Background: Individuals infected with the human immunodeficiency virus (HIV) are at increased risk of developing ischaemic stroke. The reasons for this are multifactorial, but HIV-associated vasculopathy is a potentially important cause. HIVinduced chronic inflammation may initiate endothelial dysfunction or accelerate vascular injury from other disease processes. Viral proteins such as the transactivator of transcription (Tat) are emerging role-players in HIV disease pathogenesis and have a putative role in HIV-associated endothelial dysfunction. Tat has paracrine proinflammatory effects, but its role in HIV-related stroke has not yet been investigated. Aims: The primary aim of this study was to determine whether specific Tat amino acid variants are associated with ischaemic stroke and biomarkers of inflammation and endothelial dysfunction in a group of HIV-1 subtype-C-infected individuals. In order to do so, I first determined the aetiology of stroke in these participants using clinical, biochemical and neuro-imaging data. A secondary aim of the study was to identify any HIV-related and/or other traditional stroke-related risk factors that might independently or cumulatively increase stroke risk. For comparison, these putative risk factors were also determined in a group of age-matched HIV-infected non-stroke controls. Finally, I aimed to identify any HIV-related factors and/or Tat amino acid variants that might distinguish strokes due to HIV-associated vasculopathy from other mechanisms of stroke. Methods: A case-control study was performed on 58 Subtype-C HIV-infected individuals with acute ischaemic stroke and 71 HIV-1 Subtype-C-infected non-stroke controls. Clinical, demographic, laboratory and imaging data were used to determined baseline differences between groups and to distinguish different stroke aetiologies. Exon 1 of the HIV-1 Tat protein was sequenced from peripheral blood samples of stroke participants and controls and amino acid variants were identified using viral epidemiology signature pattern analysis. Regression analyses were used to examine the correlation between residues at signature positions with biomarkers of inflammation and endothelial activation. Results: Stroke and control groups were mostly young (mean age 33 years) females (62.1% & 71.8%), and of Black African ancestry. The strokes showed a higher prevalence of some traditional cardiovascular risk factors. Individuals with strokes had a higher prevalence of antiretroviral treatment interruption (25.9% vs 0.0%, p= 0.003), lower CD4 nadir (112 vs 177.5 cells/μl, p=0.008) and CD4 count (208.5 vs 322.5 cells/μl, p=0.012) than controls. Median viral loads were elevated in both strokes and controls (4.58 & 4.13 log10 copies/ml, p=0.28). The most common causes of stroke were HIV-associated vasculopathy (43.1%) and opportunistic infections (22.4%). Two amino acid variants (proline at position 21 and histidine at position 29) were associated with acute ischaemic stroke. These positions were also associated with modulation of plasma interleukin 6 and monocyte chemoattractant protein 1 levels. Threonine at position 58 distinguished strokes due to alternative mechanisms from strokes due to HIV-associated vasculopathy. Conclusions: Two Tat protein amino acid variants are associated with stroke in HIV. The precise mechanisms by which these associations occur are not known. However, they are likely to be part of a multiple-hit phenomenon in HIV stroke pathogenesis. Tatmediated inflammation with endothelial dysfunction, HIV disease severity, treatment interruption and conventional cardiovascular risk factors probably all contribute to stroke aetiology. Thus, a multi-modal approach is needed to reduce ischaemic stroke risk in HIV infection.