Objective Incidence and manifestations of postacute sequelae of coronavirus disease 2019 (PASC) are poorly defined among immunosuppressed populations. We reported, phenotyped, and assessed risk factors for PASC in adults with systemic autoimmune diseases.

Methods Persons aged ≥ 18 years with systemic autoimmune diseases were recruited into a national, prospective observational cohort of SARS-CoV-2 vaccination and infection between December 2020 and April 2021. Serial surveys assessed vaccination status, SARS-CoV-2 infection incidence, and disease flares. Participants reporting SARS-CoV-2 infection received a questionnaire assessing symptom duration, severity, and quality of life (QOL) effect; PASC was defined as ≥ 1 symptom persisting for > 12 weeks. PASC syndromes were mapped by overlapping symptom domains. Characteristics were compared between participants who did vs did not report PASC.

Results Among 1615 participants, 590 (36.5%) reported SARS-CoV-2 infection and were sent PASC surveys, 299 (50.7%) of whom responded > 12 weeks following the reported infection. Respondents were 91.6% female, 91.2% White, median (IQR) age was 48 (40-60) years with median (IQR) 3 (2-3) vaccine doses at time of first infection. Common diagnoses included inflammatory arthritis (38.5%) and inflammatory bowel disease (14.4%). Eighty-nine of 299 (29.8%) reported PASC, with the most reported symptom domain being neurological/psychological (83.1%); 84% reported an effect on QOL. Participants with PASC reported lower number of preceding vaccines (median [IQR] 2 [2-3] vs 3 [2-3]; P < 0.001) and more reinfections (16.9% vs 5.7%; P = 0.004).

Conclusion In a large, real-world cohort, 29.8% of persons with systemic autoimmune disease reported PASC, often affecting QOL. Preceding vaccination may reduce PASC, whereas multiple infections may increase risk, supporting ongoing booster vaccine campaigns and efforts to limit breakthrough infections.

With decreasing severity of SARS-CoV-2 infections in the general population, attention has shifted to outcomes in at-risk subpopulations such as immunosuppressed persons, as well as growing recognition of long-term effects of coronavirus disease 2019 (COVID-19). In particular, some persons may develop persistence of symptoms following resolution of acute infection, a condition known as postacute sequelae of COVID-19 (PASC) or “long COVID,” which can be debilitating and affect quality of life (QOL).1

PASC definitions continue to evolve, though most consortia refer to persistence of ≥ 1 SARS-CoV-2-related symptom(s) for ≥ 4 weeks to 12 weeks following initial infection.2,3 In immunocompetent adults, PASC is characterized by myriad symptoms including fatigue, dyspnea, and neurocognitive effects (eg, “brain fog”). However, estimates of PASC prevalence, as well as symptom profiles, vary widely owing to variation in definitions, ascertainment mechanisms, and studied populations.2 PASC phenotypes in the general population have included clusters such as cardio-renal4 or postexertional malaise and fatigue.5

Risk factors for PASC also vary, such as younger age, female sex, and presence of preexisting conditions, though other studies have not found such associations.2,6-10 Although immunosuppressed status is associated with worse acute SARS-CoV-2 infection outcomes,11 and potentially high rates of PASC,10,12 rates of these outcomes in the contemporary (ie, Omicron subvariant) era are uncertain. Dilorio et al found in a cohort of persons with systemic autoimmune rheumatic diseases that 24% still had COVID-19 symptoms at 28 days and 10% still had symptoms at 90 days.10 PASC is of particular interest in this population given potential for overlapping clinical features of PASC with systemic autoimmune conditions as well as high demographic representation of younger females. Additionally, certain persons with systemic autoimmune diseases mount poor antibody responses to SARS-CoV-2 vaccination and thus are at higher ongoing risk of breakthrough infection13; correlates of PASC with antibody level are less established.14 Vaccination may still reduce PASC severity in these patients, decreasing from 54% in unvaccinated/partially vaccinated persons to 41% in persons who have received ≥ 2 mRNA or a single adenovirus-vectored (Ad26.COV2.S [recombinant]) vaccine(s) in 1 study.15

Given uncertain PASC burden in persons with systemic autoimmune disease amid ongoing virus circulation, we sought to measure PASC prevalence within a prospective cohort, focusing on associated risk factors and clinical phenotyping to characterize this chronic complication of SARS-CoV-2 infection.

Study population. Participants aged ≥ 18 years with systemic autoimmune diseases reporting ≥ 1 SARS-CoV-2 vaccine and any SARS-CoV-2 infection were recruited to a national, prospective, observational cohort between December 2020 and April 202116 and were subsequently sent a PASC survey. Informed consent was obtained electronically, with a waiver of written consent. Demographics, autoimmune disease characteristics and medications, disease activity, SARS-CoV-2 vaccination history, and incident SARS-CoV-2 infection (self-reported positive PCR or antigen test) were actively collected by cohort-wide surveys and passive reporting.

Active surveillance for SARS-CoV-2 infection was conducted through serial surveys distributed 1 week and/or 1 month following each reported immunoprophylactic event (vaccination or tixagevimab/cilgavimab). Because of high SARS-CoV-2 infection rates in the Omicron era, 2 additional surveys were distributed on February 16, 2022, and July 18, 2022. Continuous passive reporting was permitted throughout the study period.

Statement of ethics and consent. Participants provided informed consent electronically. This study was approved by the Johns Hopkins Institutional Review Board (IRB00248540).

Survey tools and definition of PASC. Participants were queried across multiple symptom domains (adapted from the US Centers for Disease Control, Atlanta, GA, symptom list), and for their duration, effect on QOL, and severity of impairment of daily activities. PASC was defined as persistence of ≥ 1 symptom for > 12 weeks following reported SARS-CoV-2 infection, aligning with the World Health Organization definition.3 Index infection was defined as the date of a single infection if only 1 infection was reported. If 2 infections were reported within 60 days of each other, they were counted as a single infection episode, with the earlier infection date as the index infection. If 2 infections occurred > 60 days apart, they were counted as separate infection episodes, and the index infection was defined as the later date.

PASC survey responses were collected from September 12, 2022, to December 18, 2022. Self-perceived function levels (no, slight, moderate, or severe problems doing usual activities) before and after SARS-CoV-2 infection were assessed.

Phenotypic profiling of PASC. Reported PASC symptoms were divided into 5 domains based on common clinical syndromes. Domains included neurological/psychological (headache, depression/low mood, loss of concentration/brain fog, loss of smell or taste), fatigue, cardiopulmonary (cough, dyspnea, chest pain), gastrointestinal (loss of appetite, nausea/vomiting, diarrhea, abdominal pain), and musculoskeletal (weakness in arm/legs, persistent muscle pain, joint pain/swelling). Any given symptom appeared in only 1 domain and a participant could report symptoms across multiple domains, each of which was recorded as present.

PASC symptom spectrum and severity were analyzed posthoc in specific patient subgroups (systemic lupus erythematosus, inflammatory arthritis, and inflammatory bowel disease [IBD]), and in patients on heavy immunosuppression (mycophenolate [MMF], rituximab [RTX], or belimumab).

Statistical methods. Demographics, clinical characteristics, and vaccines received at index infection were compared between groups with and without PASC using Wilcoxon signed-rank test for continuous variables and Fisher exact test for binary/categorical variables. The denominator of each proportion reflects the total number of participants who responded to a respective question.

Two-sided α < 0.05 was used to define statistical significance. Analyses were performed using Stata/SE 17.0 (StataCorp). Figures were created using Microsoft Excel and PowerPoint (v16.73) and BioRender (Science Suite).

Incidence rate analyses. Poisson regression analyses were performed to assess the independent associations with the outcome of PASC. Namely, among only those reporting a single infection, the independent association of the number of vaccines preinfection with report of PASC was assessed, adjusting for age at time of infection, heavy immunosuppression (MMF, RTX, belimumab), and underlying condition. This procedure was repeated among the full cohort using the same multivariable Poisson model with the addition of reported reinfection, to assess the independent association of multiple infections on PASC incidence.

Sensitivity and subgroup analyses. Analyses were repeated using (1) a liberal definition of PASC (persistent symptoms ≥ 4 weeks following SARS-CoV-2 infection), and (2) restricting to only individuals who reported an infection ≥ 14 days following ≥ 2 mRNA or adenovirus-vectored (Ad26.COV2.S [recombinant]) vaccines.

Participant characteristics. Among 1615 participants reporting ≥ 1 SARS-CoV-2 vaccine dose, 590 (36.5%) reported SARS-CoV-2 infection and received a PASC survey, of whom 442 (74.9%) responded. Among these respondents, 299 responded > 12 weeks following their reported infection, at a median (IQR) of 214 (118-260) days postinfection (Supplementary Figure S1, available with the online version of this article).

The median (IQR) participant age was 48 (40-60) years, 91.6% were female, 91.2% were White, and the most common systemic autoimmune diagnoses were inflammatory arthritis (38.5%) and IBD (14.4%; Table). The majority (90.3%) were taking ≥ 1 immunomodulatory therapy, including 52.8% reporting ≥ 2 drugs. Median (IQR) number of vaccine doses received prior to index SARS-CoV-2 infection was 3 (2-3). Of 330 infection episodes (31 reinfections), dominant variant eras were pre-Delta in 47 (14.2%), Delta in 57 (17.3%), Omicron BA.1 in 88 (26.7%), BA.2 in 133 (40.3%), and BA.5 in 5 (1.5%). Of the infections, 98.1% were mild (ie, not associated with hospitalization).

Clinical characteristics of adults with autoimmune disease who did vs did not report PASC.

Prevalence and severity of PASC. PASC was reported by 89 of 299 (29.8%) participants. The most commonly reported symptoms included fatigue (70.8%), loss of concentration/brain fog (68.5%), and joint pain/swelling (38.2%). The most common symptom domain was neurological/psychological (83.1%). Sixty-eight of 89 (76%) reported symptoms in ≥ 2 domains, including 15 of 89 (17%) reporting symptoms across all 5 domains (Figure 1A). Regarding effect of symptoms, 65 of 77 (84%) reported continued effect on QOL at the time of survey. Sixty-four of 89 (72%) reported a change in level of functioning post infection, including 37 (58%) who perceived declined functional level (eg, no problems pre-SARS-CoV-2 infection, yet at least slight problems at survey; Figure 1B). Fifty-seven of 63 (90%) attributed these changes to SARS-CoV-2 infection. Additionally, 38 of 89 (43%) reported increased disease activity (flares) requiring treatment following infection, 67% of whom (24 of 36) reported flares in the 6 months prior to infection (n = 2 did not respond).

(A) Symptom domains for 89 participants who reported PASC (persistent symptoms > 12 weeks post SARS-CoV-2 infection). (B) Impact on daily functioning for 89 participants who reported PASC. PASC: post-acute sequelae of coronavirus disease 2019.

Among patient subgroups, the most common symptoms were fatigue and brain fog, regardless of autoimmune condition. Severity of reported PASC did not differ between those on vs not on heavy immunosuppression (Supplementary Table S1, available with the online version of this article).

Associations with PASC. Demography (age, sex, race), immunosuppression, and use of antivirals was similar among those with vs without PASC (Table). In contrast, participants with PASC reported fewer vaccine doses before the index SARS-CoV-2 infection (median [IQR] 2 [2-3] vs 3 [2-3]; P < 0.001), moderately more time elapsed between most recent vaccination and index infection (median [IQR] 264 [154-363] days vs 208 [121-286] days; P = 0.006) and more reinfections (16.9% vs 5.7%; P = 0.004) compared to those without PASC.

Independent associations with PASC. Single infections were reported by 272 respondents. After adjustment, the incidence rate ratio (IRR) of reporting PASC comparing participants who received 1, 2, or 3 vaccine doses to those reporting no vaccine doses prior to their index infection, was 0.23 (95% CI 0.04-1.46; P = 0.12), 0.48 (95% CI 0.31-0.76; P = 0.002), and 0.30 (95% CI 0.19-0.46; P < 0.001), respectively.

In the reinfection analysis, the IRR of reporting PASC comparing participants who reported multiple vs single infections pre-PASC survey was 2.32 (95% CI 1.58-3.41; P < 0.001).

Subgroup analysis. Among participants with ≥ 2 vaccine doses prior to index infection (n = 264, 88.3%) vs those with < 2 vaccines (n = 35, 11.7%), PASC frequency was lower in the group with ≥ 2 vaccine doses (26.5% vs 54.3%; P = 0.001). Those with ≥ 2 vaccine doses prior to index infection showed similar symptomatology as the main analysis, with common reports of fatigue (70.0%) and loss of concentration (67.1%). The most common symptom domain was neurological/psychological (84.3%), with 75.7% reporting symptoms in ≥ 2 domains and 14.3% reporting symptoms across all 5 domains. Continued effect on QOL at the time of the survey was reported by 52 of 59 (88%) participants. Among those with ≥ 2 vaccines, there was no difference in median number of vaccine doses preceding their index SARS-CoV-2 infection among those with vs without PASC (median [IQR] 3 [2-3] doses in both groups).

Sensitivity analysis. Under the ≥ 4-week symptom definition of PASC, 203 of 403 (50.3%) reported PASC. Symptom patterns were broadly similar to those with symptoms at > 12 weeks, including frequent fatigue (66.5%) and brain fog (57.1%), yet respiratory symptoms (cough 40.4%, dyspnea 38.4%) were more frequent (Supplementary Figure S2, available with the online version of this article).

One hundred fifty-three of 203 (75.4%) participants who reported PASC symptoms at 4 weeks completed the PASC survey > 12 weeks following their index infection. Of these participants, 89 of 153 (58.2%) reported persistent PASC symptoms at > 12 weeks.

SARS-CoV-2 infection waves continue, and although rates of severe acute SARS-CoV-2 infection have decreased, it remains important to understand consequences of PASC among at-risk subpopulations. In this national study of vaccinated adults with systemic autoimmune diseases who generally experienced mild disease in the Omicron era, 30% reported PASC, of whom 84% reported enduring effect on QOL. Importantly, this suggests that PASC remains a substantial concern even for those without a severe infection course. Fatigue and neurological/psychological symptoms predominated, though phenotypic manifestations varied widely. Although clinical, demographic, and immunosuppressive medications were very similar between those with vs without PASC, those receiving more preceding vaccines and fewer reinfections reported less PASC. Encouragingly, approximately half of those reporting PASC at 4 weeks showed resolution by 12 weeks.

Estimates of PASC prevalence have ranged between 10% and 60% across varying demographics and PASC definitions.5,8-10,14,15,17-20 The relatively high PASC frequency of 30% in this cohort may be related to the potentially higher risk of PASC associated with female participants, and/or baseline autoinflammatory conditions, and/or a contribution of immunocompromised status and associated blunting of vaccine-associated protection. In keeping with certain studies in immunocompetent persons, fatigue and loss of concentration/brain fog were the most commonly reported symptoms in this cohort.2 A relatively large subgroup (17%) reported symptoms across all 5 domains, a phenotype that is not as well characterized in prior literature.4,5 This may be because of the immunoreactive profile of the included participants, with potential for exacerbation of baseline autoinflammatory conditions or development of other multisystem inflammation or dysfunction. Indeed, the potential for overlap of reported PASC symptoms with those of systemic autoimmune conditions indicates a need for careful assessment by clinicians to ascertain SARS-CoV-2 infection history and tempo of consistent symptoms. Reliably identifying PASC in this group, and parsing among phenotypes, may require dedicated mechanistic and biomarker studies.

Consistent with prior work, we found higher rates of PASC among unvaccinated or partially vaccinated participants.15 This suggests that, despite potential attenuated vaccine immune responses among immunocompromised populations,21-23 multiple vaccinations may still reduce long-term sequelae. Notably, although improved antibody response to vaccination may not correlate well with reduced PASC risk,14 other immunological responses might still provide protection against this complication. Correspondingly, our data also suggest that SARS-CoV-2 reinfections may increase the risk of PASC (ie, more “natural” viral antigenic encounters may compound direct and indirect effects on health). This association may also mark a state of heavier immunosuppression that subverts storage of functional immune memory to preceding infection or vaccination. It remains difficult, however, to disambiguate receipt of vaccine behavior and reinfection histories from accompanying measures of social behaviors and adherence to mitigation strategies. Regardless, the potential for additive risks of PASC over time highlights the importance of infection prevention efforts for at-risk populations including optimized vaccination, passive immunoprophylaxis, and access to antiviral therapies.

Limitations of this study relate to its observational design including reliance on patient report, incomplete data (such as lack of comorbidities, behavioral factors such as smoking status, or interruptions of disease-modifying antirheumatic drug therapy surrounding COVID-19 events), as well as recall and selection biases (such as persons with more severe symptoms being more likely to respond to surveys). Additionally, the cohort was relatively homogenous in demographics, vaccination status (all vaccinated), and disease severity (2% hospitalized), which may limit generalizability to other populations. Finally, the study did not include a nested comparator group with longitudinal report of autoimmune disease activity, which limits ability to distinguish PASC symptoms from those of underlying systemic autoimmune conditions. The prospective, national structure and high survey response rate within this cohort, however, reflects real-world longitudinal participant experience and recorded difficult-to-assess infection events by home testing.

In summary, PASC appears relatively common in vaccinated persons with systemic autoimmune and rheumatologic diseases, including those experiencing mild acute disease in the Omicron era. Symptoms are heterogeneous, occur in multiple organ systems, and can have significant and prolonged effects on QOL. More research in preventive and treatment strategies is necessary among immunosuppressed populations to determine an optimal approach to reduce the burden of disease. Given higher PASC incidence among those with fewer SARS-CoV-2 vaccinations and multiple infections, these findings support the ongoing recommendation for booster vaccination and infection prevention strategies in this patient population.

We acknowledge the following individuals for their assistance with this study: Brian J. Boyarsky MD, PhD, Jake A. Ruddy, MD, and Roni Shanoada, BSc, MSc.

This research was made possible with generous support of the Ben-Dov and Trokhan Patterson families. This work was supported by grant numbers K24AI144954 (DLS), K23AI157893 (WAW), and U01AI134591 and U01AI138897 (DLS) from the National Institute of Allergy and Infectious Diseases; K23AR073927 (JJP) from the National Institute of Arthritis and Musculoskeletal and Skin Diseases; R01DK132395 (AM) from the National Institute of Diabetes and Digestive and Kidney Diseases; and the Jerome L. Greene Discovery Fund (CMC, JJP). The analyses described here are the responsibility of the authors alone and do not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.

M.S. Teles and J. Brundage contributed equally as co-first authors.

J.J. Paik and W.A. Werbel contributed equally as co-senior authors.

LCS reports consultant fees from Janssen, Boehringer Ingelheim, Mallinckrodt, EMD Serono, Allogene, and ArgenX. DLS reports receiving honoraria from Sanofi, Novartis, Caredx, AstraZeneca, Bridge to Life (speaking); from Novartis, Veloxis, Mallinckrodt, Jazz, CSL Behring, Thermo Fisher, Caredx, Transmedics, Kamada, MediGO, Regeneron, AstraZeneca, Takeda (consulting); and from Novavax (advisory board). WAW reports advisory board fees from AstraZeneca and Novavax; speaker fees from AstraZeneca; and consultant fees from the US Centers for Disease Control/Infectious Diseases Society of America COVID-19 Real-Time Learning Centers for Disease Control, Atlanta, GA/Infectious Diseases Society of America COVID-19 Real-Time Learning Network and GlobalData. The remaining authors have no financial disclosures or conflicts of interest relevant to this work.

This is an Open Access article, which permits use, distribution, and reproduction, without modification, provided the original article is correctly cited and is not used for commercial purposes.