Adriëtte Oostvogels

126 Chapter 5 nearest millimetre using a Leicester portable height measure (Seca). Fat percentage was measured by arm-to-leg bioelectrical impedance analysis (Bodystat 1500 MDD machine (Bodystat Inc., Douglas, UK)), calculations were based on equations adapted from Lohman et al. and Kushner et al. 26 Fasting glucose and C-peptide were determined based on overnight fasting capillary blood samples collected with an ambulatory collection kit (Demecal: Lab Anywhere, Haarlem, The Netherlands). 27 C-peptide measurements had a detection limit of >0.34 nmol/L, 63.6% of the C-peptide levels were below the level of detection of the assay. These values were predicted with a survival analysis applying a loglogistic distribution based on sex, age and BMI of the children (see appendix). Covariates Sex and birth weight were obtained from the Youth Health Care Registration and the Dutch Perinatal Registration (PRN). Maternal age during pregnancy (years; continuous), education level (years after primary school; continuous) and ethnicity were self-reported in the pregnancy questionnaire. Ethnicity (Dutch, Turkish, Moroccan, Other Western, Other non-Western) was defined by the country of birth of the mother of the pregnant woman. Offspring’s standardized birth weight (continuous) and sex of the child were available by combining data from the pregnancy questionnaire and the Netherlands Perinatal Registry (PRN). Maternal BMI, paternal BMI and screen (computer and television) time (hours/day; continuous) at age 5-6 years was obtained from the questionnaire, a validated food frequency questionnaire was used to calculate the child’s energy intake in kilocalories (kcal, continuous) and the exact age of the child at the health check was calculated. To account for missings on covariates, covariates were imputed using MICE in SPSS accounting for all covariates included in this study. 28 Statistics Overall differences in characteristics (table 1) between the four different second- degree FHD groups were examined using chi square tests for categorical variables and one way ANOVA with a post hoc tukey alpha test for continuous variables. Significant differences with the no FHD group were reported. Metabolic outcomes at age 5-6 years of the different FHD groups were corrected for sex and age at the time of measurement (table 2). Since the residuals of C-peptide were not normally distributed, the Box-Cox method in R Package was used to find appropriate transformations. C-peptide was transformed to the reciprocal. Then, linear regression analysis was performed using three consecutive models. Model 1 adjusted