Lack of racial and ethnic disparities in mortality in minority patients hospitalised with COVID-19 in a mid-Atlantic healthcare system.

INTRODUCTION: In the USA, minoritised communities (racial and ethnic) have suffered disproportionately from COVID-19 compared with non-Hispanic white communities. In a large cohort of patients hospitalised for COVID-19 in a healthcare system spanning five adult hospitals, we analysed outcomes of patients based on race and ethnicity. METHODS: This was a retrospective cohort analysis of patients 18 years or older admitted to five hospitals in the mid-Atlantic area between 4 March 2020 and 27 May 2022 with confirmed COVID-19. Participants were divided into four groups based on their race/ethnicity: non-Hispanic black, non-Hispanic white, Latinx and other. Propensity score weighted generalised linear models were used to assess the association between race/ethnicity and the primary outcome of in-hospital mortality. RESULTS: Of the 9651 participants in the cohort, more than half were aged 18-64 years old (56%) and 51% of the cohort were females. Non-Hispanic white patients had higher mortality (p<0.001) and longer hospital length-of-stay (p<0.001) than Latinx and non-Hispanic black patients. DISCUSSION: In this large multihospital cohort of patients admitted with COVID-19, non-Hispanic black and Hispanic patients did not have worse outcomes than white patients. Such findings likely reflect how the complex range of factors that resulted in a life-threatening and disproportionate impact of incidence on certain vulnerable populations by COVID-19 in the community was offset through admission at well-resourced hospitals and healthcare systems. However, there continues to remain a need for efforts to address the significant pre-existing race and ethnicity inequities highlighted by the COVID-19 pandemic to be better prepared for future public health emergencies.

Bayesian tensor-on-tensor regression with efficient computation.

We propose a Bayesian tensor-on-tensor regression approach to predict a multidimensional array (tensor) of arbitrary dimensions from another tensor of arbitrary dimensions, building upon the Tucker decomposition of the regression coefficient tensor. Traditional tensor regression methods making use of the Tucker decomposition either assume the dimension of the core tensor to be known or estimate it via cross-validation or some model selection criteria. However, no existing method can simultaneously estimate the model dimension (the dimension of the core tensor) and other model parameters. To fill this gap, we develop an efficient Markov Chain Monte Carlo (MCMC) algorithm to estimate both the model dimension and parameters for posterior inference. Besides the MCMC sampler, we also develop an ultra-fast optimization-based computing algorithm wherein the maximum a posteriori estimators for parameters are computed, and the model dimension is optimized via a simulated annealing algorithm. The proposed Bayesian framework provides a natural way for uncertainty quantification. Through extensive simulation studies, we evaluate the proposed Bayesian tensor-on-tensor regression model and show its superior performance compared to alternative methods. We also demonstrate its practical effectiveness by applying it to two real-world datasets, including facial imaging data and 3D motion data.

A Bayesian approach for investigating the pharmacogenetics of combination antiretroviral therapy in people with HIV.

Combination antiretroviral therapy (ART) with at least three different drugs has become the standard of care for people with HIV (PWH) due to its exceptional effectiveness in viral suppression. However, many ART drugs have been reported to associate with neuropsychiatric adverse effects including depression, especially when certain genetic polymorphisms exist. Pharmacogenetics is an important consideration for administering combination ART as it may influence drug efficacy and increase risk for neuropsychiatric conditions. Large-scale longitudinal HIV databases provide researchers opportunities to investigate the pharmacogenetics of combination ART in a data-driven manner. However, with more than 30 FDA-approved ART drugs, the interplay between the large number of possible ART drug combinations and genetic polymorphisms imposes statistical modeling challenges. We develop a Bayesian approach to examine the longitudinal effects of combination ART and their interactions with genetic polymorphisms on depressive symptoms in PWH. The proposed method utilizes a Gaussian process with a composite kernel function to capture the longitudinal combination ART effects by directly incorporating individuals' treatment histories, and a Bayesian classification and regression tree to account for individual heterogeneity. Through both simulation studies and an application to a dataset from the Women's Interagency HIV Study, we demonstrate the clinical utility of the proposed approach in investigating the pharmacogenetics of combination ART and assisting physicians to make effective individualized treatment decisions that can improve health outcomes for PWH.

Common antiretroviral combinations are associated with somatic depressive symptoms in women with HIV.

OBJECTIVE: While modern antiretroviral therapy (ART) is highly effective and safe, depressive symptoms have been associated with certain ART drugs. We examined the association between common ART regimens and depressive symptoms in women with HIV (WWH) with a focus on somatic vs. nonsomatic symptoms. DESIGN: Analysis of longitudinal data from the Women's Interagency HIV Study. METHODS: Participants were classified into three groups based on the frequency of positive depression screening (CES-D ≥16): chronic depression (≥50% of visits since study enrollment), infrequent depression (<50% of visits), and never depressed (no visits). Novel Bayesian machine learning methods building upon a subset-tree kernel approach were developed to estimate the combined effects of ART regimens on depressive symptoms in each group after covariate adjustment. RESULTS: The analysis included 1538 WWH who participated in 12 924 (mean = 8.4) visits. The mean age was 49.9 years, 72% were Black, and 14% Hispanic. In the chronic depression group, combinations including tenofovir alafenamide and cobicistat-boosted elvitegravir and/or darunavir were associated with greater somatic symptoms of depression, whereas those combinations containing tenofovir disoproxil fumarate and efavirenz or rilpivirine were associated with less somatic depressive symptoms. ART was not associated with somatic symptoms in the infrequent depression or never depressed groups. ART regimens were not associated with nonsomatic symptoms in any group. CONCLUSIONS: Specific ART combinations are associated with somatic depressive symptoms in WWH with chronic depression. Future studies should consider specific depressive symptoms domains as well as complete drug combinations when assessing the relationship between ART and depression.

Clinical Outcomes Associated With Overestimation of Oxygen Saturation by Pulse Oximetry in Patients Hospitalized With COVID-19.

Importance: Many pulse oximeters have been shown to overestimate oxygen saturation in persons of color, and this phenomenon has potential clinical implications. The relationship between overestimation of oxygen saturation with timing of COVID-19 medication delivery and clinical outcomes remains unknown. Objective: To investigate the association between overestimation of oxygen saturation by pulse oximetry and delay in administration of COVID-19 therapy, hospital length of stay, risk of hospital readmission, and in-hospital mortality. Design, Setting, and Participants: This cohort study included patients hospitalized for COVID-19 at 186 acute care facilities in the US with at least 1 functional arterial oxygen saturation (SaO2) measurement between March 2020 and October 2021. A subset of patients were admitted after July 1, 2020, without immediate need for COVID-19 therapy based on pulse oximeter saturation (SpO2 levels of 94% or higher without supplemental oxygen). Exposures: Self-reported race and ethnicity, difference between concurrent SaO2 and pulse oximeter saturation (SpO2) within 10 minutes, and initially unrecognized need for COVID-19 therapy (first SaO2 reading below 94% despite SpO2 levels of 94% or above). Main Outcome and Measures: The association of race and ethnicity with degree of pulse oximeter measurement error (SpO2 - SaO2) and odds of unrecognized need for COVID-19 therapy were determined using linear mixed-effects models. Associations of initially unrecognized need for treatment with time to receipt of therapy (remdesivir or dexamethasone), in-hospital mortality, 30-day hospital readmission, and length of stay were evaluated using mixed-effects models. All models accounted for demographics, clinical characteristics, and hospital site. Effect modification by race and ethnicity was evaluated using interaction terms. Results: Among 24 504 patients with concurrent SpO2 and SaO2 measurements (mean [SD] age, 63.9 [15.8] years; 10 263 female [41.9%]; 3922 Black [16.0%], 7895 Hispanic [32.2%], 2554 Asian, Native American or Alaskan Native, Hawaiian or Pacific Islander, or another race or ethnicity [10.4%], and 10 133 White [41.4%]), pulse oximetry overestimated SaO2 for Black (adjusted mean difference, 0.93 [95% CI, 0.74-1.12] percentage points), Hispanic (0.49 [95% CI, 0.34-0.63] percentage points), and other (0.53 [95% CI, 0.35-0.72] percentage points) patients compared with White patients. In a subset of 8635 patients with a concurrent SpO2 - SaO2 pair without immediate need for COVID-19 therapy, Black patients were significantly more likely to have pulse oximetry values that masked an indication for COVID-19 therapy compared with White patients (adjusted odds ratio [aOR], 1.65; 95% CI, 1.33-2.03). Patients with an unrecognized need for COVID-19 therapy were 10% less likely to receive COVID-19 therapy (adjusted hazard ratio, 0.90; 95% CI, 0.83-0.97) and higher odds of readmission (aOR, 2.41; 95% CI, 1.39-4.18) regardless of race (P for interaction = .45 and P = .14, respectively). There was no association of unrecognized need for COVID-19 therapy with in-hospital mortality (aOR, 0.84; 95% CI, 0.71-1.01) or length of stay (mean difference, -1.4 days; 95% CI, -3.1 to 0.2 days). Conclusions and Relevance: In this cohort study, overestimation of oxygen saturation by pulse oximetry led to delayed delivery of COVID-19 therapy and higher probability of readmission regardless of race. Black patients were more likely to have unrecognized need for therapy with potential implications for population-level health disparities.

Functional Bayesian networks for discovering causality from multivariate functional data.

Multivariate functional data arise in a wide range of applications. One fundamental task is to understand the causal relationships among these functional objects of interest. In this paper, we develop a novel Bayesian network (BN) model for multivariate functional data where conditional independencies and causal structure are encoded by a directed acyclic graph. Specifically, we allow the functional objects to deviate from Gaussian processes, which is the key to unique causal structure identification even when the functions are measured with noises. A fully Bayesian framework is designed to infer the functional BN model with natural uncertainty quantification through posterior summaries. Simulation studies and real data examples demonstrate the practical utility of the proposed model.

Peripheral inflammatory subgroup differences in anterior Default Mode network and multiplex functional network topology are associated with cognition in psychosis.

INTRODUCTION: High-inflammation subgroups of patients with psychosis demonstrate cognitive deficits and neuroanatomical alterations. Systemic inflammation assessed using IL-6 and C-reactive protein may alter functional connectivity within and between resting-state networks, but the cognitive and clinical implications of these alterations remain unknown. We aim to determine the relationships of elevated peripheral inflammation subgroups with resting-state functional networks and cognition in psychosis spectrum disorders. METHODS: Serum and resting-state fMRI were collected from psychosis probands (schizophrenia, schizoaffective, psychotic bipolar disorder) and healthy controls (HC) from the B-SNIP1 (Chicago site) study who were stratified into inflammatory subgroups based on factor and cluster analyses of 13 cytokines (HC Low n = 32, Proband Low n = 65, Proband High n = 29). Nine resting-state networks derived from independent component analysis were used to assess functional and multilayer connectivity. Inter-network connectivity was measured using Fisher z-transformation of correlation coefficients. Network organization was assessed by investigating networks of positive and negative connections separately, as well as investigating multilayer networks using both positive and negative connections. Cognition was assessed using the Brief Assessment of Cognition in Schizophrenia. Linear regressions, Spearman correlations, permutations tests and multiple comparison corrections were used for analyses in R. RESULTS: Anterior default mode network (DMNa) connectivity was significantly reduced in the Proband High compared to Proband Low (Cohen's d = -0.74, p = 0.002) and HC Low (d = -0.85, p = 0.0008) groups. Inter-network connectivity between the DMNa and the right-frontoparietal networks was lower in Proband High compared to Proband Low (d = -0.66, p = 0.004) group. Compared to Proband Low, the Proband High group had lower negative (d = 0.54, p = 0.021) and positive network (d = 0.49, p = 0.042) clustering coefficient, and lower multiplex network participation coefficient (d = -0.57, p = 0.014). Network findings in high inflammation subgroups correlate with worse verbal fluency, verbal memory, symbol coding, and overall cognition. CONCLUSION: These results expand on our understanding of the potential effects of peripheral inflammatory signatures and/or subgroups on network dysfunction in psychosis and how they relate to worse cognitive performance. Additionally, the novel multiplex approach taken in this study demonstrated how inflammation may disrupt the brain's ability to maintain healthy co-activation patterns between the resting-state networks while inhibiting certain connections between them.

Peripheral inflammation is associated with impairments of inhibitory behavioral control and visual sensorimotor function in psychotic disorders.

Elevated markers of peripheral inflammation are common in psychosis spectrum disorders and have been associated with brain anatomy, pathology, and physiology as well as clinical outcomes. Preliminary evidence suggests a link between inflammatory cytokines and C-reactive protein (CRP) with generalized cognitive impairments in a subgroup of individuals with psychosis. Whether these patients with elevated peripheral inflammation demonstrate deficits in specific cognitive domains remains unclear. To examine this, seventeen neuropsychological and sensorimotor tasks and thirteen peripheral inflammatory and microvascular markers were quantified in a subset of B-SNIP consortium participants (129 psychosis, 55 healthy controls). Principal component analysis was conducted across the inflammatory markers, resulting in five inflammation factors. Three discrete latent cognitive domains (Visual Sensorimotor, General Cognitive Ability, and Inhibitory Behavioral Control) were characterized based on the neurobehavioral battery and examined in association with inflammation factors. Hierarchical clustering analysis identified cognition-sensitive high/low inflammation subgroups. Among persons with psychotic disorders but not healthy controls, higher inflammation scores had significant associations with impairments of Inhibitory Control (R2 = 0.100, p-value = 2.69e-4, q-value = 0.004) and suggestive associations with Visual Sensorimotor function (R2 = 0.039, p-value = 0.024, q-value = 0.180), but not with General Cognitive Ability (R2 = 0.015, p-value = 0.162). Greater deficits in Inhibitory Control were observed in the high inflammation patient subgroup, which represented 30.2 % of persons with psychotic disorders, as compared to the low inflammation psychosis subgroup. These findings indicate that inflammation dysregulation may differentially impact specific neurobehavioral domains across psychotic disorders, particularly performance on tasks requiring ongoing behavioral monitoring and control.

Trauma Across the Life Span and Multisystem Morbidity in Women With HIV.

OBJECTIVE: Sexual and physical abuse are highly prevalent among women living with HIV (WLWH) and are risk factors for the development of mental health and substance use disorders (MHDs, SUDs), and cognitive and medical comorbidities. We examined empirically derived patterns of trauma, MHD, and SUD, and associations with later cognitive and health outcomes. METHODS: A total of 1027 WLWH (average age = 48.6 years) in the Women's Interagency HIV Study completed the World Mental Health Composite International Diagnostic Interview from 2010 to 2013 to identify MHDs, SUDs, and age at onset of sexual and physical abuse. Then, cognitive impairment, cardiovascular/metabolic conditions, and HIV disease outcomes were assessed for up to 8.8 years. Latent class analysis identified patterns of co-occurring trauma, MHDs, and/or SUDs. Generalized estimating equations determined associations between these patterns and midlife cognitive and medical outcomes. RESULTS: Six distinct profiles emerged: no/negligible sexual/physical trauma, MHD, or SUD (39%); preadolescent/adolescent sexual trauma with anxiety and SUD (22%); SUD only (16%); MHD + SUD only (12%); early childhood sexual/physical trauma only (6%); and early childhood sexual/physical trauma with later MHD + SUD (4%). Profiles including early childhood trauma had the largest number of midlife conditions (i.e., cognitive, cardiovascular, HIV-related). Preadolescent/adolescent sexual trauma with anxiety and SUD predicted both global and domain-specific cognitive declines. Only SUD without trauma predicted lower CD4, whereas childhood trauma with MHD + SUD predicted increased CD8. CONCLUSIONS: WLWH have complex multisystem profiles of abuse, MHD, and/or SUD that predict midlife cognitive, metabolic/cardiovascular, and HIV outcomes. Understanding the interplay between these factors over time can identify risks and personalize preventative and treatment interventions.

Federated Learning for Sparse Bayesian Models with Applications to Electronic Health Records and Genomics.

Federated learning is becoming increasingly more popular as the concern of privacy breaches rises across disciplines including the biological and biomedical fields. The main idea is to train models locally on each server using data that are only available to that server and aggregate the model (not data) information at the global level. While federated learning has made significant advancements for machine learning methods such as deep neural networks, to the best of our knowledge, its development in sparse Bayesian models is still lacking. Sparse Bayesian models are highly interpretable with natural uncertain quantification, a desirable property for many scientific problems. However, without a federated learning algorithm, their applicability to sensitive biological/biomedical data from multiple sources is limited. Therefore, to fill this gap in the literature, we propose a new Bayesian federated learning framework that is capable of pooling information from different data sources without breaching privacy. The proposed method is conceptually simple to understand and implement, accommodates sampling heterogeneity (i.e., non-iid observations) across data sources, and allows for principled uncertainty quantification. We illustrate the proposed framework with three concrete sparse Bayesian models, namely, sparse regression, Markov random field, and directed graphical models. The application of these three models is demonstrated through three real data examples including a multi-hospital COVID-19 study, breast cancer protein-protein interaction networks, and gene regulatory networks.

Integrase Inhibitors are Associated with Neuropsychiatric Symptoms in Women with HIV.

OBJECTIVE: Women with HIV(WWH) are more likely to discontinue/change antiretroviral therapy(ART) due to side effects including neuropsychiatric symptoms. Efavirenz and integrase strand transfer inhibitors(INSTIs) are particularly concerning. We focused on these ART agents and neuropsychiatric symptoms in previously developed subgroups of WWH that differed on key sociodemographic factors as well as longitudinal behavioral and clinical profiles. WWH from the Women's Interagency HIV Study were included if they had ART data available, completed the Perceived Stress Scale-10 and PTSD Checklist-Civilian. Questionnaires were completed biannually beginning in 2008 through 2016. To examine ART-symptom associations, constrained continuation ratio model via penalized maximum likelihood were fit within 5 subgroups of WWH. Data from 1882 WWH contributed a total of 4598 observations. 353 women were previously defined as primarily having well-controlled HIV with vascular comorbidities, 463 with legacy effects(CD4 nadir < 250cells/mL), 274 aged ≤ 45 with hepatitis, 453 between 35-55 years, and 339 with poorly-controlled HIV/substance users. INSTIs, but not efavirenz, were associated with symptoms among key subgroups of WWH. Among those with HIV legacy effects, dolutegravir and elvitegravir were associated with greater stress/anxiety and avoidance symptoms(P's < 0.01); dolutegravir was also associated with greater re-experiencing symptoms(P = 0.005). Elvitegravir related to greater re-experiencing and hyperarousal among women with well-controlled HIV with vascular comorbidities(P's < 0.022). Raltegravir was associated with less hyperarousal, but only among women aged ≤ 45 years(P = 0.001). The adverse neuropsychiatric effects of INSTIs do not appear to be consistent across all WWH. Key characteristics (e.g., age, hepatitis positivity) may need consideration to fully weight the risk-benefit ratio of dolutegravir and elvitegravir in WWH.

Continuous-Time Decision Transformer for Healthcare Applications.

Offline reinforcement learning (RL) is a promising approach for training intelligent medical agents to learn treatment policies and assist decision making in many healthcare applications, such as scheduling clinical visits and assigning dosages for patients with chronic conditions. In this paper, we investigate the potential usefulness of Decision Transformer (Chen et al., 2021)-a new offline RL paradigm-in medical domains where decision making in continuous time is desired. As Decision Transformer only handles discrete-time (or turn-based) sequential decision making scenarios, we generalize it to Continuous-Time Decision Transformer that not only considers the past clinical measurements and treatments but also the timings of previous visits, and learns to suggest the timings of future visits as well as the treatment plan at each visit. Extensive experiments on synthetic datasets and simulators motivated by real-world medical applications demonstrate that Continuous-Time Decision Transformer is able to outperform competitors and has clinical utility in terms of improving patients' health and prolonging their survival by learning high-performance policies from logged data generated using policies of different levels of quality.

T-cell activation state differentially contributes to neuropsychiatric complications in women with HIV.

Neuropsychiatric complications are common among women with HIV (WWH). The pathophysiological mechanisms underlying these complications are not fully known but likely driven in part by immune modulation. We examined associations between T-cell activation states which are required to mount an effective immune response (activation, co-stimulation/normal function, exhaustion, senescence) and neuropsychiatric complications in WWH. 369 WWH (78% HIV RNA undetectable/<20cp/mL) enrolled in the Women's Interagency HIV Study completed neuropsychological testing and measures of depression (Center for Epidemiological Studies Depression Scale-CES-D), self-reported stress levels (Perceived Stress Scale-10), and post-traumatic stress (PTSD Checklist-Civilian Scale). Multiparametric flow cytometry evaluated T-cell activation state. Partial least squares regressions were used to examine T-cell phenotypes and neuropsychiatric outcome associations after confounder adjustment. In the total sample and among virally suppressed (VS)-WWH, CD4+ T-cell exhaustion was associated with poorer learning and attention/working memory (P's < 0.05). In the total sample, CD4+ T-cell activation was associated with better attention/working memory and CD8+ T-cell co-stimulation and senescence was associated with poorer executive function (P's < 0.05). For mental health outcomes, in the total sample, CD4+ T-cell activation was associated with more perceived stress and CD4+ T-cell exhaustion was associated with less depressive symptoms (P's < 0.05). Among VS-WWH, CD4+ senescence was associated with less perceive stress and CD8+ T-cell co-stimulation and senescence was associated with higher depression (P's < 0.05). Together, results suggest the contribution of peripheral CD4+ and CD8+ T-cell activation status to neuropsychiatric complications in WWH.

Thromboprophylaxis in people hospitalized with COVID-19: Assessing intermediate or standard doses in a retrospective cohort study.

Background and Objectives: Current clinical guidelines recommend thromboprophylaxis for adults hospitalized with coronavirus disease 2019 (COVID-19), yet it is unknown whether higher doses of thromboprophylaxis offer benefits beyond standard doses. Methods: We studied electronic health records from 50 091 adults hospitalized with COVID-19 in the United States between February 2020 and February 2021. We compared standard (enoxaparin 30 or 40 mg/day, fondaparinux 2.5 mg, or heparin 5000 units twice or thrice per day) versus intermediate (enoxaparin 30 or 40 mg twice daily, or up to 1.2 mg/kg of body weight daily, heparin 7500 units thrice per day or heparin 10 000 units twice or thrice per day) thromboprophylaxis. We separately examined risk of escalation to therapeutic anticoagulation, severe disease (first occurrence of high-flow nasal cannula, noninvasive positive pressure ventilation or invasive mechanical ventilation), and death. To summarize risk, we present hazard ratios (HRs) with 95% confidence intervals (CIs) using adjusted time-dependent Cox proportional hazards regression models. Results: People whose first dose was high intensity were younger, more often obese, and had greater oxygen support requirements. Intermediate dose thromboprophylaxis was associated with increased risk of therapeutic anticoagulation (HR, 3.39; 95% CI, 3.22-3.57), severe disease (HR, 1.22; 95% CI, 1.17-1.28), and death (HR, 1.37; 95% CI, 1.21-1.55). Increased risks associated with intermediate-dose thromboprophylaxis persisted in subgroup and sensitivity analyses varying populations and definitions of exposures, outcomes, and covariates. Conclusions: Our findings do not support routine use of intermediate-dose thromboprophylaxis to prevent clinical worsening, severe disease, or death among adults hospitalized with COVID-19.

BAGEL: A BAYESIAN GRAPHICAL MODEL FOR INFERRING DRUG EFFECT LONGITUDINALLY ON DEPRESSION IN PEOPLE WITH HIV.

Access and adherence to antiretroviral therapy (ART) has transformed the face of HIV infection from a fatal to a chronic disease. However, ART is also known for its side effects. Studies have reported that ART is associated with depressive symptomatology. Large-scale HIV clinical databases with individuals' longitudinal depression records, ART medications, and clinical characteristics offer researchers unprecedented opportunities to study the effects of ART drugs on depression over time. We develop BAGEL, a Bayesian graphical model to investigate longitudinal effects of ART drugs on a range of depressive symptoms while adjusting for participants' demographic, behavior, and clinical characteristics, and taking into account the heterogeneous population through a Bayesian nonparametric prior. We evaluate BAGEL through simulation studies. Application to a dataset from the Women's Interagency HIV Study yields interpretable and clinically useful results. BAGEL not only can improve our understanding of ART drugs effects on disparate depression symptoms, but also has clinical utility in guiding informed and effective treatment selection to facilitate precision medicine in HIV.

Immunometabolic Reprogramming in Response to HIV Infection Is Not Fully Normalized by Suppressive Antiretroviral Therapy.

BACKGROUND: HIV infection results in immunometabolic reprogramming. While we are beginning to understand how this metabolic reprogramming regulates the immune response to HIV infection, we do not currently understand the impact of ART on immunometabolism in people with HIV (PWH). METHODS: Serum obtained from HIV-infected (n = 278) and geographically matched HIV seronegative control subjects (n = 300) from Rakai Uganda were used in this study. Serum was obtained before and ~2 years following the initiation of ART from HIV-infected individuals. We conducted metabolomics profiling of the serum and focused our analysis on metabolic substrates and pathways assocaited with immunometabolism. RESULTS: HIV infection was associated with metabolic adaptations that implicated hyperactive glycolysis, enhanced formation of lactate, increased activity of the pentose phosphate pathway (PPP), decreased ß-oxidation of long-chain fatty acids, increased utilization of medium-chain fatty acids, and enhanced amino acid catabolism. Following ART, serum levels of ketone bodies, carnitine, and amino acid metabolism were normalized, however glycolysis, PPP, lactate production, and ß-oxidation of long-chain fatty acids remained abnormal. CONCLUSION: Our findings suggest that HIV infection is associated with an increased immunometabolic demand that is satisfied through the utilization of alternative energetic substrates, including fatty acids and amino acids. ART alone was insufficient to completely restore this metabolic reprogramming to HIV infection, suggesting that a sustained impairment of immunometabolism may contribute to chronic immune activation and comorbid conditions in virally suppressed PWH.