Transient hypothyroxinemia (TH) of prematurity has been correlated with poor neurodevolpmental outcome. However, the relationship between thyroid function and neonatal morbidity and mortality has not been described. Objective: to investigate the relationship between thyroid function and neonatal outcome. Methods: a review of infants <1500g admitted to our NICU from 7/93-7/96. Infants who received a minimum of 1 head ultrasound and 1 newborn screen for thyroid function were included in the analysis, n=342. One-way ANOVA with post-hoc testing was performed comparing serum thyroxine and TSH values on initial newborn screen from infants who died before hospital discharge with surviving infants, and from infants with and without intraventricular hemorrhage (IVH), and periventricular leukomalacia(PVL). Forward stepwise discriminant analysis was performed to account for likely confounding variables. Results: T4 correlated strongly with gestational age (r=.56, p<.001). Overall, 289/342 infants (85%) had TH, as defined as T4 values < 10th percentile with normal TSH values. None of the infants had true congenital hypothyroidism. T4 values of infants with PVL (n=15) were lower than infants without PVL (5.9± 3.6 vs 7.7 ± 3.8 mcg/dl, p=.2) but the difference was not statisticaly significant. Infants with IVH (n=58) had a lower T4 than infants who did not develop IVH (5.4 ± 3.4 vs 7.8 ± 3.6 mcg/dl, p=.0003). Infants who died before hospital discharge (n=24) had a lower T4 than infants discharged to home (3.4 ± 2.2 vs 7.9 ± 3.7 mcg/dl, p=.00002). There was no statistical difference in TSH values between any of the above groups. When controlling for gest. age by discriminant analysis, low T4 was not correlated with IVH. However, after controlling for gest. age, bacterial sepsis, bronchopulmonary dysplasia, and IVH, T4 remained predictive for neonatal death (p=.003). Conclusions: In our population of VLBW infants: 1) TH has an incidence of 85%, 2) very low T4 values on initial newborn screening are correlated with increased mortality. Further investigation is needed to determine whether low serum thyroxine physiologically contributes to neonatal death, or is simply an associated factor.