The Effect of Obesity on Esophageal Adenocarcinoma
Kevin J. Nattinger,Chung Yin Kong,Tristan J. Hayeck,Zehra Omer,Caihua Liang,Y. Claire Wang,Stuart J. Spechler,Pamela M. McMahon,G. Scott Gazelle,Chin Hur
DOI: https://doi.org/10.1016/S0016-5085(11)61457-1
IF: 29.4
2011-01-01
Gastroenterology
Abstract:Background: Over the last 40 years, obesity has risen at an astonishing rate in the US, nearly tripling between 1973 and 2005. The incidence of esophageal adenocarcinoma (EAC) has undergone an even sharper increase over the same time period, growing by a factor of five. Epidemiologic studies have shown that many precursors of EAC, such as Barrett's esophagus (BE), are associated with obesity, leading some to suggest that obesity may be a contributing factor in the development of EAC. The trends in EAC and obesity do not, however, show the time lag that one would expect if obesity were a significant causal factor in EAC development. The aim of our study was to determine the extent to which the rise in obesity has affected the EAC incidence between 1973 and 2005. Methods: A previously validated computer simulation model of EAC (EACMo) was adapted for this analysis. The model was used to simulate two groups in the US white male population, OBESE (BMI above 30) and NON-OBESE (BMI below 30). The study period extended from 1973 to 2005, corresponding to the National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) data. The model simulations were analyzed in two different scenarios. Scenario 1 depicted a hypothetical US population where the rates of obesity prevalence and progression to cancer did not change over the study period. Scenario 2 incorporated the rise in obesity prevalence over the same time period. Comparing the results of the two scenarios with each other and with SEER data provided an estimate of the number of EAC cases attributable to obesity. As a sensitivity analysis, we determined the odds ratio of EAC incidence associated with obesity necessary for Scenario 2 results to match SEER data. Results: The EAC incidence from each scenario as well as SEER data for each year simulated may be seen in Fig. 1, with representative points seen in Table 1. Scenario 1 resulted in 30,559 EAC cases in the US over the study period and 1,150 cases in 2005, while Scenario 2 resulted in 35,773 cases cumulatively and 1,607 in 2005. SEER incidence has 111,223 cases cumulatively and 7,173 cases in 2005, meaning that obesity accounted for 6.46% of EAC cases in the US between 1973 and 2005, and 7.59% of cases in 2005. The sensitivity analysis showed that an OR of 74 best matched the SEER incidence over the study period and 82 best matched the EAC incidence in 2005. Conclusions: The low percentage of EAC cases attributable to obesity supports the idea that obesity is unlikely to be a primary cause of the exponential rise in EAC incidence. The high ORs—between 27 and 30 times that reported in literature— required to recreate the historical SEER data by changing obesity alone further illustrates this notion. Table 1: EAC Incidence (per 100,000)