Purpose: Brain metabolites show very rapid maturation over infancy, particularly following very preterm (VPT) birth, and can provide an index of brain injury. The utility of magnetic resonance imaging (MRS, magnetic resonance spectroscopy) in predicting outcome in VPT-born infants is largely limited to 2-year outcomes. We examined the value of MRS in VPT followed longitudinally to 4 years.
Methods: MRS datasets were acquired in 45 VPT infants (< 32 weeks gestational age) longitudinally: at birth, at term-equivalent and at 4 years of age. Using LCModel analyses in a basal ganglia voxel, we investigated metabolite ratios as a function of age, brain injury and outcome. We also studied a full-term (FT) cohort at 4 years and compared group differences with outcome.
Results: We found significant age-related changes in many brain metabolites in infancy, including phosphocreatine (CR)/phosphocholine (CHO), N-acetylaspartylglutamate (NAA)/CHO, myoinositol (INS)/CHO and INS/CR; there were no significant MRS differences between VPT and FT groups at 4 years of age, or differences at 4 years as a function of early brain injury or outcome. The rate of change in metabolite ratios from VPT birth to term-equivalent age did not predict outcome in the VPT children at 4 years.
Conclusion: Brain metabolite ratios measured in VPT-born infants have shown associations with short-term outcomes, but these correlations did not extend to early childhood nor predict cognitive sequelae. The most frequently reported poor outcome in VPT-born children is cognitive difficulties starting at early school age. MRS metrics early in the infant's life do not appear to predict these longer-term outcomes.
Keywords: Children; Infants; MRS; Outcome; Very preterm born.