小妲己直播

Importance听 Neuropsychiatric symptoms have been reported as a prominent feature of postacute sequelae of COVID-19 (PASC), with common symptoms that include cognitive impairment, sleep difficulties, depression, posttraumatic stress, and substance use disorders. A primary challenge of parsing PASC epidemiology and pathophysiology is the lack of a standard definition of the syndrome, and little is known regarding mechanisms of neuropsychiatric PASC.

Observations听 Rates of symptom prevalence vary, but at least 1 PASC neuropsychiatric symptom has been reported in as many as 90% of patients 6 months after COVID-19 hospitalization and in approximately 25% of nonhospitalized adults with COVID-19. Mechanisms of neuropsychiatric sequelae of COVID-19 are still being elucidated. They may include static brain injury accrued during acute COVID-19, neurodegeneration triggered by secondary effects of acute COVID-19, autoimmune mechanisms with chronic inflammation, viral persistence in tissue reservoirs, or reactivation of other latent viruses. Despite rapidly emerging data, many gaps in knowledge persist related to the variable definitions of PASC, lack of standardized phenotyping or biomarkers, variability in virus genotypes, ascertainment biases, and limited accounting for social determinants of health and pandemic-related stressors.

Conclusions and Relevance听 Growing data support a high prevalence of PASC neuropsychiatric symptoms, but the current literature is heterogeneous with variable assessments of critical epidemiological factors. By enrolling large patient samples and conducting state-of-the-art assessments, the Researching COVID to Enhance Recovery (RECOVER), a multicenter research initiative funded by the National Institutes of Health, will help clarify PASC epidemiology, pathophysiology, and mechanisms of injury, as well as identify targets for therapeutic intervention.

This Special Communication highlights what is currently understood about neurological and psychiatric (herein neuropsychiatric) symptoms in adults that develop and persist after SARS-CoV-2 infection and considers them in the context of the pandemic using a targeted rapid review of the literature. We then introduce Researching COVID to Enhance Recovery (RECOVER),¹ a multicenter research initiative funded by the National Institutes of Health (NIH) to help identify and address the postacute sequelae of COVID-19 (PASC).

A primary challenge of parsing PASC epidemiology and pathophysiology is the lack of a standardized and biologically based definition of the syndrome. While the Centers for Disease Control and Prevention (CDC) describes post鈥揅OVID-19 conditions as occurring 4 weeks or more after infection,² the World Health Organization (WHO) definition³ requires symptoms to be present 3 months or longer after infection and to last 2 months or longer. Because the WHO definition was released October 6, 2021, most publications use heterogeneous time frames, in part because authors used pragmatic definitions prior to the CDC or WHO guidance statements. Other groups have also published PASC definition guidelines. The UK National Institute for Health and Care Excellence (NICE) put forth 2 definitions of PASC: (1) ongoing symptomatic COVID-19 lasting 4 to 12 weeks after the onset of acute symptoms and (2) post鈥揅OVID-19 syndrome for those with symptoms longer than 12 weeks after the onset of acute COVID-19.⁴ For the purposes of this communication, we use the CDC definition of symptoms occurring 4 weeks or longer from index infection.

We conducted a targeted rapid review of literature published on PubMed and PsycInfo between January 2020 and February 1, 2022, using the search terms 鈥渃hronic COVID,鈥� 鈥減ost-acute COVID,鈥� 鈥淟ong-Haul Covid,鈥� 鈥淟ong-Hauler Covid,鈥� 鈥淟ong Covid,鈥� 鈥減ost-acute sequelae,鈥� 鈥減ersistent symptom,鈥� 鈥淪ARS-CoV-2,鈥� AND 鈥減sychiatric鈥� OR 鈥減sychological鈥� OR 鈥渘eurological鈥� OR 鈥渘europsychiatric鈥� OR 鈥渕ental disorders鈥� OR 鈥渄epression鈥� OR 鈥渁nxiety鈥� OR 鈥渃ognition鈥� or 鈥渕ood disorders鈥� OR 鈥渂rain.鈥� Relevant articles were reviewed by both authors for inclusion. In this rapid review, we used transparent and reproducible search methods; however, because of time contraints and the rapidly changing nature of COVID-19 literature, source searches were limited in scope. In addition, we did not include a risk-of-bias assessment, an analysis of missing data or data heterogeneity, a meta-analysis, or a certainty assessment, as are typical in a formal systematic review.

Acute COVID-19 neuropsychiatric symptoms may persist in a subset of patients or recur later, consistent with PASC. Understanding the continuity of symptoms from the acute to postacute period is critical to identifying risk factors and mechanisms underlying PASC. For example, some symptoms may represent a static insult with residual disability (eg, stroke), while others may represent ongoing maladaptive responses, such as inflammation or autoimmune mechanisms. Other symptoms may be tied to acute illness factors but wane over time. Disaggregating etiologies hinges on evaluating the entire disease course from the acute to postacute periods as well as contextualizing symptoms in the evolving pandemic environment.

The overall prevalence of neuropsychiatric symptoms during acute COVID-19 is difficult to estimate, and differences among SARS-CoV-2 variants are still emerging. Because at least 33% of individuals infected with SARS-CoV-2 are completely asymptomatic during the acute phase of infection, the association between SARS-CoV-2 and neuropsychiatric symptoms may be unrecognized, potentially leading to underestimates of PASC after asymptomatic or mild COVID-19.⁵ Conversely, acute neurological events among hospitalized patients with severe COVID-19 are extensively documented, with prevalence rates from 14% to 33%^6,7 across heterogeneous studies. The most prevalent reported neurologic symptoms are fatigue in 33%,⁸ sleep abnormalities in 29%,⁹ headache in 23%,¹⁰ and anosmia/dysgeusia in 18%.⁷ Meta-analyses addressing psychiatric symptoms have reported pooled prevalence rates as high as 42% to 45% for depression and 37% to 47% for anxiety,^9,11 both higher than rates without infection (eg, 24% depression, 26% anxiety) (Table 1).^6-22 Although control groups are variably included, some studies suggest higher rates of ischemic stroke, hemorrhagic stroke, Guillain-Barr茅 syndrome, neuropathy, myopathy or neuromuscular junction disorder, anxiety, mood disorder, psychotic disorder, insomnia, and substance abuse disorder compared with patients with influenza or other respiratory tract infections.^16,23

Post鈥揅OVID-19 neuropsychiatric sequelae have been reported in up to 91% of patients with COVID-19¹⁴ 6 months after hospitalization and in approximately 25% of nonhospitalized individuals with COVID-19.¹² However, sequelae rates vary depending on the spectrum of complications considered, the severity of the index infection and course, the time window from initial infection, and the assessment methodology. The most commonly reported post鈥揅OVID-19 neuropsychiatric events occurring within 4 weeks to 6 months postinfection include cognitive abnormalities (4%-47%),^{12,14,15,19,20} sleep disturbances (3%-27%),^12,14,16 anxiety (7%-46%),^12-14,24 depression (3%-20%),^12-14,24 posttraumatic stress disorder (PTSD) (6%-43%),^13,24 fatigue (5%-32%),^12,14,15,17 and headache (5%-12%)^12,15 (Table 1). These rates appear to be higher than rates observed in similar patient populations without COVID-19. Among 73鈥�000 nonhospitalized patients 30 days or longer after infection, incident onset of anxiety and fear-related disorders and trauma- and stress-related disorders was significantly greater than among those without COVID-19.²⁵ Additionally, rates of stroke, dementia, mood, anxiety, psychotic, and substance use disorders were each significantly higher at 6 months than among control patients with influenza or other respiratory conditions (Table 1).¹⁵ However, not all studies have found higher rates of mood and anxiety symptoms after COVID-19 hospitalization than among comparator groups, and substantial variability exists in study methods and quality.²⁶ Studies with shorter time frames (eg, 1-3 months) tend to report higher threshold anxiety, depression, and PTSD rates than those more than 3 months postinfection, although early distress (eg, at 1 month) predicts 1-year depression, anxiety, and traumatic distress, irrespective of prior psychiatric history or gender,²⁷ supporting the need for longitudinal studies. That prior neuropsychiatric illness is also associated with higher rates of hospitalization, intensive care unit (ICU) admission, and mortality with COVID-19^16,28 highlights the potential for bidirectional risk.

Systematic research is needed to better understand risk factors for neuropsychiatric aspects of PASC. Individuals with preexisting neurological or psychiatric histories are at risk for worsening and/or recurrence with COVID-19.⁶ Although data are inconsistent, more severe COVID-19 disease and course (requirement of hospitalization, use of invasive mechanical ventilation, and/or ICU admission) are risk factors for PASC.^15,16,25 Multiomics analyses suggest that history of diabetes, initial SARS-CoV-2 viral load, autoantibodies (anti-interferon and antinuclear), and reactivation of latent Epstein-Barr virus at COVID-19 illness onset may all predict PASC.²⁹ Of note, risk factors for neuropsychiatric events that occur acutely after COVID-19 (older age, male sex, White race, COVID-19 illness severity, medical comorbidities, and past neurological and psychiatric disease)¹⁴ vary substantially from risk factors for postacute neuropsychiatric sequelae (middle age, female sex, racial and ethnic minority groups, baseline disability, fewer years of formal education, and/or a history of psychiatric disease).^14,30 Potential hypotheses include that mechanisms of injury more common in middle-aged women, such as autoimmune disease, may explain some of these differences. Social determinants of health, variability in access to health care, and community-specific stigma surrounding treatment of psychiatric illness that may include physical symptomatology may also contribute.

The preponderance of human data suggests the pathophysiology underpinning neuropsychiatric injury during acute infection is related to secondary effects of SARS-CoV-2, including hypoxemia, hyperinflammation, and hypercoagulability. Neuropathological evidence of acute hypoxic-ischemic brain injury exists in multiple COVID-19 autopsy cohorts.³¹ Additionally, elevations in proinflammatory cytokines, particularly IL-6, is a hallmark of moderate-severe acute COVID-19 known to promote endothelial dysfunction, vascular permeability, and potentially blood-brain barrier (BBB) dysfunction.³² Neuropathological data among COVID-19 decedents have revealed endothelial injury, microhemorrhages, disruption of the microvasculature basal lamina, and extravasation of fibrinogen into brain parenchyma,³³ all suggestive of BBB disruption that may be mediated by a COVID-19鈥搑elated inflammatory state. Intense inflammation in turn triggers hypercoagulability resulting in microthromboses and microvascular endothelial injury, seen in postmortem neuropathological studies.³¹ Hypercoagulability coupled with endothelial injury may lead to ischemic stroke, hemorrhagic conversion, or primary intracranial hemorrhage.³⁴ Neurodegenerative blood markers (GFAP, NFL, UCHL1) indicating injury to neurons, glia, and axons are acutely elevated in hospitalized patients with COVID-19 (and no history of neurodegenerative disease) to levels observed in control patients with Alzheimer disease but without COVID-19, particularly among patients with acute neurological symptoms, suggesting profound brain injury for some.³⁵

While direct neuronal invasion by the SARS-CoV-2 virus has been postulated as a mechanism of injury, most neuropathological evidence suggests it is not a driving factor. While SARS-CoV-2 RNA has been identified by polymerase chain reaction in regionally based samples of olfactory mucosa and surrounding tissue, in situ hybridization (more cell specific and less prone to contamination by neighboring cells) did not detect SARS-CoV-2 in any neural tissue in 2 studies.^36,37 Most other reports of SARS-CoV-2 detection by reverse transcriptase鈥損olymerase chain reaction in neural tissue either did not report levels or had low RNA copies,³¹ which may represent contamination.³⁸

Little is known regarding mechanisms of neuropsychiatric PASC, and it is likely that host, viral, and environmental factors contribute to differing degrees depending on the PASC subphenotype. Here we present several hypothetical mechanisms of PASC. First, PASC symptoms may represent static brain injury accrued during acute COVID-19. Patients with COVID-19鈥搑elated ischemic stroke, intracranial hemorrhage, or hypoxic-ischemic brain injury may have trajectories of recovery that span months to years, and many may have permanent disabilities. Additionally, critical illness of any cause is associated with long-term cognitive deficits³⁹ as well as motor deficits related to critical illness neuropathy/myopathy. In these cases, disability is largely related to secondary complications of COVID-19, rather than an ongoing insult specific to SARS-CoV-2.

Another mechanism may involve progressive neurodegeneration triggered by post鈥揅OVID-19 hypoxia, inflammation, and BBB disruption, similar to that described in other disease models, such as traumatic brain injury. Hypoxia, particularly chronic hypoxia, has been linked to early Alzheimer disease pathology by a variety of mechanisms that lead to amyloid 尾 accumulation,⁴⁰ with decreased amyloid breakdown. Neuroinflammation may additionally promote both amyloid plaque and neurofibrillary tangle formation.⁴¹ Other neurodegenerative diseases, such as Parkinson, may occur or progress at higher rates after SARS-CoV-2 infection.⁴²

Autoimmune mechanisms for neuropsychiatric PASC have also been proposed. Indeed, acute disseminated encephalomyelitis, acute necrotizing encephalomyelitis, Guillain-Barr茅 syndrome, and transverse myelitis, which are believed to be caused by molecular mimicry, have all been reported after SARS-COV-2 infection, although onset is typically within 7 to 14 days (Table 1). Autoantibodies, including anti-interferon 伪2 and antinuclear antibodies, have been correlated with postacute respiratory and gastrointestinal symptoms.²⁹ A compartmentalized central nervous system autoimmune response following SARS-CoV-2 infection has also been observed, further bolstering this hypothesis.⁴³ Some data suggest these autoantibodies precede COVID-19 and only generate clinical syndromes following SARS-CoV-2 infection.²⁹ Notably, PASC shares some symptomatic similarities with chronic fatigue syndrome, which may have autoimmune underpinnings, such as autoantibodies to neurotransmitters, changes in cytokine profiles, and decreased natural killer cell cytotoxicity.⁴⁴ Additionally, there has been an increase in diagnoses of chronic fatigue syndrome 3 to 9 months after SARS-CoV-2 infection among nonhospitalized patients.⁴⁵

Viral persistence in immune sanctuaries has been postulated as a PASC mechanism, similar to that described with HIV; this speculation is largely driven by cases of persistent viral shedding in immunosuppressed patients.⁴⁶ Alternately, reactivation of latent viruses may occur, such as herpesviruses (herpes simplex virus 1, varicella zoster, Epstein-Barr). Initial SARS-CoV-2 viral load and Epstein-Barr viremia during acute COVID-19 have been implicated in postacute memory concerns.²⁹

Both autoimmunity and viral persistence can contribute to persistent chronic inflammation in patients with PASC.⁴⁷ Persistence of proinflammatory cells, altered cytokine production, and immune-metabolic pathway disruptions may all play a role in promoting a chronic inflammatory state.⁴⁷ Furthermore, persistent inflammatory responses such as IL-6 cytokine dysregulation have been reported, consistent with decades of psychoneuroimmunology research in patients with anxiety disorders, depression, and traumatic stress鈥搑elated disorders.^27,48 Complex interactions between preexisting psychiatric disease, psychiatric medications, stress, and SARS-CoV-2 infection effects on inflammation and neuronal function have been highlighted as needing further study.⁴⁹ For example, treatment with serotonin selective reuptake inhibitors, fluoxetine or fluvoxamine in particular, may be acutely protective against COVID-19鈥搑elated mortality, with mechanistic hypotheses including reduced inflammation, decreased platelet aggregation, and functional inhibition of acid sphingomyelinase.⁵⁰

While ongoing research efforts are beginning to help elucidate the COVID-19鈥搑elated biological underpinnings such as immune-inflammatory dysregulation²⁷ across the range of neuropsychiatric PASC symptoms and severity, it should be noted that environmental factors, including stressors, social determinants of health, and resiliency factors, likely also contribute to the aforementioned mechanisms of PASC. For example, pandemic-related stressors have been shown to affect such symptoms as cognition, anxiety, depression, fatigue, and sleep and may play a larger role in generating these symptoms than SARS-CoV-2 infection itself.¹² Social determinants of health, including access to health care resources, endemic discrimination, and education level, may also place vulnerable individuals at risk for post鈥揅OVID-19 neuropsychiatric sequelae.⁵¹

Variable follow-up periods and heterogeneous grouping of neuropsychiatric symptoms and disorders in existing COVID-19 literature make it challenging to estimate the population at risk and the breadth of resources necessary to address current and future burdens of illness. The conflation of subjective symptoms, diagnosed syndromes, and objective testing abnormalities has muddied the understanding of PASC epidemiology. Short clinical observation times related to the rapid emergence of SARS-CoV-2 may in part explain the difficulties in producing a pragmatic working definition of PASC. Iterative refinement of the defining features of PASC are expected as biological mechanisms become better understood.

Even when an association between SARS-CoV-2 and neuropsychiatric disease seems plausible, causality is difficult to verify without pathological evidence or robust biomarker data. While active research is ongoing internationally, there are no diagnostic criteria available to definitively identify SARS-CoV-2 as the underlying etiology of either acute or postacute neuropsychiatric events.

Genetic variants of SARS-CoV-2 may have different neurotropic and secondary effects. Grouping together populations affected by different strains of SARS-CoV-2 may lead to difficulties in estimating risk, causality, and outcome trajectories. Limited global genotyping makes subphenotyping difficult, beyond consideration of data collection time frames and geographical trends.

Most epidemiologic data are not from the entire population at risk and may underestimate rates of neuropsychiatric PASC among populations with limited access to health care. Without a national disease screening system, calculating observed rates in relation to background expected neuropsychiatric event rates following SARS-CoV-2 infection presents a challenge.

While there are multiple cohort studies and some ongoing longitudinal studies of PASC among hospitalized patients, there is a relative dearth of data addressing PASC among those with asymptomatic, mild, or moderate COVID-19. This may be due in part to the logistical challenges of conducting research outside the hospital setting during pandemic lockdowns. Additionally, much of the current PASC literature contains stochastic or cross-sectional data. Repeated observations over time are needed to chart outcome trajectories, understand mechanisms, and identify therapeutic targets. Pre鈥揅OVID-19 cognitive, psychiatric, and functional status is often unknown or unreported, making disentangling relapses or progression of underlying disease processes vs de novo illness challenging.

Health care disparities in the acute and chronic phases of COVID-19 have been well documented among different racial and socioeconomic groups.⁵¹ Factoring pandemic-related stressors and social determinants of health into PASC research is imperative for tailoring population-level interventions.

While some studies have compared COVID-19 patients and those with other infections, few include well-characterized contemporaneous control groups. Age, sex, social determinants of health, and comorbidity-matched control patients who are SARS-CoV-2 negative, as well as patients positive for SARS-CoV-2 who do not have prolonged symptoms, are needed to adequately characterize PASC.

Some but not all studies identify older age as a risk factor for more severe COVID-19 and PASC.¹² However, aside from studies addressing multisystem inflammatory syndrome in children, there is a paucity of data on neuropsychiatric effects of SARS-CoV-2 in children, neonates, and pregnant individuals. The effect of SARS-CoV-2 infection vs pandemic-related factors on the developing brain may not be fully understood for several years.

Understanding the potential role of interventions for COVID-19 as well as new or preexisting neurologic and psychiatric treatments is also needed to fully recognize the contributors to neuropsychiatric PASC.

The RECOVER initiative is a NIH-funded multipronged approach to understanding PASC, which will address many of the aforementioned knowledge gaps. RECOVER includes clinical cohort studies spanning the age spectrum (adult, pediatric, pregnant populations), pathology studies, and big data studies that draw from electronic medical records. The adult clinical cohort study assesses disorders affecting multiple organ systems and is an ambidirectional, longitudinal meta-cohort study combining retrospective and prospective data with nested case-control studies. The study will enroll 15鈥�000 individuals with cases that meet WHO criteria for suspected, probable, or confirmed SARS-CoV-2 infection on or after March 1, 2020, as well as 2680 uninfected control patients recruited from inpatient, outpatient, and community-based settings, assuming a 25% rate of PASC among cases. The primary aims of RECOVER are to (1) characterize the incidence and prevalence of long-term sequelae of COVID-19 and describe subphenotypes, (2) characterize the clinical course and recovery trajectories of PASC and identify risk factors, and (3) identify mechanisms and pathophysiology leading to PASC and potential modifiers.

RECOVER takes a 3-tiered approach to prospective data collection, with assessments from early tiers setting gateways for later-tier testing across a broad range of potentially affected organ systems (eg, cardiology, respiratory, neurologic, and psychiatric¹). An estimated 30% of patients in tier 1 will go on to tier 2, and 20% will go on to tier 3 testing for any given symptom. For example, a participant with a positive tier 1 self-report screening measure for a psychiatric condition such as depression, generalized anxiety disorder, or PTSD is referred to tier 2 structured clinical interview, while a subset who complete tier 2 with confirmed psychiatric diagnoses are referred to magnetic resonance imaging, which is tier 3 (Table 2). A proportion of uninfected participants and infected participants without relevant symptoms will also complete tiers 2 and 3 as part of a control group. RECOVER does not set a time frame from SARS-CoV-2 infection to symptom onset, nor a symptom duration to qualify as PASC, because a primary aim is to determine a data-based, practical definition using comparator groups. RECOVER data will be accessible for researchers interested in conducting ancillary studies, and RECOVER is expected to help identify targets for future interventional trials.

Growing data support a high prevalence of PASC neuropsychiatric symptoms, but the current literature is heterogeneous with variable assessments of critical epidemiological factors. By enrolling large, diverse patient samples across the age spectrum with varying levels of illness exposure and experiences, and conducting state-of-the-art assessments, RECOVER will help clarify PASC epidemiology, pathophysiology, and mechanisms of injury, as well as identify targets for therapeutic intervention.

Accepted for Publication: May 2, 2022.

Published Online: June 29, 2022. doi:10.1001/jamapsychiatry.2022.1616

Corresponding Author: Naomi M. Simon, MD, MSc, Department of Psychiatry, NYU Grossman School of Medicine, One Park Avenue, 8th Floor, New York, NY 10016 (naomi.simon@nyulangone.org).

Author Contributions: Drs Frontera and Simon had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Both authors.

Acquisition, analysis, or interpretation of data: Both authors.

Drafting of the manuscript: Both authors.

Critical revision of the manuscript for important intellectual content: Both authors.

Administrative, technical, or material support: Frontera.

Conflict of Interest Disclosures: Dr Frontera reported grants from the National Institutes of Health (NIH), consulting support from FirstKind and Braincool, publishing fees from Thieme, and speaking fees from Physician Education Resource (PER) outside the submitted work. Dr Simon reported grants from the NIH during the conduct of the study; grants from the NIH, Patient-Centered Outcomes Research Institute, Department of Defense, American Foundation for Suicide Prevention, Cohen Veterans Network, and Vanda Pharmaceuticals; consulting fees from Vanda Pharmaceuticals, Bionomics Limited, BehavR LLC, Praxis Therapeutics, Cerevel, Genomind, and Engrail Therapeutics; spousal equity in G1 Therapeutics and Zentalis; and royalties or fees from APA Publishing, Wolters Kluwer (UpToDate), and Wiley (Deputy Editor, Depression and Anxiety) outside the submitted work. No other disclosures were reported.

Funding/Support: Both authors receive funding as co-investigators/faculty at the NYU Grossman School of Medicine for the RECOVER initiative (OTA-21-015A Post-Acute Sequelae of SARS-CoV-2 Infection Initiative: NYU Langone Health Clinical Science Core). Dr Frontera receives funding for the following COVID-19鈥搑elated grants from these NIH institutes: National Institute of Neurological Disorders and Stroke (3U24NS11384401S1), National Heart, Lung, and Blood Institute (1OT2HL161847-01), and National Institute on Aging (3P30AG066512-01).

Role of the Funder/Sponsor: The funding organization had no role in the preparation, review, or approval of the manuscript or decision to submit the manuscript for publication.

Additional Information: There was no original data collection or analyses for this Special Communication.