1421PD_PR - Moderate Alcohol Intake and Cancer: The Role of Under-Reporting
|Date||01 October 2012|
|Event||ESMO Congress 2012|
|Session||Public health and familial cancer|
|Topics|| Cancer Aetiology, Epidemiology, Prevention
A.L. Klatsky1, Y. Li2, D. Baer2, M.A. Armstrong3, N. Udaltsova3, G.D. Friedman3
Background: There is consensus that heavy alcohol drinking is related to increased risk of several cancer types. However, the role of light-moderate drinking and the existence of a threshold amount of alcohol for risk are less clear.
Methods: We exploredpossible underreporting among moderate drinkers in a cohort study of 129,987 persons. From baseline in 1978-85 through 2008 cancer was diagnosed in 15,080 persons. Cox proportional hazard models were controlled for age, sex, ethnicity, smoking, education, and body mass index. Alcohol was studied categorically, with lifelong abstainers as referent. We studied risk of any cancer and of a composite of 5 types with alcohol-associated risk in this cohort (upper airway digestive, liver, breast, lung, colo-rectal). We stratified moderate drinking categories (<1 and 1-2 drinks per day) into persons suspected of under-reporting (“suspect”) and those not suspected (“not suspect”). The “suspect” group either reported heavier intake at some other time or had an alcohol-related diagnosis (out-patient, in-patient, or death) at any time.
Results: The relative risks (RR) and 95% confidence intervals (CI) among the alcohol categories for any cancer were: exdrinkers=1.17 (1.07-1.27, p<0.001), <1 drink per day=1.05 (1.00-1,09, p=0.04), 1-2 drinks per day=1.09 (1.04-1.14, p<0.001), and ≥3 drinks per day=1.16 (1.09 = 1.24, p<001). Relationships of light-moderate drinking (<3 drinks per day) were similar in men, women, whites, blacks, Asians, never smokers, persons diagnosed within or after 10 years. For all persons reporting 1-2 drinks per day, the RR of cancer among “suspect” was 1.14 (1.04-1.25, p=0.004) and for “non-suspect it was 1.00 (0.89-1.11). For those reporting <1 drink per day the RR for “suspect" was 1.17 (1.06-1.28, p=0.001); for “non-suspect” it was 1.00 (0.90-1.09). For the alcohol-related composite the RR(CI) among “suspect” persons reporting 1-2 drinks per day was 1.41 (1.23-1.62) and for “not suspect” persons it was 1.06 (0.89-1.25), while at < 1 drink per day the RR for "suspect" was 1.39 (1.21-1.59, p< 0.00) and for "non-suspect" it was 1.03 (0.90-1.18). .
Conclusions: We conclude that apparent increased risk of cancer among light-moderate drinkers is substantially due to underreporting of intake.