Skip to main content
. 2022 May 24;9(6):166. doi: 10.3390/jcdd9060166

Table 2.

Multinomial logistic regression models of selected environmental risk factors in the control and congenital heart defect (CHD) sub-groups (septal, left ventricular outflow tract obstruction [LVOTO], and conotruncal). OR, odds ratio; CI, confidence interval.

Variable Control vs. Septal Control vs. LVOTO Control vs. Conotruncal
OR (95% CI) p adj OR (95% CI) p adj OR (95% CI) p adj
Child gender
Male vs. female 2.1 (1.1–4.1) 0.035 0.2 (0.1–0.7) 0.011 0.4 (0.1–1.2) 0.094
Maternal smoking status
Non-smoker vs. smoker 7.3 (2.1–24.9) 0.002 2.9 (0.7–11.8) 0.139 7.7 (1.8–31.9) 0.005
Non-smoker vs. ex-smoker 2.7 (1.1–6.5) 0.029 0.7 (0.2–2.1) 0.523 1.4 (0.4–5.0) 0.561
Maternal education
MSc, PhD vs. elementary school 6.1 (0.5–72.3) 0.154 4.1 (0.1–127) 0.415 1.4 (0.04–45.1) 0.865
MSc, PhD vs. vocational school 2.4 (0.4–13.1) 0.311 3.7 (0.3–47.9) 0.312 3.2 (0.3–30.7) 0.309
MSc, PhD vs. high school 2.8 (0.3–23.4) 0.345 2.8 (0.1–75.6) 0.549 NA
MSc, PhD vs. college 4.6 (0.8–26.7) 0.090 3.9 (0.2–65.9) 0.339 0.7 (0.04–13.9) 0.812
MSc, PhD vs. university 1.6 (0.3–8.6) 0.560 2.8 (0.2–35.8) 0.418 0.7 (0.06–7.2) 0.737
No. of live births 1.8 (1.2–2.6) 0.004 2.1 (1.4–3.3) 0.001 1.9 (1.1–3.1) 0.017
Family anamnesis of CHD
Negative vs. Positive 9.3 (2.3–37.0) 0.002 18.9 (3.8–90.9) 0.0003 11.1 (2.1–58.8) 0.005
Maternal chronic disease
No vs. Yes 2.8 (1.0–7.4) 0.043 3.6 (0.9–13.9) 0.062 0.9 (0.2–5.5) 0.921
Other drugs in pregnancy
No vs. Yes 2.0 (1.0–4.1) 0.055 3.6 (0.9–13.9) 0.062 0.9 (0.2–5.5) 0.921
Folate supplement initiation
No folate suppl. vs. before 3 weeks post-conception 0.3 (0.1–0.7) 0.010 0.2 (0.05–0.7) 0.011 0.6 (0.2–2.1) 0.394
No folate suppl. vs. after 3 weeks post-conception 0.4 (0.1–1.1) 0.067 0.5 (0.1–2.0) 0.350 0.4 (0.1–2.0) 0.286
Folic acid intake per month (mg) 1.0 (1.0–1.0) 0.069 1.0 (1.0–1.0) 0.417 1.0 (1.0–1.0) 0.071

Reference category. Odds ratio (OR), 95% confidence interval (CI), and adjusted p values were calculated in multinomial logistic regression models for mothers and children separately. Only variables with unadjusted p values < 0.250 were included in the multinomial logistic regression models and adjusted for co-variables. Variables with high levels of correlation were not included in the same model. NA: not applicable (the variable was not tested in the specific model).