Seok-Hoo Jeong1, Pumsoo Kim1, Sang-Wook Yi2,3, Yu Jin Kim1, Myong Ki Baeg1, Jee-Jeon Yi4. 1. Department of Internal Medicine, Catholic Kwandong University International St. Mary's Hospital, Incheon, Korea. 2. Department of Preventive Medicine and Public Health, Catholic Kwandong University College of Medicine, Gangneung, Korea. 3. Institute for Clinical and Translational Research, Catholic Kwandong University, Gangneung, Korea. 4. Institute for Occupational and Environmental Health, Catholic Kwandong University, Gangneung, Korea.
Abstract
BACKGROUND: The association between body mass index (BMI) and mortality from gastrointestinal (GI) cancer remains unclear, especially in Asian populations. METHODS: A total of 510 148 Korean adults who participated in routine health examinations during the period 2002-2003 were followed up until 2013. RESULTS: During a mean follow up of 10.5 years, 7831 individuals died of GI cancer. Various associations with BMI were found: U-curve (overall GI, colorectal, liver, and gallbladder cancer), L-curve (stomach cancer), linear (esophageal, extrahepatic bile duct [EBD], and small intestine cancer), and none (pancreatic cancer). Overall GI cancer mortality was lowest at approximately 23.5-26 kg/m2 . For cancers with linear associations, the multivariable adjusted hazard ratios per each 5 kg/m2 higher BMI were 0.53 (95% confidence interval = 0.43-0.65, esophagus), 1.19 (1.02-1.40, EBD), and 0.64 (0.41-0.999, small intestine). For cancers with U-curve or L-curve associations, the corresponding hazard ratios ≥25 kg/m2 were 1.19 (1.08-1.32, overall GI), 1.30 (1.04-1.64, colorectal), 1.28 (1.07-1.53, liver), and 1.30 (0.85-1.97, gallbladder), while in the range of <25 kg/m2 , they were 0.81 (0.76-0.87, overall GI), 0.43 (0.32-0.58, esophagus), 0.70 (0.62-0.79, stomach), and 0.77 (0.65-0.90, colorectal), and these inverse associations did not weaken after excluding the first 7 years of follow up and ever smokers. CONCLUSIONS: Both low and high BMIs were associated with excess mortality from GI cancers in Korean adults. EBD cancer had a positive association, while esophageal and small intestine cancers had inverse associations. Above 25 kg/m2 , liver and colorectal cancers had positive associations with BMI, whereas below 25 kg/m2 , stomach and colorectal cancers had inverse associations.
BACKGROUND: The association between body mass index (BMI) and mortality from gastrointestinal (GI) cancer remains unclear, especially in Asian populations. METHODS: A total of 510 148 Korean adults who participated in routine health examinations during the period 2002-2003 were followed up until 2013. RESULTS: During a mean follow up of 10.5 years, 7831 individuals died of GI cancer. Various associations with BMI were found: U-curve (overall GI, colorectal, liver, and gallbladder cancer), L-curve (stomach cancer), linear (esophageal, extrahepatic bile duct [EBD], and small intestine cancer), and none (pancreatic cancer). Overall GI cancer mortality was lowest at approximately 23.5-26 kg/m2 . For cancers with linear associations, the multivariable adjusted hazard ratios per each 5 kg/m2 higher BMI were 0.53 (95% confidence interval = 0.43-0.65, esophagus), 1.19 (1.02-1.40, EBD), and 0.64 (0.41-0.999, small intestine). For cancers with U-curve or L-curve associations, the corresponding hazard ratios ≥25 kg/m2 were 1.19 (1.08-1.32, overall GI), 1.30 (1.04-1.64, colorectal), 1.28 (1.07-1.53, liver), and 1.30 (0.85-1.97, gallbladder), while in the range of <25 kg/m2 , they were 0.81 (0.76-0.87, overall GI), 0.43 (0.32-0.58, esophagus), 0.70 (0.62-0.79, stomach), and 0.77 (0.65-0.90, colorectal), and these inverse associations did not weaken after excluding the first 7 years of follow up and ever smokers. CONCLUSIONS: Both low and high BMIs were associated with excess mortality from GI cancers in Korean adults. EBD cancer had a positive association, while esophageal and small intestine cancers had inverse associations. Above 25 kg/m2 , liver and colorectal cancers had positive associations with BMI, whereas below 25 kg/m2 , stomach and colorectal cancers had inverse associations.