An exploration of sociodemographic and psychosocial determinants of cognitive performance in a peri-urban clinic population of people with HIV in Cape Town, South Africa

Dreyer, Anna Jane

An exploration of sociodemographic and psychosocial determinants of cognitive performance in a peri-urban clinic population of people with HIV in Cape Town, South Africa

Doctoral Thesis

2023

Abstract

Introduction. Numerous studies, conducted in many different countries, report that cognitive impairment is highly prevalent in people with HIV (PWH). Such impairment can affect adherence to antiretroviral therapy (ART), and adherence is, in turn, essential for PWH to achieve viral suppression. The gold standard to confirm cognitive impairment is a neuropsychological assessment. However, accurate interpretation of neuropsychological test performance requires consideration of, for instance, how impairment is determined and how accurately the contribution of non-HIV factors to poor cognitive test performance is described. These non-HIV factors include sociodemographic variables (e.g., age, sex, educational attainment), psychosocial variables (e.g., socioeconomic status, food security, quality of life), psychiatric variables (e.g., depression, problematic alcohol use), and other medical co-morbidities. Because many existing studies of PWH do not account for (a) the fact that current quantitative methods for defining cognitive impairment may not accurately reflect HIV-associated brain injury, and (b) possible contributions of non-HIV factors to cognitive test performance, it is possible that the reported prevalence rates of cognitive impairment in PWH are inaccurate (or, at least, do not solely reflect the contributions of HIV disease to the impairment). Another uncertainty in the HIV neuropsychology literature concerns sex differences in the cognitive performance of PWH. Some recent studies suggest that women with HIV (WWH) may present with greater cognitive impairment than men with HIV (MWH). Such a sex difference is of potentially significant concern for South African clinicians because two-thirds of the population of PWH in this country are women. However, there is no definitive empirical evidence regarding whether this sex difference exists to a clinically significant degree (in South Africa, specifically, as well as globally) and what its underlying mechanisms might be. To address the knowledge gaps outlined above, this thesis set out to explore the following aims: (1) investigate sex differences in the cognitive performance of PWH by reviewing the current published literature; (2) determine if sex differences exist in a clinic sample of South African PWH; (3) determine how much variation in reported prevalence rates of HIV-associated cognitive impairment are due to the method used to define impairment, and which method correlates best with MRI biomarkers of HIVrelated brain injury in a South African sample of PWH; (4) investigate the contribution of sociodemographic and psychosocial variables, as well as HIV-disease factors and other medical and psychiatric comorbidities, to cognitive performance in a South African sample of PWH; and (5) investigate associations between cognitive performance and ART adherence in 10 a South African sample of PWH. Each of these aims was explored in a separate study. Hence, this thesis reports on findings from five separate journal manuscripts. Method. Study 1 was a systematic review and meta-analysis summarizing the findings of published studies investigating differences in cognitive performance between WWH and MWH. An extensive systematic search of the literature across several databases found 4062 unique articles of potential interest. After thorough screening of that pool of articles, 11 studies (total N = 3333) were included in the narrative systematic review and 6 studies (total N = 2852) were included in the meta-analysis. Effect sizes were calculated to estimate between-sex differences in cognitive performance, both globally and within discrete cognitive domains. Study 2 investigated sex differences in cognitive performance in a sample of PWH with comorbid MDD (N = 105). All participants were attending community clinics in Khayelitsha, a peri-urban community in Cape Town, South Africa, and were part of a larger research program for a randomised controlled trial of a cognitive-behavioral treatment for ART adherence and depression (CBT-AD). As part of this program, they completed baseline neuropsychological, psychiatric, and sociodemographic assessments. T-tests and multivariable regressions controlling for covariates compared baseline cognitive performance of WWH and MWH, both globally and within discrete cognitive domains. Study 3 applied 20 different quantitative methods of determining cognitive impairment to existing data from a different sample of PWH (N = 148). These individuals had also been recruited from community clinics in Khayelitsha, and had completed a comprehensive neuropsychological assessment and a 3T structural MRI and diffusion tensor imaging (DTI) session. Logistic regression models investigated the association between each method and HIV-related neuroimaging abnormalities. Study 4 again used data from the sample of PWH with comorbid MDD who participated in the larger CBT-AD research program. This study investigated which sociodemographic, psychosocial, psychiatric, and medical variables (as measured at baseline) were associated with baseline cognitive performance. Post-baseline, 33 participants were assigned to CBT-AD and 72 to standard-of-care treatment; 81 participants (nCBT-AD = 29) had a follow-up assessment 8 months post-baseline. This study also investigated whether, from baseline to follow-up, depression and cognitive performance improved significantly more in the participants who had received CBT-AD, and examined associations between post-intervention improvements in depression and cognitive performance. Study 5 assessed ART adherence in the same sample of PWH with comorbid MDD. Mixed-effects regression models estimated the relationship between ART adherence (as measured by both self-report and objective measures, and by degree of HIV viral suppression) with cognitive performance 11 and with other sociodemographic, psychosocial, and psychiatric variables at both baseline and follow-up. Results. Study 1: Analyses suggested that, in terms of overall cognitive functioning, there were no significant differences in cognitive performance between WWH and MWH. However, WWH did perform significantly more poorly than MWH in the domains of psychomotor coordination and visuospatial learning and memory. Additionally, the review suggested that cognitive differences between WWH and MWH might be accounted for by sex-based variation in educational and psychiatric characteristics among study samples. Study 2: Analyses suggested that, in our sample of PWH with comorbid MDD, there were no significant differences in cognitive performance between WWH and MWH. Study 3: Findings suggested that there was marked variation in rates of cognitive impairment (20– 97%) depending on which method was used to define impairment, and that none of these methods accurately reflected HIV-associated brain injury. Study 4: Analyses suggested that less education and greater food insecurity were the strongest predictors of global cognitive performance. Improvement in depression severity was not significantly associated with improved cognitive performance, except in the domain of Attention/Working Memory. Overall, factors associated with cognitive performance were unrelated to HIV disease and other medical factors. Study 5: Analyses identified poor global cognitive performance as a potential barrier to achieving HIV suppression. Conclusion. Taken together, the findings from the five studies contained within this thesis suggest that one oft-mooted sociodemographic influence on cognitive performance in PWH, sex, was not a consistent influence on such performance. However, non-biological (mainly psychosocial and socioeconomic) factors were stronger predictors of cognitive performance in PWH than medical factors (including HIV-disease variables). Current quantitative criteria for defining cognitive impairment in PWH also do not accurately reflect the biological effects of HIV in the brain. The implication of these findings is that research studies may be misclassifying PWH as cognitively impaired and consequently overestimating the prevalence of cognitive impairment in this population. When conducting clinical assessments of PWH, future research studies should measure and consider the strong influence of psychosocial and socioeconomic factors on cognitive test performance. Ideally, a diagnosis of impairment should only be made after a comprehensive clinical assessment that includes a detailed history taking. Overall, we need new criteria for defining cognitive impairment in diverse global populations of PWH. Ideally these criteria should be applicable to both research and clinical settings. Assessing for cognitive impairment among PWH and then providing 12 appropriate support could help achieve viral suppression in patients with non-optimal adherence to ART. At public policy levels, addressing larger psychosocial issues (e.g., food insecurity and low educational attainment) may also help improve cognitive performance in PWH.

Keywords

psychiatry

mental health

Reference:

Collections

PhD / Doctoral

Full item page