Critique 164: Why are the harmful effects of alcohol consumption greater among people with low education and income? — 14 May 2015

Jones L, Bates G, McCoy E, Bellis MA.  Relationship between alcohol-attributable diseaseand socioeconomic status, and the role of alcohol consumption in this relationship: a systematic review and meta-analysis.  BMC Public Health 2015;15:400.  DOI 10.1186/s12889-015-1720-7

Authors’ Abstract

Background: Studies show that alcohol consumption appears to have a disproportionate impact on people of low socioeconomic status. Further exploration of the relationship between alcohol consumption, socioeconomic status and the development of chronic alcohol-attributable diseases is therefore important to inform the development of effective public health programmes.

Methods: We used systematic review methodology to identify published studies of the association between socioeconomic factors and mortality and morbidity for alcohol-attributable conditions. To attempt to quantify differences in the impact of alcohol consumption for each condition, stratified by SES, we (i) investigated the relationship between SES and risk of mortality or morbidity for each alcohol-attributable condition, and (ii) where, feasible explored alcohol consumption as a mediating or interacting variable in this relationship.

Results: We identified differing relationships between a range of alcohol-attributable conditions and socioeconomic indicators. Pooled analyses showed that low, relative to high socioeconomic status, was associated with an increased risk of head and neck cancer and stroke, and in individual studies, with hypertension and liver disease. Conversely, risk of female breast cancer tended to be associated with higher socioeconomic status. These findings were attenuated but held when adjusted for a number of known risk factors and other potential confounding factors. A key finding was the lack of studies that have explored the interaction between alcohol-attributable disease, socioeconomic status and alcohol use.

Conclusions: Despite some limitations to our review, we have described relationships between socioeconomic status and a range of alcohol-attributable conditions, and explored the mediating and interacting effects of alcohol consumption where feasible. However, further research is needed to better characterise the relationship between socioeconomic status alcohol consumption and alcohol-attributable disease risk so as to gain a greater understanding of the mechanisms and pathways that influence the differential risk in harm between people of low and high socioeconomic status.

Forum Comments

Epidemiologists have noted for many years that one of the key factors that modifies the relation between alcohol intake and many diseases is the education, income, or other index of socio-economic status (SES) of subjects.  While differences in drinking practices (regular moderate versus binge drinking), other associated lifestyle factors, access to health care, etc., have been suggested as potential mechanisms, there has been little research directed at this association.

The present study describes this relation and attempts to ascertain the reasons why subjects at higher levels of socio-economic status appear to have fewer adverse effects of alcohol consumption.  It summarized data from 31 case-control or cohort studies, relating an overall measure of the effects of low SES (variously defined) to the risk of cancers related to alcohol, as well as to liver disease, hypertension, stroke, epilepsy, cardiac arrhythmias, and pancreatitis.

The key findings of the study demonstrate a tendency for greater risk for low-SES subjects than high-SES subjects to develop a number of diseases that are associated with alcohol consumption.  The authors suggest that some of the effects may relate to greater amounts of alcohol consumption by lower-SES subjects, or due to their higher levels of smoking.

Unadjusted results in the paper indicate a significant increase for lower SES for head and neck cancer and for stroke, while the risk of breast cancer was greater among high-SES women.  When adjusted estimates were used, the higher breast cancer risk for upper-SES women was no longer significant, but the greater risks for head and neck cancers and stroke among lower-SES subjects remained.

While the authors provide a good summary of the relation between SES and diseases, they admit that adequate data to judge the cause of these differences are not yet known.  In their discussion, they mention several possible mechanisms for poorer results in lower-SES subjects:  (1) different drinking patterns, with more binge drinking, (2) clustering of poor lifestyle factors, and (3) poorer access to health care.  They also mention that availability of social support and drinking context, such as when and with whom drinking occurs, as well as neighborhood deprivation may play a role.  Thus, while answers are not yet apparent, it is important to note that SES affects many diseases, and seems to modify the association between alcohol and certain cancers and stroke.

Specific comments on the paper by Forum reviewers:  Forum member Finkel commented: “This paper on how socioeconomic status modifies alcohol effects is of great value in focusing our attention on a critically important aspect of our interest in alcohol research — the influences of factors other than alcohol on alcohol’s effects.  This is clearly most difficult to quantify, or even to decide whether or not a given factor is or is not a confounder.  More concrete, thus more useful, information must be developed by further studies along the same lines.”

Reviewer Skovenborg agreed:  “I agree with Finkel’s comments. The issue of confounding in general and confounding of socioeconomic status in particular in alcohol research is at the same time very important to consider and very difficult to solve.  The authors write: ‘Pooling of the unadjusted findings for these studies showed a significant positive association between risk of breast cancer and high SES.’  What the authors found was a significant negative association between low SES and risk of breast cancer; that association did not remain significant with adjustment for confounding.  A possible explanation might be a difference in intake of folic acid vitamin supplements according to level of education.  However, I do like the conclusion of the authors: ‘It is unclear whether the lack of evidence on the interaction between SES and alcohol consumption implies that there is evidence of a lack of significant interactions for the conditions examined.’”

Skovenborg added: “As commented upon by the authors, a previous study in Germany (Kropp et al) found that the risk of breast cancer associated with an alcohol intake ≥19 g/d was increased in women with low levels of education [OR 3.70 (1.23 – 11.15)] but not for women with high education [OR 0.70 (0.39 – 1.27)].  More highly educated women in Germany have been reported to have a higher intake of folate, which may be a factor (Heseker et al).”   Reviewer Barrett-Connor was especially interested in the findings regarding folic acid intake and its association with level of education, which suggests that alcohol risk/benefit is related to specific nutrients.

Forum member Van Velden commented: “It is clear that alcohol consumption has different effects on people of different socioeconomic and education status.  In South Africa, people of low socioeconomic status drink alcohol of poor quality to ‘escape’ from reality, usually not with a healthy meal.  More affluent people with a higher education level drink alcohol, usually good quality wine, with a balanced meal as part of a healthy and responsible lifestyle, and seldom have binge drinking.

“There are just too many confounders to make any meaningful conclusions.  But it is obvious that we need to focus more on education.  We also need to clamp down on the distribution of alcoholic beverages of poor quality, sold in plastic containers to the poor.  This is a challenge that cannot be addressed by the industry and health authorities alone.”

Reveiwer Ellison was not too sure about differential effects according to the “quality” of the beverage, if this is a reference to price.  “For example, in terms of wine, the levels of alcohol and polyphenols in a wine do not relate to the cost of the beverage.  Certain types of wine, or those from certain parts of the world, are higher in key polyphenols (Burns et al, Corder et al), but wines that are ‘higher quality’ in terms of taste and ratings by experts may not have greater health benefits.  It is more likely that differences relate to the amount of alcohol consumed, or especially, to the drinking pattern of individuals.”

Reviewer Goldfinger stated; “We know that lower sociological class is associated with greater risk of all types of disease, both potentially alcohol-related and alcohol-independent diseases. This is an unfortunate truth and thus the focus on alcohol in this paper becomes irrelevant.”   Forum member Svilaas agreed: “There are too many confounders to make a conclusion about this topic.”

Reviewer Thelle also had difficulties with this paper.  “First, most chronic disorders are socially unequally distributed with only a few exceptions.  Breast cancer is one.  Second, alcohol consumption is also skewed with the more affluent drinking more.  Add to this that the origin of the British way to socially classify people was originally based upon the mortality rates.  Those with the highest death rates were considered belonging to the lowest social classes, and the population was categorized accordingly.  So, in some sense the association between assigned social class and disease is a self-fulfilling prophecy.  Social class induces an array of confounding variables, and I am not sure whether this paper contributes further to our understanding of the association between alcohol and health.”

Forum members noted that factors other than drinking patterns may be a cause for the greater adverse effects among subjects with lower SES.  This has been suggested by Mäkelä & Paljärvi, who reported on such differences among Finns, concluding: “Compared with non-manual workers, manual workers had a 2.06-fold hazard of alcohol-related death or hospitalisation.  Adjustment for drinking patterns explained only a small fraction of the excess hazard among manual workers.  Additionally, in each category of total consumption and in each level of the volume drunk in heavy drinking occasions, the risk of alcohol-related death and hospitalisation was higher for manual than for non-manual workers.”

Forum member Lanzmann-Petithory also thought that the type of beverage consumed may play a role: “Ruidavets et al reported that ‘the typical pattern in middle-aged men in France, is associated with a low risk of ischemic heart disease, whereas the binge drinking pattern more prevalent in Belfast confers a higher risk . . . only wine drinking was associated with a lower risk of hard coronary events, irrespective of the country.’”  Lanzmann-Petithory points out that the authors of the present paper (Jones et al) made no difference between the different types of alcoholic beverages, wine, beer, or spirit.  She added: “This should be precisely one of the key criteria.  Moreover, they made an almost pure Anglo-Saxon Scandinavian analysis: only 7 Mediterranean or wine-drinking countries studies around the 31 selected.  But relation between alcoholic beverages drinking and SES varies considerably by country of origin, by culture, especially between Anglo-Saxon countries and Mediterranean countries, following the availability and then price of different alcoholic beverages.  Wine is a luxury in some northern countries, as shown by Johansen, whereas in our studies in Nancy, France, unlike most of the studies mentioned, wine drinkers were not higher socio-economic status than other drinkers, and middle class and workers were even more likely to drink wine than the highest social status (Renaud et al).  With such shortcomings, this study cannot solve the problem of interactions between SES and the relation alcohol consumption to diseases.”

Reviewer de Gaetano had comments as well: “Recent studies from our group have shown that the current economical crisis is associated with a significant decrease of adherence to the Mediterranean Diet in a Southern Italian population (the Moli-sani study).  Such a decline is more obvious in less affluent people and is associated with higher prevalence of obesity.  Thus, adjustment for the dietary habits associated with alcohol consumption is essential.  Another  limitation of the analysis here discussed  is that only few studies reported a dose-related alcohol consumption.  The common confusion between alcohol abuse  (or high alcohol consumption) and alcohol consumption tout court is present in this paper too.”

Insights into the topic from a new publication from Denmark:  Forum member Skovenborg  identified a new publication that adds information on the topic of inequaltiy and disease.  “A new Danish study has found that smoking and alcohol use are the main explanation for the increase in social inequality in mortality since the mid-1980s (Koch et al).  The question about what drives the health gap is even more intriguing in a Nordic welfare state such as Denmark, where income inequality is low, and has declined through most of the twentieth century.  A major strength of the Koch et al study is that it is based on nationwide registers and therefore includes the whole adult Danish population aged 30 years or more over a period of 25 years.

“In the oldest Danish population, exceeding the sensible drinking limits is more frequent among the most highly educated part of the Danish population.  Even though a social gradient is seen in alcohol intake, the consequences from alcohol use, for example, hospitalization and mortality, is negatively associated with social position. The fact that alcohol-related mortality is negatively associated with social position might be due to social differences in drinking patterns, because a higher rate of non-frequent drinking is found among the Danish population with the lowest level of education.  The fact that alcohol plays an important role and that high alcohol consumption among the middle aged and the elderly is still more prevalent among the most highly educated also indicate that there might not only be a differential exposure but also a differential susceptibility to the health effects of alcohol.”  After seeing this paper, Reviewer Thelle added: “There is a susceptibility issue here, most likely due to effect modification of hitherto unrecognized factors, albeit not likely to be unknown. There is more to do.”

References from Forum review

Burns J, Mullen W, Landraul N, Teissedre P, Lean MEJ, Crozier A.  Variations in the profile and content of anthocyanins in wines made from cabernet sauvignon and hybrid grapes.  Journal of Agricultural and Food Chemistry 2002;50:4096-4102.

Corder R, Mullen W, Khan NQ, Marks SC, Wood EG, Carrier MJ, Crozier A.  Oenology: red wine procyanidins and vascular health.  Nature 2006;444:566.  doi:10.1038/444566a.

Heseker H, Adolf T, Eberhardt W, et al.  VERA-Schriftenreihe, Band III: Lebensmittel- und Naehrstoffaufnahme Erwachsener in der Bundesrepublik Deutschland. Niederkleen, Germany: Wissenschaftlicher Fachverlag Dr. Fleck, 1994.

Johansen D, Friis K, Skovenborg E and Gronbaek M.  Food buying habits of people who buy wine or beer : cross sectional study.  Br Med J 2006;332:519-522.

Koch MB, Diderichsen F, Grønbæk M, et al. What is the association of smoking and alcohol use with the increase in social inequality in mortality in Denmark? A nationwide register-based study. BMJ Open 2015;5:e006588. doi:10.1136/bmjopen-2014-006588

Kropp S, Becher H, Nieters A, Chang-Claude J.  Low-to-moderate alcohol consumption and breast cancer risk by age 50 years among women in Germany.  Am J Epidemiol 2001;154:624–634.

Mäkelä P, Paljärvi T.  Do consequences of a given pattern of drinking vary by socioeconomic status? A mortality and hospitalisation follow-up for alcohol-related causes of the Finnish Drinking Habits Surveys. J Epidemiol Community Health 2008;62:728-33.

Renaud SC, Gueguen R, Schenker J, et al. Alcohol and mortality in middle-aged men from eastern France. Epidemiology 1998;9:184–8.

Ruidavets J-B, Ducimetière P, Evans A, Montaye M, Haas B, Bingham A, Yarnell J, Amouyel P, Arveiler D, Kee F, Bongard V, Ferrières J.   Patterns of alcohol consumption and ischaemic heart disease in culturally divergent countries: the Prospective Epidemiological Study of Myocardial Infarction (PRIME).  BMJ 2010;341:c6077.

Johansen D, Friis K, Skovenborg E and Gronbaek M.  Food buying habits of people who buy wine or beer : cross sectional study.  Br Med J 2006;332:519-522.

Forum Summary

Epidemiologists have noted for many years that one of the key factors that modifies the relation between alcohol intake and many diseases is the education, income, or other index of socio-economic status (SES) of subjects.  While differences in drinking practices (regular moderate versus binge drinking), other more moderate lifestyle factors, better access to health care, etc., have been suggested as potential mechanisms, there has been little research directed at this relation.

The present study is a systematic review of published reports to investigate the relationship between SES and risk of mortality or morbidity for each alcohol-attributable condition and, where feasible, to explore alcohol consumption as a mediating or interacting variable in this relationship.  It summarized data from 31 case-control or cohort studies, relating an overall measure of the effects of low SES (variously defined) to the risk of cancers related to alcohol, as well as to liver disease, hypertension, stroke, epilepsy, cardiac arrhythmias, and pancreatitis.

The key findings of the study demonstrate a tendency for greater risk for low-SES subjects than for high-SES subjects to develop a number of diseases that are associated with alcohol consumption.  Specifically, the authors report higher and statistically significant increases in risk of head and neck cancer and of stroke associated with alcohol consumption among lower-SES subjects than among higher-SES subjects.  There was a tendency for lower risk of breast cancer among lower-SES women, but differences were not significant when adjusted for known confounders.  The authors suggest that some of the effects shown may relate to greater total amounts of alcohol consumption by lower-SES subjects, or due to their higher levels of smoking.

Forum members thought this was a valiant attempt of the authors to unravel what has been a mystery: why do people at lower levels of SES seem to have more adverse effects of alcohol consumption than subjects at higher levels of SES.  Unfortunately, as the authors admit, current scientific data do not allow firm answers.  Suggested theories for reasons why low-SES subjects may have more adverse effects from alcohol include (1) different drinking patterns, with more binge drinking, (2) clustering of poor lifestyle factors, and (3) less access to health care.  In some studies, even though the reported total alcohol intake of low- and high-SES subjects may be similar, the latter seem more likely to drink moderate amounts of alcohol on a regular basis, while lower-SES subjects are more likely to binge drink on fewer days per week.  However, other studies do not support such a finding.  Further, there may be differences according to the type of alcoholic beverage consumed, which was not considered in this study.

Forum members pointed out that differences in factors other than drinking patterns probably contribute to the greater adverse effects among subjects with lower SES.  One member suggested that lower SES affects susceptibility to many diseases, most likely due to effect modification of hitherto unrecognized factors.  The Forum concluded that there are too many confounders to make a conclusion about this topic, and much more research will be needed to understand how socio-economic factors affect the risk of disease and may modify the relation between alcohol consumption and disease.

*                     *                      *

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

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

Dag S. Thelle, MD, PhD, Senior Professor of Cardiovascular Epidemiology and Prevention, University of Gothenburg, Sweden; Senior Professor of Quantitative Medicine at the University of Oslo, Norway

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

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

Dominique Lanzmann-Petithory,MD, PhD, Nutrition/Cardiology, Praticien Hospitalier Hôpital Emile Roux, Paris, France

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

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

Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA

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

Elizabeth Barrett-Connor, MD, Chief of the Division of Epidemiology, Distinguished Professor in the Departments of Family and Preventive Medicine & Medicine, University of California, San Diego, La Jolla, CA, USA