Impact of COVID‐19 infection on cognition and its association with neurological symptoms

Abstract Objective: To characterize the cognitive profile following COVID‐19 infection and its possible association to clinical symptoms, emotional disturbance, biomarkers, and disease severity. Methods: This was a single‐center cross‐sectional cohort study. Subjects between 20‐ and 60‐year old with confirmed COVID‐19 infection were included. Evaluation was performed between April 2020 and July 2021. Patients with previous cognitive impairment and other neurological or severe psychiatric disorders were excluded. Demographic and laboratory data were extracted from the medical records. Results: Altogether 200 patients were included, 85 subjects were female (42.3%), and mean age was 49.12 years (SD: 7.84). Patients were classified into four groups: nonhospitalized (NH, n = 21), hospitalized without intensive care unit (ICU) nor oxygen therapy (HOSP, n = 42), hospitalized without ICU but with oxygen therapy (OXY, n = 107), and ICU (ICU, n = 31) patients. NH group was younger (p = .026). No significant differences were found in any test performed attending severity of illness (p > .05). A total of 55 patients reported subjective cognitive complaints (SCC). Subjects with neurological symptoms (NS) performed worse in trail making test B (p = .013), digits backwards (p = .006), letter&numbers (p = .002), symbol digit modalities test (p = .016), and Stroop color (p = .010) tests. Conclusions: OXY patients and females referred more SCC associated with symptoms of anxiety and depression. Objective cognitive performance was unrelated to SCC. No cognitive impairment was found regarding the severity of COVID‐19 infection. Results suggest that NS such as headache, anosmia, and dysgeusia during infection were a risk factor for later cognitive deficits. Tests assessing attention, processing speed, and executive function were the most sensitive in detecting cognitive changes in these patients.


INTRODUCTION
Acute respiratory syndrome coronavirus 2 (SARS-CoV-2) principally targets the respiratory tract. However, there is growing evidence that COVID-19 can also affect the central nervous system (CNS) Romero-Sánchez et al., 2020) and cause CNS injury (Helms et al., 2020). Douand et al. (2022) identified significant longitudinal effects in brain imaging showing a greater reduction in grey matter thickness and tissue-contrast in the orbitofrontal cortex and parahippocampal gyrus, changes in markers of tissue damage in regions functionally connected to the primary olfactory cortex and greater reduction in global brain size. The COVID-19 participants also showed on average larger cognitive decline between the two time points. Subjective cognitive complaints (SCC) are among the most frequent neurological symptoms (NS) reported by patients following the acute infection (Bliddal et al., 2021). Those who suffered from COVID-19 may complain of cognitive dysfunction (Almeria et al., 2020) often described as brain fog. The presence of neuropsychological deficits following SARS-CoV-2 is likely to result from multiple and interacting causes, such as a direct damage by the virus to the cortex and adjacent subcortical structures, or from psychological trauma (Ritchie et al., 2020).
Heterogeneous findings were reported in several cognitive domains, specifically in attention and executive function (Almeria et al., 2020;Altuna et al., 2021;Daroische et al., 2021;García-Sánchez et al., 2022;Hadad et al., 2022;Hampshire et al., 2021;Woo et al., 2020;Zhou et al., 2020). Boesl et al. (2021) found out that 30% of patients with SCC had pathological scores in the Montreal Cognitive Assessment test (MoCA). Crivelli et al. (2022) in a systematic review and meta-analysis found that patients recovered from COVID-19 had lower general cognition in MoCA test compared to healthy controls. Daroische et al. (2021) demonstrated that the percentage of patients with global cognitive impairment ranged from 15% in Van Den Borst et al. (2021) to 80% in Alemanno et al. (2021). A study of 279 hospitalized patients found that 34% patients reported memory loss and 28% impaired concentration approximately 3 months after the discharge (Garrigues et al., 2020). Deficits in executive functioning, processing speed, category fluency, and memory encoding were also found (Ariza et al., 2022;Becker et al., 2021). Other studies linked the relation between cognitive deficits and NS (Almeria et al., 2020;Guo et al., 2022). Beaud et al. (2021) did not find a correlation between cognitive scores and mechanical ventilation. Woo et al. (2020) did not find oxygen supplementations and pharmacological treatments to predict cognitive deficits. Other studies (Alemanno et al., 2021) found that patients who benefited from orotracheal intubation and ventilation had significantly better scores in attention compared to patients who received oxygen therapy with venturi masks. García-Sánchez et al. (2022) found that hospitalized patients had significantly lower performance in the MoCA test and in processing speed than nonhospitalized (NH) patients but hospitalization did not have a significant effect on test performance in most domains. Conflicting results in several studies make it difficult to conclude with certainty how oxygen therapy/mechanical ventilation can prevent or worsen cognitive impairment. Mazza et al. (2021) observed a high rate of cognitive deficits at 1 and 3 months, irrespective of medical severity of the illness, with just 22% of the sample showing a good performance in all domains. Executive function and psychomotor coordination were the most involved domains, followed by information processing, verbal fluency, and working memory. These effects were influenced both by the presence of psychopathology and by the systemic inflammation, confirming connection among depression, inflammation, and cognition. Hellgren et al. (2021) reported that, in some individuals, COVID-19 infection may have a negative impact on cognition that lasts at least several months after discharge with immediate and delayed memory being the indices with scores below the cutoff points. Mattioli et al. (2021) did not support the presence of cognitive impairment in a selected population of COVID-19 patients studied 4 months following the diagnosis, although they did not include patients that required oxygen therapy or intensive care unit (ICU) care. Anxiety, stress, and depression resulted to be significantly higher in COVID-19 patients than in controls. When considering ICU patients, those had a higher susceptibility of developing cognitive impairment than mild cases (Mattioli et al., 2022). Hadad et al. (2022) found that disease severity, premorbid condition, pulmonary function test, and hypoxia did not contribute to cognitive performance. In addition, there is not a clear link between the severity of the infection and the degree of cognitive impairment (Houben & Bonnechère, 2022).
High rates of psychological symptoms such as anxiety, depression, post-traumatic stress disorder (PTSD), and/or suicidal behavior were reported in general population irrespective of infectious status following previous coronavirus epidemics (Jeong et al., 2016). Rogers et al. (2020) meta-analysis found that after recovery from the SARS and MERS infection, sleep disorders, traumatic memories, emotional lability, fatigue, and impaired concentration/memory were reported in more than 15% of the patients at the follow-up period (6 weeks to 39 months). Our previous study following acute COVID-19 infection associated SCC with anxiety and depression (Almeria et al., 2020).  and  suggested that psychological distress was prominent in patients with acute sequelae after COVID-19 infection and related to objective cognitive performance, but objective cognitive performance was unrelated to cognitive complaints. Moreover,  and  found 6 months after infection that psychological distress, particularly somatic preoccupation, and depression were the most frequently reported symptoms in these participants.
In this line, studies about the mental status of COVID-19 patients showed the presence of depression, anxiety, and PTSD (Guo et al., 2020). De Lorenzo et al. (2020) study found that a quarter of their patients presented cognitive impairment in MoCA and 22.2% developed PTSD. Amanzio et al. (2021) studied the association among cognitive, physical, and behavioral prior, during and after the lockdown measures in cognitively normal aging subjects and found out that fatigue was related to mood deflections and cognitive function in terms of psychomotor speed. During and following the infection, patients were at increased risk to develop depression and anxiety symptoms (Deng et al., 2021) suggesting that psychological factors and other persisting symptoms such as fatigue and sleep disorders may play a significant role in SCC (Ceban et al., 2022;Krishnan et al., 2022).
Almost 2 years after the COVID-19 outbreak, there is growing evidence of its impact on cognitive performance. Most studies did not consider emotional functioning or addressed performance validity as well as they used small samples of patients or mainly brief cognitive test or online surveys, which are not suitable to characterize the neuropsychological profile associated with COVID-19 (Daroische et al., 2021). Sustained subclinical neuropsychological impairment could be a common sequel after COVID-19 in young adults (Woo et al., 2020) suggesting that COVID-19 could leave cognitive and emotional dysfunctions, whose underestimation may be costly in terms of long-term morbidity and mortality. Therefore, neurologist and neuropsychologist are facing an increasing number of requests for assessment and treatment of patients with cognitive squeals after COVID-19 infection (Sozzi et al., 2020). An early detection of neuropsychological manifestations and its possible association with clinical features and blood biomarkers may modify the risk of developing irreversible impairment and cognitive decline over time. Tracking the impact of COVID-19 on cognitive and psychological patient conditions has relevant implications for rehabilitation strategies and long-term assistance. The aim of this study is to characterize the clinical and neuropsychological manifestations and to report the SCC in the subacute period following COVID-19 infection.

Study design and participants
This is a consecutive case series cross-sectional study that included adult patients evaluated in a universal and free nationalized health care hospital at Hospital Universitari MútuaTerrassa (HUMT) from April 2020 to July 2021. All patients included in the study had SARS-CoV-2 infection confirmed by positive polymerase chain reaction from nasopharyngeal swab or by positive serology. Patients were between 20-and 60-year old. Subjects over 60 years of age were excluded to avoid age-related cognitive decline. Patients with previous cognitive impairment and any other manifestation of the CNS or sever psychiatric disorders with potential cognitive deficits were also excluded.
None of the participants were scheduled for disability scheme. The assessment was performed between 10 and 34 days post hospital or ambulatory discharge. The study was approved by the local ethic committee and all subjects signed the informed consent.

Data collection and definitions
Data was collected from the HUMT database, and a retrospective review of the electronic health records was performed. Demographic

Statistical analysis
Sample data and cognitive results were described assuming normal distribution, knowing its performance in larger samples, and standardized in our population. Standardized punctuations (T scores) for different cognitive tests were expressed in frequencies as an expression of pathological results in those with scores equal to or less than 30 in their T score (corresponding to 2 SD or less). Inferential tests were performed to compare cognitive performance according to other characteristics of the sample of clinical relevance. Comparisons between cohorts were analyzed using analysis of variance (ANOVA), and Levene test was used to assume or not equal variances on groups of comparison analysis. Kruskal Wallis was used when inferential test did not follow requirement of number of patients for group.
Sample was divided according to severity of illness into four categories depending on the requirement of hospitalization, oxygen therapy, and ICU admission. Four groups were created: NH (n = 21), hospitalized, no ICU nor oxygen therapy (HOSP, n = 42), hospitalized no ICU but with oxygen therapy (OXY, n = 107), and ICU (ICU, n = 31).
Regarding neuropsychological impairment groups were divided as pathologic when T scores were <30, inferior performance when T scores were between 30 and 39, normal-inferior when T scores where between 40 and 49, and normal functioning when T scores were >50.
To correlate the NS of the disease in the acute phase with the possible effect on cognition, the presence or absence of the main symptoms TA B L E 1 Demographic and clinical characteristics of patients within different severity groups

Demographic and clinical characteristics
A total of 200 patients who tested positive for SARS-CoV-2 were included in the study. Demographic and clinical characteristics are described in Table 1   Ferritin and D-dimer were not related to cognitive impairment (p > .05). Attending severity illness group, NH groups were younger (ANOVA F = 3.162, p = .026). No differences were observed in years of education (ANOVA F = 1.98, p = .117). There were more females in the NH group (70%) (χ 2 p = .001), 52% in the HOSP group, 36% in OXY, and 29% in ICU. Days of hospitalization were longer in ICU group but differences were significant in all the three hospitalized groups (ANOVA F = 75.13, p = .001 Scheffé post hoc test p = .001 in all pairs of comparison). There were no differences between groups according to illness severity for HAD anxiety (F = 1.96, p = .121) or HAD depression (F = 1.15, p = .330).

Neuropsychological findings
Neuropsychological characteristics are described in  Comparing the percentage of patients over the cutoff point for anxiety and depression symptoms on the HAD scale, the group without cognitive complaints had a 31% of subjects above the cutoff score in the anxiety scale over the 72% in the SCC group (χ 2 = 27.35, p = .001). Regarding depression scale, 11% of the subjects in the noncognitive complaints had scores above the cutoff score, whereas the SCC group had a 52% of the subjects (χ 2 = 37.85, p = .001). Figure 2 shows the curve for subjects following normal distribution. Table 3 Table 3). There are no differences in percentages between patients with or without SCC.

DISCUSSION
Our study was designed to characterize the extent of cognitive impairment and SCC in patients following COVID-19 infection. In our cohort, fever was the predominant symptom in all patients, followed by cough, fatigue, and headache as described in previous studies (Almeria et al., 2020;Li et al., 2020). Males showed higher levels of ferritin than females (Bliddal et al., 2021) but not D-dimer values. Both D-dimer and ferritin were increased in ICU group, possibly due to its association with severity of illness.
We found that the neuropsychological performance profile of COVID-19 patients, regardless of the degree of clinical severity, is   (Mattioli et al., 2022) and where cognitive decline was independent of disease severity (Hadad et al., 2022;Houben & Bonnechère., 2022). Other studies also found no differences when considering treatment with mechanical ventilation or oxygen supply (Beaud et al., 2021;García-Sánchez et al., 2022;Woo et al., 2020). Our results suggest that COVID-19 infection per se does not appear to produce a great cognitive impairment. Cognitive deficits in our patients could be related to other factors such as previous unreported cognitive impairment or concomitant cerebrovascular diseases (Romero-Sánchez et al., 2020), encephalopathy Xiang et al., 2020), and symptoms of anxiety and depression (Amanzio et al., 2021;De Lorenzo et al., 2020;Deng et al., 2021;Jeong et al., 2016;Rogers et al., 2020;Sozzi et al., 2020).
However, tests for learning ability, attention, processing speed, and Zhou et al., 2020). As we did not have a prior to COVID-19 infection neuropsychological assessment, our hypothesis is that although no cognitive impairment was found in our sample, subjects tend to appear in a lower performance range in those domains because there is a decline regarding their initial performance.
When analyzing our sample separately, that is, patients with SCC or without, both groups tend to stand in majority in a lower standard devi-  (Douaud et al., 2022).
As in our sample, infected participants had no signs of memory impairment but worsening in executive function, particularly in TMT-B (Douaud et al., 2022). Our results also support the relation between fatigue and anxiety and depressive symptoms but not with cognition (Amanzio et al., 2021;Rogers et al., 2020).
Finally, we found that SCC were significantly higher in females and in the OXY patients group, probably associated with fear of awareness of the severity of the disease. Higher scores in anxiety and depression scales were reported, suggesting a greater impact of emotional wellbeing on SCC. Other studies have already pointed out the role of anxiety, depression, PTSD, and psychological distress in cognitive performance (Amanzio et al., 2021;Deng et al., 2021;Guo et al., 2020;Jeong et al., 2016;Krishnan et al., 2022;Rogers et al., 2020;.

LIMITATIONS
One limitation of this study was the use of a retrospective chart review to collect clinical information and the absence of a standardized questionnaire to assess cognitive complaints. Another limitation was the use of a single self-report measure for anxiety and depression symptoms. We encourage future studies to include more specific questionnaires to assess emotional functioning, neuropsychiatric symptoms, and PTSD. We analyzed the effect of the biomarkers that we were allowed to obtain as a routine in our setting. For that reason, another limitation is the lack of other specific inflammatory biomarkers which could be associated with systemic inflammation and not having information about the suspected COVID-19 variant. The main limitation of this study was that we did not have any cognitive evaluation before COVID-19 that could possibly show small differences from baseline and also the lack of a control group. We tried to minimize this limitation with the exclusion of older patients who might have had other concomitant pathologies that could affect cognition, such as vascular risk factors or incipient neurodegenerative diseases, as well as any previous neurologic or severe psychiatric disorder that might affect cognition.
All test scores were corrected by standardized notes, according to normative data in our environment, thus correcting the effects of the subjects' age and education. Future studies should address these limitations.

CONCLUSIONS
The findings in the current study allow us to characterize the cognitive profile of patients after COVID-19 infection. Specifically, COVID-19 patients can report SCC immediately after hospital discharge although no cognitive impairment was found when considering biomarkers or severity of illness. However, subjects tend to appear in a lower performance range specifically in test of processing speed, executive function, attention, and working memory. Additionally, having all three NS was indicative of worse performance in those domains. Clinicians should consider the presence of NS as a risk factor for cognitive worsening and neuropsychiatric symptoms such as anxiety and depression as a risk factor for SCC. A long-term follow-up is necessary to stablish the permanence of such complaints. Strategies that include psychological and cognitive rehabilitation especially in attention and executive function should be considered.

DATA AVAILABILITY STATEMENT
All study data, including raw and analyzed data, and materials are available from the corresponding author on request.