Noreen Z Siddiqui1,2, Joline W J Beulens3,4,5, Nina van der Vliet6, Nicole R den Braver3,4, Petra J M Elders4,7, Femke Rutters3,4. 1. Amsterdam UMC, location Vrije Universiteit Amsterdam, Epidemiology and Data Science, Amsterdam, the Netherlands. n.z.siddiqui@amsterdamumc.nl. 2. Amsterdam Public Health, Health Behaviors and Chronic Diseases, Amsterdam, the Netherlands. n.z.siddiqui@amsterdamumc.nl. 3. Amsterdam UMC, location Vrije Universiteit Amsterdam, Epidemiology and Data Science, Amsterdam, the Netherlands. 4. Amsterdam Public Health, Health Behaviors and Chronic Diseases, Amsterdam, the Netherlands. 5. Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, the Netherlands. 6. National Institute for Public Health and the Environment, Bilthoven, the Netherlands. 7. Department of General Practiceand Elderly Care Medicine, Amsterdam PublicHealth, Amsterdam UMC, location VUmc, Amsterdam, the Netherlands.
Abstract
BACKGROUND: We aimed to study the mediating role of diet quality, physical activity, smoking, and alcohol intake in the association of stressful life events with visceral obesity over a seven-year period and assessed effect modification by sex and SES. METHODS: In total, 2416 participants with a mean age of 56.1 (±7.3) years, of which 51.4% were women, and 12.5% had a lower educational level from the Hoorn studies were followed for seven years. Stress was measured with a 'Serious Life Events' questionnaire, which was summed into a total score (range zero to ten events) and stratified to account for nonlinearity. Changes in visceral obesity were assessed by changes in BMI (kg/m2) and waist circumference (cm) in seven years. We used the product of coefficient approach to assess mediation of the following lifestyle factors: diet, physical activity, smoking, and alcohol intake. We analyzed associations between stressful life events and change in BMI and waist circumference with linear regression models. RESULTS: Within the low education group, we observed a significant association between ≥3 stressful life events and a change in BMI (0.60 kg/m2 (CI: 0.05, 1.14)) and waist circumference (2.23 cm (CI: 0.19, 4.48)), compared to experiencing no events. For both BMI and waist circumference, no significant associations were observed when experiencing 1 or 2 events. In the moderate to high education group, we observed only statistically significant associations for waist circumference when experiencing ≥3 stressful life events (0.86 cm (CI: 0.05, 1.41)) and not for the other event groups. Our mediation analyses showed that the proportion mediated by smoking was 13.2%, while the other lifestyle factors showed no mediating effect. CONCLUSIONS: Multiple stressful life events are associated with an increase in waist circumference and BMI in those with lower education. Smoking might play a mediating role in this association.
BACKGROUND: We aimed to study the mediating role of diet quality, physical activity, smoking, and alcohol intake in the association of stressful life events with visceral obesity over a seven-year period and assessed effect modification by sex and SES. METHODS: In total, 2416 participants with a mean age of 56.1 (±7.3) years, of which 51.4% were women, and 12.5% had a lower educational level from the Hoorn studies were followed for seven years. Stress was measured with a 'Serious Life Events' questionnaire, which was summed into a total score (range zero to ten events) and stratified to account for nonlinearity. Changes in visceral obesity were assessed by changes in BMI (kg/m2) and waist circumference (cm) in seven years. We used the product of coefficient approach to assess mediation of the following lifestyle factors: diet, physical activity, smoking, and alcohol intake. We analyzed associations between stressful life events and change in BMI and waist circumference with linear regression models. RESULTS: Within the low education group, we observed a significant association between ≥3 stressful life events and a change in BMI (0.60 kg/m2 (CI: 0.05, 1.14)) and waist circumference (2.23 cm (CI: 0.19, 4.48)), compared to experiencing no events. For both BMI and waist circumference, no significant associations were observed when experiencing 1 or 2 events. In the moderate to high education group, we observed only statistically significant associations for waist circumference when experiencing ≥3 stressful life events (0.86 cm (CI: 0.05, 1.41)) and not for the other event groups. Our mediation analyses showed that the proportion mediated by smoking was 13.2%, while the other lifestyle factors showed no mediating effect. CONCLUSIONS: Multiple stressful life events are associated with an increase in waist circumference and BMI in those with lower education. Smoking might play a mediating role in this association.