Critique 059: Association of quantity of alcohol and frequency of consumption with cancer mortality ———— 20 October 2011

Breslow RA, Chen CM, Graubard BI, Mukamal KJ.  Prospective study of alcohol consumption quantity and frequency and cancer-specific mortality in the US population.  Am J Epidemiol 2011; DOI: 10.1093/aje/kwr210.

Authors’ Abstract

Prospective associations between quantity and frequency of alcohol consumption and cancer-specific mortality were studied using a nationally representative sample with pooled data from the 1988, 1990, 1991, and 1997 – 2004 administrations of the National Health Interview Survey (n = 323,354).  By 2006, 8,362 participants had died of cancer. Cox proportional hazards regression was used to estimate relative risks.

Among current alcohol drinkers, for all-site cancer mortality, higher-quantity drinking (≥3 drinks on drinking days vs. 1 drink on drinking days) was associated with increased risk among men (relative risk (RR) = 1.24, 95% confidence interval (CI): 1.09, 1.41; P for linear trend = 0.001); higher-frequency drinking (≥ 3 days/week vs. <1 day/week) was associated with increased risk among women (RR = 1.32, 95% CI: 1.13, 1.55; P-trend < 0.001).  Lung cancer mortality results were similar, but among never smokers, results were null.  For colorectal cancer mortality, higher-quantity drinking was associated with increased risk among women (RR = 1.93, 95% CI: 1.17, 3.18; P-trend = 0.03).  Higher-frequency drinking was associated with increased risk of prostate cancer (RR = 1.55, 95% CI: 1.01, 2.38; P for quadratic effect = 0.03) and tended to be associated with increased risk of breast cancer (RR = 1.44, 95% CI: 0.96, 2.17; P-trend = 0.06).

Epidemiologic studies of alcohol and cancer mortality should consider the independent effects of quantity and frequency.

Forum Comments

Background:  Most epidemiologic studies have had data only on the amount of alcohol consumed in relation to health outcomes, including cancer.  In recent years, it has become increasingly clear that the pattern of drinking, especially the frequency of consumption, plays an important role in whether a given amount of alcohol has beneficial or harmful effects.  In particular, more frequent drinking, including daily consumption, has been shown to markedly reduce the risk of cardiovascular diseases, whereas the occasional consumption (as only week-end drinking), especially of large amounts (binge drinking), is associated with adverse health consequences.  Limited previous research has suggested that the benefits of more frequent consumption seen for cardiovascular disease may not apply to the risk of certain cancers, especially among women.

Comments on design of the present study:  The present study attempts to evaluate the separate and combined effects of the frequency of alcohol consumption and the average amount per occasion as they relate to the risk of mortality from all cancers, as well as cancer-specific mortality, in the United States.  It is based on a total of more than 300,000 subjects and over 8,000 deaths from cancer.  The focus was on total cancer deaths and mortality from lung, colorectal, prostate, and breast cancers.  

The classification of drinkers was as follows: none in the past year and never more than 12 drinks during their lifetime as “never drinkers;” 12 or more drinks in a year in their lifetime but fewer than 12 drinks in any previous year and none in the last year were classified as “lifetime infrequent drinkers,” but if they had consumed more than 12 drinks/year in any previous year but none in the last year, they were classified as “former drinkers.”  Having consumed at least one drink in the past year led to the classification of being a “current drinker,” with further classification according to the usual frequency and quantity to estimate “total alcohol consumption.”  Using the latter measurement, women consuming ≤ 3 drinks per week were considered to be light drinkers and > 3 to 7 drinks/week as moderate drinkers.  For men, the cut-points were ≤ 3 and > 3 to 14 drinks/week.

It is of some concern that the investigators included subjects with previous cancers in their analyses; usually, such subjects are excluded.  Also, the investigators state that binge drinking (yes/no) information was available for a subset of subjects, and in sensitivity analyses on this the results were “generally similar;” however, no specific data are presented.  A potentially important omission is the failure to present results by the type of beverage consumed.  While such information was not available for all subjects, with a dataset this large there should have been adequate numbers of subjects consuming different beverages to comment on beverage-specific effects. 

The initial analyses compared all levels of former or current drinking with never drinking, while the comparison of frequency and quantity per occasion were tested only among current drinkers (i.e., comparing heavier drinkers with lighter drinkers, rather than with abstainers).  All analyses were based only on baseline assessments of alcohol intake, as no follow-up data on alcohol consumption were available.

Comments on results of this study:  The initial results use never drinkers as the referent group.  For total alcohol consumption (frequency x quantity), the data indicate a significant reduction in the risk of all-site cancers (RR=0.87, CI 0.80-0.94) for light drinkers which, interestingly, is not pointed out by the authors and is not included in their abstract.  This result is driven by a 25% reduction in risk among women.  Moderate drinking consistently shows no effect, and only heavier drinking is associated with an increase in all-site cancer risk.  For site-specific cancers, an increase in risk of lung cancer was seen only for heavier drinkers, with a tendency for less cancer among light drinkers.  There was no evidence of an effect of total alcohol consumption on colorectal, prostate, or breast cancer in these analyses.

It is noted that outcomes among the large number of lifetime infrequent drinkers were almost the same in all comparisons as were those of lifetime abstainers.  This supports the inclusion of such subjects in a non-drinking category (which was not done in these analyses). 

For comparing the effects onmortality for separate categories of quantity and frequency of drinking, the authors excluded non-drinkers.  For all-site cancer and for lung cancer, these results again show an increase in risk only for heavier drinkers.  More frequent drinking suggests an increase in cancer risk among women but not among men.  For colorectal, prostate, and breast cancer, there is no clear pattern of an increase in risk from the quantity of alcohol consumption.  For frequency of drinking, again there is a suggestion of an increase in risk with more frequent drinking, although the trends are not statistically significant. 

We agree with the authors that both frequency and quantity per occasion of alcohol consumption should be evaluated.  The reported differences between men and women related to the frequency of consumption and disease outcomes deserve further investigation.   

Forum Summary

The present study from the National Institutes of Health attempts to evaluate the separate and combined effects of the frequency of alcohol consumption and the average quantity of alcohol per occasion as they relate to the risk of mortality from all cancers, as well as cancer-specific mortality.  It is based on repeated administrations of the National Health Interview Survey, with a total of more than 300,000 subjects and over 8,000 deaths from cancer, and reports on total cancer deaths and deaths from lung, colorectal, prostate, and breast cancers.

For total alcohol consumption (frequency x quantity), the data indicate a significant reduction in the risk of all-site cancers (RR=0.87, CI 0.80-0.94) which, interestingly, is not emphasized by the authors and is not included in their abstract.  Moderate drinking consistently shows no effect, and only heavier drinking is associated with an increase in all-site cancer risk.  For site-specific cancers, an increase in risk of lung cancer was seen only for heavier drinkers, with a tendency for less cancer among light drinkers.  There was no evidence of an effect of total alcohol consumption on colorectal, prostate, or breast cancer.

The authors excluded non-drinkers in a second analysis in which they used categories of usual daily quantity and of frequency of consumption in an attempt to investigate their separate effects.  For all-site cancer and for lung cancer, these results again show an increase in risk only for drinkers reporting greater amounts of alcohol.  The data also show an increase in cancer risk from more frequent drinking among women but not among men.  For colorectal, prostate, and breast cancer, there is no clear pattern of an increase in risk from quantity of alcohol consumed.  For frequency of drinking, again there is a suggestion of an increase in risk with more frequent drinking, although the trends are not statistically significant.  

The overall message of this analysis is that light to moderate alcohol intake does not appear to increase the risk of all-site cancer (and light drinking was shown in this study to be associated with a significant decrease in risk).  Similarly, light to moderate consumption was not associated with site-specific cancers of the lung, colorectum, breast, or prostate.  Heavier drinking is known to be associated with a large number of adverse health effects, including certain cancers, as was shown in this study. 

When considering cancer, alcohol consumption should not be considered in isolation, but in conjunction with, other lifestyle behaviors (especially smoking).  We agree with the authors that both quantity and frequency of consumption need to be considered when evaluating the relation of alcohol to cancer; further, beverage-specific effects need to be further evaluated.   

                                                                                    *                      *                      *

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

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

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

Giovanni de Gaetano, MD, PhD, Research Laboratories, Catholic University, Campobasso, Italy

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

Francesco Orlandi, MD, Department of Gastroenterology, Università degli Studi di Ancona, Italy

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

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

Gordon Troup, MSc, DSc, School of Physics, Monash University, Victoria, Australia