SARS-CoV-2 in the Central Nervous System: An Exhaustive Review of Postmortem Evidence

Autopsies confirm SARS-CoV-2 RNA in the brain, but damage stems mainly from indirect effects—hypoxia, clotting, and immune overdrive—not direct viral destruction.

Section 1: Executive Summary: SARS-CoV-2 and the Central Nervous System—An Autopsy-Based Inquiry

1.1 Direct Confirmation and Core Complexity

Autopsy studies of individuals who succumbed to COVID-19 have unequivocally confirmed the presence of SARS-CoV-2 viral ribonucleic acid (RNA), including messenger RNA (mRNA), within the tissues of the central nervous system (CNS).¹ This fundamental finding, established through highly sensitive molecular techniques, directly answers the primary question of whether the virus is capable of neuroinvasion. However, this confirmation serves not as a conclusion but as the gateway to a landscape of profound scientific complexity. The body of evidence accumulated from postmortem examinations worldwide reveals a highly nuanced and often paradoxical picture. The detection of viral RNA in the brain is inconsistent across different studies and patient cohorts; its quantity is frequently low, sometimes near the limit of detection; and most critically, its presence does not consistently correlate with the type or severity of neuropathological damage observed.⁴ This disconnect between the viral footprint and the resulting brain injury is a central theme that challenges simplistic models of neurovirulence and has redirected scientific inquiry towards more intricate pathogenic mechanisms.

1.2 The "Indirect Injury" Paradigm

A prevailing paradigm has emerged from the synthesis of global autopsy data: the most severe and frequent forms of brain damage observed in fatal COVID-19 cases appear to be indirect consequences of the systemic illness rather than the direct result of viral replication causing lytic destruction of brain cells.⁴ The neuropathological landscape is dominated by a triad of systemic insults. First, profound hypoxic-ischemic injury, a direct consequence of the severe acute respiratory distress syndrome (ARDS) that characterizes critical COVID-19, leads to widespread neuronal damage, particularly in vulnerable watershed vascular territories.⁶ Second, a virus-induced hypercoagulable state frequently results in microthrombosis and both microscopic and macroscopic cerebrovascular events, including ischemic and hemorrhagic strokes.⁷ Third, a dysregulated systemic immune response, often termed a "cytokine storm," can breach or sensitize the blood-brain barrier (BBB), inciting a state of potent neuroinflammation, characterized by the activation of resident immune cells like microglia and astrocytes.¹⁰ These indirect mechanisms collectively account for a significant portion of the neurological morbidity and mortality associated with the disease.

1.3 Synopsis of Key Findings and Report Structure

This report provides an exhaustive analysis of the postmortem evidence regarding SARS-CoV-2 in the CNS. It begins by dissecting the methodological approaches used to detect the virus, highlighting how techniques from polymerase chain reaction (PCR) to viral culture provide different levels of evidence for viral activity. The analysis then maps the virus's anatomical and cellular footprint within the CNS, revealing a surprising tropism for glial cells, particularly astrocytes, over neurons.² This finding is central to understanding the indirect pathways of neuronal injury. The report critically evaluates the contentious issue of whether the virus actively replicates in the brain or merely persists as residual genetic material, a distinction with profound implications for understanding both acute encephalitis and the chronic symptoms of Post-Acute Sequelae of COVID-19 (PASC), or Long COVID.¹ A detailed examination of the neuropathological changes follows, underscoring the frequent disconnect between viral load and tissue damage. The report also addresses the evolving nature of the pathogen, considering how different viral variants may possess distinct neurotropic capabilities.¹⁴ Finally, the synthesis of these findings is used to frame the key unresolved questions and future research imperatives, particularly in the context of elucidating the biological underpinnings of Long COVID and informing the development of targeted therapeutic strategies.

Section 2: The Methodological Landscape of Postmortem Neurovirology

2.1 A Spectrum of Detection: From RNA Fragments to Viable Virus

The investigation of SARS-CoV-2 within the CNS relies on a suite of sophisticated techniques, each providing a different piece of the puzzle. Understanding the capabilities and limitations of these methods is paramount to correctly interpreting the often-divergent findings in the literature.

Polymerase Chain Reaction (PCR): Reverse transcription polymerase chain reaction (RT-PCR) and its more quantitative counterpart, droplet digital PCR (ddPCR), represent the most sensitive methods for detecting viral genetic material.⁴ These techniques can amplify even minute fragments of viral RNA from tissue samples, providing a definitive "yes" or "no" answer to the question of whether viral genetic material has reached the brain. The exceptional sensitivity of ddPCR has enabled studies to detect SARS-CoV-2 RNA in a high percentage of cases, sometimes in all brain samples from a cohort, even when other detection methods yield negative results.⁴ However, a positive PCR result alone does not distinguish between RNA from a viable, replicating virus and residual, non-infectious genetic fragments left over from a cleared or abortive infection.¹⁷

In Situ Hybridization (ISH): ISH addresses a critical question that PCR cannot: where in the tissue the viral RNA is located. By using labeled probes that bind to viral RNA sequences, ISH allows for the visualization of the virus within the intact cellular architecture of the brain.³ This is crucial for determining cellular tropism—that is, identifying which specific cell types (e.g., neurons, astrocytes, microglia, endothelial cells) are harboring the virus.³ Advanced ISH techniques can also employ probes targeting the viral antisense RNA strand, which is an intermediate produced only during active viral replication. Detection of this antisense strand, or subgenomic RNA (sgRNA), provides much stronger evidence of ongoing viral replication than the detection of genomic RNA alone.¹⁶

Immunohistochemistry (IHC): IHC uses antibodies to detect specific viral proteins, such as the spike (S) or nucleocapsid (N) proteins, within tissue sections.⁴ This method answers the question of whether viral components are being actively produced by the host cell machinery. Because viral proteins are generally less stable and have a shorter half-life than RNA fragments, the presence of these proteins is a stronger indicator of a recent or ongoing infection compared to a standalone positive PCR test.⁴

Viral Culture/Isolation: The isolation of live, replication-competent virus from brain tissue is the undisputed gold standard for proving active and productive CNS infection.¹ This involves taking a tissue sample, processing it, and applying it to a culture of susceptible cells (e.g., Vero E6 cells) to see if the virus can infect these cells and produce new viral particles.¹⁶ While this provides definitive proof, it is technically challenging, and its success is rare in postmortem studies. Positive viral cultures have been reported but are generally limited to patients who died early in the disease course and from laboratories with specialized protocols.¹

2.2 The Critical Role of Autopsy Protocols

The variability in findings across autopsy studies is not solely due to biological differences between patients but is also heavily influenced by the methodologies of tissue collection and preservation.

Post-mortem Interval (PMI): The time elapsed between death and autopsy is a critical variable. RNA and proteins are labile and begin to degrade rapidly after death due to enzymatic processes. Studies with a short PMI are significantly more likely to detect these fragile molecules. The landmark NIH study that successfully detected persistent RNA and even isolated viable virus made a specific point of focusing on short PMIs, which enhanced their ability to detect and quantify the virus with high sensitivity.¹ Conversely, studies with longer PMIs may report lower detection rates, not because the virus was absent, but because the molecular evidence had degraded beyond the limits of detection.

Tissue Preservation: The method of tissue preservation profoundly impacts the types of analyses that can be performed. Formalin fixation, the standard for traditional histopathology, is excellent for preserving cellular structure but can cross-link and fragment nucleic acids and proteins, making sensitive molecular detection by PCR or ISH more challenging.¹⁵ To overcome this, leading research groups have adopted a comprehensive approach, collecting parallel samples from each brain region. Some tissue is formalin-fixed for histology and IHC, while adjacent tissue is either flash-frozen in liquid nitrogen or preserved in a stabilizing solution like RNAlater.¹ This dual approach allows for the optimal preservation of both morphology and molecular integrity, providing the most complete and reliable dataset from each case. The RECOVER initiative's autopsy study emphasizes this rapid processing and freezing of biospecimens to ensure their utility for future high-level molecular research.²¹

2.3 Interpreting Discrepant Findings: The Viral Debris Hypothesis

One of the most consistent and revealing patterns in the autopsy literature is the frequent discrepancy between results from different detection methods. It is common for studies to report that nearly all brain cases in a cohort are positive for SARS-CoV-2 RNA by PCR, while only a small fraction, or even a single case, shows evidence of viral protein by IHC.⁴ One large study of 41 patients found low levels of viral RNA by qRT-PCR but failed to detect any viral proteins.⁵

This is not merely a technical artifact but a key biological finding. The high sensitivity of PCR allows it to detect even degraded, non-viable RNA fragments, which can linger in tissues long after an active infection has been resolved by the immune system.¹ In contrast, IHC requires the presence of more abundant and structurally intact proteins, which are markers of more recent or ongoing viral gene expression.

This disparity strongly supports a model where, in many deceased patients, the brain contains residual viral material—often referred to as "viral debris" or a "viral footprint"—from a past or abortive infection, rather than being the site of a widespread, productive viral encephalitis at the time of death. This interpretation fundamentally shifts the pathogenic model. Instead of direct, lytic damage from replicating virus being the primary driver of neuropathology, the persistence of these viral antigens could act as a chronic, low-level stimulus for the CNS immune system. This concept of antigen persistence is a central hypothesis for explaining the sustained neuroinflammation and debilitating neurological symptoms, such as brain fog and chronic fatigue, that characterize Long COVID.

Section 3: Mapping the Viral Footprint: Anatomical Distribution and Cellular Tropism

3.1 Anatomical Localization in the CNS

Postmortem examinations have revealed that SARS-CoV-2 does not confine itself to a single region of the central nervous system. Instead, evidence of the virus has been found across a wide anatomical distribution, suggesting multiple potential routes of entry and dissemination. While the respiratory system, particularly the lungs and airways, consistently shows the highest viral burden, viral RNA has been detected throughout the neuroaxis.¹⁵

Key regions where SARS-CoV-2 RNA and/or protein have been identified include structures critical for vital functions, cognition, and sensory processing. The olfactory bulbs, which receive direct input from the nasal epithelium, have been implicated as a primary entry point, consistent with the high prevalence of anosmia (loss of smell) in early waves of the pandemic.⁷ From there, the virus may spread to interconnected areas. The brainstem, which controls fundamental processes like respiration and consciousness, is another frequently cited site of viral detection and associated inflammation.¹¹ A comprehensive study by the National Institutes of Health (NIH) provided detailed mapping in 11 extensively sampled brains, detecting SARS-CoV-2 RNA and protein in the hypothalamus and cerebellum of one patient, and in the cervical spinal cord and basal ganglia of others.¹ Other studies have reported viral RNA in the hippocampus, a region crucial for memory, and various cortical areas.⁷ One study of 35 autopsies found viral RNA in 33% of brain samples and, notably, in 75% of samples from the lamina cribrosa, the thin bone through which the olfactory nerve passes, further supporting the olfactory route of entry.²⁵ The widespread nature of this distribution suggests that both neuronal transport pathways and hematogenous (blood-borne) spread may contribute to CNS invasion.

3.2 Cellular Tropism: Identifying the Viral Targets

Identifying the specific cell types that SARS-CoV-2 infects within the brain is critical to understanding the mechanisms of neurological injury. Advanced techniques combining molecular detection with cellular markers have revealed a surprising and consequential pattern of cellular tropism.

Astrocytes: A robust and growing body of evidence from autopsy studies, as well as from experiments using human brain organoids and cell cultures, points to astrocytes as a principal target for SARS-CoV-2 infection within the brain.² One study found that the vast majority (over 65%) of cells positive for the SARS-CoV-2 spike protein in postmortem brain tissue were astrocytes.² This preferential infection is highly significant. Astrocytes are the most abundant glial cell type in the CNS and perform essential roles, including maintaining the integrity of the blood-brain barrier (BBB), providing metabolic fuel to neurons, regulating synaptic function, and modulating the local immune environment. Infection of these critical support cells can have devastating downstream consequences for neuronal health and overall brain function.

Neurons: The evidence for direct infection of neurons is more contentious and less consistent than for astrocytes. Some studies have reported the detection of viral RNA and protein with a staining pattern consistent with neuronal localization.³ However, several other detailed analyses, including those using sophisticated organoid models, have found that while adjacent astrocytes are robustly infected, there is minimal to no evidence of infection within neurons themselves.¹³ This suggests that much of the neuronal damage and dysfunction seen in COVID-19 may be a secondary, or "non-cell-autonomous," effect. Rather than being directly killed by the virus, neurons may suffer and die due to the loss of metabolic support and the toxic inflammatory environment created by infected astrocytes.²

Vascular and Perivascular Cells: The cells that form and surround the brain's blood vessels are also implicated as viral targets. Viral RNA has been detected in endothelial cells lining the brain's microvasculature, and some studies show evidence of viral replication in these cells.¹⁸ In other autopsy series, viral antigens have been specifically localized to perivascular macrophages, immune cells that reside in the space around blood vessels.⁴ This localization is of paramount importance, as it places the virus directly at the BBB interface. Infection at this site can disrupt the barrier, promote the formation of microthrombi, and serve as a nidus for initiating and propagating neuroinflammation by allowing inflammatory cells and molecules from the blood to enter the brain parenchyma.

3.3 Mechanisms of Cellular Entry: Beyond ACE2

A pivotal discovery in the neuropathology of COVID-19 was the realization that the canonical viral entry receptor, angiotensin-converting enzyme 2 (ACE2), is expressed at very low or even undetectable levels on most brain cells, especially astrocytes and neurons.² This presented a paradox: how does the virus infect brain cells if its primary receptor is absent?

This question spurred research that identified several alternative or co-receptors that mediate viral entry into CNS cells. Strong evidence now points to the roles of Neuropilin-1 (NRP1), Basigin (CD147), and Dipeptidyl Peptidase 4 (DPP4).² Studies have shown that astrocytes express these receptors and that blocking them can inhibit SARS-CoV-2 infection in culture.¹³ This finding is crucial because it provides a molecular mechanism to explain the observed tropism for astrocytes and other CNS cells, resolving the paradox of neuroinvasion in an ACE2-low environment. It demonstrates the virus's adaptability in using different molecular doorways to enter different cell types throughout the body.

3.4 The Astrocyte-Centric Model of Neurological Injury

The preferential tropism of SARS-CoV-2 for astrocytes, rather than neurons, fundamentally reframes the understanding of COVID-19's neurological consequences. Initial hypotheses logically assumed that a neurotropic virus would primarily infect and kill neurons, leading to a classic viral encephalitis. However, the data from multiple advanced studies using co-staining and organoid models have largely refuted this, instead pointing toward an astrocyte-centric model of indirect injury.

The causal chain unfolds as follows:

1. SARS-CoV-2 gains entry to the CNS, likely via olfactory or hematogenous routes.

2. Using alternative receptors like NRP1 and CD147, the virus preferentially infects astrocytes.¹³

3. The infection triggers a response in the astrocytes, causing them to become "reactive." This reactive state involves profound changes in their gene expression and metabolism.²

4. These dysfunctional astrocytes fail in their supportive roles. They may reduce the production of essential metabolites, like lactate, needed to fuel neurons.²

5. Simultaneously, the infected astrocytes begin to secrete a cocktail of inflammatory and potentially neurotoxic factors.²

6. Neurons, now starved of metabolic support and bathed in a toxic inflammatory milieu, become dysfunctional and may undergo apoptosis (programmed cell death).

This cascade of "bystander damage" provides a powerful explanation for how significant neurological symptoms, including the cognitive deficits and "brain fog" of Long COVID, can arise even with a low viral load in the brain and minimal direct infection of neurons. This model has critical therapeutic implications, suggesting that strategies aimed at protecting neurons might need to focus on modulating astrocyte function, restoring metabolic balance, and taming neuroinflammation, rather than relying solely on antiviral drugs, especially in the post-acute phase of the disease.

3.5 Summary of Key Autopsy Studies

To synthesize the diverse and evolving data from postmortem studies, the following table summarizes the findings from several key publications. It highlights the differences in patient cohorts, methodologies, and conclusions, providing a panoramic view of the evidence.

Section 4: The Central Question of Replication: Viral Persistence, Abortive Infection, or Active Encephalitis?

The mere presence of viral RNA in the brain, while significant, does not fully describe the virus's activity. A central and contentious question in the neuropathology of COVID-19 is whether SARS-CoV-2 establishes a productive, self-sustaining infection in the CNS, or if its presence is more transient or abortive. The evidence is complex, pointing to a spectrum of possibilities that may depend heavily on the timing of analysis relative to the onset of infection.

4.1 Evidence Supporting Active Replication

Several lines of evidence provide compelling support for the hypothesis that SARS-CoV-2 can, at least in some cases, actively replicate within the brain. The most definitive proof comes from the rare but successful isolation of viable, replication-competent virus from postmortem brain tissue. A landmark NIH study, for instance, cultured live virus from the thalamus of one deceased patient, providing irrefutable evidence that the virus was not only present but infectious.¹ In total, this study successfully isolated viable virus from 45% of non-respiratory specimens tested from patients who died early in their illness, including from the brain.¹

Further strong evidence comes from the detection of subgenomic RNA (sgRNA). These molecules are intermediates produced during the replication of coronaviruses and are not packaged into new viral particles. Their presence is therefore considered a reliable marker of active viral transcription and replication. Multiple studies have detected sgRNA in brain tissue samples from deceased patients, indicating that the viral life cycle was proceeding within the CNS.¹⁶

Finally, histological studies have identified foci of infection and replication, particularly within astrocytes, lending further support to the idea of a productive infection.² Together, these findings demonstrate that SARS-CoV-2 is not merely a passive passenger in the CNS but has the capacity to actively replicate within it.

4.2 Evidence Against Widespread Replication

Juxtaposed with the evidence for replication is a substantial body of work from other comprehensive autopsy series that found little to no evidence of it. Many studies have concluded that there is sparse evidence for robust viral replication in the human brain and have questioned the neurotropic potential of SARS-CoV-2, suggesting that the observed neurological symptoms must arise from mechanisms other than direct CNS infection.⁴

The most powerful argument against widespread, ongoing replication at the time of death is the frequently observed disparity between PCR and IHC results. As previously discussed, the detection of viral RNA by PCR is common, but the detection of viral proteins by IHC is much rarer.⁴ If the virus were actively and widely replicating, one would expect to find abundant viral proteins being produced. The frequent absence of these proteins strongly suggests that in many patients, especially those who survive the initial phase of the illness, viral replication in the brain is either at very low levels, has been effectively suppressed by the immune response, or represents an "abortive" infection where the virus enters cells but fails to complete its replication cycle.

4.3 The Concept of Viral Persistence

These seemingly contradictory findings can be reconciled by a model that incorporates both the timing of infection and the concept of viral persistence. The landmark NIH study that provided some of the strongest evidence for replication also delivered the strongest evidence for long-term persistence, detecting viral RNA in the brain up to 230 days after the initial onset of symptoms.¹

This suggests a dynamic process. It is plausible that active viral replication in the CNS occurs primarily during the early, acute phase of the infection. This is supported by the observation that viable virus is most successfully isolated from patients who died within the first two weeks of their illness.¹ In patients who survive this acute phase, the immune system may largely clear the productive infection from the brain. However, the clearance may be incomplete. The CNS is considered an immunologically distinct site, and it is possible that viral RNA and proteins can persist in these tissues for months, long after active replication has ceased.

This "hit-and-stay" model posits that the virus enters the CNS and may replicate early on (the "hit"), but what is detected in many autopsy cases, particularly those with a longer disease course, is the lingering footprint of this initial invasion (the "stay"). This persistent presence of viral antigens, even in the absence of replication, could be sufficient to provoke a chronic, low-grade inflammatory response, providing a compelling biological basis for the long-term neurological symptoms seen in PASC.

4.4 The Critical Role of Disease Timing

The debate over replication versus persistence underscores a critical variable: the time of death relative to the onset of infection. The virological and pathological state of the brain in a patient who dies from fulminant COVID-19 after 10 days is likely very different from that of a patient who survives the acute illness but succumbs to complications two months later.

This temporal dynamic has significant therapeutic implications. If active replication is primarily an early event, then antiviral therapies, such as Paxlovid, would likely have the greatest potential to mitigate neurological damage when administered during the acute phase of COVID-19. For patients with Long COVID, whose symptoms persist for months or years, therapeutic strategies may need to shift away from targeting active replication and toward resolving the consequences of viral persistence, such as chronic neuroinflammation, autoimmune dysregulation, and cellular dysfunction. This highlights the urgent need for autopsy studies specifically on patients with a confirmed diagnosis of Long COVID to determine if and where viral reservoirs persist.²¹

Section 5: The Pathological Aftermath: Correlating Viral Presence with Brain Injury

While detecting the virus is a critical first step, understanding the damage it causes or is associated with is the ultimate goal of neuropathological investigation. Autopsy studies of COVID-19 patients have revealed a consistent, albeit varied, constellation of brain injuries. Strikingly, the most common and severe of these pathologies are often better explained by the systemic effects of the disease than by direct viral action in the brain.

5.1 A Catalogue of Neuropathological Findings

Postmortem examinations have documented a wide spectrum of damage within the CNS of deceased COVID-19 patients. These findings can be broadly categorized into three major groups.

Hypoxic-Ischemic Injury: This is arguably the most frequent and widespread finding reported in autopsy series.⁶ It manifests histologically as eosinophilic or "red" neurons, particularly in brain regions highly vulnerable to oxygen deprivation, such as the hippocampus, cerebellum (Purkinje cells), and watershed zones of the cerebral cortex.⁷ This type of damage is the classic hallmark of brain injury resulting from a lack of oxygen and blood flow, and its high prevalence in COVID-19 autopsies is a direct reflection of the severe ARDS and respiratory failure that often precipitate death in these critically ill patients.⁶

Vascular and Thromboembolic Pathology: A second dominant feature is damage to the brain's vasculature. This ranges from microscopic findings, such as perivascular hemorrhages and microthrombi (tiny blood clots) occluding small vessels, to large, macroscopic ischemic and hemorrhagic strokes.⁷ These vascular insults are a direct neurological consequence of the systemic hypercoagulability and endotheliitis (inflammation of the blood vessel lining) that are well-established hallmarks of severe COVID-19.⁸ The presence of these clots obstructs blood flow, leading to ischemic infarcts (tissue death from lack of blood), while damage to vessel walls can lead to hemorrhages.

Neuroinflammation: The third key finding is a robust inflammatory response within the CNS. This is characterized by the widespread activation of microglia, the brain's resident immune cells, a state known as microgliosis.¹¹ Activated microglia change their shape and can form small clusters or nodules, particularly in the brainstem.¹¹ Alongside microgliosis, there is often reactive astrogliosis, where astrocytes also become activated and proliferate.⁶ Additionally, many cases show perivascular inflammatory infiltrates, where immune cells from the bloodstream, predominantly macrophages and CD8+ T-cells, accumulate in the space around blood vessels.⁹

5.2 The Disconnect: Viral Burden vs. Tissue Damage

A critical and recurring observation in the autopsy literature is the frequent lack of a direct spatial correlation between the presence of the virus and the location of the most severe tissue damage. The comprehensive NIH study, for example, explicitly noted that they observed "little evidence of inflammation or direct viral cytopathology outside the respiratory tract," including in the brain, "despite substantial viral burden".¹ Other studies have similarly reported that the quantity of detectable viral RNA in brain tissue did not correlate with the extent or type of histopathological changes observed.⁵

This disconnect is profoundly important. If the virus were causing brain damage primarily through direct infection and destruction of cells (a cytopathic effect), one would expect to see the most severe injury and inflammation precisely where the virus is most abundant. The frequent absence of this correlation strongly implies that other, indirect mechanisms are the primary drivers of neuropathology in many cases.

5.3 Brain as Victim: The Systemic Disease Fallout

The combination of the specific types of damage observed and the disconnect between viral location and injury severity has led to a unifying hypothesis: the dominant neuropathological signature of fatal COVID-19 is that of systemic disease fallout. The brain, in this model, is often a "victim" of the chaos engulfing the rest of the body, rather than the primary "target" of viral attack.⁷

The evidence for this is compelling. The most common findings—hypoxic injury and thromboembolic events—are directly and mechanistically linked to ARDS and systemic coagulopathy, respectively, which are defining features of severe COVID-19.⁶ Furthermore, the pattern of neuroinflammation, particularly microglial activation, has been shown to be comparable to that seen in patients who die from non-COVID-19 sepsis.²⁴ This suggests that this inflammatory response may be a general reaction of the brain to a state of severe systemic illness and cytokine release, rather than a specific response to the presence of SARS-CoV-2 within the brain parenchyma itself.

This paradigm does not negate the fact that SARS-CoV-2 can and does enter the brain. However, it posits that the most life-threatening and widespread neurological damage in acute, fatal cases is often driven by these powerful indirect effects. This has significant clinical implications, reinforcing the critical importance of aggressively managing systemic complications—such as respiratory failure, hypoxia, and coagulation—as a primary strategy for neuroprotection in patients with severe COVID-19. It also suggests that the neuroinflammatory state in these patients may be triggered by signals originating from the periphery, such as inflammatory cytokines crossing a compromised BBB, as much as, or even more than, by the local presence of the virus.

Section 6: An Evolving Pathogen: Neurotropism and the Impact of SARS-CoV-2 Variants

The neurological risk profile of COVID-19 is not a static entity. SARS-CoV-2 is an RNA virus that continuously evolves, and the emergence of new variants of concern has been a defining feature of the pandemic. Emerging evidence from experimental models and clinical observation suggests that these variants are not uniform in their ability to invade and affect the central nervous system. This evolving neurotropism means that findings from the initial wave of the pandemic may not be fully generalizable to infections caused by later variants, complicating long-term prognostication and public health planning.

6.1 Differential Neurovirulence

Studies utilizing sensitive in vivo and in vitro models have begun to dissect the differential neurovirulence of various SARS-CoV-2 lineages. A key model system has been the K18-hACE2 transgenic mouse, which expresses the human ACE2 receptor and is highly susceptible to SARS-CoV-2, often developing lethal neuroinvasion. Research using this model has yielded striking results. While the early Omicron BA.1 variant was found to be attenuated, causing less severe disease and lower mortality in these mice compared to ancestral strains, subsequent Omicron sublineages demonstrated a concerning reversal of this trend.¹⁴

Specifically, studies have shown that the Omicron BA.5 and XBB variants are significantly more pathogenic in K18-hACE2 mice than BA.1. These later variants caused fulminant brain infection and high rates of mortality, a neurovirulent phenotype reminiscent of the original ancestral strains.¹⁴ This suggests that while the Omicron lineage as a whole may have adapted to be less pathogenic in the lungs, certain subvariants may have reacquired or even enhanced their potential for neuroinvasion and lethal brain infection in this specific model system.

6.2 Changes in Cellular Tropism and Entry Mechanisms

The molecular basis for these differences in neurovirulence appears to lie, at least in part, in changes to the viral spike protein, which dictates cellular tropism and the efficiency of viral entry. For example, the attenuated phenotype of the Omicron BA.1 variant was linked to less efficient cleavage of its spike protein at the S1/S2 site, which leads to a shift in its preferred method of cellular entry. It became less reliant on the cell surface protease TMPRSS2 and more dependent on endosomal entry pathways, a change that may have reduced its ability to efficiently infect certain cell types.²⁹

Conversely, studies comparing the original Wuhan-Hu-1 (WA1) strain with the Delta and Omicron variants in hamster models revealed distinct patterns of tropism within the nasal cavity. While the WA1 and Delta variants readily infected the olfactory neuroepithelium—the layer of neurons responsible for smell—the Omicron variant showed a marked transition in tropism away from these neurons and towards the respiratory epithelium.³³ This finding from animal models aligns well with clinical observations that anosmia was a much more common and pronounced symptom of infection with earlier variants compared to Omicron. Furthermore, the Delta variant demonstrated a greater propensity to infect deeper, submucosal cell types, such as the mucus-producing Bowman's glands, which may have contributed to more severe local tissue damage.³³

6.3 Implications of an Evolving Neurological Profile

The evidence that different SARS-CoV-2 variants possess distinct neurotropic and neurovirulent properties has several critical implications. Firstly, it underscores that COVID-19 is not a single, monolithic disease in terms of its neurological impact. The spectrum of potential neurological complications, from anosmia to severe encephalitis, may shift with the dominant circulating variant.

Secondly, it highlights the limitations of generalizing findings from one phase of the pandemic to another. Autopsy studies from the first wave, which predominantly involved unvaccinated individuals infected with ancestral strains, provide an invaluable baseline but may not fully capture the neuropathology associated with later variants in a largely vaccinated or previously exposed population.

Finally, this evolving threat necessitates continuous, variant-specific research and surveillance. Public health agencies and clinicians must remain adaptable and vigilant, recognizing that the neurological risk profile of COVID-19 can change. Understanding the molecular mechanisms behind these shifts in tropism and virulence is essential for predicting the potential impact of future variants and for developing variant-proof therapeutic and preventive strategies.

Section 7: Clinical Horizons: Implications for Long COVID and Future Research Imperatives

The wealth of data from postmortem studies provides more than just a grim catalogue of end-stage disease; it offers crucial insights into the potential biological mechanisms driving the chronic and often debilitating symptoms of Long COVID. By bridging the gap between acute pathology and chronic illness, these findings are helping to frame the most urgent questions for future research and inform the development of rational therapeutic strategies.

7.1 A Pathological Basis for Long COVID

The neuropathological findings in acute fatal COVID-19 align remarkably well with the leading hypotheses proposed to explain the neurological and cognitive symptoms of Long COVID, such as brain fog, memory problems, chronic fatigue, and dysautonomia.

Viral Persistence: The demonstration that SARS-CoV-2 RNA and proteins can persist in the brain and other tissues for months after the acute infection provides a direct and tangible basis for the "viral reservoir" hypothesis of Long COVID.¹ This persistent presence of viral antigens, even if non-replicating, could act as a chronic source of immune stimulation, preventing the body's systems from returning to homeostasis. The RECOVER initiative's autopsy study is specifically designed to investigate this, aiming to pinpoint where in the body these reservoirs exist and whether the virus within them is active or dormant.²¹

Chronic Neuroinflammation: The widespread microglial activation and astrogliosis seen in acute COVID-19 autopsies are the pathological signature of neuroinflammation.¹¹ If this inflammatory state is not fully resolved, it could persist for months or years, becoming a chronic condition. Persistent microglial activation is known to disrupt synaptic function, impair adult neurogenesis in critical areas like the hippocampus, and contribute to the neuronal loss and cognitive decline that are central to the Long COVID experience.³²

Endothelial and Vascular Dysfunction: The prominent vascular damage, microthrombosis, and endotheliitis observed at autopsy provide a clear structural foundation for the vascular disruption hypothesis of Long COVID.⁷ Persistent inflammation and damage to the brain's microvasculature could lead to chronic hypoperfusion (reduced blood flow), BBB leakiness, and ongoing endothelial dysfunction, all of which have been linked to cognitive deficits and other neurological symptoms in living patients.³⁴

7.2 Unresolved Questions and Future Research Directions

Despite the progress made, autopsy studies have illuminated several critical gaps in our knowledge that must be addressed by future research.

The CNS as a True Reservoir: A paramount question is whether the CNS can serve as a true, long-term reservoir for SARS-CoV-2. Answering this requires a concerted effort to perform rapid-response autopsies on well-characterized patients who had been suffering from Long COVID.²¹ Such studies are essential to determine if persistent virus in the brain is a common feature of neurological PASC and if it is associated with specific pathological changes.

Impact of Vaccination, Variants, and Reinfection: The vast majority of early, detailed autopsy reports are from unvaccinated individuals infected with ancestral strains of the virus.¹ It is crucial to understand how prior vaccination, infection with newer variants, and multiple reinfections alter the landscape of neuropathology.³⁷ This requires ongoing autopsy programs that can correlate pathological findings with detailed clinical and virological histories.

Application of Advanced Methodologies: The future of postmortem research lies in the application of cutting-edge molecular techniques. Single-cell and spatial transcriptomics, which can map gene expression changes in individual cells while preserving their anatomical context, will be transformative.³⁹ These methods can move beyond simple histology to define the precise molecular pathways that are dysregulated in specific cell populations (e.g., astrocytes, microglia, endothelial cells) in the COVID-19 brain, revealing novel targets for therapy.

Long-Term Neurodegenerative Risk: A profound and unsettling question is whether the neuroinflammation and neuronal injury precipitated by COVID-19 can act as an initiator or accelerator of long-term neurodegenerative diseases, such as Alzheimer's disease and Parkinson's disease.⁵ Answering this will require decades of longitudinal follow-up of COVID-19 survivors, correlated with eventual autopsy studies to search for the tell-tale protein aggregates (e.g., amyloid-beta, tau, alpha-synuclein) that define these conditions.

7.3 Informing Therapeutic Strategies

Ultimately, the value of autopsy research lies in its ability to guide the development of effective treatments. The findings from COVID-19 postmortem studies point toward several promising therapeutic avenues for neurological Long COVID.

The strong evidence for chronic neuroinflammation and immune dysregulation provides a solid rationale for clinical trials of anti-inflammatory and immunomodulatory agents.⁴² Therapies that can specifically target and quell the activation of microglia and astrocytes could be particularly beneficial.

The viral persistence hypothesis, if further substantiated by studies of Long COVID patients, supports the investigation of antiviral agents.⁴³ Even if replication is low-level, eliminating the source of persistent antigens could allow the immune system to reset.

The astrocyte-centric model of injury suggests that neuroprotective agents or therapies aimed at restoring normal astrocyte metabolic function could help shield neurons from bystander damage.²⁷ Finally, novel mechanistic insights, such as the discovery that the SARS-CoV-2 nucleoprotein may disrupt neurotransmission by sequestering neuropeptides, have opened the door to entirely new therapeutic strategies. This finding provides a potential mechanism of action for drugs like low-dose naltrexone, which have shown some anecdotal and clinical benefit in Long COVID, by suggesting they may work by restoring balance to these disrupted neuropeptide systems.⁴³

In conclusion, the postmortem examination of the brain in COVID-19 has been instrumental in revealing the complex interplay between direct viral invasion and indirect systemic injury. While SARS-CoV-2 RNA is frequently found in the CNS, the resulting neuropathology is multifaceted and often dominated by the secondary effects of hypoxia, coagulopathy, and neuroinflammation. These findings have not only reshaped our understanding of the acute disease but are now providing a critical foundation for deciphering the enigmatic and persistent neurological consequences of Long COVID, paving the way for the next generation of research and therapeutic innovation.

Acknowledgement

I acknowledge the assistance of Gemini AI in the preparation of the subject research plan, the execution of the research, and the preparation of this report.

Works cited

1. Autopsies show COVID-19 virus in brain, elsewhere in body - CIDRAP, accessed August 10, 2025, https://www.cidrap.umn.edu/covid-19/autopsies-show-covid-19-virus-brain-elsewhere-body

2. Morphological, cellular, and molecular basis of brain infection in COVID-19 patients | PNAS, accessed August 10, 2025, https://www.pnas.org/doi/10.1073/pnas.2200960119

3. SARS-CoV-2 infection and persistence in the human body and brain ..., accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC9749650/

4. SARS-Cov-2 infection and neuropathological findings: a report of 18 cases and review of the literature - PubMed Central, accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC10170054/

5. COVID-19 and Vaccination in the Setting of Neurologic Disease ..., accessed August 10, 2025, https://www.neurology.org/doi/10.1212/WNL.0000000000012578

6. SARS‐CoV‐2 and the brain: A review of the current knowledge on neuropathology in COVID‐19 - PMC - PubMed Central, accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC8420197/

7. COVID-19 and the Brain: The Neuropathological Italian Experience ..., accessed August 10, 2025, https://www.mdpi.com/2218-273X/12/5/629

8. Hemorrhagic and thrombotic manifestations in the central nervous system in COVID-19: A large observational study in the Brazilian Amazon with a complete autopsy series | PLOS One - Research journals, accessed August 10, 2025, https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0255950

9. Postmortem neuropathology in COVID‐19: An update - PMC - PubMed Central, accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC10579998/

10. What Can We Still Learn from Brain Autopsies in COVID-19? - PubMed, accessed August 10, 2025, https://pubmed.ncbi.nlm.nih.gov/37023787/

11. Neuropathology in COVID-19 autopsies is defined by microglial activation and lesions of the white matter with emphasis in cerebellar and brain stem areas - PubMed Central, accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC10374362/

12. Immune-Mediated Mechanisms of COVID-19 Neuropathology - Frontiers, accessed August 10, 2025, https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2022.882905/full

13. Tropism of SARS-CoV-2 for human cortical astrocytes | PNAS, accessed August 10, 2025, https://www.pnas.org/doi/10.1073/pnas.2122236119

14. SARS-CoV-2 omicron BA.5 and XBB variants have ... - Frontiers, accessed August 10, 2025, https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2023.1320856/full

15. Large COVID autopsy study finds SARS-CoV-2 all over the human body - New Atlas, accessed August 10, 2025, https://newatlas.com/science/covid-autopsy-study-virus-brain-body/

16. Study shows SARS-CoV-2 infection, replication and persistence in ..., accessed August 10, 2025, https://www.news-medical.net/news/20221216/Study-shows-SARS-CoV-2-infection-replication-and-persistence-in-human-brain-tissues.aspx

17. SARS-CoV-2, SARS-CoV-1 and MERS-CoV viral load dynamics, duration of viral shedding and infectiousness – a living systematic review and meta-analysis, accessed August 10, 2025, https://www.periodicos.capes.gov.br/index.php/acervo/buscador.html?task=detalhes&id=W3045863127

18. Multisystemic cellular tropism of SARS-CoV-2 in autopsies of COVID-19 patients, accessed August 10, 2025, https://sciety-labs.elifesciences.org/articles/by?article_doi=10.1101/2021.06.03.21258241

19. Autopsies Show COVID-19 Virus in Brain and Elsewhere in Body - Neuroscience News, accessed August 10, 2025, https://neurosciencenews.com/covid19-brain-tissue-22153/

20. When Two Worlds Collide: The Contribution and Association Between Genetics (APOEε4) and Neuroinflammation (IL-1β) in Alzheimer's Neuropathogenesis - MDPI, accessed August 10, 2025, https://www.mdpi.com/2073-4409/14/15/1216

21. RECOVER's tissue pathology (autopsy) study makes unique ..., accessed August 10, 2025, https://recovercovid.org/news/recovers-tissue-pathology-autopsy-study-makes-unique-contributions-long-covid-research

22. SARS-CoV-2 and the Nervous System: From Clinical Features to Molecular Mechanisms, accessed August 10, 2025, https://www.mdpi.com/1422-0067/21/15/5475

23. Potential mechanism of SARS‐CoV‐2‐associated central and ..., accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC9349396/

24. Neuropathological findings in COVID-19: an autopsy cohort - Oxford Academic, accessed August 10, 2025, https://academic.oup.com/jnen/article/82/1/21/6820985

25. Long-Term Persisting SARS-CoV-2 RNA and Pathological Findings: Lessons Learnt From a Series of 35 COVID-19 Autopsies, accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC8865372/

26. Neuroinvasion and Neurotropism by SARS-CoV-2 Variants in the K18-hACE2 Mouse, accessed August 10, 2025, https://www.mdpi.com/1999-4915/14/5/1020

27. In SARS-CoV-2, astrocytes are in it for the long haul | PNAS, accessed August 10, 2025, https://www.pnas.org/doi/10.1073/pnas.2209130119

28. Multisystemic cellular tropism of SARS-CoV-2 in autopsies of COVID-19 patients | medRxiv, accessed August 10, 2025, https://www.medrxiv.org/content/10.1101/2021.06.03.21258241v1.full-text

29. Do SARS-CoV-2 Variants Differ in Their Neuropathogenicity? | mBio - ASM Journals, accessed August 10, 2025, https://journals.asm.org/doi/10.1128/mbio.02920-22

30. Postmortem brain 7T MRI with minimally invasive pathological correlation in deceased COVID-19 subjects - PubMed Central, accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC8760871/

31. Neurovascular injury with complement activation and inflammation in COVID-19 | Brain, accessed August 10, 2025, https://academic.oup.com/brain/article/145/7/2555/6621999

32. Coronavirus Does Not Infect the Brain But Still Inflicts Damage, accessed August 10, 2025, https://www.cuimc.columbia.edu/news/coronavirus-does-not-infect-brain-still-inflicts-damage

33. Evolution of nasal and olfactory infection characteristics of SARS-CoV-2 variants - JCI, accessed August 10, 2025, https://www.jci.org/articles/view/174439

34. Neurological sequelae of long COVID: a comprehensive review of diagnostic imaging, underlying mechanisms, and potential therapeutics - Frontiers, accessed August 10, 2025, https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2024.1465787/full

35. SARS-CoV-2-Mediated Neuropathogenesis, Deterioration of ..., accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC10581113/

36. Development of a long Covid postmortem tissue bank with pathological & molecular analyses - PolyBio Research Foundation, accessed August 10, 2025, https://polybio.org/projects/development-of-a-long-covid-postmortem-tissue-bank-with-pathological-molecular-analyses/

37. Neuropathology of COVID-19 (neuro-COVID): clinicopathological update - PMC, accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC7861505/

38. Long COVID or Post-COVID-19 Condition: Past, Present and Future ..., accessed August 10, 2025, https://www.mdpi.com/2076-2607/11/12/2959

39. SARS-CoV-2 promotes microglial synapse elimination in human brain organoids - PMC, accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC9533278/

40. Single-cell sequencing of brain tissues reveal the central ... - Frontiers, accessed August 10, 2025, https://www.frontiersin.org/journals/pharmacology/articles/10.3389/fphar.2022.971017/full

41. Characterizing cell type specific transcriptional differences between the living and postmortem human brain | medRxiv, accessed August 10, 2025, https://www.medrxiv.org/content/10.1101/2024.05.01.24306590v1.full

42. Neurological sequelae of long COVID: a comprehensive review of diagnostic imaging, underlying mechanisms, and potential therapeutics - PubMed Central, accessed August 10, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC11881597/

43. Long Neuro-COVID-19: Current Mechanistic Views and Therapeutic ..., accessed August 10, 2025, https://www.mdpi.com/2218-273X/14/9/1081

44. Strategies to Manage Post-COVID Neurological Symptoms - Michigan Neurology Associates, accessed August 10, 2025, https://www.michiganneurologyassociates.com/blog/strategies-to-manage-post-covid-neurological-symptoms