The association between depressive symptoms and neurocognitive impairment in people with well-treated HIV in Switzerland

Background: Depression may contribute to neurocognitive impairment (NCI) in people with HIV (PWH). Attributing NCI to depression rather than to HIV is complicated as depression may be both a causal factor and an effect of NCI. This study aimed to determine the association between depressive symptoms and NCI among PWH with well-controlled infection.

Methods: The Neurocognitive Assessment in the Metabolic and Ageing Cohort study is an ongoing, prospective, longitudinal study of PWH aged ≥45 years old nested within the Swiss HIV Cohort Study. Neurocognitive Assessment in the Metabolic and Ageing Cohort study participants underwent neurocognitive assessment and grading of depressive symptoms using the Centre for Epidemiological Studies Depression Scale. Neurocognitive impairment categories were defined using Frascati criteria. Participants with NCI related to neurological or psychiatric confounders other than depression were excluded. The cross-sectional association between the Centre for Epidemiological Studies Depression score and neurocognitive impairment was examined taking Centre for Epidemiological Studies Depression score as a continuous variable and then as a binary variable using two score thresholds, 16 and 27.

Results: Excluding 79 participants with confounding factors, 902 participants were studied: 81% were men; 96% had plasma viral loads <50 copies/ml; 35% had neurocognitive impairment; 28% had Centre for Epidemiological Studies Depression scores ≥16. Higher Centre for Epidemiological Studies Depression scores were associated with female sex (p = 0.0003), non-Caucasian origin (p = 0.011) and current/past intravenous drug use (p = 0.002). Whilst neurocognitive impairment was associated with higher Centre for Epidemiological Studies Depression scores, the Centre for Epidemiological Studies Depression score was a poor predictor of having neurocognitive impairment (area under the ROC curve 0.604). Applying a Centre for Epidemiological Studies Depression score threshold of 16 predicted the presence of neurocognitive impairment with a sensitivity of 38.3% (specificity 77.2%), increasing the threshold to 27 lowered sensitivity to 15.4% (specificity 93.6%).

Conclusion: In this large cohort of PWH in Switzerland, we did not observe a Centre for Epidemiological Studies Depression score threshold that was sensitive in predicting neurocognitive impairment. As neurocognitive impairment was however associated with higher Centre for Epidemiological Studies Depression scores, the data support the screening for and treatment of depression among PWH diagnosed with neurocognitive impairment.