Cognitive dysfunction in post-COVID-19 condition: Mechanisms, management, and rehabilitation

Marika Möller, Kristian Borg, Christer Janson, Maria Lerm, Johan Normark, Katarina NiwardSeptember 2023, Journal of Internal Medicine Volume 294, Issue 5 p. 563-581

Abstract

The long-term effects of COVID-19 on cognitive function have become an area of increasing concern. This paper provides an overview of characteristics, risk factors, possible mechanisms, and management strategies for cognitive dysfunction in post-COVID-19 condition (PCC).

Prolonged cognitive dysfunction is one of the most common impairments in PCC, affecting between 17% and 28% of the individuals more than 12 weeks after the infection and persisting in some cases for several years. Cognitive dysfunctions can be manifested as a wide range of symptoms including memory impairment, attention deficit, executive dysfunction, and reduced processing speed. Risk factors for developing PCC, with or without cognitive impairments, include advanced age, preexisting medical conditions, and the severity of acute illness. The underlying mechanisms remain unclear, but proposed contributors include neuroinflammation, hypoxia, vascular damage, and latent virus reactivation not excluding the possibility of direct viral invasion of the central nervous system, illustrating complex viral pathology.

Introduction

Lingering symptoms after COVID-19 are not uncommon, but symptoms with significant impact on quality of life and functioning following an acute SARS-CoV-2 infection affect a minority of all COVID-19 cases [1–4]. However, as there are more than half a billion cases of COVID-19, the coronavirus pandemic has led to a huge influx of people experiencing persistent symptoms in need of acute care and rehabilitation. Furthermore, vaccination prior to SARS-CoV-2 infection is not a fully reliable protector against post-COVID-19 condition (PCC) [3, 5]. According to the World Health Organization (WHO), PCC is a condition that occurs in adult individuals with a history of probable or confirmed SARS-CoV-2 infection, usually 3 months from the onset of COVID-19 with symptoms that last for at least 2 months and are not explained by an alternative diagnosis. Thus, the key when diagnosing PCC is to exclude other relevant differential diagnoses [6]. A similar definition was presented by WHO in February 2023 for PCC in children and adolescents, with some differences in the clinical picture compared with adults [7]. Studies on and descriptions of PCC uncover a wide spectrum of clinical manifestations involving different organ systems—including but not limited to severe fatigue, shortness of breath, cognitive dysfunction, palpitations, chest pain, prolonged fever, anosmia or ageusia, vertigo and tinnitus, headache, peripheral neuropathy, skin rash, and joint pain or swelling [4, 8, 9]. Symptoms included in PCC may appear in clusters or be monosymptomatic. Furthermore, they may fluctuate or relapse over time and often have an impact on everyday functioning [6]. For most people suffering from PCC, the symptoms will gradually resolve over time, but for some, they will linger on for many months and even years after the initial infection [1, 10–14].

Collaborative research to deepen the understanding of risk factors, mechanisms, and prognosis of PCC is urgently needed as well as to promote the development of preventive and therapeutic strategies. Today, progress has been made to reduce the risk of developing PCC, but specific medical treatment is still lagging. This review will focus on cognitive aspects of PCC in patients with an initial non-critical SARS-CoV-2-infection, and cover characteristics, risk factors, possible mechanisms and provide an approach to management and rehabilitation. Physiological aspects of PCC are covered in a recent article by Fedorowicz et al. [15]. For those experiencing long-term post-COVID-19 symptoms, the possible explanatory mechanisms may not be the same as in the critically ill patients treated in mechanical ventilation, and this review will not discuss post-intensive care syndrome in COVID-19 patients.

Characteristics of cognitive symptoms in PCC

The most frequently reported symptoms in PCC are fatigue and cognitive symptoms of “brain fog,” memory problems, decreased concentration and impaired attention [16–18]. In a large meta-analysis on symptom reports more than 12 weeks following COVID-19 infection, the pooled proportion of individuals experiencing cognitive impairment was 22% (95% Confidence Interval [CI]: 17%–28%) [19]. Another meta-analysis on symptom reports more than 1 year after diagnosis, reported a pooled prevalence of 19% regarding difficulties with memory and concentration [20]. Substantial heterogeneity across studies has been observed in the aforementioned studies. The cognitive symptoms in PCC often fluctuate and, in some cases, worsen over time [21]. The term “brain fog” is not a medical condition and lacks a scientific definition. Rather it is an umbrella term used to describe the difficulty in thinking clearly, which is commonly seen in various disorders that affect processing speed and attention functions. Patients experiencing brain fog often describe themselves as disorganized or confused, having trouble focusing, or struggling finding words [22]. Comorbid conditions are common, including sleep disturbances, anxiety, and depression [16, 17, 19].

In a comprehensive analysis of retrospective propensity-score matched cohort studies spanning a 2-year period involving 1,284,437 patients diagnosed with SARS-CoV-2, Taquet et al. discovered that the risks of mood and anxiety disorders diminished after 1–2 months and did not exhibit a notable increase after 2 years. However, the risks of cognitive deficit, dementia, psychotic disorders, and seizures remained elevated even 2 years after the infection. Additionally, Taquet et al. observed no differences in risks of neurological and psychiatric outcomes between infections caused by the Omicron (B.1.1.529) or Delta (B.1.617.2) variants at the 3-month follow up. The study included both hospitalized and non-hospitalized patients, although the results were not stratified by severity of illness [11]. In a cohort consisting of 65 COVID-19 patients treated in the intensive care unit (ICU) at Danderyd University Hospital in Stockholm, we discovered no major disparities in cognitive outcome between patients treated during the first pandemic wave compared to those treated in subsequent waves, despite the longer duration of ICU stay during the first wave (26 vs. 9 days) (Möller 2023, personal communication). In a prospective multicenter cohort study in Sweden [14], we found a high prevalence of persistent cognitive symptoms 6 months after infection both in non-hospitalized (168/283, 59%) and hospitalized patients (119/151, 79%). In a long-term follow-up population-based study of 165 patients diagnosed with PCC at 4 months post-discharge from any hospital in Region Östergötland in Sweden [10], we found that the majority of patients (139/165, 84%) reported persisting problems affecting everyday life at the 2-year follow-up, and approximately half of the patients who were on sick leave related to PCC at 4 months after infection were still on sick leave. The most common persisting symptoms were cognitive, sensorimotor, and fatigue symptoms, and we observed no difference between individuals treated in the ICU and non-ICU-treated individuals. In patients with initial mild infection, the correspondence between self-rated cognitive symptoms and outcome of neuropsychological assessment is less clear. Some follow-up studies report no significant differences between individuals 4 months after COVID-19 diagnosis and healthy controls [23], whereas studies on symptomatic patients do find differences between those with verified COVID-19 infection and those without in terms of lower performance compared with demographically matched norms on tests of attention and working memory [16]. Larger studies with appropriate control groups are lacking.

It is not clear whether lingering symptoms included in PCC differ from what can occur after other respiratory viral infections. In a large retrospective cohort study of nearly 250,000 survivors of COVID-19, neurological and psychiatric outcomes after 6 months were compared with those of patients with influenza [11]. The risk of intracranial hemorrhage and ischemic stroke was higher in the COVID-19 group, and there was a preliminary association with dementia, anxiety, and mood disorders as well as insomnia. The risk increased with the severity of the COVID-19 infection but was also found in patients who did not require hospitalization. Furthermore, the reported long-term cognitive symptoms and fatigue are not specific for PCC. Rather, they resemble what often can be seen in several other conditions such as mild traumatic brain injury [24], stroke [25], or multiple sclerosis [26]—that is, decreased attention and memory functions, reduced processing speed, and fatigue. The phenotypes of these disabilities are possibly related to the functional anatomy of the brain, although the underlying mechanisms behind different conditions may be different. There is also an overlap with myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) as further presented below. Common to all these conditions is that subjective cognitive complaints may not be associated with objective cognitive test results. Moreover, psychological distress can interfere with self-reported symptoms [16, 27, 28]. The most common cognitive symptoms after COVID-19 are further described later.

Cognitive symptoms

Attention functions enable the ability to process information from our surroundings and are considered to be hierarchical in nature [29]. Focused attention and sustained attention are regarded as fundamental attention functions. Higher levels of attention rely on executive functions and encompass alternating, selective, and divided attention. Despite its name, working memory is also considered a function of attention [29, 30] and plays a significant role in encoding memory and retrieving information stored in long-term memory [31]. Given that attention functions serve as fundamental cognitive processes and subsystems for other cognitive functions, they are crucial for managing our everyday life. Impaired attention functions, even in the cases of mild impairments, directly affect performance in both everyday tasks and work life [32].

Attention functions are frequently associated with processing speed. From a neuroanatomical perspective, the thalamus plays a crucial role as a hub in networks that support processes related to attention, information processing, memory, and executive functions [33]. However, in the studies carried out so far on COVID-19, it has not been determined which of the attention functions is most affected in PCC, nor to what extent other cognitive dysfunctions are independent or linked to decreased attention functions. Impairments in attention, working memory, and executive functions often have secondary effects on tests assessing memory encoding and retrieval. Consequently, episodic memory can indirectly be affected by the type of neurological damage caused by viral diseases, primarily through reductions in attention and processing speed required for encoding [34]. In the context of PCC, it is still not fully established whether the results observed in memory tests stem from primary difficulties in memory storage or if they are secondary effects resulting from impaired attention and/or working memory.

Fatigue

Fatigue is a prominent symptom in both acute COVID-19 and PCC. Prevalence rates for post-COVID fatigue range from 32% to 46% in different studies [17, 19, 35] and in meta-analysis of 1-year follow-ups between 18% and 39% [20]. However, fatigue is a multifactorial and vaguely defined symptom present in various conditions including neurological disorders [36], chronic pain [37], and depression [38]. Post-infectious fatigue has also been reported after other viral epidemics [39]. In most studies, fatigue is subjectively reported using self-assessment scales designed to capture a low energy level not in proportion to the individual’s activity level and that is not alleviated by normal rest or sleep [40, 41]. Currently, there is no validated fatigue scale specifically for post-COVID fatigue. As COVID-19 is a novel condition, it is not evident whether the fatigue experienced in PCC is equivalent to and shares the same underlying mechanisms as fatigue in neurological conditions. In neurological conditions, decreased attention [42], slowed processing speed [43], and fatigability [44] have been linked to the experience of fatigue but have also shown significant correlations with depression [45] and sleep disorders [46]. How eg. sleep disturbances, depression or pain are related to self-rated fatigue in PCC are not clear. Nonetheless, in a study by Calabria et al. on 136 COVID-19 patients with cognitive complaints—assessed with both neuropsychological tests and Modified Fatigue Impact Scale—clinically significant fatigue was reported in 82.3% of the participants on average 8 months after SARS-CoV-2 infection (both hospitalized and non-hospitalized patients). Different types of fatigue were related to distinct psychological and cognitive factors [47].

Risk factors for PCC

A recent meta-analysis by Tsampasian et al. [5] stated that female sex (odds ratio [OR], 1.56; 95% CI, 1.41–1.73), age (OR, 1.21; 95% CI, 1.11–1.33), high BMI (OR, 1.15; 95% CI, 1.08–1.23), and smoking (OR, 1.10; 95% CI, 1.07–1.13) were associated with an increased risk of developing symptoms of PCC also in non-hospitalized cases. The presence of comorbidities and previous hospitalization including ICU admission were found to be associated with even higher risk of PCC (OR, 2.48; 95% CI, 1.97–3.13 and OR, 2.37; 95% CI, 2.18–2.56, respectively). However, this meta-analysis refers to symptoms in general and not separating cognitive symptoms. Clinically, we have noted a higher percentage of women (approx. 70%) seeking care for cognitive post-COVID symptoms after an initial non-critical acute SARS-CoV-2 infection at the cognitive post-COVID clinic at Danderyd University Hospital (Möller, personal communication). The incidence of PCC is estimated to be higher among hospitalized COVID-19 cases compared with non-hospitalized [8, 20]. Thus, the initial disease severity is an important factor. In one of our studies in Sweden [14], hospitalization was the main risk factor for developing PCC and reduced health-related quality of life 6 months after infection. Markers of systemic inflammation are associated with persisting fatigue and cognitive symptoms with marked functional impairment [19], and reports indicate a large proportion (40%–80%) of hospitalized patients experiencing post-COVID sequalae with neuropsychiatric symptoms [19, 48, 49]. In line with this, vaccination against SARS-CoV-2—which for most people confers less risk of severe COVID-19—seems to decrease the risk of developing PCC after infection [3, 5].

Psychiatric comorbidity

Mental conditions—that is, disorders affecting the central nervous system such as major depressive disorder and autism spectrum—have been reported to be an increased risk for infection of COVID-19 [11, 50, 51]. However, in the study by Merzon et al. [52], there was no increase of depression in the COVID-19 cohort. People with attention deficit hyperactivity disorder (ADHD) have been reported to have a higher risk for contracting COVID-19 infection, especially the patients without pharmacological treatment [52]. Merzon et al. [52] speculated that pharmacotherapy for ADHD moderated the risk due to an increase of awareness of preventive actions, and Cortese et al. [53] encouraged ADHD treatment in order to reduce the spread of COVID-19. The study by Heslin et al. [50] reported an increased risk for ADHD patients to contract COVID-19 infection but also an increased risk to contract influenza, which may indicate that this finding is not specific for COVID-19. However, ADHD patients had a reduced risk of severe outcomes such as hospitalization and mechanical ventilation [50], which may be due to the age of the patients. This was contradicted by data showing that the association of ADHD with less-severe COVID-19 outcomes was only significant for the older age group [50]. Different confounders were discussed considering biological effects as an explanation via pharmacotherapy and elevated inflammatory processes that may induce alterations in the immune and antiviral systems [50]. Thus, it is still not obvious whether the increased rate of mental disorders in COVID-19 is due to a biological or to a social factor.

Myalgic encephalomyelities and similarities with PCC

PCC shares symptoms with ME/CFS which is characterized by mental fatigue and fatigability as well as cognitive difficulties lasting more than 6 months [54] and post-exertional symptom exacerbation (PESE) also referred to as post-exertional malaise (PEM), defined as exacerbation of symptoms after minimal cognitive, physical, emotional, or social activity, or activity that was previously tolerated [55]. Symptoms typically intensify 12–72 h following activity and can last for days or even weeks, sometimes leading to a relapse [56]. PESE can contribute to the episodic nature of disability in PCC, often presenting as unpredictable fluctuations in symptoms and function [56].

ME/CFS is characterized by WHO as a neurological disorder, but the etiology of ME/CFS is unknown. More than half of the patients report onset of symptoms after a virus infection [57]. Jason et al. (2021) compared symptoms of the patients with PCC to those of patients with ME/CFS, and the PCC symptoms were less intense with the exception of orthostatic problems. The symptoms of the patients with PCC decreased with time in contrast to the symptoms of the patients with ME/CFS. However, this was not the case with problems such as trouble forming words, difficulties focusing and absent mindedness which remained unchanged in the patients with PCC [58]. In our PCC outpatient clinic at Danderyd University Hospital, we find patients with PCC with a considerable overlap of ME/CFS (Borg and Möller 2023, personal communication), which is corroborated by findings of the studies of Kedor et al. [59] and Twomey et al. [60] reporting that around half of the patients with PCC were estimated to meet the criteria of ME/CFS. A cytokine increase has been reported in ME/CFS [61] and also found in PCC [62]. In the study by Haffke et al. [63], endothelial dysfunction of the same kind as found in ME/CFS was reported in PCC. Furthermore, Peluso et al. [64, 65] reported reactivation of viruses including Epstein-Barr virus (EBV) also described in ME/CFS [66]. Thus, one may speculate that PCC and ME/CFS have a common etiology or that they represent post-viral conditions with a common pathogenesis.

Explanatory models for long-term cognitive symptoms after COVID-19

Today, there is still insufficient knowledge of what causes cognitive dysfunctions in patients with PCC and even less knowledge of the most effective rehabilitation efforts. Different pathophysiological events explaining remaining symptoms after infection of SARS CoV-2 virus have been discussed [67], but in this context, it is important to bear in mind that some brain functions are more readily affected than others [68]. Acquired knowledge is considered resistant to brain injuries, whereas the ability to solve novel problems, attention functions, and episodic memory are more susceptible to both external disruptions such as brain injuries [69] as well as to inner states [70]. Below we discuss different possible mechanisms behind PCC including applicable theories for long-term cognitive symptoms in PCC (see Fig. 1).

Details are in the caption following the image — **Fig. 1**Open in figure viewer PowerPointPossible pathophysiological mechanism in cognitive impairment of post-COVID-19 condition (PCC).

Direct injuries during acute SARS-CoV-2 infection

As an obligate intracellular parasite without a metabolism of its own, the virus uses the host nucleic acid machinery to reprogram the host for reproduction and induces cell apoptosis, autophagy, and programmed necrotic cell death [71]. The level of cellular damage is related to the level of viremia which in turn is beholden to virulence—that is, the infectiveness, the effectiveness in control over cellular metabolism, and ability to avoid innate and acquired immunity [72, 73]. The ability to exploit the Angiotensin Converting Enzyme-2 (ACE2) protein as a receptor for entry into the cell to commandeer cellular machinery is central to inducing cell injury [74]. ACE2 is an essential protein in cell membranes that catalyzes angiotensin II to I and is involved in the renin–angiotensin–aldosterone system essential for blood pressure control and homeostasis [75]. Viral-induced cell death results in loss of ACE2 action in the entire body [76, 77]. Direct damage at the organ level includes multiple sites of injury, including brain stem inflammation and hypothalamic-pituitary axis dysfunction [78]. It is known that the SARS-CoV-2 virus can infect neural cells [79]. Several mechanisms have been proposed despite scant data supporting neuroinvasive routes [80], especially as ACE2 and the transmembrane serine protease 2 (TMPRSS2) receptor—an essential co-receptor for SARS-CoV-2 cell entry—are not expressed by the cells in CNS [81] but rather by brain endothelial cells. Hypoxia has been shown to alter the dynamics in ACE2 and TMPRSS2 expression in this cell type [82]. Viral access to these cells may affect endothelial function and in turn cause neural tissue injury [83].

Hypoxia: central and peripheral

Cardiorespiratory fitness as well as peak oxygen uptake and left ventricular ejection fraction has been shown to be associated with cognitive function [84] and cognitive impairment in patients with chronic heart failure [85]. Hypoxia is one possible mechanism of cognitive symptoms after COVID-19. Silent hypoxia and poor oxygenation despite relatively minor parenchymal involvement is common especially in the early stages of COVID-19 [86, 87]. The patients often have a high respiratory drive and dyspnea, but the respiratory rate is often normal. Lung computed tomography imaging, lung gas volume, and respiratory mechanics are also nearly normal. Pulmonary embolism is common in COVID-19, with a higher prevalence in critically ill patients [88]. The underlying mechanism behind the hypoxia is primarily altered lung perfusion possibly caused by a combination of pulmonary embolism and ventilation-perfusion mismatching [89, 90]. Alveolar wall damage could also contribute to the hypoxia. This mechanism is supported by the fact that many patients with COVID-19 have a decreased diffusing capacity (DLco), and this impairment often remains at follow-ups [91]. In one study of hospital survivors with COVID-19, no improvement in 6 MWT and DLco was found between 6 and 12 months [92]. In a 2-year study, steady improvement was seen in symptoms, health status, and psychological status after hospitalization due to COVID-19, but still sequelae were relatively common as was DLco impairment [13].

Microvascular inflammation and microthrombosis

The cognitive impairment in COVID-19 could also be related to microvascular inflammation. Severe COVID-19 is characterized by microvascular dysfunction. SARS-CoV-2 triggers endothelial exocytosis, which results in microvascular thrombosis and inflammation [93] leading to endothelial dysfunction and vascular leak. Postmortem analysis of pulmonary capillaries in COVID-19 cases demonstrated capillary plugging from fibrous microthrombi, suggesting that systemic thrombosis is a primary mechanism of pulmonary COVID-19 infection [94]. The endothelial dysfunction seen in COVID-19 can also be related to hemoglobin released from damaged erythrocytes that scavenges endothelial NO and thereby contributes to oxidative stress and hypoxia [95]. A case study reported improvement in a patient with PCC who was seropositive for a number of autoantibodies targeting G-protein-coupled receptors (GPCRs), and who was treated with an aptamer directed to inactivate GPCR autoantibodies. The authors noted measurable improvement of microcirculation in the eyes and of the general PCC symptoms and proposed that the impaired microcirculation observed in PCC is linked to GPCR autoantibodies [96].

A small study has reported amyloid-like microclots, up to a diameter of 200 μm, in 11 patients with PCC compared with healthy controls [97]. The clots were resistant to fibrinolysis, and because of their size, they may cause occlusion of capillaries and ischemic damage to the neurons [98]. The small cohort sizes of the study, the heterogeneity, and the difficulty in the characterization of the condition have to be considered when evaluating the importance of microthrombosis to PCC. We have not found any large prospective observational clinical studies that could confirm this finding. A study from the same authors (n = 25 cases of PCC) showed biomarker evidence of thrombotic endotheliitis in the patients in which these markers are suggested to be indicators of PCC [99]. However, the dysregulation of procoagulant and fibrinolytic pathways in patients with PCC has been observed elsewhere as well [100]. There is also an aspect of autoimmunity that may play a role in the formation of microclots in PCC. In a study by Zuo et al., 52% of patients treated in hospital due to COVID-19 exhibited prothrombotic autoantibodies targeting phospholipids and phospholipid-binding proteins [101]. This has to our knowledge not been confirmed in matched case-control studies of PCC.

Ongoing systemic inflammation, neuroinflammation, and autoimmune aspect

High levels of autoantibodies, including against ACE2, were found in patients with PCC and were inversely correlated with antibodies against COVID-19 antibodies [102]. Another possible mechanism that has been proposed is that inflammation in the airways and lung causes neuroinflammation [103]. In a paper by Fernandez-Castaneda et al., mice were infected in a way that limited the infection to the respiratory system [104]. This resulted in increased levels of cytokines in the cerebrospinal fluid (CSF), whereof one was C–C motif chemokine ligand 11 (CCL11)—a chemokine associated with cognitive impairment [105]. The authors also found impaired hippocampal neurogenesis, decreased oligodendrocytes, and myelin loss in the infected mice compared to control mice. Interestingly, similar findings were found in human postmortem SARS-CoV-2 infection. Furthermore, elevated CCL11 levels were found in humans with persisting cognitive symptoms post-COVID [104]. Taken together, these findings indicate that airway inflammation causes inflammation of the nervous system through release of systemic chemokines such as CCL11 [103].

Low-level persistence of SARS-CoV-2 viral antigens in CNS

The presence of viral antigens in the CNS may trigger an immune response with cytokine activation, leading to inflammation damaging the nervous system. Low levels of viral RNA have been detected by PCR in some autopsy studies [106–108], but evidence of causality between virus and COVID-associated neuropathology is controversial. Many studies have failed to identify SARS-CoV-2 in the autopsied brain or in CSF of live patients hospitalized due to COVID-19 [109–114]. In patients suffering from PCC, data are even more scarce and limited to case reports presenting residual viral antigens in CSF or tissues [115, 116]. In sum, there is so far insufficient evidence to conclude that direct or persisting CNS infection by SARS-CoV-2 itself explains the long-term neurological and/or neuropsychiatric symptoms of COVID-19.

Reactivation of latent herpesviruses

In PCC, signs of reactivated latent virus infections, such as EBV, human herpesvirus (HHV)-6, and HHV-7, have been observed [64, 66, 117], and in vitro studies show mitochondrial fragmentation and impaired energy metabolism as a result of the response to reactivated herpesvirus [118]. In a cohort of 280 adults with prior SARS-CoV-2 infection, lingering fatigue and cognitive dysfunction were independently associated with serological pattern, suggesting recent EBV reactivation [64]. In another study, suggesting COVID-19 inflammation-induced EBV reactivation as a possible driver of PCC, two-thirds of the patients with persisting symptoms post-COVID versus 10% in the control group were positive for EBV reactivation [119]. Furthermore, Verma et al. suggested a synergistic relationship between SARS-CoV-2 and EBV by demonstrating that lytic replication of EBV induces increased ACE2 expression in cultures of epithelial cells, thereby enhancing the entry of SARS-CoV-2 in the cells [120].

Epigenetic response

There is growing evidence to suggest that viral and bacterial infections can lead to epigenetic changes in the host, which can affect the host’s immune response and the severity of the infection [121–123]. Epigenetic changes refer to alterations in the way genes are expressed rather than changes in the underlying DNA sequence itself [124]. COVID-19 has been shown to induce epigenetic changes such as a modified DNA methylome also in healthy convalescents [125, 126]. More recently, evidence shows that PCC is associated with certain DNA methylation changes [127, 128]. In one of these studies [127], strong epigenetic modifications were reported in genes involved in the electron transport chain in mitochondria. Of note, DNA methylation changes to genes involved in mitochondrial function have also been reported in ME/CFS [129, 130] with similarities to PCC as described above. In addition, our unpublished data point toward enrichment of EBV infection-related pathways in the altered epigenomes of patients with PCC, further corroborating the notion that EBV reactivation may be mechanistically involved in PCC.

Correlates to imaging findings

The abovementioned explanatory models may have the potential to contribute to residual cognitive symptoms and fatigue. These effects may not necessarily manifest as detectable organic brain injuries. However, various outcomes have been reported in magnetic resonance imaging (MRI) and positron emission tomography studies of brain volume and metabolism in patients after COVID-19 infection. In a study by Martini et al. [131], severe hypometabolism was seen during the acute infection but after 7–9 months postinfection, no hypometabolism was detected. Rather, brain hypermetabolism was detected in the brainstem, cerebellum, hippocampus, and amygdala and was correlated to the inflammatory status [131]. Martini et al. (2022) suggested a synergistic effect of virus-mediated inflammation and transient hypoxia leading to dysfunction of the fronto-insular cortex. On the other hand, Dressing et al. [132] reported no significant changes of regional cerebral glucose metabolism, and functional imaging showed no distinct pathological changes in a long-term phase after COVID-19. Furthermore, cognitive testing showed minor abnormalities on an individual level. In a study of 56 patients with mild-to-moderate COVID-19 and 37 controls, Bispo et al. [133] found no differences in neuropsychological performance or gray matter volume between the patients with mild-to-moderate COVID-19 and the controls, but the patients with COVID-19 had lower fiber density in the association, projection, and commissural tracts. However, fatigue scores, reaction time, and visual memory tests correlated with microstructural changes in the COVID-19 group, suggesting possible brain substrate underlying the symptoms during recovery (approximately 3 months) from COVID-19 [133].

In a large study with data from the UK Biobank, 785 participants had been imaged twice using MRI, and 401 of the participants tested SARS-CoV-2-positive between the two scans. The average time between diagnosis and second scan was 141 days. The remaining 384 persons were included as controls. The study identified not only changes in olfactory pathways and orbitofrontal cortex but also a greater reduction in global brain size and on average a greater cognitive decline between the two time points among SARS-CoV-2 cases compared with the controls. As the study used preinfection imaging data, preexisting risk factors being misinterpreted as disease effects were reduced. The effects were still observed after excluding patients who had been hospitalized. The authors concluded that the limbic brain imaging results may be the in vivo hallmarks of a degenerative spread of the disease through olfactory pathways, of neuroinflammatory events, or of the loss of sensory input due to anosmia [134]. These results are interesting because in functional MRI studies, fatigue has been related to disturbed attention functions and altered connectivity in fronto-striatal networks in other patient groups [135]. Furthermore, in normal aging, decreased mental processing speed has been correlated to decreased connectivity in cortico-striatal networks in the brain [136]. As for COVID-19 fatigue, regional hypo- and hypermetabolism were found in several areas in the frontal lobe, and smaller volumes of the frontal lobe areas were correlated with fatigue [137]. However, radiological findings in COVID-19 should still be interpreted with caution as differences in outcome may be due to such factors as severity of initial infection, time since illness, methodological differences, or demographics.

Management and rehabilitation of cognitive dysfunction after COVID-19

Cognitive rehabilitation

Studies on cognitive rehabilitation for patients with persisting symptoms after COVID-19 are scarce [146, 147]. A proof-of-concept study drawing upon principles used in acquired brain injury (ABI) rehabilitation, and the rehabilitation of psychiatric populations has demonstrated potential benefits of computerized cognitive training [148]. As presented in WHO recommendations [149], specific evidence for training programs targeting COVID-19-related cognitive impairments predominantly relies on experiences and knowledge gained from cognitive dysfunctions caused by other factors. Given that fatigue and mental health symptoms overlap or tend to present in clusters, these factors as well as demographic factors should be considered when designing rehabilitation programs. The WHO recommends an individualized approach using a combination of education, skills training on self-management strategies, and cognitive training. Furthermore, training in the use of assistive products and environmental modifications is suggested as they apply to daily functioning.

Functional cognitive retraining includes systematic interventions that gradually increase in difficulty with the aim of either (1) retraining reduced functions directly by targeting the underlying processes or (2) implementing specific compensatory behaviors that help mitigate the impact of reduced function indirectly. The type of impairment determines which rehabilitation principle should be used. Regardless of the choice of training method, all cognitive rehabilitation is based on the premise that the individual is motivated to engage in training. For optimal treatment outcomes and sustained results, it is important to provide opportunities for generalization of strategies to everyday activities by incorporating metacognitive aspects (reflecting on one’s own cognitive processes) in the training as well as including strategies for emotional regulation, endurance in training, and motivation [150]. However, some impairments—such as poor memory functions and certain executive functions—cannot be functionally trained. In such cases, compensatory training must be offered.

Compensatory training involves communicating alternative behaviors or action sequences to circumvent difficulties and optimizing performance by using strategies that enhance skills for functional abilities and minimize situations that may lead to repeated failures. This approach relies on learning to employ a range of external and internal self-management strategies to establish a changed pattern of behavior. These can vary from simple everyday tips, such as using post-it notes to compensate for poor memory, to cognitively demanding processes [150, 151].

Below, principles are described for the rehabilitation of certain specific functional impairments commonly reported by patients with PCC. These are based on existing knowledge about rehabilitation of acquired brain injuries. As the suffering is great among those affected, we need to offer safe interventions while evaluating these carefully and taking evidence-based medicine into account [152]. As some patients may experience severe fatigue and worsened symptoms with intensive cognitive interventions, close monitoring of intensity levels and careful observation for any signs of deterioration are important. On the other hand, compensatory strategies should be possible to implement safely. To effectively implement the rehabilitation methods mentioned below, we suggest that the patient rehabilitation program is person-centered and individualized. It is important to start with a medical examination by a physician to rule out underlying medical conditions that need to be addressed first and to rule out that the patient is suffering from extensive PEM/PESE, making demanding rehabilitation efforts inappropriate. Next, a neuropsychological assessment should be conducted by a neuropsychologist to determine the patient’s cognitive dysfunctions and underlying cognitive processes that contribute to the dysfunctions, as well as to determine cognitive resources that can be used to compensate for impairments that cannot be trained. Based on the result of the neuropsychological assessment and evidence from other conditions, specific training methods (described and further visualized in Figs. 2 and 3) can be offered by therapists with expertise in cognitive rehabilitation. The rehabilitation can be guided by one or more therapists depending on the complexity. However, it is also important to remember that PCC can be very complex, and patients can have one or more cognitive dysfunctions, ranging from mild to severe. In addition, the patient may suffer from other somatic ailments or psychological symptoms which more or less also affect the cognitive difficulties. These need to be treated in parallel and to be taken into account when planning further rehabilitation. Which health professions should be involved in the rehabilitation process may vary depending on the type of rehabilitation but also differ both nationally and internationally depending on economic and cultural factors. Most often, psychologists, occupational therapists, and speech therapists are involved in cognitive rehabilitation. A proposal for rehabilitation therapy is visualized in Fig. 3.

Attention and processing speed

In ABI, specific attention training that includes metacognitive strategy training has been shown to be effective in retraining function and developing compensatory strategies that promote generalization to everyday activities [150, 153]. One well-evaluated attention training method is Attention Process Training, which is based on a cognitive model of attention [154]. Computer-based training methods targeting working memory have also been shown to be effective if conducted with the support of a therapist who provides continuous follow-up and feedback during the training. However, when used in isolation, computer-based training has limited evidence of improvements to other functional areas [153]. Training interventions targeting decreased attention and processing speed have demonstrated potential effect on fatigue in other patient groups [155]. However, considering that many PCC patients are easily fatigued, it may be preferable to initially prioritize compensatory strategies until it is determined whether the patient can tolerate functional training. For patients with reduced information processing speed or impaired ability to manage time, Time Pressure Management has been recommended as a compensatory strategy for patients following an ABI [153].

Memory

If memory problems are attributed to reduced memory span or working memory leading to, for example, decreased prospective memory, functional training targeting these underlying functions is recommended (see above). However, when memory storage itself is primarily affected, functional training alone lacks supporting evidence. Nonetheless, there is some evidence for strategy training as a compensatory approach to address impaired memory. Memory strategy training encompasses various techniques, some of which are tailored to specific difficulties, whereas others aim to enhance awareness of the consequences of memory impairment in practical situations. Specific training may focus on remembering names, developing routines, or establishing systems for organizing personal belongings. These strategies can also incorporate the use of aids that provide an overview or electronic memory aids that offer reminder functions. Patients with mild memory impairments may have better cognitive reserves, enabling them to employ internal strategies such as association techniques for compensation [153].

Executive functions

Executive impairments have a range of manifestations including both subtle behavioral changes such as difficulties in emotional regulation and self-monitoring, as well as specific impairments such as difficulties in problem-solving skills, activity regulation (initiation and inhibition), decision-making, and evaluation. Training of executive functions targets the enhancement of integrative processes that regulate goal-oriented and purposeful behavior aiming for the patient to acquire strategies to compensate for various controlled situations. Through increased awareness of problematic situations, specific strategies are employed and subsequently generalized to other activities in daily life, supported by the use of metacognitive skills. The training is centered around tasks that involve structured planning, generating alternative solutions, practicing feedback, and the ability to appropriately modify chosen plans.

Several methods have been evaluated for specific executive impairments. The highest evidence for training problem-solving skills and the ability to reach set goals is seen with metacognitive strategy training in combination with multimodal feedback [153].

Fatigue

Evidence-based treatments for fatigue are limited [156]. As fatigue is a multidimensional condition with many contributing factors, the methods also need to be multidimensional and preferably team-based. Treatment should be preceded by an investigation of possible primary and secondary causes of the patient’s fatigue, such as general health, depression, sleep disturbances, medication side effects, pain, other fatigue-inducing illnesses (e.g., anemia and hypothyroidism), and physical condition. The patient’s balance between rest and activity level should also be mapped. Once underlying causes are identified, pragmatic and individualized treatments can be implemented. As with cognitive rehabilitation, the interventions can be psychoeducational, compensatory, or focused on functional training.

Physical exercise

Physical exercise and fitness could be helpful for fatigue in neurological diseases [157, 158]. The effect is less certain in ME/CFS and systemic exertion intolerance disease for which exercise increases fatigue [159]. According to WHO living guidelines, a cautious return to physical exercise training is suggested via a combination of education and skills-training on energy conservation techniques [149]. Physical activity should be guided by the presence and severity of symptoms and can be titrated up and down to avoid exacerbating symptoms. It is important to exclude PEM before commencing exercise therapy and consider careful monitoring for PEM both during and after exercise [160, 161]. For patients with symptoms of PESE/PEM, the stabilization of symptoms needs to be secured before increasing activities [162], and graded exercise should not be offered [65].

Pulmonary rehabilitation has been evaluated as a method for improving outcome after COVID-19 in several studies, and it has been shown to improve physical capacity and health-related quality of life in patients with persistent symptoms [163]. Other studies have shown that exercise-based rehabilitation programs also improve dyspnea, anxiety, and muscle strength [147]. According to reports before the COVID-19 pandemic, aerobic training can improve cognitive function [164]. There is some support that pulmonary rehabilitation programs improve mental and cognitive function after COVID-19 [165, 166]. In one of these studies, 64 patients with persistent symptoms after COVID-19 underwent a 6-week interdisciplinary individualized pulmonary rehabilitation program. Apart from an improvement in the primary effect variable, 6-min walk distance, there was also an improvement in the Fatigue Assessment Scale [166]. In another study, a 12-week rehabilitation program was associated with a significant decrease in the proportion of patients reporting having a cognitive deficit [165]. However, both studies were observational studies without a control group.

Pharmacological treatment

Implementation of systemic corticosteroids as standard of care in the management of severe COVID-19 during the early phase of the pandemic was based on evidence from clinical trials. In patients hospitalized with COVID-19, the use of dexamethasone resulted in reduced 28-day mortality among those who were receiving either invasive mechanical ventilation or oxygen [167]. In a systematic review, corticosteroid therapy was associated with clinical recovery and a significantly shortened length of ICU hospitalization compared to those not treated with corticosteroids [168]. There are also a few observational studies indicating that treatment with systemic steroids confers faster improvement in lung function and resolution of radiological findings [169, 170]. Regarding cognitive function, Manera et al. assessed cognitive function with Mini-Mental State Examination (MMSE) and found that treatment with systemic steroids was associated with improved MMSE scores in a subgroup of patients that had both COVID-19 and concurrent other infections [171].

Antiviral therapy has become increasingly important in the management of COVID-19 as an early treatment of mild and moderate disease to prevent high-risk populations from progression to severe illness [172, 173]. Remdesivir and nirmatrelvir/ritonavir are the most frequently used antiviral drugs, and in a recent preprint study, patients with acute COVID-19 who received nirmatrelvir were 26% less likely to develop PCC regardless of their vaccination status and prior history of SARS-CoV-2 infection [174].

Antithrombotic treatment is recommended for patients with acute severe COVID-19 infection due to increased risk of venous and arterial thrombosis [175, 176]. However, the role of antithrombotic treatment in PCC remains unclear.

Identified knowledge gaps

Specific risk factors associated with the development of cognitive dysfunction in PCC need to be further explored, such as why women seem to be overrepresented in this group. There is no specific cure for PCC generally nor for cognitive dysfunction specifically. Rather, treatment is focused on symptom management and individualized rehabilitation. Larger interventional studies regarding the effect on cognitive functions in PCC are still lacking. There are some ongoing studies—for example, hyperbaric oxygen treatment in PCC (ClinicalTrials.gov Identifier: NCT04842448) and the effect of nirmatrelvir/ritonavir on individuals with severe PCC (ClinicalTrials.gov Identifier: NCT05823896)—but these are small. Preferably, multicenter studies are needed. Potential underlying subgroups that may respond differently to treatment and/or specific rehabilitation must be identified by careful consideration of the study population recruitment regarding severity of initial infection, when PCC investigation is initiated, certain clusters of PCC symptoms, immune status, comorbidities, demographics, and—if possible—virus variant. In addition, there is still a lack of clear clinical markers indicating which patients are suitable for which interventions, a prerequisite for successful rehabilitation in clinical contexts. Pharmacological treatment regarding cognitive dysfunction also needs to be evaluated, as well as whether treatment of SARS-CoV-2 or reactivated viruses can produce secondary beneficial effects on cognition.

Conclusions

Prolonged cognitive dysfunction is a common impairment affecting individuals with PCC. Risk factors for PCC in general include female sex, age, preexisting medical conditions, and the severity of the acute illness. Proposed mechanisms contributing to PCC and the cognitive impairments include neuroinflammation, hypoxia, vascular damage, latent viral reactivation, and direct viral invasion of the central nervous system. Managing cognitive dysfunction in PCC requires a multidimensional approach including a neuropsychological examination and individualized rehabilitation. Although evidence specific to COVID-19-related cognitive impairments is limited, interventions based on established practices for other neurological conditions can be implemented. The WHO recommends education, skills training, cognitive exercises, assistive products, and environmental modifications. Functional training carefully monitored for intensity is recommended for people who do not suffer from PEM. Further research is essential for evidence-based interventions specific to COVID-19-related cognitive impairments.