Abstract
Objectives:
To examine associations of immune biomarkers with neurocognition among Ugandan Youth living with Perinatally acquired HIV (YPHIV) virally suppressed on dolutegravir (DTG) versus Youth Without HIV (YWoH).
Methods:
In this cross-sectional study of 101 age- and sex-matched youth in Kampala, Uganda (49 YPHIV, 52 YWoH), neurocognition was assessed using NeuroScreen. Biomarkers were measured using LEGENDplex™ and ELISA. Group differences were assessed using Wilcoxon rank-sum tests. Associations were evaluated using Spearman correlations and multivariable linear regression.
Results:
Median age was 16.9 years [IQR 15.1–17.1]. YPHIV performed worse across all neurocognitive domains (P ≤ 0.035). IFAB, MIP-3α, and RANTES were higher among YPHIV (P ≤ 0.024). In adjusted YPHIV models, myeloperoxidase was associated with worse processing speed (β = −3.24, 95% CI = −4.89, −1.58), learning and memory (β = −3.53, 95% CI = −6.48, −0.59), and global z-scores (β = −1.64, 95% CI = −2.78, −0.51); soluble CD14 (sCD14) with worse processing speed (β = −0.78, 95% CI = −1.55, −0.01); β-d-Glucan (BDG) with worse learning and memory (β = −0.35, 95% CI = −0.68, −0.03).
Conclusions:
Ugandan YPHIV demonstrated worse neurocognitive performance. Biomarkers linked to HIV-related mortality and cardiovascular disease (MPO, sCD14, BDG) were associated with poorer performance.
Keywords: Neurocognition, Perinatally acquired HIV, Inflammatory markers, Vascular markers, Chemokines
1. Introduction
Although combination antiretroviral therapy (cART) has significantly improved life expectancy, Youth living with Perinatally acquired HIV (YPHIV), who represent a substantial proportion of the roughly 1.5 million adolescents living with HIV globally, continue to experience disproportionately high rates of neurocognitive impairment [1–3]. Multiple mechanisms have been proposed to explain this effect, including persistent viral replication in the central nervous system (CNS), impaired gut barrier and microbial translocation, activation of cellular immune pathways, CNS damage before the initiation of cART, neurotoxic effects of cART, and socioeconomic factors [4–7]. Prior studies of children living with HIV have demonstrated associations between monocyte activation and proinflammatory markers in blood and cerebral spinal fluid (CSF) with neurocognitive performance [8]. In this cohort, we have previously observed persistent activation of T cells and monocytes among Ugandan YPHIV [9, 10] which was associated with poorer neurocognitive performance relative to Youth Without HIV (YWoH) [11]. Additionally, T-cell activation in this population has been linked to subclinical cardiovascular disease [12].
The present study builds on prior studies by evaluating an expanded panel of inflammatory, chemokine, gut microbial translocation, and vascular markers in the setting of 100% viral suppression on Dolutegravir (DTG) using a tablet-based neurocognitive assessment tool. The new integrase inhibitor, DTG, has become a cornerstone of first-line cART worldwide since its introduction in 2013 [13]. DTG achieves high concentrations in the CNS [14], however there are conflicting data as to whether DTG offers benefit in neurocognition [15, 16]. Additional studies are needed to clarify the neurocognitive trajectory of pediatric patients treated with DTG–based regimens.
The primary objective of this study was to assess associations between markers of vascular and immune activation with neurocognitive performance among both YPHIV and controls using general linear regression models. A secondary objective was to assess socioeconomic factors and their impacts on neurocognitive performance. Gaining insight into how inflammation, microvascular dysfunction, and neurocognition relate to one another may help shape clinical screening practices and lay the groundwork for future interventions in YPHIV.
2. Methods
2.1. Study design
This study is a cross-sectional analysis from an observational cohort study of YPHIV and YWoH prospectively enrolled at the Joint Clinical Research Center in Kampala, Uganda between January 2023 to March 2024. Approval for the study was acquired from the Research Ethics Committee in Uganda, the Ugandan National Council of Science and Technology, the New York State Psychiatric Insti tute IRB, and the University Hospitals Cleveland Medical Center IRB, Cleveland, OH. Parent/caregivers gave written informed consent and participants gave assents. Age- and sex-matched participants between the ages of 10–18 years were recruited from the same area, including various clinics and communities in Kampala, Uganda. A parent study of these same participants showed both groups had similar socioeconomic profiles [9, 12]. The control group excluded individuals with exposure to HIV in utero; accordingly, siblings of YPHIV participants were not recruited. Enrollment criteria for the YPHIV group included youth with documented HIV-1 infection who have a cumulative duration of cART of at least 2 years, have been on the same regimen for at least 6 months, and have an HIV-1 RNA level < 50 copies/ml. The YWoH group included youth with no documented HIV infection, who tested negative for HIV at every study visit, and had no known medical condition requiring chronic use of prescription medications. Participants were excluded if they had an acute infection (i.e. malaria, helminthiasis, acute pneumonia, meningitis, active tuberculosis), had a history of meningitis in the past 3 months, had a history of cerebral malaria, were pregnant, or had an intent to become pregnant during the study period. Children with diabetes were also excluded. For NeuroScreen (NS) testing, only participants who speak English and/or Luganda were included.
2.2. Socioeconomic measures
The participants or primary caregiver were given a 12-item questionnaire that incorporated the WHO STEP-wise Approach to NCD Risk Factor Surveillance (STEPS) instrument [17] and the US-AID Demographic and Health Surveys Wealth Index [18]. STEPS assesses non communicable diseases (NCDs) primarily through standardized surveillance of NCD risk factors, rather than through definitive clinical diagnosis. It is designed for population-level monitoring, especially in low- and middle-income settings. The US-AID Wealth Index assesses a household’s cumulative living standard including types of water access and sanitation facilities. Living below extreme poverty line was defined as living with < $2.15 a day as set by the World Bank as the International Poverty Line before June 2025 [19].
2.3. Plasma markers
Blood was drawn after an 8-hour fast and plasma and serum were cryopreserved for shipment to Cleveland Medical Center, Cleveland, Ohio. Samples were stored at 80 °C and were thawed once, immediately before running commercially sourced LEGENDplex™ and ELISA assays. The LEGENDplex™ Human Vascular Inflammation Panel (BioLegend, San Diego, CA) one is a multiplex bead-based assay kit used to simultaneously quantify 13 human proteins (CRP, Cystatin C, ICAM-1, IGBP4, MMP-2, MMP-9, MPO, MRP8/14, Myloglobin, NGAL, OPN, SAA, and VCAM-1) (See Supplement 1 for full list of abbreviations). Similarly, the LEGENDplex™ Human Proinflammatory Chemokine Panel 1 (BioLegend, San Diego, CA) was used to quantify 13 human chemokines (ENA-78, Eotaxin, GROα, I-TAC, IL-8, IP-10, MCP-1, MIG, MIP-1α, MIP-1β, MIP-3α, RANTES, and TARC). Plasma markers of systemic inflammation, gut integrity, and bacterial translocation were measured by ELISA. Specifically, β-d-Glucan (BDG, My-biosource Inc. CA), is a polysaccharide cell wall component of most fungal species. Intestinal Fatty Acid Binding Protein (IFAB, R&D Systems, Minneapolis, MN) is a marker of enterocyte inflammation and intestinal barrier dysfunction. Interleukin 6 (IL-6), interferon-γ-Inducible Protein 10 (IP-10), Tumor Necrosis Factor-α receptor I (TNFR1), and Tumor Necrosis Factor-α receptor II (TNFR2) are all systemic inflammation markers (R&D Systems, Minneapolis, MN). Soluble CD14 (sCD14) and soluble CD163 (sCD163) are markers of monocyte activation (R&D Systems, Minneapolis, MN). Lipopolysaccharide binding protein (LBP) is a measure of microbial translocation (Hycult Biotech, The Netherlands). Finally, d-dimer is a measure of coagulation (Diagnostica Stago, France). To ensure quality control, 10% of the samples were run in duplicate. All assays were performed in Dr. Funderburg’s lab at The Ohio State University, Columbus, OH. Laboratory personnel were blinded to individual HIV status.
2.4. Neurocognitive assessment
A tablet-based tool, NeuroScreen (NS), was utilized to assess neurocognition. The test was offered in the English or Luganda language. NS has previously been validated for use in Luganda speaking populations and among HIV populations, including YPHIV [20–22]. The battery of 12 tests in the NS assessment evaluated five neuropsychological domains affected in HIV. Test battery components making up each domain include (1) processing speed (two trail-making test sequencing paradigms, two visual discrimination tasks, and a number input task), (2) working memory (a number span forward and backward task), (3) learning and memory (two trials of learning five words followed by 5-minute delayed recall), (4) executive functioning (an alternating trail-making test sequencing paradigm), and (5) motor speed (tests of finger-tapping with dominant and nondominant hands). During analysis, the scores of timed tests were reverse scored so that higher scores indicated better performance. Raw scores were converted to z-scores relative to the YWoH control group. A global z-score was also calculated for each participant by averaging the z-scores for each of the 12 tests.
2.5. Statistical analysis
Wilcoxon rank-sum and chi-square tests were used to compare all variables between groups (YWoH vs YPHIV). Spearman correlation was used to assess pairwise correlations between NS and biomarkers. General linear regression models were applied to evaluate associations of biomarkers with neurocognitive test performance while adjusting for covariates of age, gender, education, income, dietary diversity, electricity, and clean water source. Covariate selection was performed a priori based on known relevance to neurocognitive performance and biomarker expression. All statistical analyses were performed using SAS Enterprise Guide 7.15HFC (SAS Institute Inc., Cary, NC).
3. Results
3.1. Baseline characteristics
This analysis included a total of 101 participants (52 YWoH and 49 YPHIV) (Table 1). The median age of participants was 16.9, IQR [15.1, 17.1] years and 58% were girls. The YWoH cohort was younger compared to the YPHIV cohort (16.6, IQR [15.7, 17.8] v. 17.4, IQR [16.3, 18.1], P = 0.033). One YPHIV participant was on Darunavir boosted Ritonavir and DTG (DRV/r + DTG) and the rest of the YPHIV were on Tenofovir Disoproxil Fumarate (TDF), Lamivudine (3TC), and DTG (TLD). These participants have been followed since 2017 as part of the larger longitudinal parent study [12] and 85% of participants have remained virally suppressed (< 50 copies/ml) at each annual visit. The transition to DTG in Ugandan children started in 2020 and was completed by 2023. Prior regimens included protease inhibitors and non-nucleoside reverse transcriptase inhibitors. Median cART duration for YPHIV was 173.5 months (148.0, 185.0) and 92% initiated cART between the ages of 2 and 8. The median nadir CD4 cell count among YPHIV at cross-sectional analysis was 604 and HIV-1 viral load was undetectable (< 50 copies/ml) in all participants. Median WHO clinical stage—a four-level classification of HIV disease severity—was two (mild symptomatic), with only seven participants meeting criteria for stage 4 (AIDS-defining illness), suggesting that most participants had achieved immune recovery. Compared to YWoH, YPHIV had less access to electricity (90.4% v. 75.5%, P = 0.046), higher economic insecurity (defined as earning < $2.15 a day) (9.6% v. 28.6%, P = 0.015), and a lower dietary diversity (8.0, IQR [7.0, 9.0] v. 7.0, IQR [6.0, 8.0], P = 0.049).
Table 1.
Baseline characteristics of Ugandan youth without HIV (YWoH) and Ugandan youth with perinatally acquired HIV (YPHIV).
| HIV status | |||
|---|---|---|---|
| YWoH (n = 52) | YPHIV (n = 49) | P-value | |
| Age (years) | 16 (15, 17) | 17 (16, 18) | 0.033 |
| Female | 32 (62%) | 27 (55%) | 0.512 |
| Male | 20 (39%) | 22 (45%) | 0.512 |
| HIV history | |||
| HIV RNA status (<50 c/ml) | 49 (100%) | ||
| Antiretroviral Duration (in years) | 15 (13, 16) | ||
| CD4 percent | 37 (32, 43) | ||
| CD4 absolute value | 981 (721, 1249) | ||
| cART regimen: TDF/3TC/DTGa | 48 (98%) | ||
| Socioeconomics | |||
| Access to clean water source | 47 (90%) | 41 (84%) | 0.314 |
| Electricity at home | 47 (90%) | 37 (76%) | 0.046 |
| Education | |||
| Primary school | 10 (20%) | 16 (35%) | 0.123 |
| Secondary school | 41 (80%) | 29 (63%) | 0.123 |
| University | 0 (0%) | 1 (2%) | 0.123 |
| Extreme povertyb | 5 (10%) | 14 (29%) | 0.015 |
| Food insecure | 1 (2%) | 0 (0)%) | 0.329 |
| Dietary diversity | 8 (7, 9) | 7 (6, 8) | 0.049 |
| Vital signs | |||
| Body Mass Index | 20 (18, 22) | 20 (19, 22) | 0.265 |
| Diastolic blood pressure | 68.5 (62, 74) | 67 (62, 70) | 0.180 |
| Systolic blood pressure | 116 (109, 121) | 113 (108, 118) | 0.083 |
| Tanner stage | 4 (3, 5) | 4 (4, 5) | 0.301 |
| Waist:Hip ratio | 0.8 (0.8, 0.9) | 0.8 (0.8, 0.8) | 0.276 |
Categorical variables presented as total (percent total).
Ordinal variables presented as median (IQR).
Antiretroviral therapy (ART): Tenofovir Disoproxil Fumarate (TDF), Lamivudine (3TC), and Dolutegravir (DTG).
Extreme poverty defined as living with < $2.15 a day as set by the World Bank as the International Poverty Line.
3.2. Plasma biomarkers by HIV status
Compared to YWoH, YPHIV plasma contained higher concentrations of the gut integrity marker IFAB (1899.2 pg/ml [1314.3, 2807.9] v. 2434.6 pg/ml [1848.6, 3631.6], P = 0.004) and higher concentrations of chemokines Macrophage Inflammatory Protein-3α (MIP-3α) (48.2 pg/ml [32.1, 71.7] v. 56.1 pg/ml [42.9, 89.3], P = 0.024) and Regulated upon Activation, Normal T-cell Expressed and Secreted (RANTES) (708.1 pg/ml [429.2, 1133.9] v. 1027.1 pg/ml [692.7, 1430.3], P = 0.013) (Table 2). The acute phase protein Serum Amyloid A (SAA) was the only marker significantly elevated in YWoH compared to YPHIV (454504 pg/ml [223522, 859142] v. 223522 pg/ml [147042, 411298], P = 0.016).
Table 2.
Immune biomarkers in Ugandan youth without HIV (YWoH) and Ugandan youth with perinatally acquired HIV (YPHIV).
| HIV status | |||
|---|---|---|---|
| YWoH (n = 52) median (IQR) | YPHIV (n = 49) median (IQR) | P-value | |
| Gut/translocation markers | |||
| BDG (pg/ml) | 948.5 (734.0, 2237.4) | 865.2 (665.7, 2316.1) | 0.246 |
| IFAB (pg/ml) | 1899.2 (1314.3, 2807.9) | 2434.6 (1848.6, 3631.6) | 0.004 |
| Immune assay | |||
| CD163 (ng/ml) | 501.8 (407.1, 755.2) | 551.3 (457.2, 756.5) | 0.353 |
| D-Dimer (ng/ml) | 543.2 (362.3, 917.7) | 518.4 (306.4, 790.7) | 0.091 |
| IL-6 (pg/ml) | 0.8 (0.6, 1.3) | 0.8 (0.5, 2.1) | 0.394 |
| IP-10 (pg/ml) | 74.5 (57.5, 109.5) | 83.7 (61.0, 123.5) | 0.091 |
| LBP (ng/ml) | 11319 (8091.5, 16356) | 10153 (6222.3, 14138) | 0.095 |
| Oxidized LDL (av) | 61608 (51576, 76425) | 63025 (53274, 72220) | 0.411 |
| sCD14 (ng/ml) | 1590.0 (1373.8, 1756.8) | 1494.2 (1306.9, 1743.1) | 0.192 |
| TNFR1 (pg/ml) | 759.4 (658.0, 855.6) | 733.6 (623.3, 829.7) | 0.163 |
| TNFR2 (pg/ml) | 2058.0 (1726.4, 2531.4) | 2092.7 (1706.7, 2569.9) | 0.469 |
| Vascular markers (all pg/ml) | |||
| CRP | 699221 (379509, 1.59 × 106) | 611030 (300477, 1.33 × 106) | 0.218 |
| Cystatin C | 239820 (123738, 286460) | 266685 (211247, 340459) | 0.051 |
| ICAM-1 | 39072 (14226, 149397) | 14226 (3553.5, 67376) | 0.060 |
| IGFBP4 | 179883 (102659, 311703) | 140765 (104090, 226548) | 0.119 |
| MMP-2 | 477449 (210946, 693675) | 261224 (210946, 693675) | 0.055 |
| MMP-9 | 379355 (67539, 427530) | 155756 (52025, 427530) | 0.096 |
| MPO | 270927 (109310, 270927) | 270927 (109310, 270927) | 0.257 |
| MRP8/14 | 1.64 × 106 (373662, 2.04 × 106) | 486150 (288677, 2.04 × 106) | 0.128 |
| Myoglobin | 50613 (44004, 70060) | 48115 (33209, 65838) | 0.118 |
| NGAL | 71868 (55821, 96247) | 69391 (49564, 97186) | 0.288 |
| OPN | 238662 (56815, 239433) | 134992 (60822, 238662) | 0.107 |
| SAA | 454504 (223522, 859142) | 223522 (147042, 411298) | 0.016 |
| VCAM-1 | 237272 (232549, 405513) | 287431 (223515, 430493) | 0.336 |
| Chemokines (all pg/ml) | |||
| ENA-78 | 994.0 (641.7, 1481.0) | 863.5 (510.4, 1720.8) | 0.384 |
| Eotaxin | 86.6 (61.4, 116.8) | 93.6 (73.9, 122.4) | 0.160 |
| GROα | 552.5 (336.2, 724.1) | 467.9 (217.5, 800.4) | 0.260 |
| I-TAC | 77.0 (52.2, 184.8) | 79.6 (58.6, 135.5) | 0.399 |
| IL-8 | 119.6 (100.1, 165.6) | 103.7 (101.3, 161.1) | 0.376 |
| MCP-1 | 86.2 (54.4, 117.0) | 82.6 (55.7, 135.3) | 0.275 |
| MIG | 421.0 (312.3, 609.5) | 464.3 (362.1, 701.6) | 0.089 |
| MIP-1α | 153.6 (116.4, 271.4) | 247.1 (116.4, 247.1) | 0.321 |
| MIP-1β | 15.0 (9.5, 24.2) | 15.0 (9.5, 23.8) | 0.471 |
| MIP-3α | 48.2 (32.1, 71.7) | 56.1 (42.9, 89.3) | 0.024 |
| RANTES | 708.1 (429.2, 1133.9) | 1027.1 (692.7, 1430.3) | 0.013 |
| TARC | 343.1 (233.8, 423.8) | 267.9 (129.9, 619.9) | 0.434 |
See Supplement 1. for definitions of immune, vascular, and chemokine abbreviations.
3.3. NeuroScreen performance by HIV status
The global z-score for NS was lower among the YPHIV group compared to controls (0.0 [−0.3, 0.3] v. −0.5 [−0.8, 0.1], P = 0.001) (Figure 1). In addition, the YPHIV group scored lower on various neurocognitive sub-domains. The YPHIV cohort scored lower on 4/5 tests for processing speed (number speed test (0.1 [−0.5, 0.8] v. −0.1 [−1.0, 0.5], P = 0.034), trail making test #3 (0.2 [−0.8, 0.8] v. −0.2 [−1.1, 0.4], P = 0.031), visual discrimination test #1 (0.1 [−0.6, 0.7] v. −0.6 [−1.5, 0.4], P = 0.035), and visual discrimination test #2 (0.2 [−0.6, 0.6] v. −0.4 [−1.2 v. 0.2], P = 0.004)). In addition, the YPHIV cohort scored lower on 1/2 tests for motor speed (finger tapping nondominant hand (0.1 [−0.6, 0.6] v. −0.3 [−1.6, 0.2], P = 0.004)), 1/2 tests for learning and memory (verbal list learning (0.8 [−0.6, 0.8] v. −0.6 [−2.1, 0.8], P = 0.005)), and 2/2 tests for working memory (number span forward (0.4 [−0.7, 0.4] v. −0.7 [−0.7, 0.4], P = 0.002) and number span backward (−0.1 [−0.1, 1.2] v. −0.1 [−1.4, −0.1], P = 0.038)). The YPHIV cohort also scored lower on 1/1 test of executive function (trail making test #2 (0.4 [−0.4, 0.7] v. −0.3 [−1.0, 0.4], P = 0.018).
Figure 1.
NeuroScreen z-scores by neurocognitive domain in Ugandan Youth Without HIV (YWoH) and Ugandan Youth with Perinatally acquired HIV (YPHIV). Boxplots show domain-specific and global performance on NeuroScreen tablet-based test battery. YPHIV demonstrate significantly lower scores across all domains (P ≤ 0.035).
After adjusting for socioeconomic covariates of age, gender, education, income, dietary diversity, electricity, and clean water source, positive HIV status remained significantly correlated with worse working memory (β = −0.414, 95% CI: −0.738, −0.089, P = 0.013), motor speed (β = −0.430, 95% CI: −0.843, −0.016, P = 0.042), and learning and memory (β = −0.380, 95% CI:−0.718, −0.042, P = 0.028), but not executive functioning (β = −0.275, 95% CI: −0.669, 0.119, P = 0.168) or processing speed (β = −0.325, 95% CI: 0.055, −0.658, P = 0.055).
3.4. NeuroScreen and biomarkers
Among all participants, the monocyte activation marker sCD14 and the chemoattractant Interferon-inducible T cell Alpha Chemoattractant (I-TAC) correlated with an overall worse global z-score (r = −0.216 to −0.214). In subgroup analysis, among YPHIV only, sCD14, I-TAC, MPO, and d-Dimer correlated with a worse overall global z-score (r = −0.443 to −0.318). In contrast, among YWoH only, SAA and TNFR1 correlated with a worse overall global z-score (r = −0.374 to −0.328).
In separate, linear regression models of all participants, adjusting for age, gender, education, income, dietary diversity, electricity, and clean water source, no biomarkers remained statistically associated with overall worse global z-scores. Among YPHIV only, MPO remained associated with worse global z-score (β = −1.64, 95% CI = −2.78, −0.51), as well as processing speed (β = −3.24, 95% CI = −4.89, −1.58), and learning and memory (β = −3.53, 95% CI = −6.48,−0.59) (Figure 2). Additionally, among YPHIV only, sCD14 was associated with worse processing speed (β = −0.78, 95% CI = −1.55, −0.01) and the fungal translocation marker BDG was associated with worse learning and memory (β = −0.35, 95% CI = −0.68,−0.03). Among YWoH participants only, MPO was associated with worse working memory (β −2.47, 95% CI = −4.58,−0.36). Additionally, IGFBP4, MMP-9, and TNFR1 were associated with worse processing speeds among YWoH (β = −2.92, 95% CI = −4.84,−0.99, β = −1.37, 95% CI = −2.43,−0.31, and β = −2.19, 95% CI = −3.95,−0.43, respectively).
Figure 2.
Adjusted models for immune and vascular biomarkers among Ugandan Youth Without HIV (YWoH) and Ugandan Youth with Perinatally acquired HIV (YPHIV). BDG (1,3-β-d-glucan), IGFBP4 (Insulin-like Growth Factor Binding Protein 4), MMP-9 (Matrix Metalloproteinase-9), MPO (Myeloperoxidase), sCD14 (soluble CD14), TNFR1 (Tumor Necrosis Factor Receptor 1). Linear regression models adjusted for age, gender, education, income, dietary diversity, electricity, and clean water source.
3.5. NeuroScreen and socioeconomic status markers
Without adjustments, female gender was associated with worse processing speed and motor speed, while older age was associated with better working memory but worse learning and memory. Access to clean water was associated with better executive function and motor speed, and access to electricity with better working memory. After adjusting for age, gender, income, dietary diversity, electricity, clean water source, and HIV status, secondary school completion was associated with better performance in all five neurocognitive domains (P ≤ 0.05).
4. Discussion
This analysis extends prior work by showing that Ugandan YPHIV, even with durable viral suppression on contemporary DTG-based therapy, demonstrate lower neurocognitive test performance globally and across multiple domains of the NS battery. After controlling for socioeconomic covariates, YPHIV had lower scores in working memory, motor speed, and learning and memory compared to YWoH. Previous research has also demonstrated lower neurocognitive performance in YPHIV [1, 8, 11], though the underlying mechanisms remain unclear.
The present findings add to the literature by identifying associations between biomarkers reflecting vascular processes, immune activation, and gut integrity and neurocognitive performance. Specifically, higher levels of MPO, a marker of vascular inflammation and a key mediator of oxidative stress, were associated with worse processing speed, learning and memory, and global cognition. MPO contributes to the generation of reactive oxygen species (ROS), and oxidative stress has been proposed as a central mechanism in neurocognitive dysfunction, potentially through pathways involving neuronal injury, blood–brain barrier disruption, and neuroinflammation [23]. Similarly, elevated sCD14, a marker of monocyte activation, was associated with poorer processing speed and has previously been linked to increased morbidity and mortality in adults living with HIV [24]. In addition, BDG, a marker of fungal translocation and gut barrier dysfunction, was associated with worse learning and memory and has been correlated with neurocognitive impairment in prior HIV studies [7]. However, given the cross-sectional design of the present study, associations should be interpreted cautiously and do not establish mechanistic pathways.
Although the absolute values of these biomarkers were not significantly higher among YPHIV compared to YWoH, substantial variability existed within the YPHIV cohort. Associations between MPO, sCD14, and BDG and worse neurocognitive performance were specific to YPHIV, indicating that relative differences in immune activation among virally suppressed youth may still be clinically relevant. These findings underscore that between-group comparisons and within-group associations address distinct questions and may yield different, yet complementary, insights.
Similarly, IGFBP4, MMP-9, and TNFR1 were associated with poorer processing speeds specifically in the YWoH cohort, suggesting that inflammatory pathways linked to neurocognition may differ between YWoH and YPHIV. Prior studies of this cohort has shown that, even among YWoH, markers of oxidative stress (oxidized LDL), monocyte activation (sCD14), and activated immune cell populations increased over time [9], indicating that non-HIVrelated inflammatory exposures may also contribute to neurocognitive variability in this setting.
A secondary aim was to examine associations between socioeconomic factors and neurocognition. Compared to YWoH, the YPHIV cohort reported poorer access to electricity, greater economic insecurity, and worse dietary diversity. Lower socioeconomic status, orphanhood, and undernutrition have all been linked to neurocognitive deficits in YPHIV in Sub-Saharan Africa [4]. A meta-analysis of neurocognitive outcomes in YPHIV found that processing speed deficits were more severe in countries with lower Gross National Income per capita, even when comparing sociodemographically matched controls [25]. Nevertheless, our study finds that despite adjusting for socioeconomic factors, living with HIV remained significantly correlated with worse working memory, motor speed, and learning and memory. These findings support a potential role for both biological and social factors in shaping neurocognitive outcomes in YPHIV.
Further data are needed to clarify how poorer performance on neurocognitive testing translates into symptomatic or clinically meaningful impairment. While our assessment tools did not directly measure symptomatic impairment, prior research suggests that lower scores on neurocognitive tests correlate with a higher risk of mortality [26]. In the current study, YPHIV were approximately 1 year older than YWoH, yet a higher proportion remained enrolled in primary school (35% vs 20%, P = 0.139). This pattern—older adolescents with HIV attending lower grade levels—may suggest functional impact related to neurocognitive differences. However, educational attainment is influenced by multiple contextual, social, and structural factors, and these findings should therefore be interpreted cautiously.
This study addresses a critical gap by assessing neurocognition in adolescents receiving contemporary, DTG-based cART, reflecting current first-line treatment in the setting of sustained viral suppression. In adults, initiation and switch studies of DTG-based cART have reported improvements in global cognition and sleep, with some outcomes approaching those of HIV-negative control [27, 28], although findings have been mixed, with some studies suggesting attenuation of prior cognitive gains [16]. In children and adolescents, the open-label ODYSSEY trial across Africa, Europe, and Asia demonstrated a 40% reduction in treatment failure with DTG compared to standard care but did not evaluate neurocognition [13]. Future randomized trials comparing neurocognitive outcomes among participants receiving DTG versus alternative regimens are unlikely, given that integrase strand inhibitors like DTG are expected to remain a mainstay of long-term therapy worldwide. However, longitudinal assessment of neurocognitive trajectories, including comparisons with participants who are HIV-exposed, but uninfected, and HIV-unexposed, may provide valuable insights.
This study is strengthened by its analysis of a comprehensive set of inflammatory biomarkers and recruitment of participants from an under-studied population. However, the study has several limitations. The sample size was relatively small, and cohorts were incompletely age matched. The study also lacks data on maternal factors that may influence neurocognitive development, such as breastfeeding and perinatal complications. Although most participants demonstrated preserved immune recovery, data on early childhood CNS infections, in utero cART exposure, and prior neurologic insults were not available. Therefore, potential legacy effects of earlier HIV disease or other unmeasured factors cannot be excluded when interpreting this study’s results. In addition, CSF data were not included and would likely provide a more direct measure of CNS processes. Nonetheless, previous research suggests circulating cytokines in people living with HIV may access the brain and contribute to neuroinflammation through peripheral–central immune crosstalk [29]. A limitation to NS is that raw scores were converted into z-scores and compared relative to YWoH scores, since reference values have not yet been established for this population. Similarly, clinical reference ranges for many inflammatory and vascular biomarkers are not well defined in pediatric populations or in sub-Saharan African settings; therefore, biomarker concentrations were interpreted relative to the study population rather than classified as pathologically elevated.
Taken together, these findings demonstrate that neurocognitive test performance remained lower in YPHIV despite durable viral suppression on cART, underscoring the multifactorial nature of neurodevelopmental outcomes in this population. The observed associations between markers of immune activation, vascular inflammation, and neurocognitive performance highlight biologically plausible pathways that may contribute to neurocognitive differences in YPHIV. These findings are hypothesis generating and support the need for longitudinal studies to determine whether these pathways play a causal role or represent residual effects of earlier HIV-related or non-HIV-related exposures. Findings also suggest that ongoing attention to cognitive function could be important in adolescent HIV care, particularly during the transition to adulthood. They also point to the potential value of scalable neurocognitive screening tools, such as NS, in Sub-Saharan Africa to help identify youth who may benefit from additional evaluation or support. Given that CSF sampling for inflammatory markers is invasive and not feasible for routine clinical use, our results further suggest that plasma biomarkers may reflect CNS processes and could warrant exploration as more practical tools for monitoring. Finally, these findings provide a rationale for future studies to investigate whether adjunctive therapies targeting inflammation and vascular injury might play a role in preventing or mitigating neurocognitive impairment in YPHIV.
5. Conclusion
Ugandan YPHIV on DTG-based cART, despite achieving durable viral suppression, demonstrated lower neurocognitive test performance compared to negative peers without HIV. Associations between, MPO, sCD14, and BDG and neurocognition suggest that immune activation, vascular processes, and gut barrier integrity may be relevant to these outcomes. These findings are associative and support further investigation into the biological pathways underlying neurocognitive differences in youth growing up with HIV.
Supplementary Material
Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.ijid.2026.108723.
Funding
This work was supported by the National Institute of Allergy and Infectious Diseases [grant numbers U01AI168630] to NF, NL and SDF, and the National Institute of Child Health and Development [grant numbers R01HD113450] to RR, VM and SDF.
Footnotes
Ethical approval
Approval for the study was acquired from the Research Ethics Committee in Uganda, the Ugandan National Council of Science and Technology, the New York State Psychiatric Institute, and the IRB of the University Hospitals Cleveland Medical Center, Cleveland, OH. Parent/caregivers gave written informed consent and participants gave assents.
Declaration of generative AI and AI-assisted technologies in the writing process
During the preparation of this work the author(s) used ChatGPT to improve readability. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
Declaration of competing interest
The authors have no competing interests to declare.
References
- [1].Horton RH, Mcintosh A, Ostinelli EG, Harriss E, Fazel M. Neuropsychiatric outcomes in children and adolescents with perinatally acquired HIV: a systematic review and meta-analysis. J Acquir Immune Defic Syndr 2025; 98:411–28. doi: 10.1097/QAI.0000000000003595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Adolescent HIV treatment. UNICEF DATA n.d. https://data.unicef.org/topic/hivaids/adolescent-hiv-treatment/; [accessed 7 December 2025].
- [3].Slogrove AL, Sohn AH. The global epidemiology of adolescents living with HIV: time for more granular data to improve adolescent health outcomes. Curr Opin HIV AIDS 2018; 13:170–8. doi: 10.1097/COH.0000000000000449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Musindo O, Krabbendam L, Mutahi J, García MP, Bangirana P, Kumar M. Neurocognitive deficits and socioeconomic risk factors among children and adolescents living with HIV in sub-Saharan Africa: a systematic review. Child Adolesc Psychiatry Ment Health 2022; 16:31. doi: 10.1186/s13034-022-00465-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Ellis RJ, Marquine MJ, Kaul M, Fields JA, Schlachetzki JCM. Mechanisms underlying HIV-associated cognitive impairment and emerging therapies for its management. Nat Rev Neurol 2023; 19:668–87. doi: 10.1038/s41582-023-00879-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Williams ME, Stein DJ, Joska JA, Naudé PJW. Cerebrospinal fluid immune markers and HIV-associated neurocognitive impairments: a systematic review. J Neuroimmunol 2021; 358:577649. doi: 10.1016/j.jneuroim.2021.577649. [DOI] [PubMed] [Google Scholar]
- [7].Hoenigl M, de Oliveira MF, Pérez-Santiago J, Zhang Y, Morris S, McCutchan AJ, et al. (1→3)-β-D-glucan levels correlate with neurocognitive functioning in HIV-infected persons on suppressive antiretroviral therapy: a cohort study. Medicine (Baltimore) 2016; 95:e3162. doi: 10.1097/MD.0000000000003162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Williams ME, Janse Van Rensburg A, Loots DT, Naudé PJW, Mason S. Immune dysregulation is associated with neurodevelopment and neurocognitive performance in HIV pediatric populations—a scoping review. Viruses 2021; 13:2543. doi: 10.3390/v13122543. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Dirajlal-Fargo S, Strah M, Ailstock K, Sattar A, Karungi C, Nazzinda R, et al. Persistent immune activation and altered gut integrity over time in a longitudinal study of Ugandan youth with perinatally acquired HIV. Front Immunol 2023; 14:1165964. doi: 10.3389/fimmu.2023.1165964. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Dirajlal-Fargo S, Albar Z, Bowman E, Labbato D, Sattar A, Karungi C, et al. Increased monocyte and T-cell activation in treated HIV+ Ugandan children: associations with gut alteration and HIV factors. AIDS L Engl 2020; 34:1009–18. doi: 10.1097/QAD.0000000000002505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Dirajlal-Fargo S, Sattar A, Strah M, Karungi C, Gumikiriza-Onoria JL, Santoro AF, et al. Immune activation is associated with neurocognitive performance in ugandan adolescents living with HIV. J Acquir Immune Defic Syndr 2024; 97:296–304. doi: 10.1097/QAI.0000000000003483. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Dirajlal-Fargo S, Zhao C, Labbato D, Sattar A, Karungi C, Longenecker CT, et al. Longitudinal changes in subclinical vascular disease in Ugandan youth with human immunodeficiency virus. Clin Infect Dis Publ Infect Dis Soc Am 2023; 76:e599–606. doi: 10.1093/cid/ciac686. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Turkova A, White E, Mujuru HA, Kekitiinwa AR, Kityo CM, Violari A, et al. Dolutegravir as first- or second-line treatment for HIV-1 infection in children. N Engl J Med 2021; 385:2531–43. doi: 10.1056/NEJMoa2108793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Letendre SL, Mills AM, Tashima KT, Thomas DA, Min SS, Chen S, et al. ING116070: a study of the pharmacokinetics and antiviral activity of dolutegravir in cerebrospinal fluid in HIV-1-infected, antiretroviral therapy-naive subjects. Clin Infect Dis Publ Infect Dis Soc Am 2014; 59:1032–7. doi: 10.1093/cid/ciu477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Brkić-Jovanović N, Karaman M, Andrić V, Marić D, Brkić S, Bugarski-Ignjatović V. Neurocognitive profile in HIV subjects on INSTI-regimen- one year follow up: is there room for optimism? PLOS ONE 2024; 19:e0306278. doi: 10.1371/journal.pone.0306278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Michael H, Kibuuka H, Owuoth J, Sing’oei V, Maswai JK, Gervas RR, et al. Neurocognitive effects of transitioning to dolutegravir-based antiretroviral therapy in people with HIV in sub-Saharan Africa. Open Forum Infect Dis 2026; 13 (suppl_1):ofaf695.439. doi: 10.1093/ofid/ofaf695.439. [DOI] [Google Scholar]
- [17].The World Health Organization STEPS instrument. [accessed 26 February 2025] https://www.who.int/publications/m/item/standard-steps-instrument.
- [18].The DHS program - wealth-index-construction n.d. https://dhsprogram.com/topics/wealth-index/Wealth-Index-Construction.cfm; [accessed 26 February 2025].
- [19].What is the $2.15 poverty line, and based on this new measure, how many people are living in extreme poverty in the world? World Bank data n.d. https://datahelpdesk.worldbank.org/knowledgebase/articles/746163-what-is-the-2-15-poverty-line-and-based-on-this; [accessed 29 December 2025].
- [20].Robbins RN, Kluisza L, Liu J, Santoro AF, Raymond J, Ngyuen N, et al. Construct validity supports use of a novel, tablet-based neurocognitive assessment for adolescents and young adults affected by perinatal HIV from vulnerable communities in the United States. AIDS Behav 2021; 25:1185–91. doi: 10.1007/s10461-020-03099-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Robbins RN, Santoro AF, Ferraris C, Asiedu N, Liu J, Dolezal C, et al. Adaptation and construct validity evaluation of a tablet-based, short neuropsychological test battery for use with adolescents and young adults living with HIV in Thailand. Neuropsychology 2022; 36:695–708. doi: 10.1037/neu0000851. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Asiedu N, Mwesiga EK, Akena D, Morrison C, Gumikiriza-Onoria JL, Nanteza A, et al. Evaluating construct and criterion validity of NeuroScreen in assessing neurocognition among hospitalized Ugandan first-episode psychosis patients. Schizophr Res Cogn 2023; 32:100276. doi: 10.1016/j.scog.2022.100276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Louboutin J-P, Strayer D Role of oxidative stress in HIV-1-associated neurocognitive disorder and protection by gene delivery of antioxidant enzymes. Antioxidants 2014; 3:770–97. doi: 10.3390/antiox3040770. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Sandler NG, Wand H, Roque A, Law M, Nason MC, Nixon DE, et al. Plasma levels of soluble CD14 independently predict mortality in HIV infection. J Infect Dis 2011; 203:780–90. doi: 10.1093/infdis/jiq118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Dahmani S, Rakhmanina N, Jiang X. Cognitive impairment in children and adolescents living with perinatal HIV disease in the ART era: a meta-analysis. eClinicalMedicine 2024; 72:102602. doi: 10.1016/j.eclinm.2024.102602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Yu J, Haynie DL, Sundaram R, Gilman SE. Adverse childhood experiences, neurocognitive functions, and long-term mortality risk. JAMA Netw Open 2025; 8:e2531283. doi: 10.1001/jamanetworkopen.2025.31283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Rodrigues J, Lowton K. Neurocognitive improvement in HIV-positive patients treated with dolutegravir-based regimens. South Afr J Psychiatry SAJP J Soc Psychiatr South Afr 2023; 29:2071. doi: 10.4102/sajpsychiatry.v29i0.2071. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Nightingale S, Dreyer AJ, Thomas KGF, van Zyl G, Decloedt E, Naude PJW, et al. Cognitive performance, neuropsychiatric symptoms, and cerebrospinal fluid viral control following programmatic switch from efavirenz-based to dolutegravir-based antiretroviral therapy in South Africa (CONNECT): a prospective cohort study. Lancet HIV 2024; 11:e680–9. doi: 10.1016/S2352-3018(24)00209-1. [DOI] [PubMed] [Google Scholar]
- [29].Hong S, Banks WA. Role of the immune system in HIV-associated neuroinflammation and neurocognitive implications. Brain Behav Immun 2015; 0:1–12. doi: 10.1016/j.bbi.2014.10.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.