To assess for possible missed hypothyroidism in infants of very low birth weight (VLBW) whose initial newborn screening (NBS) was within normal reference range. We analyzed serum thyroid-stimulating hormone (TSH) obtained at 36weeks of corrected gestational age or at hospital discharge if earlier (retest TSH) in infants with VLBW in the neonatal intensive care unit to determine the prevalence and factors associated with retest TSH ≥5 mU/L, a concentration requiring close follow-up for hypothyroidism. Utility of alternative cut-offs for NBS TSH also was assessed. A total of 398 infants, median gestational age 29 (range 22-36) weeks, birth weight 1138 (470-1498) g, were included in this study. Retest TSH was obtained at 49.5 (12-137) days after birth. Median retest TSH was 3.1 (0.5-27.9) mU/L. Seventy-three (18.3%) of the cohort had retest TSH ≥5 mU/L. Adjusting NBS cut-off to ≥15 or ≥10 mU/L identified <50% of infants with TSH ≥5 mU/L, resulting in 6% false positives and >70% false negatives. Multiple regression modeling indicated that 35% of variance in retest TSH value was explained by NBS TSH concentration, birth weight, and gestational age, all P<.01. Retesting for hypothyroidism at 36weeks of corrected gestational age in infants with VLBL and normal NBS could identify infants who require ongoing surveillance until thyroid function has been definitively ascertained. Adjusting NBS TSH cutoffs is not a valid option for identifying potential hypothyroidism in infants with VLBW because of lack of sensitivity and unacceptable false-positive and false-negative rates.