Results aid development of predictive tool for Barrett's esophagus among older men

SAN DIEGO — Findings from a recent study aided investigators in the development of an online tool to predict the absolute probability of Barrett’s esophagus in older men, according to information presented at the 2012 Digestive Disease Week Annual Meeting.

In a case-control study, researchers evaluated two cohorts of patients with Barrett’s esophagus (BE): one consisting of men aged 50 to 79 years undergoing colonoscopy screening for colorectal cancer recruited to receive upper endoscopy (n=822), and the other consisting of men of a similar age who received an initial diagnosis of BE via clinically indicated endoscopy within 1 month of recruitment (n=80). Patients screened for colorectal cancer who were not diagnosed with BE served as the control. Participants also took a questionnaire in order to assess the presence of GERD. The study was designed to determine how factors including waist-to-hip ratio (WHR), age and tobacco impact BE in patients with and without GERD symptoms.

Of the patients who received upper endoscopy, 155 (19%) reported GERD symptoms occurring at least once a week, and 70 were diagnosed with BE (15% of patients with GERD symptoms and 7% of those without). The mean BMI was 30.0 kg/m², and the mean WHR was 1.002 within this cohort. The group also was significantly less likely to have GERD symptoms than the previously diagnosed group (34% vs. 81%, P<.01). BE was associated with increased GERD frequency (adjusted OR=2.18; 95% CI, 1.25-3.81) and duration (adjusted OR per 10-year increments=1.31; 95% CI, 0.904-1.86) within the upper endoscopy group. Both groups had a median BE length of 2.0 cm.

Across both groups, 89 patients had BE with GERD and 61 had BE without GERD. Researchers found an association between BE and WHR (adjusted OR for 3^rd vs. 1^st tertile=2.15; 95% CI, 1.30-3.57), age (adjusted OR per 10 years=1.82; 95% CI, 1.38-2.40) and tobacco use (adjusted OR for 35 or more pack-years vs. 0 pack-years=2.45; 95% CI, 1.53-3.93) among the entire cohort.

Joel H. Rubenstein, MD, MSc, FACG, FASGE

“Barrett’s esophagus is substantially prevalent in a population of older, mostly overweight men, even among those without GERD symptoms,” Joel H. Rubenstein, MD, MSc, FACG, FASGE, Center for Clinical Management Research, Ann Arbor Veterans Affairs Medical Center in Ann Arbor, Mich., told Healio.com. “We found that GERD frequency and duration are only weak predictors of Barrett’s esophagus in unselected men. A model that includes GERD frequency, age, [WHR] and pack-years of cigarette smoking is substantially better at predicting [BE] than GERD symptoms alone.”

The researchers used the results from this study to help develop a model for predicting the absolute probability of BE in men aged 50 to 79 years. The resulting Michigan Barrett’s Esophagus pREdiction Tool (M-BERET) is based on a patient’s age, frequency of GERD symptoms (weekly vs. nonweekly), tobacco use according to pack years and WHR, and is designed to more accurately establish an individual’s likelihood of developing BE than a model only using GERD frequency and duration. ROC-curve analysis indicated that M-BERET had a larger area under the curve than the GERD-based model (0.7243 vs. 0.6073, P<.001).

“We chose to use waist-to-hip ratio instead of circumference based on this data,” Rubenstein said. “We have created a quick clinical tool for predicting Barrett’s esophagus, and it is available on the Web … the model still needs to be validated in another population, however.”

For more information:

Rubenstein JH. #316: Risk Factors for Barrett’s Esophagus Among Men without GERD Symptoms: Results from the Newly Diagnosed Barrett’s Esophagus Study. Presented at: the 2012 Digestive Disease Week Annual Meeting; May 19-22, San Diego.