Critique 188: Alcohol-attributable cancer in New Zealand . . . 28 June 2016
Connor J, Kydd R, Maclennan B, Shield K, Rehm J. Alcohol-attributable cancer deaths under 80 years of age in New Zealand. Drug and Alcohol Review 2016; pre-publication. DOI: 10.1111/dar.12443
Introduction and Aims. Cancer deaths made up 30% of all alcohol-attributable deaths in New Zealanders aged 15–79 years in 2007, more than all other chronic diseases combined. We aimed to estimate alcohol-attributable cancer mortality and years of life lost by cancer site and identify differences between Māori and non-Māori New Zealanders.
Design and Methods. We applied the World Health Organization’s comparative risk assessment methodology at the level of Māori and non-Māori subpopulations. Proportions of specific alcohol-related cancers attributable to alcohol were calculated by combining alcohol consumption estimates from representative surveys with relative risks from recent meta-analyses. These proportions were applied to both 2007 and 2012 mortality data.
Results. Alcohol consumption was responsible for 4.2% of all cancer deaths under 80 years of age in 2007. An average of 10.4 years of life was lost per person; 12.7 years for Māori and 10.1 years for non-Māori. Half of the deaths were attributable to average consumption of <4 standard drinks per day. Breast cancer comprised 61% of alcohol-attributable cancer deaths in women, and more than one-third of breast cancer deaths were attributable to average consumption of <2 standard drinks per day. Mortality data from 2012 produced very similar findings.
Discussion and Conclusions. Alcohol is an important and modifiable cause of cancer. Risk of cancer increases with higher alcohol consumption, but there is no safe level of drinking. Reduction in population alcohol consumption would reduce cancer deaths. Additional strategies to reduce ethnic disparities in risk and outcome are needed in New Zealand.
While alcohol consumption of all levels is associated with risk for only a relatively small percentage of total cancer deaths (4.2% in this study), there appear to be an increasing number of articles on the dangers of alcohol in terms of cancer. The present study compares population-based surveys of alcohol intake to population statistics for deaths from a number of types of cancer in New Zealand, and attempts to judge whether there are differences in estimates of risk for Māori and non- Māori subjects.
Forum members had a number of concerns with this paper. Data on exposure to alcohol (or other risk factors) for individual subjects were not available for those who did, or did not, die of cancer. The method used was to base estimates of the proportion of the risk of each type of cancer, shown in other studies to be attributable to alcohol consumption, to the New Zealand population. Obviously, the attributable proportion related to any exposure will depend on what estimates are chosen. As stated by Forum member McCormick: “The quality of any such analysis will depend on the veracity of the ‘facts’ underpinning it and the assumptions used. The drinking patterns of Māori and non- Māori in New Zealand have been shown to be very different (Bramley et al), and the application of the same ‘attributable’ estimates to both groups may not be justified.”
Other Forum members noted that the study does not account for the type of alcohol consumed, the role of diet or social and economic factors, data that would be necessary to interpret the disparity in cancer deaths among different ethnic groups in New Zealand. As added by reviewer Thelle, “The investigators apply relative risk estimates from meta-analyses, and that is fair enough, but the study is essentially an ecological study and not a population-based study on single individuals. No confounders are taken into account.”
Forum member Barrett-Connor summarized the many difficulties in interpreting this study: “Conclusions are based on very small population isolates, with perhaps atypical drinking habits, or unknown diet & lifestyle. Thus I cannot interpret the methods of the New Zealand study — it is too small a sample of people whose culture, lifestyle, and diet are not well described, and it is impossible to interpret the results. Questions remain, such as ‘Do the participants work outside all day performing heavy physical labor?’ ‘What types of alcohol did they drink?’ ‘How else do the Māori and non- Māori subjects differ?’ I cannot comment on statistics either, as there is not enough information about other behaviors/lifestyle.”
It was noted that the paper does not refer to a recent article by Rehm et al that admits that previous estimates of population-attributable risk between alcohol and adverse cardiovascular outcomes have been markedly over-estimated in the past. We are not aware of updates of alcohol-attributable estimates regarding cancer. In all the tables and figures in the present paper, no data are presented for the number of non-drinkers who developed the various cancers.
As pointed out by Forum member Thelle: “The main issue in the data set is of course the low numbers without any suggestion of the random errors around these figures. The investigators should have shown us the population at risk for exposed and unexposed, and the corresponding incidence rates in order to estimate the relative risk. With these low numbers it is likely that there is large variation affecting the confidence intervals.”
In this study the type of beverage or the pattern of drinking (regular, in binges, etc.) are not known for the subjects who developed (or did not develop) cancer. As reviewer Stockley noted: “The authors suggest in their discussion that ‘recent evidence supports little impact of drinking pattern on the cancer risk for light to moderate drinkers’ (referring to an article by Cao et al). However, Cao et al state in their conclusions: ‘Light to moderate drinking is associated with minimally increased risk of overall cancer. For men who have never smoked, risk of alcohol related cancers is not appreciably increased for light and moderate drinking (up to two drinks per day). However, for women who have never smoked, risk of alcohol related cancers (mainly breast cancer) increases even within the range of up to one alcoholic drink a day.’”
Stockley continues: “Allen et al conclude that: ‘Low to moderate alcohol consumption in women increases the risk of certain cancers. For every additional drink regularly consumed per day, the increase in incidence up to age 75 years per 1000 for women in developed countries is estimated to be about 11 for breast cancer, 1 for cancers of the oral cavity and pharynx, 1 for cancer of the rectum, and 0.7 each for cancers of the esophagus, larynx and liver, giving a total excess of about 15 cancers per 1000 women up to age 75.’”
Reviewer Keil had a number of observations: “The authors do not provide us with the population at risk for exposed and unexposed, and therefore the calculations presented are somehow ‘in the air.’ I calculated 5,786 cancer deaths in New Zealand for 2007 of which 243 are attributable to alcohol (4.2%). It would have been much better had the authors provided this figure from the official mortality statistics of New Zealand. Again, one can calculate 507 breast cancer deaths in New Zealand women, 71 of these 507 attributable to alcohol of all intake levels. The alcohol-attributable factor of 14 % is much too high according to my knowledge of the literature; this figure is more around 5-6 %. But again, it is difficult to follow the calculations. The absolute number of breast cancers in New Zealand women occurring in the alcohol intake group < 20 grams of alcohol per day turns out to be 26, according to the authors’ calculations. Any individual case of breast cancer and cancer death is one too many, but I do not consider this absolute number of 26 a major public health hazard for New Zealand, especially in light of the 5,786 total cancer deaths in 2007.”
Forum member Skovenborg also had a number of concerns about the paper: “According to ‘Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective, 2007,’ cirrhosis is an essential precursor of liver cancer caused by alcohol. Nevertheless data presented in this paper attributes 11 deaths of liver cancer in men with a consumption of >0 – <40 g/day and 2 deaths of liver cancer in women with a consumption of >0 – <20 g/day, although development of liver cirrhosis is unlikely at these modest consumption levels. The authors of this paper did not control for smoking as a confounder. In the ‘Million Women Study’ (Allen et al), increasing alcohol intake was not associated with an increased risk of cancers of the upper aero-digestive tract in never smokers or past smokers, but was strongly associated with an increased risk among current smokers.” Skovenborg also noted that the authors state that recent evidence supports little impact of drinking pattern on the cancer risk for light to moderate drinkers. He adds: “In The Danish Nurse Cohort Study, however, weekend consumption and binge drinking seemed to be related to an additional increased risk of breast cancer.”
Reviewer Ellison was especially concerned about the lack of information on the effects of alcohol on other conditions: “A key weakness of this paper is that it makes no mention of the effects of alcohol on total mortality. As has been repeatedly stated, before disease-specific data can be used for advising the public about alcohol intake, it is important to present the full picture, which almost always shows lower total mortality risk for moderate drinkers. Further, it is troubling that all drinkers up to 40 grams/day for men (3 ½ to 4 typical drinks) and up to 20 grams/day for women (about 1 ½ to 2 drinks) are included in the lowest alcohol category, and no reference is made to men or women with consumption at levels recommended as ‘safe’ in many countries, including the USA (up to 2 drinks/day for men and up to 1 drink/day for women). Thus a true dose-response curve cannot be estimated.” Forum member Thelle agreed: “The authors have not collected the data themselves but relied upon the New Zealand health surveys; these surveys are very detailed with regard to alcohol consumption, and should enable the authors to use a more graded scale. The category 0.01-39.9 covers around 60% of the population and there must be a large variation within that group.”
The very long Discussion is primarily a treatise on how the public must be told of the dangers of cancer from any alcohol consumption; it focuses on health policy recommendations and very little on the limitations of the data and the study. The authors end up making very broad recommendations based on very small numbers of subjects. One reviewer referred to it as more of a “polemic” than science.
References from Forum review
Allen NE, Beral V, Casabonne D, et al. Moderate alcohol intake and cancer incidence in women. J Natl Cancer Inst 2009;101:296-305.
Bramley D, Broad J, Harris R, Reid P, Jackson R. Alcohol Burden of Disease and Disability Group. Differences in patterns of alcohol consumption between Māori and non- Māori in Aotearoa (New Zealand). N Z Med J 2003;116 (1184):U645.
Cao Y, Willett WC, Rimm EB, Stampfer MJ, Giovannucci EL. Light to moderate intake of alcohol, drinking patterns, and risk of cancer: results from two prospective US cohort studies. BMJ 2015;351:h4238.
Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective, 2007, available at http://www.cancerinstitute.org.au/cancer_inst/publications/pdfs/pm-2008-03_alcohol-as-a-cause-of-cancer.pdf.
Mørch LS, Johansen D, Thygesen LC, Tjønneland A, et al. Alcohol drinking, consumption patterns and breast cancer among Danish nurses: a cohort study. European Journal of Public Health 2007;17:624–629.
Rehm J, Shield KD, Roerecke M, Grnel G. Modelling the impact of alcohol consumption on cardiovascular disease mortality for comparative risk assessments: an overview. BMC Public Health 2016;16:363. DOI 10.1186/s1 2889-016-3026-9.
Previous scientific research has shown that heavy alcohol consumption is a major risk factor for upper aero-digestive cancers, and even light drinking increases slightly the risk of breast cancer in women. The present study is based on a very small number of cases of cancer in New Zealand, tabulated separately for Māori and non-Māori subjects, and applies estimates of alcohol effects from other population-based studies. Besides having so few cases, the investigators had no individual data (on the pattern of alcohol consumption, type of beverage, smoking or other lifestyle habits, socio-economic status, etc.) on subjects who did, or did not, develop these cancers. It is not even known whether or not the specific subjects who developed these cancers consumed alcohol.
It has repeatedly been emphasized that disease-specific death rates must be interpreted in light of the effects of the exposure (in this case, alcohol consumption) on other causes of death as well, especially on total mortality risk. The large majority of studies have shown that moderate drinking clearly reduces the risk of most cardiovascular diseases, diabetes, and other of the diseases of ageing, as well as the risk of total mortality. However, the estimated effects of alcohol on total mortality are not included in this paper.
The very long Discussion in this paper is primarily a treatise on how the public must be told of the dangers of cancer from any alcohol consumption; it focuses on health policy recommendations and very little on the limitations of the data and the study. The authors end up making very broad recommendations based on very small numbers of subjects. Limitations to this study suggest that it adds little to our current understanding of the relation of alcohol consumption to the risk of cancer and other diseases.
Reference: Connor J, Kydd R, Maclennan B, Shield K, Rehm J. Alcohol-attributable cancer deaths under 80 years of age in New Zealand. Drug and Alcohol Review 2016; pre-publication. DOI: 10.1111/dar.12443
* * * * *
The following members of the International Scientific Forum on Alcohol Research provided comments for this critique:
Ulrich Keil, MD, PhD, Professor Emeritus, Institute of Epidemiology & Social Medicine, University of Muenster, Germany
Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark
Fulvio Ursini, MD, Dept. of Biological Chemistry, University of Padova, Padova, Italy
Dag S. Thelle, MD, PhD, Department of Biostatistics, Institute of Basic Medical Sciences, University of Oslo, Norway; Section for Epidemiology and Social Medicine, Sahlgrenska Academy, University of Gothenburg, Sweden
David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa
Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA
Creina Stockley, PhD, MSc Clinical Pharmacology, MBA; Health and Regulatory Information Manager, Australian Wine Research Institute, Glen Osmond, South Australia, Australia
Ross McCormick, PhD, MSc, MBChB, Professor Emeritus, The University of Auckland; former Associate Dean, Faculty of Medical and Health Sciences, The University of Auckland, Auckland, New Zealand
Yuqing Zhang, MD, DSc, Clinical Epidemiology, Boston University School of Medicine, Boston, MA, USA
Dominique Lanzmann-Petithory, MD, PhD, Nutrition Geriatrics, Hôpital Emile Roux, APHP Paris, Limeil-Brévannes, France
Tedd Goldfinger, DO, FACC, Desert Cardiology of Tucson Heart Center, University of Arizona School of Medicine, Tucson, AZ, USA
Elizabeth Barrett-Connor, MD, Distinguished Professor, Division of Epidemiology, Department of Family Medicine and Public Health and Department of Medicine, University of California, San Diego, La Jolla, CA USA
R. Curtis Ellison, MD, Professor of Medicine & Public Health, Boston University School of Medicine, Boston, MA, USA