Impact of inflammatory response in the acute phase of COVID-19 on predicting objective and subjective post-COVID fatigue

doi:10.21203/rs.3.rs-4374986/v1

The biological predictors of objective and subjective fatigue in individuals with post-COVID syndrome remains unclear. This study aims to ascertain the predictive significance of the immune response measured during the acute phase of SARS-CoV-2 infection on various dimensions of fatigue 6–9 months post-infection. We examined the association between immune markers obtained from the serum of 54 patients (mean age: 58.69 ± 10.90; female: 31%) and objective and subjective chronic fatigue using general linear mixed models. Level of IL-1RA, IFNγ and TNFα in plasma and the percentage of monocytes measured in the acute phase of COVID-19 predicted physical and total fatigue. Moreover, the higher the concentration of TNFα (r=-0.40 ; p = .019) in the acute phase, the greater the lack of awareness of cognitive fatigue 6–9 months post-infection. These findings shed light on the relationship between acute inflammatory response and the persistence of both objective and subjective fatigue.

SARS-CoV-2

Fatigue

Immunity

COVID-19

Immunology

Long COVID

Inflammation

The symptom of fatigue is a clinical entity very often invoked by patients suffering from various pathologies ^1,2. This is particularly true post-COVID condition ³. In fact, fatigue is one of the main symptoms that can last for more than a year after SARS-CoV-2 infection ¹. Moreover, symptoms of post-COVID fatigue were associated, beyond 7 months after the infection, with structural cerebral changes in the thalamus and basal ganglia ⁴. The mechanisms that perpetuate fatigue in post-COVID condition are still poorly understood, although some studies have highlighted similarities with myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS)⁵, including neuroinflammatory responses with plasma cytokines elevation (e.g. IL-1, TNFα, IFNγ) and/or immune cellular variations (e.g. CD4) ⁶. According to Komaroff and Lipkin ⁷, these inflammatory variations could be concomitant with those encountered in post-COVID condition in individuals presenting symptoms of persistent fatigue ⁷. The hypothesis of acute systemic immune dysregulation is therefore an appealing perspective to explain the appearance and persistence of post-COVID fatigue symptoms ^8,9. In particular, monocytes, IL-1, interferon-gamma and TNFα appear to be interesting markers. In this respect, Cervia-Hasler, et al. ¹⁰ have demonstrated that monocyte-platelet aggregates were associated with persistent symptoms (including fatigue) beyond one year after infection. On the other hand, reviews by Davis, et al. ³, and Ceban, et al. ⁸ described that plasma level elevation of cytokines such as IL-1, TNFα, and IFNγ were also associated with persistent symptoms of fatigue and cognitive impairment at 6- and 13-months post-infection.

Curiously, as revealed by Ceban, et al. ⁸ it exists a discrepancy between objective and subjective fatigue sequelae after COVID-19 (e.g., Subjective fatigue is similar to verbal complaints in interviews or on self-reported questionnaires, whereas objective fatigue can be measured by attentional fluctuations and does not depend on a subjective assessment of internal state¹¹). This discrepancy in awareness of disorders suggests that several post-COVID phenotypes may exist. Based on this principle of a continuum of awareness of cognitive disorders after SARS-CoV-2 infection, Nuber-Champier, et al. ^{12 13} observed that anosognosia of memory disorders measured 6–9 months post-infection was associated with elevated levels of monocytes and TNFα measured during the acute phase. These patients who were anosognosic of their cognitive difficulties also showed the most impaired cognitive performance 6–9 months after infection ¹⁴. This idea of a distinct phenotype is a hypothesis also relayed by Hartung, et al. ¹⁵ proposing two types of post-COVID trajectory defined by a different pathophysiology, accompanied on the one hand by the persistence of cognitive sequelae and on the other by the persistence of symptoms of fatigue 9 months after infection. Thus, the question of awareness of fatigue disorders in the context of COVID-19 appears to be central to understanding post-COVID sequelae and the follow-up that will be necessary in the future.

Finally, apart from the fact that fatigue is often analysed as a single entity ⁸ whereas it is a complex mechanism defined by different dimensions ¹⁶, the dimensions of fatigue observed 6–9 months after SARS-CoV-2 infection have not been studied through the prism of immune variables measured during the acute phase. In this context, the first aim of this work was to highlight the predictive (and associative) capacity between peripheral inflammatory markers measured during the acute phase (selected from the literature presented: IL-1β, IL-1RA, IFNγ, TNFα and monocyte) with different dimensions of fatigue (physical, social, cognitive and psychological) measured 6–9 months post-infection. The second objective, based on work linking awareness of cognitive impairment to inflammatory states ^12,17, was to demonstrate the association and predictive capacity of immunity measured during the acute phase on awareness of cognitive fatigue measured 6–9 months post-infection.

With these objectives in mind, and based on the existing literature on the relationship between inflammatory markers and fatigue symptoms ^17,18, we hypothesized that higher levels of monocytes and pro-inflammatory plasma cytokines during the acute phase of COVID-19 would correlate with increased physical, social, psychological, and cognitive fatigue scores measured 6–9 months post-infection. Furthermore, we proposed that acute pro-inflammatory markers would predict dimensions of fatigue 6–9 months post-infection. Similarly, we hypothesized that high levels of inflammatory response measured during the acute phase of COVID-19 are associated with (and predict) greater difficulty in awareness of cognitive fatigue 6–9 months post-infection.

2.1 Sociodemographic and clinical data

In this study, we included middle-aged patients hospitalized in intermediate or intensive care, with no medical history that might have cognitive or immune repercussions (see Table 1). The results obtained by subgroup according to the type of hospitalization in the acute phase are available in supplementary material 1–7.

Table 1

Socio-demographic and clinical data from the sample of patients assessed 6–9 months after SARS-CoV-2 infection.
	Patients included in the study
	N = 54
Mean age in years (± SD)	58.69 (± 10.90)
Education level (1/2/3)	2/18/34
Sex (F/M)	17/37
Number of patients who required intermediate/intensive care in the acute phase	32/22
Mean days of hospitalization (± SD)	21.98 (± 25.29)
Mean days between positive RT-PCR test and collection of immunological data (± SD)	1.98 (± 3.62)
Diabetes (Yes/No)	8/46
History of respiratory disorders (Yes/No)	9/45
History of cardiovascular disorders (Yes/No)	10/44
History of neurological disorders (Yes/No)	0/54
History of psychiatric disorders (Yes/No)	2/52*
History of cancer (Yes/No)	0/54
History of severe immunosuppression (Yes/No)	0/54
History of developmental disorders (Yes/No)	0/54
Note. Education level: 1 = compulsory schooling, 2 = post-compulsory schooling, and 3 = university degree or equivalent. RT-PCR: reverse transcription polymerase chain reaction; SD: standard deviation; Sex F: female and M: mal. Types of history of respiratory disorders: asthma, chronic bronchitis; Types of history of cardiovascular disorders: previous infarction, valve pathology, atrial pathology and heart failure; *Types of history of psychiatric disorders: minor depressive episodes more than 10 years ago.

2.2 Cytokines plasma levels and monocyte % measured during the acute phase of COVID-19

The levels of immune markers (cytokines and monocytes) are presented in Table 2 in logarithmically untransformed form.

Table 2

Immune markers of patients with COVID-19 on admission to hospital.
Immune markers	Plasma cytokines concentration (N = 39) and blood monocytes proportion (N = 54) on day 1 of COVID-19 related hospitalization – Median [95%CI]
TNFα (pg/ml)	3.8 [3.50 ; 5.58]
IL-1Ra (pg/ml)	4464.84 [4715.70 ; 8432,12]
IFNγ (pg/ml)	1.65 [.97 ; 2.30]
IL-1β (pg/ml)	0.62 [.46 ; 1.03]
Monocytes% (percentage of white blood cells)	6.95 [5.07 ; 7.78]
Note. IFNγ: interferon gamma; IL: interleukin; TNFα: tumor necrosis factor alpha.

2.3 Association between plasma cytokines concentration, monocytes % measured in the acute phase of COVID-19 and fatigue dimensions 6-9 months post-infection

a) Total fatigue

We observed significant negative associations between total fatigue scores measured 6–9 months post-infection and TNFα levels (r=-0.44 ; p = .006) and also with IL-1RA levels (r=-0.34 ; p = .036) (see Fig. 1) measured during the acute phase.

Figure to be placed here

Figure 1. Plasma TNFα levels measured during the acute phase in relation to chronic fatigue percentage.

b) Cognitive fatigue

Blood monocyte percentage among white blood cells measured during the acute phase were significantly negatively associated with cognitive fatigue scores measured 6–9 months post-infection (r=-0.35 ; p = .009) (see Fig. 2).

c) Physical fatigue

TNFα plasma levels measured during the acute phase were significantly negatively associated with physical fatigue scores measured 6–9 months post-infection (r=-0.42 ; p = .008).

d) Social and psychological fatigue

None of the results were significant. The results are available as supplementary material 8.

2.4 Prediction of fatigue dimensions 6–9 months post-infection by inflammation measured during the acute phase of SARS-CoV-2 infection

a) Total fatigue

The model for predicting total fatigue, which included variables such as sex, age, plasma levels of TNFα, IL-1RA, IL-1β, IFNγ and the percentage of blood monocytes measured during the acute phase, significantly selected IL-1RA (F = 4.46; p = .044; 95%CI [-0.73; -0.01]) and IFNγ (F = 5.02; p = .034; 95%CI [0.026; 0.59]) plasma levels to predict total fatigue 6–9 months post-infection.

b) Cognitive fatigue

None of the results were significant. The results are available as supplementary material 9.

c) Physical fatigue

The same predictive model as for the other fatigue dimensions and total fatigue was significant for the predictive ability of IFNγ plasma levels (F = 6.39; p = .018; 95%CI [.076; .73]) measured during the acute phase in predicting physical fatigue 6–9 months post-infection.

d) Social and psychological fatigue

None of the results were significant. The results are available as supplementary material 9.

Figure to be placed here

Figure 2. Monocytes percentage measured during the acute phase in relation with chronic cognitive fatigue percentage.

2.5 Association between immunity measured during the acute phase and awareness of cognitive fatigue 6–9 months post-infection

We observed a negative correlation between TNFα levels measured during the acute phase and SAD scores of cognitive fatigue obtained 6–9 months post infection (r=-0.40 ; p = .019) (see Fig. 3).

2.6 Prediction of immunity measured during the acute phase on awareness of cognitive fatigue 6–9 months post-infection

The regression model concerning the prediction of the SAD scores of cognitive fatigues, which included variables such as sex, age, levels of TNFα, IL-1RA, IL-1β, IFNγ and the percentage of monocytes measured during the acute phase, was significant for the IL-1RA levels (F = 6.12; p = .048; 95%CI [-3.13;-0.17]).

Figure to be placed here

Figure 3. TNFα levels measured during the acute phase in relation with self-appraisal discrepancy scores of cognitive fatigue (6–9 months post-infection).

In this study, we examined the link between the immune response during the acute phase of COVID-19 and the severity of subjective fatigue observed 6–9 months later. We found inverse correlations between plasma levels of TNFα and IL-1RA during the acute phase and the overall fatigue scores measured 6–9 months post-infection. Additionally, TNFα plasma levels during the acute phase of COVID-19 were inversely associated with physical fatigue scores 6–9 months post-infection, and the percentage of blood monocyte among white blood cells during the acute phase of COVID-19 was inversely associated with cognitive fatigue scores. Our analysis further revealed that IL-1RA and IFNγ plasma levels during the acute phase of COVID-19 predicted total fatigue scores 6–9 months post-infection. IFNγ plasma level measured during the acute phase of COVID-19 predicting physical fatigue 6–9 months post-infection. Notably, higher TNFα concentrations during the acute phase of COVID-19 were associated with a higher lack of awareness of cognitive fatigue 6–9 months post-infection and IL-1RA predicted fatigue awareness scores.

These results, suggest that inflammation reactions during the acute phase of COVID-19 may influence different long-term cognitive and fatigue profiles ^12,17. On the one hand, people with high levels of inflammation during the acute phase would develop more significant cognitive sequelae in the long term, observable via problems with awareness of deficits, memory impairment ¹² and executive impairment ^3,19. On the other hand, people with a low level of inflammation in the acute phase would develop an increased subjective sensitivity in the long term to the sensation of fatigue and symptoms of depression ^3,19. Our observations therefore do not fully corroborate the results observed in the literature describing a linear relationship between fatigue and inflammation levels. We observe here acute inflammatory mechanisms that could follow an inverted-U curve, with an over- or under-optimal acute reaction generating distinct sequelae. These observations are made possible by the prism of awareness of cognitive disorders/fatigue measured via the discrepancy between subjective complaints and objective measurements. This discrepancy between subjective complaint and objective measurement has already been mentioned in Ceban et al,⁸ where it was shown that a greater proportion of patients appeared to have cognitive difficulties when using objective measures compared with subjective measures. At the immune level in animals, McAfoose et al ²⁰, established an inverted-U relationship between inflammatory cytokine markers and cognitive performance in memory and appraisal. According to these authors, a basal level of inflammatory cytokines is necessary for good cognitive function, but an excess or a low level would have neurotoxic consequences. Thus, contrary to the prediction that a high inflammatory state in the acute phase would be associated with high subjective chronic fatigue, we observe that a hypo-inflammatory state in the acute phase is associated with high subjective chronic fatigue and hyper-inflammation with cognitive difficulties in fatigue awareness. In this way, an optimised acute phase immune response to SARS-CoV-2 would produce few or no symptoms of long COVID ²¹. This interpretation is in line with Hartung, et al. ¹⁵ who suggested the possibility of 2 observable phenotypes in the post-COVID condition. First, a phenotype of patients displaying high fatigue and a phenotype of patients displaying cognitive disorders. We showed here that awareness of disorders observed 6–9 months post-infection and inflammatory levels measured in the acute phase of infection may be markers that could allow phenotyping patient trajectories in the context of COVID-19.

On this continuum of awareness of disorders, we had previously shown that levels of TNFα and monocytes measured during the acute phase of COVID-19 were associated with memory disorders and awareness of these disorders 6–9 months post-infection ²². This anosognosia of memory impairment was accompanied by hypoconnectivity of subcortical, cerebellar and hippocampal regions 6–9 months post-infection ¹². In the context of post-COVID fatigue, Heine, et al. ⁴ also demonstrated morphometric changes in the basal ganglia 7–8 months post-infection. Other studies are currently showing the implication of TNFα plasma levels on the increased risk of developing a post-COVID condition in the longer term ²³. Furthermore, it would appear that the persistence of high levels of TNFα, IL-6 and IL-1β seems to persist beyond the acute phase of COVID-19, in some cases two years after infection ²¹. This post-COVID trajectory could be associated with accelerated brain ageing ²⁴ and the emergence of neurodegenerative pathologies ²⁵.

Contrary to anosognosia, at the other end of the awareness continuum for disorders are individuals with severe, non-objectified complaints who exhibit a sub-optimal acute inflammatory response at the onset of infection. This idea of under-optimisation of the immune response is in line with the observations of Kervevan, et al. ²⁶ who distinguished between patients with reduced immune responses and patients with increased immune responses in the post-COVID condition. It would also seem judicious to take into account that the pandemic, socio-economic conditions and pre-existing vulnerabilities could contribute to the appearance of an exacerbated awareness of cognitive disorders and post-COVID fatigue, whatever the severity of the disease in the acute phase of COVID-19 ²⁷. As shown by Miller and Maner ²⁸ in other infectious contexts, pre-existing cognitive biases could make part of the population vulnerable to over-interpreting the danger represented by the contagiousness of the environment and the disease and activate parallel behavioural and immune reactions. Finally, between these two conditions of patients, there would exist a third category of people presenting subjective complaints which are objectified by the corroboration of processing speed capacities (defining a form of objective fatigue) ²⁹.

Finally, the results of this study reveal the importance of distinguishing subtypes of fatigue because each fatigue dimension maintains specific links with immune markers ^16,30. Dimensions of fatigue could explain variance in fatigue and cognitive symptoms between post-COVID individuals ³¹. The common denominator able to predict physical and total fatigue is IFNγ. The associations suggest that specific mechanisms are at work between IL-1RA and total fatigue; monocyte percentage and cognitive fatigue. Curiously, TNFα levels are associated with both total and physical fatigue. IFNγ levels may be related to subjective awareness of physical state while IL-1RA to a more global measure of fatigue dimensions. This raises several questions, such as how the different dimensions of fatigue evolve over time in parallel with individuals' socio-economic status and intrinsic vulnerabilities. Numerous synergies between fatigue, cognitive disorders, and the trajectory of neurodegenerative pathologies or myalgic encephalomyelitis remain to be investigated.

From the point of view of hospitalization subgroups, we observed that the blood monocyte percentage of patients hospitalised in intermediate care were associated with total and physical fatigue measured 6–9 months post-infection. Inflammatory responses measured in the acute phase were not predictive of chronic fatigue scores in this subsample. Among those hospitalised in intensive care, TNFα plasma levels were associated with total, cognitive and social fatigue scores measured 6–9 months post-infection. Curiously, IFNγ and IL-1RA plasma levels measured in the acute phase predicted total, cognitive, and social chronic fatigue scores. Interestingly, IFNγ was the only marker to predict chronic physical fatigue. These results suggest that the viral load has long-term repercussions, although the severity of the acute respiratory form does not fully explain the long-term cognitive and fatigue repercussions after infection.

This study has limitations, primarily the relatively small sample size for which we have applied statistical corrections. Although we have shown that our statistical power was sufficient we are aware of the limitations of our sample size. We suggest that studies with larger sample sizes should be carried out in the field of research investigated here in order to make this information more generalizable. In line with this limitation, the restriction of the analyses to hospitalized patients also limits the generalisation of the study to the general population. A second limitation relates, of course, to the discrepancy between subjective and objective fatigue, which must be a normal process up to a certain stage of difference. The first difficulty lies in the lack of definition of objective fatigue and, in fact, its measurement. A second difficulty lies in the quantification and categorisation of the normal process of discrepancy between subjective and objective complaints. Furthermore, although the French-language measure of subjective fatigue EMIF-SEP (MFIS in English) was initially validated on a population with multiple sclerosis, its use has already been extended to the COVID-19 study ³². Finally, the retrospective nature of the analyses could give rise to variance linked to confounding factors concerning the time between the acute phase and the 6–9 months post-infection measurements. This limitation is also a strength, as it allows us to visualise the effects of baseline immunity without SARS-CoV-2 treatment and its impact on long-term cognition.

In summary, fatigue is a clinical entity present in the majority of cases of post-COVID conditions. Circulating blood monocytes and certain cytokines secreted during the acute phase of COVID-19 are associated with different dimensions of fatigue, in particular physical and cognitive fatigue 6–9 months post-infection. In addition, fatigue awareness, defined as the difference between subjective and objective cognitive fatigue, appears to be associated with TNFα concentration and predicted by IL-1RA concentrations measured in the acute phase of SARS-CoV-2 infection. The evolution of immunity and dimensions of fatigue over time will need to be investigated in the future.

4.1 General procedure

We extracted a selection of data from hospitalized patients from the COVID-COG cohort (described below): data on innate immunity and cytokine measured during the acute phase, as well as data on objective and subjective fatigue measured 6–9 months post-infection. We then calculated a self-appraisal discrepancy score to quantify the awareness of fatigue, and finally we tested the association and predictive value of immunity measured during the acute phase on the various dimensions of chronic fatigue. The study was conducted in accordance with the Declaration of Helsinki, and the study protocol was approved by the cantonal ethics committee of Geneva (CER-02186).

4.2 COVID-COG cohort

The COVID-COG cohort ³³ is made up of 121 patients recruited on the basis of strict selection criteria, notably the absence of prior neurological, psychiatric, cancer, neurodevelopmental pathology, pregnancy or age over 80 years. SARS-CoV-2 infection had to be confirmed by a positive polymerase chain reaction (PCR) test from a nasopharyngeal swab and/or positive serological results. The patients in the cohort were divided into patients hospitalized in intensive care with mechanical ventilation (N = 24), patients hospitalized without mechanical ventilation (N = 48) and patients who did not require acute hospitalisation (N = 49). All the groups were comparable in terms of socio-demographic aspects. All participants had performed an exhaustive set of neuropsychological tests measuring memory, executive, instrumental and attentional processes, as well as a set of psychiatric questionnaires.

4.3 Participants included in the study

From the COVID-COG cohort, which initially consisted of 121 patients, we retained 54 hospitalized patients (N = 32 in conventional care and N = 22 in intensive care) with leukocyte and fatigue data (see Table 1). Of these 54 patients, 39 had samples that could be analysed for cytokine quantification (see Fig. 4).

Figure to be placed here

Figure 4. Study flowchart.

4.4 Measurement of subjective fatigue

We measured total fatigue and the sub-dimensions of subjective fatigue using the French version of the EMIF-SEP questionnaire ³⁴. This validated scale consists of 40 items, including 10 items for the cognitive dimension, 13 items for the physical dimension,13 items for the social and 4 items for psychological dimension. The raw scores are then transformed into percentages so that the dimensions of fatigue are statistically comparable. The higher the fatigue percentage, the more severe the subjective fatigue.

4.5 Measurement of objective cognitive fatigue

In order to measure cognitive fatigue objectively, we used T scores for the standard deviation of reaction time in the Test of Attentional Performance (TAP) sustained attention subtest ³⁵. Higher T-scores corresponded to better performance on the task. This has been frequently used as a measure of fatigue in conditions such as multiple sclerosis ^11,36, but also in the context of comparing an objective and subjective fatigue score ³⁷.

4.6 Self-appraisal discrepancy (SAD) of cognitive fatigue

Before subtracting the objective cognitive fatigue scores from the subjective scores, we inverted the subjective fatigue score scale by applying a subtraction percentage. Thus, for example, 80% of reported cognitive fatigue corresponded to 20% of fatigue preservation. The measures (objective and subjective) thus measured fatigue difficulties in the same direction: the higher the score, the less fatigue there was.

We weighted the subjective scale against the objective scale. Finally, in order to measure the difference between objective cognitive fatigue complaints minus subjective cognitive fatigue complaints, we subtracted the T scores of the standard deviation of the TAP sustained attention reaction time ³⁵ from the percentage of subjective cognitive fatigue from the EMIF-SEP questionnaire ³⁴.

4.7 Analysis of cytokines and monocytes

The analysis of cytokines (pg/ml) of TNFα, interleukin (IL) − 1RA, IL-1β, interferon gamma (IFNγ), was made using commercially available multiplex bead immunoassays (Fluorokine MAP Multiplex Human Cytokine Panel, R&D Systems, Minneapolis, USA) and read using a Bioplex 200 Array Reader (Bio-Rad Laboratories, Hercules, CA, USA) and Luminex® xMAP™ technology (Luminex Corporation, Austin, TX, USA). Percentage of blood monocytes was evaluated with Piccolo Xpress (Sysmex, Switzerland) tools (see Table 2). The short period of time between blood collection, processing (cell analysis on fresh blood), and freezing (plasma) did not result in any sample alteration (< mean 79 hours).

4.8 Statistical power

We set the β type 2 error at .80 and, since our hypotheses were formulated in a specific sense, the threshold for the α relationship was set at .025. Finally, we estimated a correlation coefficient on the observed relationship between TNFα levels, cell concentrations (monocytes) and post-COVID cognitive symptoms obtained in Nuber-Champier, et al. ^12,22 and Cervia-Hasler, et al. ¹⁰.

Samples for cytokine analysis

The standard normal deviate for α = Z_α = 1.9600

The standard normal deviate for β = Z_β = 0.8416

C = 0.5 * ln[(1 + r)/(1-r)] = 0.6625

Total sample size = N = [(Z_α+Z_β)/C]2 + 3 = 21

Samples for cell analysis

The standard normal deviate for α = Z_α = 1.9600

The standard normal deviate for β = Z_β = 0.8416

C = 0.5 * ln[(1 + r)/(1-r)] = 0.4001

Total sample size = N = [(Z_α+Z_β)/C]2 + 3 = 52

Based on the calculation made by Sb, et al. ³⁸, the necessary sample size is estimated at 52 participants for the cellular analyses and 21 participants for the cytokine analyses.

4.9 Statistical analysis

As the values of the immune response data, in particular cytokine values, were close to 0, we performed a logarithmic (log) transformation of these variables. In addition, given the distribution of the behavioural data, we applied non-parametric statistical tests.

To test our first hypothesis postulating an association between inflammatory variables (TNFα, IFNγ, IL-1β, IL-1RA, Monocytes) measured during the acute phase and the different dimensions of fatigue (total, physical, cognitive, social, psychological) at 6–9 months post infection, we performed Spearman correlations.

Then, with a view to testing the hypothesis of the predictive capacity of immune variables measured during the acute phase on fatigue scores obtained at 6–9 months post-infection, we performed generalized linear mixed models (GLMM) gamma considering inflammatory marker levels, age and sex as predictor variables of the prediction models of fatigue dimensions at 6–9 months post-infection.

Finally, to test the hypothesis of an association and predictive capacity of immunity measured during the acute phase on cognitive fatigue awareness scores (SAD), we performed Spearman correlations and GLMM, with inflammatory marker levels, age and sex as predictor.

We applied false discovery rate (FDR) corrections for all the analyses.

false discovery rate (FDR), generalized linear mixed models (GLMM), logarithmic (log), myalgic encephalomyelitis (ME/CFS), free/cued recall paradigm (RLRI), intensive care unit (ICU), interferon (IFN), interleukin (IL), polymerase chain reaction (PCR), reverse transcription polymerase chain reaction (RT-PCR), Self-appraisal discrepancy (SAD), severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), Test of Attentional Performance (TAP), tumor necrosis factor alpha (TNFα).

Acknowledgements

The present research was supported by Swiss National Science Foundation (SNSF) funding to JAP (PI) and FA (Co-PI) (grant no. 220041).

Author contributions

A. Nuber-Champier: Contributed to the writing, analysis and examination of the patients in the study.

G. Breville: Contributed to the writing and examination of the cytokines in the study.

P. Voruz: Contributed to the writing and examination of the patients in the study.

I. Jacot de Alcântara: Contributed to the writing and examination of the patients in the study.

P.H. Lalive: Contributed to the neuro-immunological expertise, writing and proofreading of the analyses and interpretations.

G. Allali: Contributed to the neurological expertise, writing and proofreading of the analyses and interpretations.

L. Benzakour: Contributed to the psychiatric expertise and proofreading.

K.-O. Lövblad: Contributed to the expertise in neuroimaging and proofreading.

O. Braillard: Contributed to the coordination of the patients and proofreading.

M. Nehme: Contributed to the coordination of the patients and proofreading.

M. Coen: Contributed to the coordination of the patients and proofreading.

J. Serratrice: Contributed to the coordination of the patients and proofreading.

J.-L Reny: Contributed to the coordination of the patients and proofreading.

J. Pugin: Contributed to the coordination of the patients and proofreading.

I. Guessous: Contributed to patient coordination, statistical epidemiology and proofreading.

B.N. Landis: Contributed to the coordination of the patients and proofreading.

A. Griffa: Contributed to the neuroimaging expertise and proofreading.

D. Van De Ville: Contributed to the neuroimaging expertise and proofreading.

F. Assal: Contributed to the writing of the overall project, proofreading and scientific direction.

J.A. Péron: Contributed to the writing of the overall project, proofreading and scientific direction.

Competing interests

no conflicts of interest to be declared.

Seeßle, J. et al. Persistent symptoms in adult patients 1 year after coronavirus disease 2019 (COVID-19): a prospective cohort study. Clinical infectious diseases 74, 1191-1198 (2022).
Mazza, M. G. et al. Prevalence, trajectory over time, and risk factor of post-COVID-19 fatigue. Journal of psychiatric research (2022).
Davis, H. E., McCorkell, L., Vogel, J. M. & Topol, E. J. Long COVID: major findings, mechanisms and recommendations. Nature Reviews Microbiology, 1-14 (2023).
Heine, J. et al. Structural brain changes in patients with post-COVID fatigue: a prospective observational study. Eclinicalmedicine 58, 101874 (2023).
Bateman, L. et al. in Mayo Clinic Proceedings. 2861-2878 (Elsevier).
Morris, G. & Maes, M. A neuro-immune model of myalgic encephalomyelitis/chronic fatigue syndrome. Metabolic brain disease 28, 523-540 (2013).
Komaroff, A. L. & Lipkin, W. I. Insights from myalgic encephalomyelitis/chronic fatigue syndrome may help unravel the pathogenesis of postacute COVID-19 syndrome. Trends in Molecular Medicine 27, 895-906 (2021).
Ceban, F. et al. Fatigue and cognitive impairment in Post-COVID-19 Syndrome: A systematic review and meta-analysis. Brain, behavior, and immunity 101, 93-135 (2022).
Azzolino, D. & Cesari, M. Fatigue in the COVID-19 pandemic. The Lancet Healthy Longevity 3, e128-e129 (2022).
Cervia-Hasler, C. et al. Persistent complement dysregulation with signs of thromboinflammation in active Long Covid. Science 383, eadg7942 (2024).
Weinges-Evers, N. et al. Correlation of self-assessed fatigue and alertness in multiple sclerosis. Multiple Sclerosis Journal 16, 1134-1140 (2010).
Nuber-Champier, A. et al. Acute TNFα levels predict cognitive impairment 6-9 months after COVID-19 infection. Psychoneuroendocrinology, 106104 (2023).
Nuber-Champier, A. et al. Monocytosis in the acute phase of SARS-CoV-2 infection predicts the presence of anosognosia for cognitive deficits in the chronic phase. Brain, Behavior, & Immunity-Health 26, 100511 (2022).
P. Voruz, A. C., I. Jacot de Alcântara, A. Nuber-Champier, G. Allali, L. Benzakour, M. Thomasson, P. Lalive, K-O. Lövblad, O. Braillard, M. Nehme, M. Coen, J. Serratrice, J. Pugin, I. Guessous, B.N. Landis3,10, D. Adler, A. Griffa, D. Van De Ville, F. Assal & J.A. Péron. Functional connectivity underlying cognitive and psychiatric symptoms in post-COVID syndrome: is anosognosia a key determinant? . Brain communications (accepted).
Hartung, T. J. et al. Fatigue and cognitive impairment after COVID-19: A prospective multicentre study. EClinicalMedicine 53 (2022).
Billones, R., Liwang, J. K., Butler, K., Graves, L. & Saligan, L. N. Dissecting the fatigue experience: A scoping review of fatigue definitions, dimensions, and measures in non-oncologic medical conditions. Brain, behavior, & immunity-health 15, 100266 (2021).
Taquet, M. et al. Acute blood biomarker profiles predict cognitive deficits 6 and 12 months after COVID-19 hospitalization. Nature medicine, 1-11 (2023).
Monje, M. & Iwasaki, A. The neurobiology of long COVID. Neuron (2022).
Voruz, P. et al. Functional connectivity underlying cognitive and psychiatric symptoms in post-COVID-19 syndrome: is anosognosia a key determinant? Brain Communications 4, fcac057 (2022).
McAfoose, J. & Baune, B. Evidence for a cytokine model of cognitive function. Neuroscience & Biobehavioral Reviews 33, 355-366 (2009).
Altmann, D. M., Whettlock, E. M., Liu, S., Arachchillage, D. J. & Boyton, R. J. The immunology of long COVID. Nature Reviews Immunology 23, 618-634 (2023).
Nuber-Champier, A., Voruz, P., Jacot de Alcântara, I., Breville, G., Allali, G., Lalive, P. H., Assal, F., & Péron, J. A. . Monocytosis in the acute phase of SARS-CoV-2 infection predicts the presence of anosognosia for cognitive deficits in the chronic phase. Brain, behavior, & immunity - health 26, doi:https://doi.org/10.1016/j.bbih.2022.100511 (2022).
Peluso, M. J., Abdel-Mohsen, M., Henrich, T. J. & Roan, N. R. in Seminars in Immunology. 101873 (Elsevier).
Mavrikaki, M., Lee, J. D., Solomon, I. H. & Slack, F. J. Severe COVID-19 is associated with molecular signatures of aging in the human brain. Nature Aging, 1-8 (2022).
Li, C., Liu, J., Lin, J. & Shang, H. COVID-19 and risk of neurodegenerative disorders: A Mendelian randomization study. Translational psychiatry 12, 1-6 (2022).
Kervevan, J. et al. Divergent adaptive immune responses define two types of long COVID. Frontiers in Immunology 14, 1221961 (2023).
Hammerle, M. B. et al. Cognitive complaints assessment and neuropsychiatric disorders after mild COVID-19 infection. Archives of Clinical Neuropsychology 38, 196-204 (2023).
Miller, S. L. & Maner, J. K. Overperceiving disease cues: the basic cognition of the behavioral immune system. Journal of personality and social psychology 102, 1198 (2012).
Martin, E. M. et al. A hypoarousal model of neurological post-COVID syndrome: the relation between mental fatigue, the level of central nervous activation and cognitive processing speed. Journal of Neurology 270, 4647-4660 (2023).
Campos, M. C. et al. Post-viral fatigue in COVID-19: A review of symptom assessment methods, mental, cognitive, and physical impairment. Neuroscience & Biobehavioral Reviews 142, 104902 (2022).
Voruz, P. et al. Persistence and emergence of new neuropsychological deficits following SARS-CoV-2 infection: A follow-up assessment of the Geneva COVID-COG cohort. Journal of Global Health 14 (2024).
Matias-Guiu, J. A. et al. Neuropsychological predictors of fatigue in post-COVID syndrome. Journal of Clinical Medicine 11, 3886 (2022).
Voruz, P. et al. Frequency of abnormally low neuropsychological scores in post-COVID-19 syndrome: the Geneva COVID-COG cohort. Archives of Clinical Neuropsychology (2022).
Debouverie, M., Pittion-Vouyovitch, S., Louis, S. & Guillemin, F. Validity of a French version of the fatigue impact scale in multiple sclerosis. Multiple Sclerosis Journal 13, 1026-1032 (2007).
Zimmermann, P. & Fimm, B. Test for attentional performance (TAP). PsyTest, Herzogenrath 1995, 76-77 (1995).
Claros-Salinas, D. et al. Induction of cognitive fatigue in MS patients through cognitive and physical load. Neuropsychological rehabilitation 23, 182-201 (2013).
Neumann, M. et al. Modulation of alertness by sustained cognitive demand in MS as surrogate measure of fatigue and fatigability. Journal of the Neurological Sciences 340, 178-182 (2014).
Sb, C. H., Browner, W., Grady, D., Newman, T. & Gaertner, R. Designing clinical research: an epidemiologic approach. (2013).

The authors have declared there is NO conflict of interest to disclose

SupplXXmentaryinformation.docx

Impact of inflammatory response in the acute phase of COVID-19 on predicting objective and subjective post-COVID fatigue

Status:

Version 1

Abstract

Figures

1 Introduction

2 Results

2.1 Sociodemographic and clinical data

2.2 Cytokines plasma levels and monocyte % measured during the acute phase of COVID-19

3 Discussion

4 Method

4.1 General procedure

4.2 COVID-COG cohort

4.3 Participants included in the study

4.4 Measurement of subjective fatigue

4.5 Measurement of objective cognitive fatigue

4.6 Self-appraisal discrepancy (SAD) of cognitive fatigue

4.7 Analysis of cytokines and monocytes

4.8 Statistical power

4.9 Statistical analysis

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1