Table 3. Prespecified sensitivity analyses.

Outcome	Placebo N = 1,063	Dextrose N = 1,070	aRR	95% CI	p-Value
Excluding protocol deviations
NICU admission	91/1,017 (9.0%)	95/1,013 (9.4%)	1.06	0.80, 1.39	0.70
Hypoglycemia	431/1,017 (42.4%)	368/1,013 (36.3%)	0.86	0.77, 0.96	0.01
Excluding babies who did not receive assigned study gel
NICU admission	96/1,050 (9.1%)	99/1,043 (9.5%)	1.05	0.80, 1.37	0.74
Hypoglycemia	445/1,050 (42.4%)	382/1,043 (36.6%)	0.87	0.78, 0.96	0.01
Including only glucose oxidase measurements
Hypoglycemia—overall	425/1,060 (40.1%)	378/1,072 (35.3%)	0.88	0.79, 0.98	0.02
Hypoglycemia—vaginal births	245/633 (38.7%)	189/609 (31.0%)	0.80	0.69, 0.94	0.01
Hypoglycemia—cesarean section births	180/427 (42.2%)	189/460 (41.1%)	0.97	0.84, 1.14	0.74
Including only glucose oxidase measurements and excluding protocol deviations
Hypoglycemia	412/1,017 (40.5%)	351/1,013 (34.7%)	0.86	0.77, 0.96	0.01
Including only glucose oxidase measurements and excluding babies who did not receive assigned study gel
Hypoglycemia	423/1,050 (40.3%)	363/1,043 (34.8%)	0.86	0.78, 0.97	0.01

Data are n (%) or mean (SD). Adjustments are for multiple births, study site, and primary reason for risk of hypoglycemia. Hypoglycemia is blood glucose concentration < 2.6 mmol/l.

aRR, adjusted relative risk; CI, confidence interval; NICU, neonatal intensive care unit.