Background:
Recent findings from the UK Biobank revealed that healthy adults who later became infected with SARS-CoV-2 had lower brain volumes in regions involved in risk-taking behavior and olfaction compared to individuals who did not become infected. We examined if similar pre-existing differences in brain regions correspond to SARS-CoV-2 infection among people with HIV (PWH) receiving suppressive ART.
Methods:
Participants included adult Thai MSM enrolled in the acute HIV (AHI) cohort (RV254/SEARCH010) in Bangkok, Thailand. Participants underwent 3T MRI and clinical assessments (i.e., HIV disease metrics, cognitive testing, and self-reported mood and substance use). ART initiation occurred within 5 days of the MRI (median=same day). Regional brain volumes were summed across hemispheres and corrected for head size. Brain volumes and clinical indices were compared between participants with laboratory confirmed SARS-CoV-2 and those without a diagnosis of SARS-CoV-2 following ART initiation. Machine learning was utilized to identify variables at the time of enrollment into the cohort that predicted subsequent SARS-CoV-2 infection status.
Results:
112 participants were included in the analysis. All study participants achieved viral suppression after ART and received SARS-CoV-2 vaccinations. Fifty-four participants became infected with SARS-CoV-2 during the observation period (median=79 weeks from ART initiation). Study participants who became infected with SARS-CoV-2 after ART had lower volumes at the time of enrollment in several subcortical brain regions with the most pronounced effect in the pallidum (p=.025). There were no associations between brain volumes and ratings of mood, demographics, or HIV disease indices. SARS-CoV-2 infection was two-fold higher among individuals who reported use of amyl nitrites (i.e., poppers) during “chemsex”. Machine learning with repeated cross validation revealed that lower orbital and medial frontal lobe, anterior cingulate, pallidum, vermis, and olfactory volumes, worse motor function, and higher education collectively predicted co-infection status (average AUC of 85%).
Conclusions:
Study findings point toward a risk phenotype for SARS-CoV-2 infection among PWH defined by pre-existing differences in brain volumes relevant to risk-taking behavior, emotion, and neuroHIV as well as behavioral factors such as inhalant use and lack of social distancing during “chemsex”.
Table 1. Demographic and HIV disease indices.