Brain fog of post-COVID-19 condition and Chronic Fatigue Syndrome, same medical disorder?

Patients aged 18–85 years with a sufficient understanding and communication skills, with post-COVID-19 condition and with ME/CFS were recruited from those attending the Neurology Department at Cruces University Hospital. Pregnancy and/or lactation, severe trauma, alcoholism, drug addiction, severe heart disease and/or radiological diagnosis of brain structural pathology (tumors, cysts, and malformations) were considered exclusion criteria.

Patients diagnosed with post-COVID-19 condition met the criteria proposed by the NICE guidelines, in which signs and symptoms that develop during or after the infection consistent with COVID-19 continued for more than 12 weeks and were not explained by an alternative diagnosis [24]. For the diagnosis of acute COVID-19, the valid diagnostic methods were a positive nasal PCR, the detection of IgG and/or IgM antibodies against SARS-CoV-2 or a medical report supporting the diagnosis, especially for those patients of the first Spanish wave (from the beginning of the pandemic until June 2020, Spain) who did not undergo diagnostic microbiological tests.

Of the subjects diagnosed with post-COVID-19 condition, 42 patients were infected in the first Spanish wave (between the start of the pandemic and the end of June 2020) in which the predominant variant was SEC8. On the other hand, 20 participants were infected in the second Spanish wave and 8 in the third (between July-December 2020 and from December 2020 to March 2021, respectively). In both, the second and third waves, the predominant variant in Spain was 20E (EU1). The remaining sample was infected in the fourth wave (mid-March to the end of June 2021). The Alpha variant was predominant on those dates.

For patients to be diagnosed with post-COVID-19 condition, any of the following symptoms had to be present at least 3 months after the infection and persisting for at least 2 months [18]: physical fatigue, mental fatigue, palpitations, sensory symptoms and/or dysautonomic symptoms. The exclusion criteria in this group were respiratory disease lasting 12 weeks after the infection, having been admitted to an intensive care unit and/or having had severe bilateral pneumonia or other severe disease manifestations requiring hospitalization. Patients with ME/CFS should have been previously diagnosed by a professional, or meet the criteria mentioned [2]. Those patients with concomitant diseases that could influence the results, as well as those who had previously received some immunomodulatory treatment, were excluded. Regarding the vaccination of these participants, only two patients with a post-Covid-19 condition were previously vaccinated. The vaccinations of these two participants were 2 and 15 days before infection.

For the diagnosis of CFS/ME, we used the criteria proposed by Fukuda et al. [2], in which disabling chronic fatigue (persistent or intermittent) is present for at least 6 months and not explained by any alternative cause, was the main and necessary symptom. In addition, at least four secondary symptoms (odynophagia, myalgias, polyaltralgias, sleep disorder, concentration and memory deficits, lymphadenopathy, headaches, and post-exertional malaise lasting more than 24 h) had to be present.

Neuropsychologic and neuropsychiatric evaluation was performed by an experienced neuropsychologist and neurologist team. Age, sex, years of education and clinically significant variables were recorded for all participants. Overall cognition screening was performed with the Montreal Cognitive Assessment (MoCA). A complete neuropsychological evaluation was carried out to assess the following cognitive domains: attentional verbal and working memory (Digits from Wechsler Adult Intelligence Scale IV [WAIS-IV]), visual attention (Trail Making Test A [TMT A]), sustained attention (Touluose-Piéron Revised [TP-R]), alternating attention (Trail Making Test B [TMT B]) verbal fluency (animals and P), processing speed (Symbol Digit Modality Test [SDMT]) and Salthouse Perceptual Comparison Test [SPCT], cognitive flexibility (Modified Wisconsin Card Sorting Test [M-WCST]), verbal memory (Hopkins Verbal Learning Test- Revised [HVLT-R], visual memory (Brief Visuospatial Memory Test-Revised [BVMT-R]), visuoconstructive capacity (Taylor Complex Figure Test [TCF]), visual perception (Benton Judgment Of Line Orientation [JLO], inhibitory capacity (Stroop Test) and abstraction (similarities from WAIS-IV).

The neuropsychological and neuropsychiatric evaluation was performed in a single day, lasting approximately one hour and a half. To avoid cognitive fatigue, the DSQ was completed by the patients at home.

Statistical analyses were carried out using IBM SPSS Statistics for Windows, version 23.0 (IBM SPSS, Armonk, NY, USA).

We analyzed the differences between groups in the DSQ questionnaire with a Chi square test that allows comparing the distribution of the results in a qualitative variable. The explanatory figure of the DSQ was made using the means of the frequency and severity variables, to see the symptom pattern in each pathology.

Regarding the neuropsychological results, the data were transformed to scalar scores (SS) or to typified scores provided by the test manual. A descriptive analysis and frequencies were performed for the SS of the cognitive tests. We set a cut-off point for cognitive tests, and thus for cognitive impairment, at a scalar score of six. A SS less than six means being more than 1.67 standard deviation below the mean compared to people of the same age and education.

For the cognitive variables’ raw data that fulfilled the normality assumption, the Student's t test was carried out to compare the means between groups. In the case of the tests that did not meet this assumption, the U Mann–Whitney test was performed. The raw data of the cognitive test was also transformed into Z scores for each group to create cognitive composites. These cognitive composites were created as followed: general cognition was composed with the MoCA; verbal fluency by the verbal fluency of animals and words with P; processing speed was composed of the SPCT and SDMT tests; attention domain was made up of the total hits on the TP-R, the Global Index of Attention and Perception (GIAP) of the TP-R, direct digits and the TMT A; verbal memory with the HVLT-R; visual memory by the BVMT-R and the memory of TCF; visuoconstructive ability was composed with the copy of TCF and visuospatial ability with the Benton JLO; and finally, the executive functions were composed with TMT-B, indirect digits (WAIS-IV), the Word-Color subtest of the Stroop test and the M-WCST. We also transformed the neuropsychiatric assessment's results into Z-scores to compare both groups. For these comparisons we used Student’s t-test.

A comparison of means was also made between those ‘probably infected with the SEC8 variant’ (first wave) and with ‘probably infected with the 20E (EU1) variant’ (second and third wave) for neuropsychiatric and cognitive variables, transforming raw scores into Z scores for neuropsychiatric and cognitive variables and comparing them with Student’s t-test. Patients from the fourth wave (possible Alpha variant) were not included due to the scarcity of the sample.

For the BSIT in the post-COVID-19 condition group, we divided the results into three groups following the BSIT normative data, taking into account performance, sex, and age: normal, relatively abnormal, and abnormal. We convert the raw data into Z-scores and create the same cognitive composites only for the post-COVID group, comparing the scores between BSIT groups with the Student's t-test.

The demographic and clinical data are shown in Table 1. No significant differences were found in age. There were statistically significant differences in education level (U = 1191.00; p = 0.046) with the post-covid group having more years of formal education. Statistically significant differences were also found in the proportion of women between groups (χ² = 8.29; p = 0.004), 92.9% of ME/CFS and 69.9% of post-COVID-19 patients were women. Most of the patients were Caucasian, being only one patient in the post-COVID-19 group an another in ME/CFS who did not identify with any of the U.S. Office of Management and Budget’s race categories. These differences are not expected to have an impact on the results, as these were corrected for age, education and sex.

Initially, we analyzed the most prevalent symptoms in both conditions. Figure 1 shows the frequency and severity of the symptoms evaluated in the DSQ with a comparison between both groups. Fatigue, the feeling of heaviness or exhaustion when exercising, post exertional malaise, difficulty sleeping and mental fatigue were the most prevalent symptoms. Significant differences between the frequency and severity showed a more severe disease course in the ME/CFS group (Fig. 1). These significant differences were found in the frequency of abdominal pain (χ² = 11.39; p = 0.022), contractions (χ² = 10.14; p = 0.038), instability/lack of balance (χ² = 14.30; p = 0.006), dizziness or fainting (χ² = 13.53; p = 0.009), unintentional weight loss/gain (χ² = 12.91; p = 0.012), sweaty hands (χ² = 18.51; p = 0.001), hot sensation (χ² = 18.47; p = 0.001), cold sensation (χ² = 10.11; p = 0.039) and sore throat (χ² = 9.54; p = 0.049). Significant differences were also found in the severity of some of the symptoms evaluated, including the feeling of unrefreshing sleep (χ² = 9.85; p = 0.043), bloating (χ² = 10.65; p = 0.031), abdominal pain (χ² = 10.93; p = 0.027), contractions (χ² = 12.87; p = 0.012), muscle weakness (χ² = 10.6; p = 0.039), the ability to maintain attention (χ² = 15.13; p = 0.004), problems with depth perception (χ² = 10.03; p = 0.040), nausea (χ² = 11.72; p = 0.020), instability/lack of balance (χ² = 16.11; p = 0.003), dizziness or fainting (χ² = 14.62; p = 0.006), weight loss/gain (χ² = 13.32; p = 0.010), sweaty hands (χ² = 18.12; p = 0.001), feeling of having a high temperature (χ² = 26.91; p = 0.000) or low temperature (χ² = 9.78; p = 0.044), flu-like symptoms (χ² = 15.30; p = 0.004) and discomfort/nausea from certain odors (χ² = 11.25; p = 0.024).

Regarding the cognitive performance of the groups, statistical analysis of cognitive outcomes revealed that the most predominant affected cognitive domain was the sustained attention capacity (Fig. 2), which was detected in 56.2% of patients with post-COVID-19 condition vs. 83.3% of patients with ME/CFS, with statistically significant differences between both groups in the TP-R test (Table 1), being ME/CFS patients the most affected ones. These deficits were followed by processing speed, memory and ability to learn verbal material.

The differences between groups in cognitive domains were analyzed using Z-scores (Fig. 3). The scores of the neuropsychological and neuropsychiatric evaluations imply a worse cognitive performance the lower the Z. There were statistically significant differences in attention (t = 2.12; p = 0.037) and visuospatial ability (t = 2.12; p = 0.037) with lower scores in the ME/CFS group. The ME/CFS group performed worse in most cognitive domains, except for visuoconstructive ability and verbal memory.

The Z-scores analyzed for neuropsychiatric variables show significantly worse mental health in ME/CFS patients (Fig. 3). SF-36 questionnaire showed that ME/CFS patients perceived more physical limitations (t = 2.70; p = 0.008), lower energy (t = 2.29; p = 0.000), poorer mental health (t = 2.27; p = 0.025), more impact of health on their social interactions (t = 3.57; p = 0.001), and a worse general health perception (t = 3.87; p = 0.000). ME/CFS patients had also greater levels of anxiety trait (t = -3.35; p = 0.001), more depressive symptoms (t = -3.28; p = 0.001), and a greater suicidal ideation (t = − 3.17; p = 0.002) compared to the post-COVID-19 patients (Table 2).

We found statistically significant correlations between post-COVID-19 patient’s cognitive performance and BSIT raw data. Taking into account that the ability to smell varies by sex and age, the raw data were corrected according to these two variables, creating three groups of results: normal, relatively abnormal and abnormal, as proposed in the BSIT manual. The ‘normal’ BSIT group was composed of 53 patients, 3 patients were in the ‘relatively abnormal’ group, and 15 in the ‘abnormal’ group. The ‘normal’ group had a mean of 9.96 (SD = 0.99), the ‘relatively abnormal’ group had a mean of 7.67 (SD = 0.58), and the ‘abnormal’ group had a mean of 6.47 (SD = 0.99) on the BSIT. This test’s results correlated with general cognition (MoCA; r = 0.24, p = 0.043), visual processing speed (SPCT 3; r = 0.24, p = 0.043, and SDMT; r = 0.29, p = 0.014), verbal memory (HVLT-R 1–3; r = 0.23; p = 0.049), visual memory (BVMT-R discrimination index; r = 0.26; p = 0.031), attentional capacity (TMT A; r = -0.39; p = 0.001) and visuospatial ability (Benton JLO; r = 0.35; p = 0.003). The relatively abnormal group was discarded due to the small sample, as well as the absence of statistically significant differences with the rest of the groups. Statistically significant differences were found between the groups with normal results and those with an abnormal result in general cognition (t = 2.43; p = 0.020), attention (t = 2.46; p = 0.050), verbal memory (t = 2.14; p = 0.036), visual memory (t = 2.29; p = 0.025), visuospatial perception (t = 3.04; p = 0.014), and abstraction capacity (t = 3.10; p = 0.005) (Fig. 4).

We also analyzed correlations between cognitive performance and neuropsychiatric status (Table 3). Table 3 shows correlations between cognitive and neuropsychiatric variables ranged from Rho = 0.30 to Rho = 0.53 (p ≤ 0.01). In the post-COVID-19 condition group, sleep quality, fatigue and pain correlated with general cognition, phonological verbal fluency, visual processing speed, long-term verbal memory, attention in the word Stroop subtest and inhibition capacity in the word-color Stroop subtest. Anxiety levels correlated with general cognition, visual processing speed, attention capacity (TMT A), sustained attention and visual recognition. Visual recognition also correlated with the depressive symptoms in GDS. In the ME/CFS patients, anxiety levels were the ones that presented the most statistically significant correlations with the different cognitive domains, more specifically, they seem to influence the capacity for phonological verbal fluency, attention (TMT A), sustained attention and cognitive flexibility (perseverative errors in M-WCST).

Correlations between neuropsychiatric outcomes are shown in Table 4. Data range from Rho = 0.25 to Rho = 0.70. Patients with post-COVID-19 condition presented lower independence and general health (Karnofsky scale) the greater the physical limitations, pain and fatigue. In these patients, the greater the depressive symptomatology, the worse the quality of sleep and the perception of their health. Fatigue levels, in turn, correlated mostly with physical problems. On the other hand, the ME/CFS group showed less independence and general health (Karnofsky scale) the greater the physical limitations, pain and anxiety. The perception of health correlated inversely with anxious and depressive symptomatology. Fatigue levels correlated mainly with anxiety- state. In both groups, mental health (SF-36) correlated with anxiety (STAI-state) and depression (GDS). Pain and physical limitations were also related in both groups. Limitations in maintaining a social life due to health issues were also related to fatigue and physical limitations, likewise, depressive symptomatology and anxiety symptoms were related to suicidal ideation in both groups.

Taking into account the correlations between education level, age, disease duration, cognition and neuropsychiatric variables, we made a stepwise linear regression analysis. Education, physical problems, pain, fatigue, sleep quality, depressive symptoms, anxiety and suicidal ideation, largely explained the variance of the cognitive deficits found in both groups, with a range of 3.8–37.1% explained variance (Fig. 5).

Post-COVID-19 condition patients’ education explained the highest percentage of variance in the cognitive performance (from 5.1% to 37.1%). Physical problems explained 28.5% and 16.2% of the variance in attention (Stroop test and TMT-A, respectively) and 20.9% of visual processing speed (SPCT). Pain (SF-36) was the variable that most explained the variance of sustained attention (TP-R errors) (8%), also explaining 5.8% of impulsivity (TP-R Impulsivity Control Index).

On the other hand, visual cognitive tests' performance was explained mainly by the limitation due to emotional problems (SF-36), which explained 12.2% of the variance of visual recognition (BVMT-R Discrimination Index), and fatigue (MFIS), which explained 10.9% of visual memory (TCF 3' memory), 9.2% of phonological verbal fluency (letter P), and 7.2% of visuoconstructive capacity (TCF copy).

Finally, sleep quality (PSQI) explained 15.7% of executive functions (M-WCST perseverative errors), anxiety state (STAI) explained 12.9% of sustained attention (TP-R) and 6.9% of general cognition (MoCA), and suicidal ideation (C-SSRS) explained 18.9% of the visual perception capacity variance (Benton JLO).

Regarding the ME/CFS group, energy/fatigue (SF-36) explained 28.6%. of the variance of sustained attention (TP-R errors). Sleep quality (PSQI) 20.8% of semantic verbal fluency (animals), anxiety levels (STAI-state) 28.6% of executive functions (M-WCST perseverative errors), and suicidal ideation (C-SSRS) 24.5% of immediate verbal memory capacity (HVLT-R trial 1).

Both conditions presented slow processing speed, deficient sustained attention and verbal memory impairment [29]. Statistically significant differences were found in attention and visual perception, with the ME/CFS group presenting the largest impairment. Therefore, we can understand that the brain fog could have similar cognitive deficits in both groups, being it mainly a reduction in the attentional capacity and a lower processing speed. This mental fatigue or brain fog seems to be closely related to physical fatigue.

Deficits in the attentional capacity may be due to different factors depending on the condition. Our results pointed out that physical fatigue, pain, anxiety, suicidal ideation, sleep quality and fatigue levels were related with cognitive impairment. Specifically, fatigue, physical problems, pain and sleep quality were the most prevalent correlations in the cognitive performance of post-COVID-19 group. In the ME/CFS patients, anxiety, sleep quality, energy levels and suicidal ideation were the most prevalent correlations. In the group of patients with ME/CFS group, anxiety inversely correlated with executive functions (M-WCST perseverative errors) [30, 31]. Energy levels of the SF-36 related to the number of mistakes made in the sustained attention test. Even so, pain did not explain the cognitive performance in ME/CFS group. Some of these patients may suffer from fibromyalgia, in this case, it would be interesting to verify if it is a subgroup of patients with greater cognitive impairment associated with this disease. With regard to the post-COVID-19, those with more physical problems and more pain appeared to perform worse on some cognitive domains. Higher levels of pain were related with poorer long-term visual memory and number of errors in sustained attention test. Chronic pain causes a decrease in attentional capacity and processing speed [30, 31], prolonged pain may also cause a reduction of the gray matter which also could lead to the worsening of the general cognitive performance [32]. In addition, we must add the cognitive deterioration related to the olfactory capacity in post-COVID-19 patients. Other studies have revealed the relationship between prolonged hyposmia after SARS-CoV-2 infection and cognitive impairment [33‐35]. According to our results, general cognition, processing speed, abstraction capacity and visuospatial capacity are related to olfactory function in these patients.

Sleep quality also correlated the performance in cognitive flexibility the post-COVID-19 condition group while in the ME/CSF group, it related to the semantic verbal fluency. All these tests require the involvement of the prefrontal lobe for its adequate performance, so that the quality of sleep could be closely related to the prefrontal activity. Previous studies have already shown the implication of sleep quality in adults without cognitive problems [36, 37]. In the group of patients with ME/CFS, poorer sleep quality was related to worse performance on executive function tasks, although it was not related to working memory. Previous studies had already revealed the relationship between poorer sleep quality and performance in this type of task [37].

Disease duration did not correlate with any of the cognitive tests, nonetheless, some studies associate ME/CFS with small fiber neuropathies, so the absence of progressive cognitive impairment does not imply the absence of neurodegeneration, since the damage of small fibers can progress and thus affect various functions as sudomotor, genitourinary, cardiovascular, pupillary or gastrointestinal functions [38, 39]. The evaluation of damage in the autonomic nervous system and the denervation of small fibers could be useful to discern phenotypes of patients. Some syndromes caused by the alteration of the vegetative nervous system, such as Postural Orthostatic Tachycardia Syndrome (POTS), are common in ME/CFS [9]. This syndrome can be caused by an autoimmune disease or be caused by an infectious disease, as recently seen after SARS-CoV-2 infection [40]. Other studies showed that some of ME/CFS patients notice a progressive worsening over time compared to the majority of patients with this condition, who refer a stable disease path with some fluctuations or relapses [41]. Therefore, it would be convenient to carry out a dysautonomic study that would allow us to better discern the dysautonomic symptoms associated with both syndromes and possible phenotypes in both conditions. In the case of post-COVID-19 condition, we are facing a unique situation for studying post-infectious fatigue since its inception, for this, it would be crucial to carry out a follow-up study in these patients and assess the progression and course of the disease.

Regarding the neuropsychiatric status, ME/CFS patients showed greater anxiety, depressive symptoms, suicidal ideation and a general worse perception of their health. These results along with the cognitive performance showed a more severe disease curse in ME/CFS.

The level of functionality in patients with post-COVID-19 was mainly related to a more physical than a neuropsychiatric sphere, correlating mainly with physical limitations, pain, and physical and mental fatigue. In the case of ME/CFS, the level of functionality was mostly related to physical capacity, pain and anxiety. Therefore, pain reduction and physical recovery should be objective variables in these populations, trying to avoid a sedentary lifestyle that can produce greater physical deconditioning. Both anxious and depressive symptoms should be evaluated and treated in both conditions, since both are closely related to suicidal ideation [42‐44].

Although the relevance information of these results, this study presented limitations. We found a disparity of the sample in terms of the percentage of women. The absence of a control group must also be taken into account. Despite this, these limitations were partially corrected by converting the cognitive tests results to normalized scores. The study participants did not have genome sequencing tests to determine the variant they suffered from, so we can only deduce the probability of having been infected by a specific variant taking into account the date of infection. No analysis was performed comparing the results between vaccinated and unvaccinated post-COVID-19 patients due to the small sample of vaccinated participants. Likewise, taking into account the predominance of Caucasians and unvaccinated participants in the study, we cannot know if the symptomatology varies between subjects of other races and vaccinated people. Future studies could clarify these questions. On the other hand, a larger sample of patients with different levels of hyposmia could help discern the cognitive impairment associated with the olfactory function in post-COVID-19 patients.