Table 4.

Multivariate binary logistic regression analysis (Germany-born patients only).

	Model 1 (n = 59) Normal vs. mildly abnormal development			Model 2 (n = 46) Normal vs. severe brain injury			Model 3 (n = 48) Normal vs. severe brain injury
Variables	Odds ratio	95% CI	p-value	Odds ratio	95% CI	p-value	Odds ratio	95% CI	p-value
Diagnosis (transient vs. persistent CHI)	n.s.			n.s			n.s.
Birth weight (g)	1.001	1.0–1.002	0.002	n.s.			n.s.
Delay between first symptoms and first BG measurement	E			71.7	2.2–2379	0.017	E
Pancreatic surgery	E			n.s.			n.s.
Hypoglycemic seizures	E			n.s.			12.9	1.9–86.4	0.008
Birth at lower-level hospital	E			n.s.			n.s.
Lowest BG < 20 mg/dl	E			134.3	1.3–14361	0.004	E
Lowest BG (mg/dl)	E			0.85*	0.73–0.98	0.035	E

Only significant variables from univariate analysis were entered into the multivariate logistic regression models. Patients born abroad were excluded from the analysis due to the high rate of adverse neurodevelopment in this small subgroup. Data are presented as odds ratio for adverse neurodevelopment with 95% confidence interval and p-value, if significant <0.05.

BG, blood glucose; CHI, congenital hyperinsulinism; OR, odds ratio; CI, confidence interval; n.s., not significant; E, excluded in model.

Model 1: included the following risk factors: diagnosis, birth weight, percentile. Backward logistic regression was performed.

Model 2: included the following risk factors: diagnosis, birth weight, delay between first symptoms and first BG measurement, pancreatic surgery, hypoglycemic seizures, birth at lower-level hospital, lowest BG < 20 mg/dl. Backward logistic regression was performed. *When lowest BG < 20 mg/dl is exchanged for lowest BG value in mg/dl (metric variable), the result is also significant.

Model 3: included the following risk factors: diagnosis, birth weight, pancreatic surgery, hypoglycemic seizures, birth at lower-level hospital. Backward logistic regression was performed.