While high active antiretroviral therapy (HAART) has lengthened the life expectancy of patients infected with human immunodeficiency virus (HIV), the risk of cognitive impairment in the aging HIV+ subgroup has continued to increase (1, 2). Premature aging have been hypothesized as new risk factors for HIV associated neurocognitive disorders (HAND) (3). The current study was undertaken to investigate the relation between the neurocognitive changes among elderly HIV+ subjects and the brain metabolic markers as measured by 7T MRS and PET using 18F AV45, a specific tracer targeting amyloid plaques, the first hallmark in Alzheimer’s disease (4).Thirty two elderly HIV+ subjects (24 male, with mean age of 59.8± 5.8) were stratified into 4 groups according to their cognitive status, using the Frascati criteria (5). Seven HIV+ individuals with normal cognition, 8 with asymptomatic neurocognitive impairment (ANI), 10 with mild neurocognitive disorder (MND) and 7 with HIV associated dementia (HAD) were evaluated. All subjects were receiving combination antiretroviral therapy. Twelve HIV- controls (11 male, with mean age of 62.0± 8.8) were also included. Using a 7.0T Philips ‘Achieva’ scanner and 32-channel head coil, brain MRI and single voxel STEAM spectra (TR/TE=3000/14 msec) were acquired from the left frontal white matter (FWM), basal ganglia (BG), Precuneus (PC), Posterior cingulate cortex (PCC) and hippocampus (Hippo) with and without water suppression. The voxel sizes ranged from 8 to 15 cc (Figure 1). Spectra were analyzed using LCModel (6) and quantified in millimolar (mM) concentrations, i.u. relative to the unsuppressed water signal. Metabolite concentrations and ratios relative to creatine (Cr) were calculated. The same subjects had undergone18F AV45 PET imaging on a high resolution research tomography (HRRT) PET scanner. Ten mCi of 18F AV45 tracer was injected. Seventy-nine dynamic PET scans were acquired. Standardized uptake value ratio (SUVR) relative to cerebellum images were calculated from 50 to 70 min post tracer injection (Figure 2). Brain Regions of interest (ROI) were manually drawn on the co-registered MRI and applied to SUVR images. The data was not normally distributed; therefore, comparisons of the groups’ medians and interquartile (IQR) ranges were evaluated for significant differences using non-parametric median tests as well as Spearman correlations.

There were no differences in age among our comparison groups.

On comparing HIV- versus HIV+, HIV+ showed significant decrease in FWM tNAA (NAA+NAAG), BG tNAA/Cr and Glu/Cr [(median (IQR) = 8.0 (7.8-8.5) mM in HIV+ versus 7.7 (7.3 -8.0) mM in HIV- , P=0.05], [(median (IQR) = 1.14 (1.05-1.24) in HIV+ versus 1.23 (1.15 -1.42) in HIV- , P=0.01], and [(median (IQR) = 1.07 (0.99-1.1) in HIV+ versus 1.25 (1.01 -1.27) in HIV-, P=0.04] respectively. HIV+ showed significant increase in Precuneus GABA/Cr [(median (IQR) = 0.2 (0.18-0.26) in HIV+ versus 0.18 (0.15 -0.19) in HIV-, P=0.007]. Precuneus myo-inositol (mI) was higher in HIV+ as compared to HIV-; however, it did not reach statistical significance. In regards to PET imaging, the only significant finding was higher AV45 SUVR in the Precuneus in HIV+ as compared to the HIV- subjects (P= 0.007). On correlating both imaging modalities, there was significant positive correlation of Precuneus SUVR AV45 uptake with MRS mI/Cr and total Cho (GPC+PCH)/Cr (P= 0.007, 0.023 respectively).

On comparing HIV+ patients by cognitive status, there were significant decrease in FWM tNAA, NAA, NAA/Cr and posterior cingulate tNAA/Cr with increasing degrees of cognitive impairment (P=0.008, 0.02, 0.03, 0.028 respectively), (Figure 3, 4, 5). There were no significant differences in PET AV45 SUVR among the 4 groups.

The current study confirms previous findings namely reduced FWM tNAA, NAA, NAA/Cr in in symptomatic HAND (MND and HAD) as compared to ANI HIV+ patients (7, 8, 9), besides adding new findings in other brain regions namely post cingulate tNAA/Cr among HIV+ with significant association with the degree of neurocognitive impairment (3). In this study, the increases in BG Glu/Cr and Precuneus GABA/Cr in HIV+ subjects were not associated with the degree of cognitive impairment; therefore, they may be regarded as early markers of HIV infection rather than markers of neurocognitive impairment progression. The higher AV45 uptake in the Precuneus and the significant positive correlation of Precuneus SUVR AV45 uptake with MRS mI/Cr and tCho/Cr can be interpreted as increased amyloid burden associated with microglial activation in the Precuneus of Elderly HIV+ patients. In conclusion, 7T MRS metabolites changes of tNAA, NAA, NAA/Cr and Glu/Cr can be reliable biomarkers for assessing the stage of HIV seroconversion and the degree of HAND among elderly HIV+ patients and can be used as baseline levels to follow up premature aging and disease progression.