2.7. Statistical Analysis

We performed univariate analysis to describe the sample characteristics. Categorical variables were presented as frequencies and percentages, and continuous variables as mean ± standard deviation or median (p25–p75). We compared the frequencies of sociodemographic, maternal, and neonatal characteristics between adolescents and adults using the Chi-square test in the bivariate analysis. Unadjusted Poisson regression models were performed to analyze the association of GWG, age, pBMI as independent variables with maternal and neonatal outcomes as dependent variables. We ran three different models: in M1, the variables GWG and pBMI were introduced in different (independent) models; in M2, both variables, GWG and pBMI, were included in the same model as main effects; and in M3, both variables, GWG and pBMI, were included in the same model, with interaction with age.

All models were adjusted by socioeconomic level, initiation of prenatal care, parity, and history of diseases. The reference groups were adolescents, adequate GWG, and normal weight. Statistical significance was considered with a p-value < 0.050. All analyses were performed using the software Stata/v.SE16.1 (College Station, TX, USA) [28].