In order to identify abnormal neurodevelopment it is critical to understand normal growth. Throughout childhood the brain develops structurally, functionally as well as biochemically. Few longitudinal studies have examined how imaging measures change with age in normal developing children.

Magnetic resonance spectroscopy (MRS) measures localized brain metabolism providing information about brain maturation. N-acetyl-aspartate (NAA) is associated with neuronal density and integrity, and previous studies showed constant levels with age ^1,2 and age-related increases ^3-7 across different regions in childhood. GPC + PCh (choline) is a marker of cell membrane density and integrity that is consistently found to be constant in childhood.^1,2 Cr + PCr (creatine) is related to energy metabolism and is expected to be stable in childhood.^1-3 Several studies have examined age related-profiles of metabolite levels in healthy children.^1-5 These studies are limited by wide age ranges, small sample sizes, cross-sectional analysis and focus primarily on metabolites ratios.

The purpose of this study is to determine normal maturational trajectories in a representative pediatric population in South Africa from 5 to 9 years of age. In South Africa, 95% of HIV-positive pregnant women and 68% of HIV-exposed infants have been receiving antiretroviral therapy (ART). ^8,9 As such, part of the local pediatric population includes HIV-exposed, uninfected (HEU) children. The risks of exposure to ART and HIV in utero to brain development are expected to be minimal, however the long-term consequences have not been established. ^10,11 We hypothesize that NAA levels will increase with age and creatine and choline levels will remain constant across all children.

Subjects. Children studied were a subset of a longitudinal neuroimaging study, including HEU and HIV-unexposed, uninfected (HUU) children. Participants included twenty-three 5-year-old (mean age ± standard deviation = 5.6 ± 0.4 years; 12 boys; 13 Xhosa/10 Cape Coloured; 11 HEU/12 HUU), forty-five 7-year-old (7.2 ± 0.1 years; 25 boys; 36 Xhosa/9 Cape Coloured; 18 HEU/27 HUU), and twenty-two 9-year-old (9.2 ± 0.1 years; 14 boys; 13 Xhosa/9 Cape Coloured; 11 HEU/11 HUU) children. Nine children were imaged at all ages and fourteen children at two time points. All HEU children were exposed to antiretroviral treatment for prevention of mother-to-child transmission.

MRI. The protocol included a high-resolution T1-weighted acquisition for voxel placement and segmentation as well as single voxel 1H-MRS in the midfrontal gray matter (MFGM) with a real-time motion and B0 corrected ¹² point resolved spectroscopy (PRESS) sequence (TR/TE 2000/30 ms, 64 averages, Scan Time: 2:16 min) performed on a Siemens 3T Allegra Head Scanner (Siemens, Erlangen, Germany) in Cape Town, South Africa. SPM12 was used for voxel segmentation to determine tissue type percentages for water concentration and partial volume calculations. Eddy current compensation and frequency/phase correction were performed. Absolute metabolite levels were calculated with LCModel.

Statistics. The R-programming language was used for statistical analysis. To study the relationship between metabolite level and age, we used a linear mixed effects regression model to account for repeated measures in some children. Gender, HIV-exposure, ethnicity, metabolite standard deviation and signal-to-noise ratio were all included as confounds.

NAA levels increased across all children with age (slope = 0.26; p < 0.0001), as shown in figure 1. Choline levels remained constant with age (figure 2). Contrary to our hypothesis, creatine levels increased with age (slope = 0.13; p = 0.004), shown in figure 3. There were no gender, HIV exposure or ethnicity effects.

Our results demonstrate age specific changes in both creatine and NAA levels in normal developing children. The increased NAA levels may be interpreted as healthy neuronal growth in the region. The increase in creatine levels was surprising, in particular since creatine levels are assumed to be constant with age and commonly used in metabolite ratios. The increase in creatine with age suggests an increase in energy metabolism associated with neurodevelopment in the MFGM.

HIV exposure was not a significant confound in the model, providing additional support that ART/HIV exposure in utero does not impact healthy growth in this brain region. However, ART exposure has subsequently increased from 2 to 3 drugs in utero and for longer periods.

Lastly, the lack of gender or ethnicity effects implies these results can be generalized to a wide pediatric population.

The trajectories of NAA, choline and creatine provide a benchmark of normal development in a pediatric population against which pathology and abnormal development may be compared. Caution must be exercised when using metabolite ratios in pediatric populations given the increased creatine levels observed over this age range.