Objectives To investigate association between lifestyle factors and 5-year body weight gain among apparently healthy middle-aged male workers.

Methods A occupational cohort in an electronic-manufacturing factory was established by using a medical checkup program in 2002, and was followed up with a health examination in 2007. Baseline comprehensive questionnaires were applied for collection of lifestyle behaviours, and abdominal ultrasound was performed for evaluation of hepatic steatosis. Male worker with complete data of baseline lifestyle factors, and laboratory examination, including ALT, ferritin, were included in analysis. Cross-sectional analyses in 2002 were conducted by stepwise multiple linear regressions for exploring significant association between lifestyle factors and log-ferritin. Stepwise linear regression models for 5-year body weight gain were used for identifying significant lifestyle factors with predictive value.

Results A total of 1096 male workers with a baseline mean (SD) age of 32.5 (6.0) were included for analysis. Log-ferritin was positively associated with red meat intake, and negatively associated with blood donation, regular exercise at leisure time, and vegetarian diet. 5-year body weight gain was associated with 12-hour shift work, and vegetarian diet after adjustment for age. In subgroup analysis among overweight males (BMI> = 25) with non-alcoholic steatohepatitis (NASH), blood donation was significantly associated with less body weight gain in 5 years.

Conclusions 12-hour shift work, vegetarian diet may increase body weight in male workers. behaviours of health promotion, such as blood donation, regular exercise at leisure time, were associated with lower level of ferritin, a marker of total body iron load. Blood donation was the strongest factor associated with lower body weight gain within 5 years, especially among males with NASH. Physicians should recognise related behavioural factors for hyperferritinemia and body weight gain, and give high-risk individuals health counselling based on personalised risk profiles.