Relation between dietary pattern analysis (Principal component analysis) and body mass index: A 5-year follow-up study in a Belgian military population

Abstract

Introduction Increasing body mass index (BMI) has been related to many chronic diseases. Knowledge of nutritional determinants of BMI increase may be important to detect persons at risk. Methods A longitudinal prospective study design was used in 805 Belgian soldiers. Daily nutrition was recorded with a validated food-frequency questionnaire. Weight and height were recorded from medical military data and principal component analysis was used to detect dietary patterns. Results During the 5 years follow-up, mean BMI increased from 25.8 (±3.3) kg/m2 to 27.1 (±3.6) kg/m2 (p<0.05). Consequently, the prevalence of being overweight and obesity increased from 46.2% and 9.6% to 51.6% and 19.9% (p<0.05), respectively. Mean (SD) weight gain differed between the BMI categories at baseline with a respective weight gain of 3.8 (±3.1) kg for normal weight at baseline, 4.2 (±3.2) kg for overweight and 5.1 (±3.4) kg for obesity (p for trend <0.05). Three dietary patterns were detected by principal component analysis: Meat, Sweet and Healthy dietary pattern. In energy-unadjusted and adjusted linear regressions, no dietary pattern was associated with BMI increase. Conclusions No specific dietary pattern was related to BMI increase. Prevention of obesity should focus on total energy intake at all BMI categories.