“…For each of the 5 continuous outcome variables we used multivariable generalized linear models with averaged continuous LAPS as the exposure. We considered models adjusted for (1) age 1 , 2 , 3 , 4 , 50 , 51 , 52 and (2) age, BMI, 53 , 54 , 55 , 56 marital status, 57 , 58 , 59 , 60 number of stressful events during past 6 months, presence of life-threatening events, history of physical and/or sexual childhood abuse, 61 , 62 , 63 , 64 alcohol consumption, 65 , 66 physical activity, 67 , 68 , 69 , 70 BSSNI, 53 , 59 , 71 , 72 and FMI. 73 , 74 , 75 , 76 , 77 , 78 We performed analogous analyses for the 4 defined dichotomous outcomes using logistic regression models.…”