Associations of body composition with regional brain volumes and white matter microstructure in very preterm infants.

Objective: To determine associations between body composition and concurrent measures of brain development including (1) Tissue-specific brain volumes and (2) White matter microstructure, among very preterm infants at term equivalent age.

Design: Prospective observational study.

Setting: Single-centre academic level III neonatal intensive care unit.

Patients: We studied 85 infants born <33 weeks' gestation.

Methods: At term equivalent age, infants underwent air displacement plethysmography to determine body composition, and brain MRI from which we quantified tissue-specific brain volumes and fractional anisotropy (FA) of white matter tracts. We estimated associations of fat and lean mass Z-scores with each brain outcome, using linear mixed models adjusted for intrafamilial correlation among twins and potential confounding variables.

Results: Median gestational age was 29 weeks (range 23.4-32.9). One unit greater lean mass Z-score was associated with larger total brain volume (10.5 cc, 95% CI 3.8 to 17.2); larger volumes of the cerebellum (1.2 cc, 95% CI 0.5 to 1.9) and white matter (4.5 cc, 95% CI 0.7 to 8.3); and greater FA in the left cingulum (0.3%, 95% CI 0.1% to 0.6%), right uncinate fasciculus (0.2%, 95% CI 0.0% to 0.5%), and right posterior limb of the internal capsule (0.3%, 95% CI 0.03% to 0.6%). Fat Z-scores were not associated with any outcome.

Conclusions: Lean mass-but not fat-at term was associated with larger brain volume and white matter microstructure differences that suggest improved maturation. Lean mass accrual may index brain growth and development.