POS0478 ASSOCIATION BETWEEN BODY SHAPES AND BODY SHAPE TRAJECTORIES, AND THE RISK OF RHEUMATOID ARTHRITIS IN THE FRENCH E3N COHORT

Abstract

Background:Several cohort and case-control studies have suggested that overweight or obesity is associated with the risk of Rheumatoid Arthritis (RA). Associations were based on Body Mass Index (BMI), although this measurement does not reflect the fat distribution [1].Objectives:To study the relationships between anthropometric measurements and the risk of RA in women involved in the E3N cohort.Methods:E3N is an ongoing French prospective cohort that enrolled 98,995 women aged 40-65 years in 1990. Women completed mailed questionnaires every 2-3 years on lifestyle and health-related information. A total of 698 incident RA cases have been validated among 78,452 women [2]. Available anthropometric measurements include birth height and weight, height and weight (collected at baseline and regularly updated during follow-up), and age-related body shapes (BS). Women had to identify the silhouette among 8 BS that best described their BS at 8 years, puberty, 20-25 years, 30-35 years, and study baseline. BS trajectories (from 8 to 30-35 years) were constructed using Nagin’s approach to group-based trajectory modeling that identifies 6 different trajectories [3]. Hazard ratios (HRs) and 95% confidence intervals (CIs) for the risk of incident RA were estimated using Cox proportional hazards regression models with age as the time scale. Models were first adjusted for known risk factors of RA (model 1), and then multi-adjusted (model 2).Results:Taking lean BS as reference, medium BS at puberty was associated with an increased risk of RA [HR=1.23 (95% CI 1.0-1.5)], in the fully adjusted model (table 1), as was large BS at baseline [HR =1.32 (95% CI 1.1-1.6)] (in model 1). Obesity (BMI>30 kg/m2) was marginally associated with RA in model 1 [HR=1.30 (95% CI 1.0-1.7)], but the association was no longer statistically significant in the multi-adjusted model, taking normal BMI [18-25 kg/m2[as the reference.Birth weight and height, BS (at 8, 20-25 and 30-35 years), and BS trajectories were not significantly associated with RA in any model.Conclusion:In E3N cohort, medium body shape at puberty was associated with an increased risk of RA independently of the BMI and smoking exposure.Table 1.Cox proportional-hazards analysis of anthropometric factors associated with RARANon-casesMODEL 1MODEL 2HRs (95%CI)ptrendHRs (95%CI)ptrendBaseline Body mass index (kg/m2)<18101,7970.75 (0.4-1.4)0.05320.85 (0.4-1.6)0.6165 [18-25[46250,330RefRef [25-30]17119,7161.10 (0.9-1.3)1.00 (0.8-1.2)> 30555,9111.30 (1.0-1.7)1.13 (0.8-1.6)Body shape at pubertyLean33240,485Ref0.0937Ref0.1490Medium17917,9031.24 (1.0-1.5)1.23 (1.0-1.5)Large14715,9841.14 (0.9-1.4)1.12 (0.9-1.4)Body shape at baselineLean35944,718Ref0.0055Ref0.1743Medium20020,0411.20 (1.0-1.4)1.16 (0.9-1.4)Large1109,8381.32 (1.1-1.6)1.20 (0.9-1.6)ptrend: p for trend. Ref: reference.Model 1 adjusted for age, smoking (past/current/never), passive smoking during childhood and/or adulthood (ever/never), educational level (<high–school, up to 2 years of university, ≥ 3 years of university).Multi-adjusted model 2 included model 1 + body mass index (<18, 18–25, 25–30, >30 kg/m2), body shape at puberty, body shape at baseline adjusted for age, baseline physical activity (in quartiles), age at menarche (<13, 13–15, ≥15 years), age at menopause (≤45, 45–53, ≥53 years), age at the first pregnancy (<22, 22–27, ≥27 years), number of full-term pregnancies (≤1, 2, ≥3), and duration of premenopausal use of progestogen (0, 0–24, >24 months).