HIV increases sleep-based brain age despite antiretroviral therapy

Abstract

Study objectives: Age-related comorbidities and immune activation raise concern for advanced brain aging in people living with HIV (PLWH). The brain age index (BAI) is a machine learning model that quantifies deviations in brain activity during sleep relative to healthy individuals of the same age. High BAI was previously found to be associated with neurological, psychiatric, cardiometabolic diseases, and reduced life expectancy among people without HIV. Here, we estimated the effect of HIV infection on BAI by comparing PLWH and HIV- controls.

Methods: Clinical data and sleep EEGs from 43 PLWH on antiretroviral therapy (HIV+) and 3,155 controls (HIV-) were collected from Massachusetts General Hospital. The effect of HIV infection on BAI, and on individual EEG features, was estimated using causal inference.

Results: The average effect of HIV on BAI was estimated to be +3.35 years (p < 0.01, 95% CI = [0.67, 5.92]) using doubly robust estimation. Compared to HIV- controls, HIV+ participants exhibited a reduction in delta band power during deep sleep and rapid eye movement sleep.

Conclusion: We provide causal evidence that HIV contributes to advanced brain aging reflected in sleep EEG. A better understanding is greatly needed of potential therapeutic targets to mitigate the effect of HIV on brain health, potentially including sleep disorders and cardiovascular disease.

Keywords: EEG; HIV; brain age; machine learning; sleep.

Figure 1.

Flowchart of HIV+ participants inclusion and exclusion. N is the number of HIV+ participants eligible following each evaluation step. 43 HIV+ participants with diagnostic studies of sufficient quality were ultimately selected for analysis. CPAP, continuous positive airway pressure.

Figure 2.

Brain ages among HIV+ and HIV− participants. Scatter plot showing each participant’s Chronological Age (CA), the age at the time of the sleep study, versus the Brain Age (BA), the sleep EEG-predicted age. The solid line represents BA = CA, or BAI = 0. Above and below the line are indicated as the BAI > 0 and BAI < 0 regions, respectively.

Figure 3.

BAI is elevated by HIV after adjusting for potential confounders. (A) Causal diagram of the variables. An arrow from a variable X to another variable Y indicates our assumption that X causally affects Y. C is the set of covariates. HIV represents the presence or absence of the exposure to HIV infection. BAI represents the outcome variable, the BAI. The red arrow represents the effect of interest, which is measured as the difference in the expected potential outcomes of BAI in the presence and absence of HIV. (B). Bar chart showing the expected potential outcome of BAI in the absence (HIV−) and presence (HIV+) of HIV. Error bars depict the standard error of the mean (SEM; HIV− = 0.18 years, HIV+ = 1.43 years). The difference in expected potential outcomes of BAI is significant (p < 0.05), indicated by asterisks.

Figure 4.

Individual EEG features underlying brain age are altered by HIV. Rows show features by sleep stage: (A, B) REM. (C, D) N1. (E, F) N2. (G, H) N3. (A,C,E,G) Volcano plots of significance level of changes in potential outcome of BAI features versus log₂ fold change. Dotted lines represent the significance threshold for a FDR of 0.1. (B, D, F, H) Bar charts comparing potential outcomes of specific BAI features in the presence (HIV+) and absence (HIV−) of HIV. Only the significant features are shown. Solid horizontal black lines show SEM.

Figure 5.

Hypnograms and spectrograms of representative HIV− and HIV+ participants. (A) HIV−. (B) HIV+. The x-axis is time since the sleep EEG recording in hours. The upper panel in each subplot shows the trajectory of sleep stages; the lower panel shows the spectrogram with y-axis being frequency in Hz.