Abstract A74: Cohort study of obesity, tobacco, and alcohol and risk of neoplastic progression to esophageal adenocarcinoma in Barrett's esophagus

Abstract

Abstract Background: Over the past three decades, esophageal adenocarcinoma (EA) incidence in the US and many other developed countries has increased dramatically. Despite improvements in diagnostic and therapeutic techniques, EA patients have a poor prognosis and, on average, survive less than a year after diagnosis. Persons diagnosed with Barrett's esophagus (BE), a metaplastic condition that confers a 30- to 100-fold increase in risk, are typically followed in a surveillance program involving periodic endoscopy with biopsies. However, a vast majority of persons with BE never develop EA and there are no medical, surgical or lifestyle interventions that have been proven to safely lower risk of progressing to EA. Methods: We investigated whether baseline measures of obesity, tobacco or alcohol use could predict progression to EA in cohort of 411 BE patients followed prospectively. Data were collected by personal interview and measurements by trained staff. Hazard ratios (HR) for baseline body mass index (BMI), waist-to-hip ratio (WHR), cigarette use and alcohol intake were estimated using Cox proportional hazards regression while adjusting for age, sex and non-steroidal anti-inflammatory drug (NSAID) use. Effect modification was examined through stratified analyses and the inclusion of interaction terms in regression models. Results: At baseline, 39% of the cohort had a BMI over 30; their mean WHR was 0.95; 64% were current or former cigarette smokers; and 83% reported alcohol use. The main analyses focused on 397 persons with at least 5 months of follow-up (median 78.6 months, 33,635 person-months), in whom 45 developed EA. There was no association between BMI at baseline and risk of progression to EA after adjusting for age, gender, smoking and NSAID use. In contrast, a suggestion of increased risk was observed in the highest quartile of WHR, compared to the lowest, (HR = 1.61; 95% CI 0.67-3.89), which was apparent in males (HR = 1.68; 95% CI 0.66-4.30) but not in females (HR = 0.95; 95% CI 0.05-18.92). Among males, the adjusted WHREA association was also observed to be higher among younger persons (<61 years) and those who had never smoked cigarettes. After adjusting for age, gender, WHR and NSAID use, EA risk among ever-smokers was 1.57 (95% CI 0.78-3.14) with a statistically significant trend with pack-years of smoking (p-value 0.03); the HR in those in the highest tertile of pack-years was 2.29 (95% CI 1.04-5.07) compared to never-smokers. In contrast, there was no evidence of increasing EA risk with increasing alcohol intake, whether examined as intake of all alcoholic drinks or by beverage type (beer, wine, hard liquor). Conclusions: Abdominal obesity (as measured by WHR), but not BMI, was associated with a modest, but not statistically significantly higher risk of EA in this cohort of BE patients. This association was stronger among males, younger persons and non-smokers. We observed a significant trend of increased EA risk with pack-years of smoking. Continued follow-up of this and other cohorts is needed to further understand these relationships; until such data are available, current recommendations for EA prevention in persons with BE should focus on reducing abdominal obesity and smoking cessation. Citation Information: Cancer Prev Res 2010;3(12 Suppl):A74.