian Falls Prevention Curriculum©, Vancouver, British Columbia. Canada. Injury Research and Prevention Unit (unpublished data).

WHO Global Report on Falls Prevention in Older Age, World Health Organization 2007, ISBN 978 92 4 156353 6.

Forum Summary

To judge the relation of reported alcohol consumption to the risk of subsequent falls requiring hospitalization, the authors of this paper used data from the EPIC-Norfolk study, a prospective population-based cohort study in Norfolk, UK. In that study, there were a total of 25,637 community-dwelling adults, aged 40–79 years when recruited. The main outcome was the first hospital admission due to a fall over a follow-up period of about 11.5 years.

The key results of the study were that 19.2 % of non-drinkers were hospitalized for a fall during follow up, compared with 13.7% of low drinkers, 10.9 % of moderate drinkers, and 12.3% of heavy drinkers.   In addition, the cumulative incident rates for falls at 121-180 months of follow up were 11.08% for teetotalers, 7.53% for low drinkers, 5.91% for moderate drinkers, and 8.20 % for heavy drinkers.

Forum members considered this to be a well-done study, with the analysis adjusting for a large number of factors that were potential confounders of the relation of alcohol intake to the risk of falls. We agree with the conclusions of the authors: “Moderate alcohol consumption appears to be associated with a reduced risk of falls hospitalization, and intake above the recommended limit is associated with an increased risk,” although we note that the increase in risk for heavier drinkers only appeared after essentially all covariates were added to the analysis. Obviously, residual confounding is always a possibility, but the present analyses indicating some protection against falls among light-to-moderate drinkers are supported by some well-done previous research.

The results of this study do not support the claim that any alcohol increases the risk of falls in older adults. It appears that this is not a concern for light-to-moderate drinkers who, in this study, had a lower risk of falls requiring hospitalization than did non-drinkers.

*                      *                      *                      *                      *                      *

Comments on this critique by the International Scientific Forum on Alcohol Research have been provided by the following members:

Giovanni de Gaetano, MD, PhD, Department of Epidemiology and Prevention, IRCCS Istituto Neurologico Mediterraneo NEUROMED, Pozzilli, Italy

R. Curtis Ellison, MD, Professor of Medicine, emeritus, Section of Preventive Medicine & Epidemiology, Boston University School of Medicine, Boston, MA, USA

Ramon Estruch, MD, PhD, Hospital Clinic, IDIBAPS, Associate Professor of Medicine, University of Barcelona, Spain

Harvey Finkel, MD, Hematology/Oncology, Retired (Formerly, Clinical Professor of Medicine, Boston University Medical Center, Boston, MA, USA)

Tedd Goldfinger, DO, FACC, Desert Cardiology of Tucson Heart Center, University of Arizona School of Medicine, Tucson, AZ, USA

Fulvio Mattivi, MSc, Department of Cellular, Computational and Integrative Biology – CIBIO and C3A, University of Trento, Italy

Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark

Creina Stockley, PhD, MSc Clinical Pharmacology, MBA; Adjunct Senior Lecturer at the University of Adelaide, Australia

Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway

Pierre-Louis Teissedre, PhD, Faculty of Oenology–ISVV, University Victor Segalen Bordeaux 2, Bordeaux, France

Fulvio Ursini, MD, Dept. of Biological Chemistry, University of Padova, Padova, Italy

David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa

Andrew L. Waterhouse, PhD, Department of Viticulture and Enology, University of California, Davis, USA