Fatigue can influence the development of late‐onset pain in post‐COVID‐19 syndrome: An observational study

Late‐onset pain is frequent following COVID‐19, and many pathogenetic mechanisms have been proposed. Identifying the main features of patients may help in designing tailored rehabilitative interventions.


| INTRODUCTION
Coronavirus disease (COVID-19) is an infection caused by the SARS-CoV-2 virus resulting in various pathology phenotypes characterized by different symptoms severities.It is currently a global healthcare issue with disabling consequences on people's condition and quality of life (Rahman et al., 2021;World Health Organization, 2023).After acute symptoms resolution, long-term effects such as fatigue, dyspnoea, cardiovascular, and neuropsychiatric disorders were observed in many subjects, the so-called "postacute sequelae of COVID-19" (Nalbandian et al., 2021).
Causes of pain persistence after COVID-19 infection are poorly established, and different pathogenetic mechanisms have been proposed.Nociceptor sensitization may be due to extensive systemic hyperinflammation following COVID-19 infection (McFarland et al., 2021).Furthermore, the impact of cytokine storms after severe disease must be considered a potential driving factor for developing neuropathies and may contribute to the development of pain after acute COVID-19 infection resolution (McFarland et al., 2021).The sustained increase in central cytokines and chemokines leads to widespread pain at several body locations, suggesting that neuroinflammation drives pain (Goudman et al., 2021).
A useful screening tool to identify late-onset post-COVID-19 pain could be the COVID-19 Yorkshire Rehabilitation Scale (C19-YRS) (O'Connor et al., 2022).The C19-YRS represents a multidisciplinary assessment of post-COVID-19 syndrome (PCS), based on symptoms and functional difficulties reported by COVID-19 survivors (Halpin et al., 2021).The C19-YRS was already used in a cohort of Italian COVID-19 inpatients and was able to differentiate new-onset pain from pre-infection conditions (Straudi et al., 2022), without providing any information regarding the nature of pain.The emergence of precision medicine represents the gold standard in post-COVID-19 pain management (Fernández-de-las-Peñas et al., 2023) and the identification of a specific pain mechanism is necessary to provide a tailored rehabilitative intervention (Fernández-de-las-Peñas, Nijs, et al., 2022).
Several predisposing factors have been proposed as characterizing people with post-COVID-19 pain (Fernández-de-las-Peñas et al., 2023) but no consensus was reached about mechanism underlying pain development in the chronic phase.Recently, Fernández-de-las-Peñas et al. tried to define objective criteria to characterize pain attributes in post-COVID-19 subjects, identifying three major phenotypes (nociceptive, neuropathic, and nociplastic pain) based on diagnostic and symptoms presentation (Fernández-de-las-Peñas, Nijs, et al., 2022).Central sensitization as pathologic mechanism of nociplastic pain has already been hypothesized in post-COVID-19 patients (Goudman et al., 2021;McFarland et al., 2021) and the presence of pain hypersensitivity can further support this idea (Kosek et al., 2021).However, no objective measures of altered pain processing have been tested in post-COVID-19 patients and the presence of hyperalgesia and allodynia may contribute to predict sensitization mechanisms (Baroni et al., 2020).For these reasons, the primary aim of this paper was to identify possible demographic and pathological predisposing factors of late-onset pain in people 1 year after COVID-19 infection.Furthermore, we aimed to underline clinical characteristics linked to different subgroups of patient with post-COVID-19 pain, identifying pathological signs related to a prevalent pain mechanism.

| METHODS
This observational study was approved by the local Ethical Committee (Comitato Etico di Area Vasta Emilia Centro, CE-AVEC, protocol code: EM66 2022_539/2020/ Oss/AOUFe_EM1) and registered on Clini caltr ials.gov (NCT04615390).The "Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement" for cohort studies was used as reporting guidance (von Elm et al., 2014).
We enrolled a convenience sample of patients aged over 18 hospitalized at the Ferrara University Hospital with a diagnosis of COVID-19 (World Health Organization, 2020), with rehabilitation needs during the acute phase, and with an increase in pain intensity (estimated by a gain of ≥2 points at the C19-YRS) at 52 weeks from the infection's onset compared to the pre-COVID-19 condition.We excluded patients who presented severe cognitive impairments or could not give informed consent.We also excluded patients who presented re-infection by SARS-CoV-2 virus or other medical conditions that need hospitalization during the time of observation.A group of healthy subjects matched for age and sex was also included.
The research team screened patients, and informed consent was requested after explaining the study procedures.In order to identify predisposing factors of late-onset pain, the following demographic and anamnestic data were collected: age, sex, disease severity (World Health Organization, 2020), length of stay (LOS) in the acute phase, admission or not to Intensive Care Units (ICU), hypertension, obesity, dyslipidaemia, chronic respiratory diseases (asthma, chronic obstructive pulmonary disease), gastropathy, diabetes, Critical Illness Neuro-Myopathy (CINM) as a complication during the acute stay, chronic medications for pain management.Furthermore, all the subjects were monitored through periodic screening of PCS using C19-YRS at 12, 26, and 52 weeks after symptoms onset.We focused on the subgroup of items potentially informative or linked to the pain experience; thus, we registered scores related to fatigue, pain/discomfort, anxiety, depression, mobility, usual activities, and global health.

| Assessment of pain and related signs and symptoms
For the subjects with an increase in pain intensity of two or more points at 52 weeks, a specific pain assessment was performed with a view to identifying a prevalent pain mechanism.The instruments used were: (i) a body chart where the patient has to identify all the painful areas.(ii) the Numeric Pain Rating Scale (NPRS), an 11-point numeric scale ranging from 0 to 10 where higher scores indicate greater pain intensity (Hawker et al., 2011).(iii) the Leeds Assessment of Neuropathic Symptoms and Signs (LANSS), a pain questionnaire with dichotomous responses and sensory testing.The LANSS maximum score is 24 and a score of ≥12 indicates a possible neuropathic mechanism (Bennett, 2001).(iv) The Central Sensitization Inventory (CSI), a self-report questionnaire used to investigate the presence of central sensitization (Mayer et al., 2012).The patient scores each answer on a scale of 0 (never) to 4 (always); the CSI score ranges from 0 to 100 and a score of ≥40 indicate the presence of central sensitization (Mayer et al., 2012).(v) The Pain Catastrophizing Scale (PCS), a self-report questionnaire where patients have to rate how often they experience specified painrelated feelings on a 5-point Likert scale, from 0 (not at all) to 4 (always) (Sullivan et al., 1995).The PCS score ranges from 0 to 52 and a score of >30 represents a clinically significant level of pain catastrophization (Sullivan et al., 1995).(vi) The Tampa Scale of Kinesiophobia (TSK), a self-report checklist where the patient assesses on a 4-point Likert scale his fear of movement.The TSK score ranges from 17 to 58 and a score of ≥37 suggest the presence of kinesiophobia (Dupuis et al., 2023).The evaluation of the pressure pain threshold (PPT) and temporal summation (TS) through an algometer (Commander Algometer, JTECH Medical) on the upper trapezius and rectus femoris bilaterally was performed in COVID-19 patients and age-and sex-matched controls to analyse the nature and characteristics of pain.All the tests were administered with the subjects lying down; for the PPT, four stimuli were provided with an interstimulus interval of 10 s in each area, while for TS, ten stimuli were given with an interval of 5 s in the painful side only (for the control group, in the dominant side).
In the patients included, we also analysed the psychological well-being at 52 weeks from the infection's onset through: (i) the Beck Anxiety Inventory (BAI), a self-report questionnaire for measuring the severity of anxiety (Beck et al., 1988).Each answer can be scored on a scale value from 0 (not at all) to 3 (severely); the BAI score ranges from 0 to 63 and higher total scores indicate more severe anxiety symptoms (Beck et al., 1988).(ii) The Patient Health Questionnaire-9 (PHQ-9), a self-report questionnaire for measuring depression severity (Kroenke et al., 2001).Frequency of depressive symptoms can be scored on a scale value from 0 (not at all) to 3 (nearly every day); the PHQ-9 score ranges from 0 to 27 and higher total scores indicate more severe depressive symptoms (Kroenke et al., 2001).(iii) The Impact Event Scale-Revised (IES-R), a self-report measure of subjective distress caused by traumatic events (Weiss & Marmar, 1997).Every item can be rated on a 5-point scale ranging from 0 ("not at all") to 4 ("extremely"); the IES-R score ranges from 0 to 88 and a score of ≥33 indicates a probable Post Traumatic Stress Disorder (PTSD) (Creamer et al., 2003).

| Statistical analysis
Descriptive statistics were used to characterize the sample.Continuous variables were reported as the means and standard deviations, and dichotomous variables were reported as counts and percentages.Baseline characteristics were compared among people with and without pain using the Wilcoxon rank-sum or Pearson's chi-square test.Subjects with pain were divided according to the pain mechanism, and the presence of psychological disorders and differences between groups were analysed.The Wilcoxon rank-sum test was used to test differences in PPT and pain intensity at the first and tenth stimulation of the TS test in people with post-COVID-19 pain and healthy subjects.C19-YRS domain scores were compared between groups at the three assessment points using the Wilcoxon rank-sum test due to non-normal data distribution.Multiple regression model was run to predict pain development at 52 weeks from C19-YRS domains at the 12 and 26 week assessment between people with and without pain, as well as the effect of age and sex.Statistical analysis was performed using STATA 13.1; the significance level was set to 0.05.

| RESULTS
One hundred thirteen inpatients with rehabilitative needs after COVID-19 infection were considered for eligibility.Sixty-seven patients completed the C19-YRS evaluation for post-COVID-19 symptoms at 52 weeks (mean age = 66.4 ± 10.6 years, 45 males and 22 females; no one vaccinated against COVID-19).Of these, 20 subjects presented increased pain intensity ≥2 points at the 52-week C19-YRS pain assessment compared to the pre-COVID condition and the 12-and 26-week evaluations (Figure 1).Looking at the individual trend of pain, 14 patients who showed an increase in pain intensity at 52 weeks showed a maintenance or reduction of pain intensity between 12 and 26 weeks and developed de novo pain at 52 weeks (three of them started without pain before infection).Conversely, 17 of the patients without pain at 52 weeks showed pain intensity different from zero at 12 and/or 26 assessments.After 52 weeks, subjects with and without pain were similar in demographic and clinical characteristics (Table 1).Comparison of C19-YRS domains at the three time points (12, 26, 52 weeks) revealed significant differences between subjects with and without pain at 26 weeks that were not recorded in the 12-week assessment.Remarkably, subjects with pain showed significantly worse outcomes in terms of fatigue, anxiety, mobility, and ability to perform usual daily activities and generally decreased global health perception.The decrease in all these domains, except for usual activities, was retained at the 52-week evaluation (Figure 2).Multiple linear regression was used to test whether C19-YRS domain scores at 12 and 26 weeks, as well as age and sex, significantly predicted the onset of pain approximately 1 year following infection.The overall regression was statistically significant for C19-YRS domain scores at 26 weeks (R 2 = 0.38, F (7, 58) = 5.16, p < 0.000); fatigue significantly predicted pain onset (β = 0.54, p = 0.002), no effects of age (β = 0.07, p = 0.694) and sex (β = 0.87, p = 0.243) were recorded.No significance was found for C19-YRS domain scores at 12 weeks, age and sex (R 2 = 0.19, F (7, 58) = 1.92, p = 0.082).
Eighteen of the 20 subjects with pain at 52 weeks agreed to participate in the specific pain assessment to identify possible pain mechanisms and clinical features of the symptoms.Most patients presented with more than one painful body site.Sixteen different body regions were identified as painful by the sample, with a frequent bilateral presentation.The most frequent body regions involved in pain were the shoulder, knee, and lumbar spine (Figure 3).A mean intensity of pain of 6.0 ± 1.9 was recorded between subjects using the NPRS, with a minimum of 3 and a maximum of 10; most of the sample did not show possible neuropathic or nociplastic mechanisms evaluated through the LANSS and CSI, respectively (Figure 4).The results of the clinical and psychological assessments are reported in Table 2.We did not find significant differences in pain intensity by stratifying F I G U R E 1 C19-YRS pain score before COVID-19 infection and at the three-time follow-up (12, 26, and 52 weeks) in the two samples.* p < 0.01.patients according to LANSS and CSI scores and intensity of psychological disorders (Table 2).Comparing subjects with pain and a sample of age-and sex-matched healthy subjects revealed no differences in pressure pain threshold and temporal summation of pain (Table 3).
Pain is one of the most disabling consequences of COVID-19 (Yang et al., 2022).Our study found that almost one out of three patients hospitalized for COVID-19 developed pain 52 weeks after symptoms resolution.Increased pain sensation following viral infection is not a prerogative of COVID-19 (Kawai et al., 2014;Watson et al., 2020;Wimalasiri-Yapa et al., 2020).Nevertheless, the novelty of our funding is that pain intensity fluctuates during the first year following COVID-19 infection, with an increase after 12 weeks, a reduction at 26 weeks, and a new increase at 52 weeks.A recent meta-analysis showed that almost 10% of people infected by COVID-19 experience pain during the first year following infection, and an increase in pain intensity was recorded 60 days after symptoms onset and a reduction 180 days later (Fernández-de-Las-Peñas, Navarro-Santana, et al., 2022).The symptoms in our sample seem to be different from typical pain with a sensitization mechanism.Following IASP criteria for nociplastic pain, it can be defined as such if symptoms last at least three months (Kosek et al., 2021).In our case, even when pain was reduced in the earlier phases post infection resolution (12 and 26 weeks), pain intensity increased at 52 weeks.We can speculate that the development of pain long after COVID-19 resolution may be due to new mechanisms developed months after infection that are not imputable to inflammatory agents responsible for nociceptive pathway stimulation and central sensitization (Fernández-de-Las-Peñas, Cancela-Cilleruelo, et al., 2022).People with and without pain at the oneyear assessment were similar in demographic and clinical features.Nevertheless, they differed at the 26-week evaluation in fatigue perception, anxiety, general mobility, and ability to perform daily activities.Our analysis shows that fatigue is the factor most responsible for developing late-onset pain.The relationship between pain and fatigue was previously investigated in people with chronic pain syndromes, and a unidirectional connection was proposed (Van Damme et al., 2018).However, the possibility that fatigue may be involved in pain onset was already hypothesized, and a bidirectional  relationship cannot be ruled out (Lenaert et al., 2018).Pain perception in our sample seems to be influenced by fatigue.A reduced pain threshold was recorded in people suffering from chronic fatigue syndromes, probably due to sensitization of the central nervous system (Nijs et al., 2012;Winger et al., 2014).Quantitative sensory testing applied to our sample did not reveal hyperalgesia compared to a cohort of healthy subjects, helping to undermine idea that one-year post-COVID-19 pain may be due to sensitization mechanisms.Furthermore, pain was not recorded during the 26-week assessment, suggesting that fatigue may predict pain development.This causal relationship may open the doors to new treatment approaches in pain management, targeting fatigue for late-onset pain treatment.
Another sign of a nociplastic mechanism of pain is the regional distribution of symptoms (Kosek et al., 2021).Pain drawings of subjects who reported pain at 52 weeks showed a discrete distribution of painful sites, ruling out the idea of widespread pain.
The attempt to classify subjects as affected by central sensitization or neuropathic pain through CSI and LANSS did not provide a univocal interpretation.Looking at the scores achieved by people with pain at 52 weeks, most of the sample can be classified as affected by "not probably" nociplastic or neuropathic pain.Moreover, the subjects who showed lower CSI and LANSS scores had higher pain intensity.The idea that only one causal mechanism could explain pain is still highly controversial.Recently, Fernández-de-las-Peñas et al. tried to phenotype post-COVID-19 pain to tailor the most successful treatment (Fernández-de-las-Peñas, Nijs, et al., 2022).However, distinguishing between nociceptive, nociplastic, or neuropathic pain is challenging, and overlapping pain mechanisms will likely make classification more complex (Shraim et al., 2021).Looking at our sample, the presence of one phenotype, as evaluated through CSI or LANSS, does not exclude another.Furthermore, the CSI score is used to assess symptoms of central sensitization and not to diagnose nociplastic pain, whose identification should always be made through the assessment of all the criteria identified by IASP researchers (Nijs et al., 2021).
The fourth requirement that should be satisfied for classifying the subject as having "possible nociplastic pain" is the presence of pain hypersensitivity.Assessment of the pressure pain threshold and temporal summation of pain in our sample did not reveal significant differences compared to age-and sex-matched healthy subjects.Similar findings were recently found by Peterson et al., who assessed pain sensitivity in symptomatic and asymptomatic individuals affected by COVID-19 (Peterson et al., 2022).No differences were found between them and a control group of healthy subjects, confirming that previous COVID-19 infections have no effects on pressure pain sensitivity.Compared to Peterson's trial, our sample was made of older subjects with more severe symptoms, suggesting that age and severity of infection did not affect pain sensitivity.The assessment of pain-related disorders in people with post-COVID-19 pain was inconclusive.The analysis of behavioural and psychological aspects of the sample revealed that only the minority showed psychosocial factors typically related to central sensitization, such as psychological disorders, kinesiophobia, and catastrophizing.All these psychosocial factors are frequently observed in people with chronic pain syndromes and are mainly related to the intensity and maintenance of pain and disability (Rogers & Farris, 2022).The uncertainty about the presence of typically chronic pain-related disorders seems to sustain the idea about the doubtful sensitization mechanism behind post-COVID pain.
All these findings support the idea that post-COVID-19 pain cannot be classified as nociplastic.The identification of a "post-COVID pain" model, as suggested by Fernández-de-las-Peñas, needs to consider all the components of the pain related to different overlapping phenotypes (Fernández-de-Las-Peñas, Cancela-Cilleruelo, et al., 2022).Evaluating possible predisposing factors such as fatigue, general impaired mobility, and inability to perform usual daily activities a time before pain onset may help identify subjects at risk of developing pain after 1year.A comprehensive approach following a biopsychosocial model must consider all possible factors related to pain development, acting on the components of a vicious circle where fatigue and mood disorders play a crucial role in pain development and maintenance.It is noted that all post-COVID-19 rehabilitation programmes aim to manage symptoms related to respiratory disease to reduce fatigue and breathlessness.Nevertheless, introducing aerobic and conditioning strategies may help prevent pain development.Our findings on the management of pain syndromes with a long-lasting duration of symptoms combined with fatigue and reduced ability to perform usual daily activities may open the doors to new ways of prevention and treatment.
Interpretation of our findings must take into account several limitations.The use of PPT and temporal summation as only objective evaluation of pain cannot completely exclude the presence of other neurological conditions as small fibre neuropathy.Furthermore, the simultaneous occurrence of pain and fatigue may be related to autonomic nervous system dysfunction, a frequent consequence of SARS-CoV-2 virus infection (Lo, 2021).Dysautonomic symptoms are frequent following viral infections, and approximately 2.5% of post-COVID-19 patients suffer from autonomic dysfunction (Carmona-Torre et al., 2022).Despite this, we didn't use any specific screening tool for testing the presence of dysautonomia in our sample (Bisaccia et al., 2021).Finally, the small sample size does not allow us to draw definitive conclusions on the mechanism underlying one-year post-COVID-19 pain.Although an overlapping of different phenotypes may be hypothesized, it should be verified in a more significant cohort of subjects.

| CONCLUSION
This observational study aimed to assess the incidence of one-year post-COVID-19 pain to identify potential contributing factors.Almost one out of three patients hospitalized for COVID-19 infection developed pain 1 year later.Fatigue seems to be the most predisposing factor in late-onset pain, but an overlapping of different mechanisms may be identified as responsible for post-COVID pain.Therefore, a comprehensive approach following a biopsychosocial model must be considered for patients' therapeutic management.Further studies are needed to verify our findings, but the results of our research may open the doors to new strategies for the prevention and treatment of patients where pain is combined with fatigue and a reduced ability to perform daily activities.T A B L E 3 Comparison of QST between people with pain at 52 weeks and healthy controls.

F
I G U R E 2 C19-YRS score at threetime follow-up.* p < 0.05.

without pain at 52 weeks (n = 47) People with pain at 52 weeks (n = 20) Total (n = 67) p
Demographic and clinical details of the sample.
T A B L E 1Abbreviation: CINM, critical illness neuromyopathy.

People with pain at 52 weeks Healthy controls p
: QST, quantitative sensory testing; PPT, pressure pain threshold; TS, temporal summation. Abbreviations