POST-ACUTE SEQUELAE AND ADAPTIVE IMMUNE RESPONSES IN PEOPLE LIVING WITH HIV RECOVERING FROM SARS-COV-2 INFECTION

Abstract

Background:

Limited data are available on the long-term clinical and immunologic consequences of SARS-CoV-2 infection in people with HIV (PWH).

Methods:

We measured SARS-CoV-2-specific humoral and cellular responses in people with and without HIV recovering from COVID-19 (n=39 and n=43, respectively) using binding antibody, surrogate virus neutralization, intracellular cytokine staining, and inflammatory marker assays. We identified individuals experiencing post-acute sequelae of SARS-CoV-2 infection (PASC) and evaluated immunologic parameters. We used linear regression and generalized linear models to examine differences by HIV status in the magnitude of inflammatory and virus-specific antibody and T cell responses, as well as differences in the prevalence of PASC.

Results:

Among PWH, we found broadly similar SARS-CoV-2-specific antibody and T cell responses as compared with a well-matched group of HIV-negative individuals. PWH had 70% lower relative levels of SARS-CoV-2-specific memory CD8+ T cells (p=0.007) and 53% higher relative levels of PD-1+ SARS-CoV-2-specific CD4+ T cells (p=0.007). Higher CD4/CD8 ratio was associated with lower PD-1 expression on SARS-CoV-2-specific CD8+ T cells (0.34-fold effect, p=0.02). HIV status was strongly associated with PASC (odds ratio 4.01, p=0.008), and levels of certain inflammatory markers (IL-6, TNF-alpha, and IP-10) were associated with persistent symptoms.

Conclusions:

We identified potentially important differences in SARS-CoV-2-specific CD4+ and CD8+ T cells in PWH and HIV-negative participants that might have implications for long-term immunity conferred by natural infection. HIV status strongly predicted the presence of PASC. Larger and more detailed studies of PASC in PWH are urgently needed.

Brief Summary:

We identified potentially important differences in SARS-CoV-2-specific CD4+ and CD8+ T cells in PWH and HIV-negative participants that might have implications for long-term immunity conferred by natural infection. HIV status strongly predicted the presence of post-acute sequelae of COVID-19.

BACKGROUND

The intersection between the SARS-CoV-2 and HIV epidemics has gained increased attention [1]. While early studies did not show differences in acute COVID-19 associated with HIV status [2,3], larger studies show that people with HIV (PWH) are at higher risk for adverse outcomes [4,5]. Data on SARS-CoV-2-specific adaptive immune responses in PWH remain sparse, however, with one study showing less robust immune responses among PWH [6] but another [7] suggesting similar responses. Furthermore, there is growing recognition of the clinical burden of post-acute sequelae of SARS-CoV-2 infection (PASC, including “long COVID”) [8], but this condition remains poorly understood, especially in PWH.

A combination of factors might underlie PASC [9–14]. Higher prevalence of certain socioeconomic factors and comorbidities among PWH, along with differences in immune responses to SARS-CoV-2 [6,7] and persistent inflammation and immune dysregulation in the presence of antiretroviral therapy (ART) [15–18], may make PWH selectively vulnerable to developing this condition. For these reasons, examination of PASC in PWH is urgently needed.

Here, we sought to test the hypothesis that PASC would be more prevalent in PWH with a history of COVID-19 prior to vaccination and that these individuals would have reduced SARS-CoV-2-specific immune responses in comparison to HIV-negative individuals recovering from COVID-19.

METHODS

Informed consent

The study was approved by the University of California, San Francisco (UCSF) Institutional Review Board. Participants provided written informed consent.

Participants

Volunteers were enrolled in the Long-term Impact of Infection with Novel Coronavirus (LIINC) study (NCT04362150) [19,20]. Briefly, participants with a positive nucleic acid amplification test for SARS-CoV-2 were enrolled 21 or more days following symptom onset. Recruitment occurred through self- and clinician referrals; all PWH testing positive for SARS-CoV-2 at two UCSF-based HIV clinics were invited to participate. At each visit, participants completed an interview regarding prior and current COVID-19-attributed symptoms, medical history, and quality of life.

We selected all PWH who enrolled prior to receipt of a SARS-CoV-2 vaccine (n=39) and compared them with a randomly selected group of HIV-negative individuals (n=43) with a similar distribution of age, sex, COVID-19 hospitalization, and time since infection [9,21–23]. Participants were assessed at the visit closest to 16 weeks post-infection (median 112 days [IQR: 91–129]).

Clinical assessment

Our methodology for assessing PASC is described elsewhere [20]. Briefly, participants were queried regarding the presence of 32 symptoms derived from the U.S. Centers for Disease Control (CDC) list of COVID-19 symptoms (https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/symptoms.html) and the Patient Health Questionnaire Somatic Symptom Scale [24] were at each visit. Symptoms were considered related to COVID-19 if they were new or worsened since SARS-CoV-2 infection; stable, chronic symptoms that pre-dated SARS-CoV-2 infection were not included.

We defined PASC as any COVID-19-attributed symptom that was present during a visit >6 weeks following SARS-CoV-2 infection [25,26]. We conducted sensitivity analyses, comparing those with 3 or more symptoms to those with fewer than 3 symptoms or with no symptoms.

Biospecimen collection

Peripheral blood mononuclear cells (PBMCs), plasma and serum were cryopreserved prior to testing. PBMCs were isolated using Ficol-Paque in SeptMate Tubes, frozen in heat-inactivated FBS and 10% dimethyl sulfoxide (DMSO), and stored in liquid nitrogen.

Antibody assays

Virus-specific antibody responses were measured using the Pylon COVID-19 total antibody assay (ET Health) and a validated surrogate virus neutralization test (sVNT) [27]. The lower limit of detection was a sVNT reciprocal titer <10 and anti-RBD IgG <10 relative fluorescence units.

T cell responses by Intracellular Cytokine Staining (ICS)

We performed ICS [22] with the following modifications. Two peptide megapools containing SARS-CoV-2-derived peptides experimentally determined to be recognized by CD4+ or CD8+ T cells were used for 18-hour stimulations (1 μg/ml/peptide; CD4-E with 280 and CD8-E with 454 T cell epitopes) [28]. Additional detail and a complete list of antibodies are listed in Supplementary Methods. All samples were acquired on a BD LSR-II analyzer and analyzed with FlowJo X. The gating strategy is shown in Supplementary Figure 1.

Markers of inflammation

In a subset with additional specimens, the fully automated HD-X Simoa platform was used to measure plasma biomarkers including monocyte chemoattractant protein 1 (MCP-1), Cytokine 3-PlexA (IL-6, IL-10, TNFα), interferon gamma-induced protein-10 (IP-10), and interferon-gamma (IFNγ).

Statistical methods

For comparison of humoral responses, we log-transformed sVNT and total antibody values to satisfy assumptions of a normal distribution. We used linear regression to examine differences by HIV status, adjusting for days since SARS-CoV-2 infection, age, sex, and COVID-19 hospitalization. Differences and 95% confidence intervals (CI) are presented as fold-changes in the geometric mean. CD4/CD8 ratio was then added as a covariate to examine if it was associated with magnitude of responses. To examine percent differences in cellular immune responses, we used generalized linear models from the binomial distribution with bootstrapped standard errors and adjusted for the same potential confounders as above. For inflammatory markers, we fit linear regression models following log-transformation. To examine differences in PASC by HIV status, we used logistic regression adjusting for the same potential confounders.

RESULTS

Participants

All participants were diagnosed with COVID-19 between March and December 2020 (pre-Delta and Omicron). The groups were similar in age, sex, and race/ethnicity (Table 1). The majority of PWH were men, reflecting the local HIV epidemic [29]. The vast majority of PWH were on antiretroviral therapy including integrase strand transfer inhibitors (34/39, 87%) and nucleoside reverse transcriptase inhibitors (37/39, 95%); non-nucleoside reverse transcriptase inhibitors (3/39, 8%) and protease inhibitors (2/39, 5%) were rare. PWH more commonly reported medical comorbidities including heart and lung disease. Most were managed as outpatients during acute COVID-19. All previously hospitalized participants required supplemental oxygenation and one in each group required mechanical ventilation. No participant received SARS-CoV-2-specific treatment.

Table 1. Characteristics of study cohort.

Values reported as median (IQR) unless otherwise specified. IQR, interquartile range. Note: Autoimmune disease reported as one case of hypothyroidism and one case of Raynaud’s disease. Cancer reported as prostate cancer controlled with hormone therapy, ocular melanoma, treated Kaposi sarcoma, and resected renal cancer. Plasma HIV RNA values were >50 in two participants (87 copies/mL and 28,118 copies/mL). All hospitalized participants in each group required oxygen support and one hospitalized participant in each group required mechanical ventilation. None reported receiving SARS-CoV-2-targeted therapy or steroids during the hospitalization.

Characteristic	All (n=82)	HIV+ (n=39)	HIV− (n=43)
Demographic characteristics
Age	52 (44–58)	54 (46–58)	51 (43–58)
Sex assigned at birth, n (%)
Female	10 (12)	2 (5)	8 (19)
Male	72 (88)	37 (95)	35 (81)
Race and ethnicity, n (%)
Hispanic/Latino	26 (32)	14 (36)	9 (21)
White	36 (43)	15 (38)	25 (60)
Black/African American	(10)	7 (18)	1 (2)
Asian	4 (5)	1 (3)	3 (7)
Pacific Islander/Native Hawaiian	3 (4)	2 (5)	1 (2)
Not Provided	1 (1)	0 (0)	1 (2)
Tobacco use history, n (%)
Ever smoker	28 (34)	21 (54)	7 (16)
Current smoker	11 (13)	10 (26)	1 (2)
Clinical characteristics
Concurrent medical conditions, n (%)
Autoimmune disease*	2 (2)	1 (3)	1 (2)
Cancer (with treatment received within 2 years prior to COVID-19 diagnosis)	4 (5)	3 (8)	1 (2)
Diabetes	7 (9)	4 (10)	3 (7)
Heart disease	4 (5)	4 (10)	0 (0)
Pulmonary disease	8 (10)	5 (13)	3 (7)
HIV-related laboratory parameters
CD4+ T cell count (cells/uL)	639 (486–844)	596 (404–740)	670 (594–918)
CD4/CD8 ratio	1.18 (0.78–1.86)	0.94 (0.51–1.10)	2.00 (1.52–2.32)
Plasma HIV RNA <50 copies/mL, n (%)	-	37 (95)	-
Body mass index	28.1 (24.8–31.0)	28.8 (25.1–32.5)	27.1 (23.9–30.3)
Hospitalized during acute COVID-19, n (%)	11 (14)	4 (10)	7 (17)
Time since COVID-19 symptom onset (days)	112 (91–129)	118 (85–129)	111 (94–131)

The median time between COVID-19 onset and assessment was similar between groups (117 vs 11 days, p=0.58; Table 2). Initial severity was also similar (13% vs 17% hospitalized, p=0.41). Participants reported a wide array of PASC symptoms (Table 2). A larger proportion of PWH reported certain symptoms such as fatigue, gastrointestinal and certain neurocognitive symptoms, and issues with sleep.

Table 2. Symptoms reported at late follow-up among those with PASC.

Values reported as n (%). Participants were systematically asked about 32 individual symptoms at the late follow-up visit, which took place a median of 112 days from initial COVID-19 symptom onset.

Symptoms reported at late follow-up	All (n=82)	HIV+ (n=39)	HIV− (n=43)
Constitutional
Fatigue	26 (32)	16 (42)	10 (23)
Subjective fever	1 (1)	1 (3)	0 (0)
Chills	2 (2)	2 (5)	0 (0)
Objective fever	0 (0)	0 (0)	0 (0)
Upper Respiratory
Rhinorrhea	13 (16)	5 (13)	8 (19)
Sore throat	2 (2)	0 (0)	2 (5)
Cardiopulmonary
Cough	7 (9)	3 (8)	4 (9)
Shortness of breath	20 (25)	11 (29)	9 (21)
Chest pain	7 (9)	3 (8)	4 (9)
Palpitations	10 (12)	4 (11)	6 (14)
Fainting	0 (0)	0 (0)	0 (0)
Gastrointestinal
Diarrhea	7 (9)	5 (13)	2 (5)
Nausea	9 (11)	5 (13)	4 (9)
Loss of appetite	4 (5)	3 (8)	1 (2)
Abdominal pain	9 (11)	5 (13)	4 (9)
Vomiting	1 (1)	0 (0)	1 (2)
Constipation	0 (0)	0 (0)	0 (0)
Genitourinary
Menstrual cramps	0 (0)	0 (0)	0 (0)
Dyspareunia	0 (0)	0 (0)	0 (0)
Rash	8 (10)	7 (18)	1 (2)
Musculoskeletal
Myalgia	14 (17)	9 (24)	5 (12)
Back pain	5 (6)	5 (13)	0 (0)
Joint pain	11 (14)	8 (21)	3 (7)
Anosmia/Dysgeusia	14 (17)
Neurologic
Headache	14 (17)	9 (24)	5 (12)
Concentration problems	24 (30)	16 (42)	8 (19)
Dizziness	11 (14)	5 (13)	6 (14)
Balance problems	9 (11)	6 (16)	3 (7)
Neuropathy	9 (11)	7 (18)	2 (5)
Vision problems	11 (14)	8 (21)	3 (7)
Parosmia	2 (2)	0 (0)	2 (5)
Trouble Sleeping	20 (24)	13 (34)	7 (16)

SARS-CoV-2-specific antibody responses

Antibody responses were similar between groups. In models incorporating days since infection, age, sex, and COVID-19 hospitalization, HIV status was not a predictor of humoral responses as measured by SARS-CoV-2-specific antibody binding (1.31-fold higher; 95%CI: 0.70–2.46; p=0.40; Fig. 1a) and surrogate virus neutralization (1.01-fold higher; 95% CI: 0.63–1.63; p=0.95; Fig. 1b). Hospitalization for COVID-19 was associated with 4.29-fold higher titers of binding antibodies (95% CI:1.74–10.57; p=0.002) and 2.57-fold higher titers of surrogate viral neutralization (95% CI: 1.33–4.98; p=0.005).

Figure 1. SARS-CoV-2-specific immunologic parameters among people with and without HIV recovering from COVID-19.

A. Total SARS-CoV-2 antibodies. B. Surrogate SARS-CoV-2 neutralization titres. C. Frequency of IFNγ+ SARS-CoV-2-specific memory CD4+ and CD8+ T cells. D. PD-1 expression on SARS-CoV-2-specific CD4+ and CD8+ T cells. Bars represent median and interquartile ranges. **p<0.01; *** p<0.001 in covariate adjusted linear regression modeling.

SARS-CoV-2-specific T cell responses

In models incorporating days since infection, age, sex, and COVID-19 hospitalization, HIV status was not a predictor of the magnitude of interferon gamma (IFNɣ)-producing SARS-CoV-2-specific memory CD4+ T cell responses (1.14-fold higher; 95%CI: 0.76–1.71; p=0.55; median: 0.070% vs. 0.068%; Fig. 1c). However, PWH had 70% lower relative levels of SARS-CoV-2-specific memory CD8+ T cells (0.30-fold 95% CI: 0.13–0.72; p=0.007; median: 0.016% vs 0.034%; Fig. 1c).

PWH exhibited higher levels of PD-1 expression on SARS-CoV-2-specific memory CD4+ T cells in adjusted analyses (1.18-fold higher; 95% CI: 1.07–1.30; p=0.001; median: 65% vs. 57.1%; Fig. 1d), but no significant differences in PD-1 expression on SARS-CoV-2-specific CD8+ T cells (1.21-fold higher; 95% CI: 0.83–1.76; p=0.33; median: 25.0% vs. 24.1%; Fig. 1d).

Effect of CD4/CD8 ratio

The CD4/CD8 ratio was not predictive of binding antibody levels (p=0.30) or surrogate virus neutralization (p=0.61). Higher ratios were associated with 67% lower frequency of SARS-CoV-2-specific CD4+ T cells (0.33-fold; 95% CI: 0.19–0.97; p=0.04) and 36% lower SARS-CoV-2-specific PD-1 expression (0.64-fold 95% CI:0.57–0.97; p=0.03). Notably, higher CD4/CD8 ratios were associated with 66% lower PD-1 expression on SARS-CoV-2-specific CD8+ T cells (0.34-fold 95% CI: 0.13–0.87; p=0.02). There was a trend toward a similar finding among HIV-negative individuals (0.70-fold; 95% CI: 0.47–1.04; p=0.08).

Relationship between antibody and T cell immune responses

We observed strong correlations between binding (r=0.33, p=0.008) and surrogate viral neutralization responses (r=0.33, p=0.007) and between binding and surrogate viral neutralization responses and SARS-CoV-2-specific CD4+ T cells (r=0.41, p<0.001 and r=0.42, p<0.001, respectively).

Markers of systemic inflammation

In PWH compared to HIV-negative individuals, mean IL-6 levels were 1.55-fold higher (95% CI: 1.06–2.26; p=0.02), mean IP-10 levels were 1.31-fold higher (95% CI: 1.06–1.62; p=0.01), and TNFα levels were 1.26-fold higher (95% CI: 1.08–1.47; p=0.003).

Post-acute sequelae

PWH had 4.01-fold higher odds of PASC (95% CI: 1.45–11.1; p=0.008; 82.8% vs. 54.4%), in a model controlling for time since infection, hospitalization, and age. This was maintained when defining PASC as 3 or more symptoms in comparison to fewer than 3 symptoms (adjusted odds ratio (AOR) 2.72; 1.08–6.88; p=0.03; 59.8% vs. 33.6%). PWH reported more symptoms overall (median 3 [IQR 1–6] versus median 1 [IQR 0–5], p=0.02), and those with HIV had a 1.91-fold higher number of PASC symptoms than those without HIV (p=0.02).

Antibody and T cell responses did not correlate with PASC (Fig. 2a–d). In models adjusting for HIV status, higher PD-1 expression on total memory CD4+ T cells, but not memory CD8+ T cells, was independently predictive of lower odds of PASC (Fig. 2e–f). However, this finding appeared to be driven by differences in overall PD-1 expression between the HIV-positive and -negative groups. Interestingly, there appeared to be a trend toward higher PD-1 expression among those with PASC in stratified analyses (Fig. 2e–f). There was no relationship between PD-1 expression on SARS-CoV-2-specific CD4+ or CD8+ T cells and PASC (AOR 0.23; 95% CI: 0.01–6.31; p=0.39 and AOR 0.20; 0.01–3.50; p=0.27 respectively; Fig. 2g–h).

Figure 2. SARS-CoV-2-specific immune responses among people with and without PASC, stratified by HIV status.

A. Total SARS-CoV-2 antibodies. B. Surrogate viral neutralization titers. C. SARS-CoV-2-specific CD4+ T cells. D. SARS-CoV-2-specific CD8+ T cells. E. PD-1 expression on total memory CD4+ T cells. F. PD-1 expression on total memory CD8+ T cells. G. PD-1 expression on SARS-CoV-2-specific memory CD4+ T cells. H. PD-1 expression on SARS-CoV-2-specific memory CD8+ T cells. Bars represent median and interquartile ranges.

Some inflammatory markers were associated with increased odds of PASC (Fig. 3a–e). After adjusting for HIV status, the odds of PASC in the study population increased 1.18-fold for each 10% increase in IP-10 (AOR 1.18; 95% CI:1.01–1.38; p=0.04); and 1.10-fold for each 10% increase in IL-6 (AOR 1.10; 95% CI: 1.01–1.21; p=0.04); there was a trend in increased PASC with higher TNFα levels (AOR 1.19; 95% CI: 0.98–1.46; p=0.08).

Figure 3. Plasma markers of immune activation among people with and without PASC, stratified by HIV status.

A. Interleukin-6 levels. B. TNF-alpha levels. C. Interferon-gamma levels. D. IP-10 levels. E. MCP-1 levels. Bars represent median and interquartile ranges. ^aThe odds of PASC of the entire study population increased 1.18-fold for each 10% increase in IP-10 (AOR 1.18; 95% CI:1.01–1.38; p=0.04); and ^b1.10-fold for each 10% increase in IL-6 (AOR 1.10; 95% CI: 1.01–1.21; p=0.04) from adjusted linear regression. ^cAmong PWH, there were increased odds of PASC with each 10% increase in IP-10 levels (AOR 1.06; 95% CI: 1.00–1.11; p=0.05).

Among PWH, there were increased odds of PASC with each 10% increase in IP-10 levels (AOR 1.06; 95% CI: 1.00–1.11; p=0.05), and a trend for increased PASC with higher TNFα levels (AOR 1.20 per 10% increase; 95% CI: 0.97–1.49; p=0.09), but not IL-6 (p=0.64). This analysis was limited by the small number of individuals with HIV who reported full recovery.

DISCUSSION

We observed that HIV status was strongly associated with PASC, raising concerns that this condition might be common among PWH recovering from COVID-19. Higher levels of inflammation were associated with PASC. Finally, we observed differences in SARS-CoV-2-specific CD4+ and CD8+ T cells that might have implications for long-term immunity conferred by natural infection. This study adds to the limited data examining SARS-CoV-2-specific immune responses in PWH and underscores the need for larger and more detailed studies of PASC in PWH.

While there are massive efforts underway to understand the prevalence and pathophysiology of PASC, data among PWH are limited [30]. One study suggested that COVID-19 severity in PWH was associated with PASC [31], however it did not find an association with CD4+ T cell count, viral load, demographics or comorbidities, and did not compare PWH with HIV-negative individuals. Another study suggested that HIV was one factor associated with PASC among those requiring emergency department or hospital-based care [32], but included only 10 PWH and did not include biological measurements. While our cohort cannot estimate the population-level prevalence of PASC, the observation that persistent SARS-CoV-2-attributed symptoms were highly prevalent in PWH and that the adjusted odds of PASC were four-fold as high as in well-matched HIV-negative comparators was striking. Large-scale studies in which HIV can be examined as a predictor of PASC are urgently needed.

PASC may be driven, at least in part, by residual or ongoing inflammation following SARS-CoV-2 infection [9,10]. ART-treated HIV is a chronic inflammatory condition associated with persistent immune activation [15–18]. Further immune perturbations related to COVID-19 may therefore lead to a higher prevalence of PASC among PWH. Additional factors could also predispose PWH to PASC, such as autoimmunity [33], localized tissue inflammation [34], human herpesvirus reactivation [13], and microvascular dysfunction [14]. Other comorbidities including substance use and metabolic disorders may further contribute. Regardless of mechanism, our observation suggests that PASC may be especially common in PWH and emphasizes the need for studies of PASC in this population.

Data comparing SARS-CoV-2-specific adaptive immune responses in people with and without HIV remain limited. Given the association between the presence of potent, durable immune responses and protection from disease upon re-exposure, it is critical to understand how HIV may modulate protective immunity. Furthermore, there is evidence that SARS-CoV-2 can cause chronic infection in certain immunocompromised individuals, including those with advanced HIV infection [35]. There is ongoing concern that PWH are less likely to develop and maintain protective immunity. While some studies suggested lower humoral responses in PWH [6], we and others [36–38] have not observed this.

Data on cellular immune responses to SARS-CoV-2 infection in PWH are even more limited. A single high-quality study has shown similar T cell responses between PWH on ART and HIV-negative individuals [7]. Our findings contribute three key observations regarding SARS-CoV-2-specific cellular immune responses in PWH. First, using peptide pools that include optimal SARS-CoV-2 epitopes spanning the proteome, we found that PWH had lower SARS-CoV-2-specific CD8+ T cell responses. This difference was previously observed in non-Spike-specific T cell responses in PWH, and may indicate that PWH have impaired capacity to mount a protective CD8+ T cell response upon re-infection, particularly with heterologous variants with immune-evading mutations in the spike protein. It is also possible that PWH have expansion of other antigen-specific CD8+ T cells (e.g., CMV-specific) thereby diluting the SARS-CoV-2-specific pool as the denominator was total non-naive CD8+ T cells. Second, we found that SARS-CoV-2-specific CD4+ T cells in PWH had higher expression of the co-inhibitory receptor PD-1, suggesting they may have impaired functionality upon re-encountering infection. Alternatively, PWH may have more SARS-CoV-2 antigen exposure leading to a more exhausted phenotype. Third, we found that a higher CD4/CD8 ratio - which can sometimes be optimized with early ART initiation [39] - was associated with lower expression of PD-1 on SARS-CoV-2-specific CD8+ T cells.

This study was small and underpowered to make comparisons within PWH. Future studies could explore questions regarding differences in post-COVID cellular immunology and cytokine patterns in relation to ART regimen, potential changes in HIV gene expression or viral transcription especially in light of recent observations that viral “blips” may be common post-COVID [40], and concomitant latent viral infections such as human herpesviruses (e.g., Epstein-Barr virus and cytomegalovirus), which are common among PWH and may contribute to PASC. Given the nature of recruitment, the high prevalence of PASC should not be considered to represent the population-level prevalence, which is likely lower [41]. Data on other potential confounders such as comorbid mental health and substance use were unavailable. Inflammatory marker data was not available on all participants, although sample availability was based on the timing of collection and is not expected to bias the results; a wider array of markers including general inflammatory markers (e.g., C-reactive protein, complements, and autoantibodies) should be considered in future analyses. Finally, our population was mostly male, stable on ART, and had strong immune reconstitution, and care should be taken when extrapolating to populations with less access or adherence to ART or those with advanced HIV.

Our analysis provides compelling preliminary evidence suggesting an urgent need to better understand the epidemiology and pathophysiology of PASC within PWH. Such efforts may lead to targeted interventions to prevent or treat PASC among this special population of interest.