Exploring multidimensional characteristics in cervicogenic headache: Relations between pain processing, lifestyle, and psychosocial factors

Abstract Objective Although multidimensional interventions including physiotherapy, psychology, and education are generally recommended in managing headache, and to prevent chronification, such approach is lacking in cervicogenic headache (CeH). Therefore, exploring CeH within a biopsychosocial framework is deemed an essential first step. Methods Non‐randomized cross‐sectional design to compare pain processing, lifestyle, and psychosocial characteristics between 18 participants with CeH (CeH group) (40.2 ± 10.9 years) and 18 matched controls (control group) (39.2 ± 13.1 years). Pain processing characteristics included degree of central sensitization (Central Sensitization Inventory), and (extra)‐cephalic pressure pain thresholds (kPa/cm²/s). Lifestyle characteristics included sleep quality (Pittsburgh Sleep Quality Index), physical activity, screen time, and sedentary time (hours a week). Psychosocial characteristics included degree of depression, anxiety and stress (Depression Anxiety Stress Scale‐21), and quality of life (Headache Impact Test‐6). Results Pain processing characteristics: More (p = .04) participants in the CeH group showed higher degrees of central sensitization compared to the control group. Lower (p < .05) (extra)‐cephalic pressure pain thresholds were revealed in the CeH group compared to the control group for each muscle. Lifestyle and psychosocial characteristics: Compared to the control group, sleep quality and headache‐related quality of life were worse (p < .0001) in the CeH group. Severe to extreme stress was experienced by more participants in the CeH group (p = .02). Further, significant relations between pain processing and (1) lifestyle characteristics and (2) psychosocial characteristics were seen in the CeH group. Conclusion Exploring multidimensional characteristics in CeH exposed relations between pain processing, lifestyle, and psychosocial characteristics. These novel findings fill a gap in the current scientific literature, and highlight the need for outcome research targeting lifestyle and psychosocial factors.


INTRODUCTION
The International Classification of Headache Disorders 3 (ICHD) defines cervicogenic headache (CeH) as: "Headache caused by a disorder of the cervical spine and its components (bony, disc and/or soft tissue elements) usually but not invariably accompanied by neck pain" (Headache Classification Committee of the International Headache Society, 2013). General recommendations for the nonpharmacological non-invasive management of CeH can be summarized as: Manual therapy applied to the cervical and thoracic spine whether or not combined with low load endurance exercises, trigger point therapy, combination of static and dynamic cervical, scapulothoracic strengthening and endurance exercises, and low-load craniocervical and cervicoscapular endurance exercises (Côté et al., 2019;Fernandez et al., 2020;Gross et al., 2016;Luedtke et al., 2016;Varatharajan et al., 2016). Despite these recommendations, there is limited evidence regarding the effectiveness of such interventions on pain intensity, frequency, duration, and disability in the long-term (Fernandez et al., 2020). Further, results from a systematic review with meta-analysis revealed that pooling results of studies on manually managing CeH did not support significant effects of such therapies on headacheintensity, -disability, and -frequency (Coelho et al., 2019). Based on these inconsistencies it was already proposed that not all patients with CeH benefit from particular interventions, and that not all interventions are appropriate (Fernández-De-Las-Peñas & Cuadrado, 2014).
Although non-pharmacological interventions could play an important role in managing CeH, scientific evidence for success of such interventions remains contradictory (Coelho et al., 2019;Falsiroli Maistrello et al., 2019). One in four patients with CeH is non-responsive to musculoskeletal physiotherapy. According to physiotherapists, such non-responsiveness can be related to augmented pain processing in the central nervous system, and presence of negative psychosocial factors (Liebert et al., 2013;Moore et al., 2017).
It was recently stated that local therapy, exclusively addressing the upper-cervical spine, was ineffective if signs of central sensitization (CS) are already present. In such state, local therapy might act as a minimal peripheral nociceptive stimulus maintaining CS (Fernández-De-Las-Peñas et al., 2020;Woolf, 2011). Headaches which are merely mediated by a peripheral nociceptive source, also known as bottom-up sensitizer (e.g., musculoskeletal disorder of the upper-cervical), can be addressed by targeting that peripheral source through manual therapy and/or local exercises (Fernández-De-Las-Peñas et al., 2020). However, if additionally central mechanisms, also known as top-down sensitizers (e.g., anxiety, stress), are involved, clinical management needs to shift towards a multimodal approach including physical, psychological, cognitive, lifestyle and educational dimensions directed at normalizing CS (Bialosky et al., 2018;Fernández-De-Las-Peñas et al., 2020). Such dimensions are known influencers on therapy-responsiveness in other musculoskeletal syndromes such as non-specific chronic low back pain (O'sullivan, 2012). Integrating both bottom-up and top-down interventions in order to modulate pain at different levels of the nervous system therefore seems advised in complex pain disorders such as CeH (Fernández-De-Las-Peñas et al., 2020;Lluch Girbés et al., 2015).
Disturbed pain processing has already been reported in patients with chronic CeH (Chua et al., 2011). However, it is yet to determine if such disturbances are also present in patients with episodic CeH, and the factors which influence it. Identifying maladaptive pain processing in episodic CeH is relevant in the context of headache chronification since headache-frequency is related to sensitization (Buchgreitz et al., 2006).
Yet, scientific results on relations between pain processing, lifestyle and psychosocial characteristics are, to the best of our knowledge, lacking in episodic CeH. Such relations however need to be explored since they relate to chronification of pain (Borsook et al., 2018). To fill this gap, we propose to evaluate patients with episodic CeH within a biopsychosocial framework. Therefore, the aim of this study was to analyze relations between pain processing, lifestyle, and psychosocial characteristics in patients with episodic CeH.

Design
This study is designed as follows: Non-randomized cross-sectional comparison of pain processing, lifestyle, and psychosocial characteristics between a CeH group and matched healthy control group; Analysis of relations between pain processing, lifestyle, and psychosocial characteristics within a CeH group and matched healthy control group.

Sample size
Sample size calculation was not performed to establish a treatmenteffect, but rather to assess feasibility in the context of an exploratory study. A priori sample size calculation (G*Power 3.1.9.4, Kiel Germany) to detect differences between groups, based on cervical pressure pain thresholds (PPT) (mean, standard deviation), resulted in a required sample size of 15 participants per group (power 80%; α = .05) (Chua et al., 2011).

Diagnostic criteria
A Any headache fulfilling criterion C B Clinical and/or imaging evidence of a disorder or lesion within the cervical spine or soft tissues of the neck, known to be able to cause headache C Evidence of causation demonstrated by at least two of the following: 1. Headache has developed in temporal relation to the onset of the cervical disorder or appearance of the lesion 2. Headache has significantly improved or resolved in parallel with improvement in or resolution of the cervical disorder or lesion 3. Cervical range of motion is reduced and headache is made significantly worse by provocative maneuvers 4. Headache is abolished following diagnostic blockade of a cervical structure or its nerve supply D Not better accounted for by another ICHD-3 diagnosis

Participants and ethics
The neurological staff of the headache department of the AZ Vesal-

Outcomes, measurements, and instruments
Headache, pain processing, lifestyle, and psychosocial characteristics were the outcomes of interest.

Headache characteristics
Headache characteristics, that is, headache-intensity (mean 100 mm Visual Analogue Scale (VAS) per attack during a month), duration (mean hours per attack during a month), and frequency (days per month) were extracted from the headache diary of the Belgian Headache Society (2020). Additional information on referred pain from the neck (yes/no) was obtained through a customized questionnaire and anamnesis.

Lifestyle characteristics
Sleep quality was assessed via the Dutch Pittsburgh Sleep Quality Index (PSQI) which is a standardized, valid, and reliable self-reported onemonth recall questionnaire (Marinus et al., 2003;Mollayeva et al., 2016). We refer to the systematic review and meta-analysis by Mol- indicates poor sleep quality (Buysse et al., 1989;Smyth, 2008).

Physical activity (mean hours a week), screen-time (mean hours a
week computer-use), and sedentary-time during free-time and work (mean hours a week) were extracted from a customized one-week recall questionnaire.

Impact of headache on quality of life was assessed with the Dutch
Headache Impact Test-6 (HIT-6) (Buse et al., 2012). The HIT-6 shows good test re-test reliability (ICC 0.80), and internal consistency (Cronbach's α 0.79) (Kosinski et al., 2003;Martin et al., 2004). The HIT-6 evaluates the impact of headache on daily activities: the ability to function at work, school, home, and in social situations. Questions are answered by: "never," "rarely," "sometimes," "very often," and "always," scores are 6, 8, 10, 11, and 13, respectively. The impact of headache on daily life depends on the total score, which varies between 36 and 78. Scores are interpreted as follows: ≤49 means no to little impact of headache on daily life, between 50 and 55 indicates that headache seems to affect daily life, between 56 and 59 means headache has a significant impact on daily life, and ≥60 indicates that headache has a very heavy impact on daily life (Kawata et al., 2005;Martin et al., 2004).

Test procedure
A condition to be measured was a score of <3 on the 11-point Numeric Pain Rating Scale on the test day. Participants were asked not to take analgesics, muscle relaxants, and caffeine-containing beverages Questionnaires were completed before PPT measurements. The questionnaires used allow all kinds of health practitioners (e.g., physiotherapists) to screen for psychosocial risk factors in patients with musculoskeletal pain (Matheson et al., 2020).  (Chesterton et al., 2007;Walton et al., 2011). The principal researcher performed the test procedure for both the CeH group and control group. All outcomes were evaluated by the principal researcher (physiotherapist with degree in manual therapy, >10-year experience).  (Rosenthal, 1996). sleep duration (hours/night) were presented as means (SD), and compared between groups by using the Mann-Whitney test.

Research
Research Question 3: Are headache, pain processing, lifestyle, and psychosocial characteristics related in both groups?
Linear regression models (simple, multiple, ordinal) or contingency tables (Fisher's exact test) were composed depending on the dependent and independent variables, and the conditions (cfr. described above) to analyze possible relations. Variables were selected based on a priori hypotheses and entered in the regression model, leading to many hypotheses being tested. A backward stepwise approach was used to downsize the model and obtain the best model fit (i.e., smallest mean square of the error, VIF < 4).

Demographics and group characteristics (Table 2)
Demographics and group characteristics were comparable between groups. Age, BMI, their interaction (age*BMI), level of education, and employment did not significantly influence pain processing, lifestyle, or psychosocial outcome variables (appendix a).

3.2
Headache characteristics (Table 3) Participants with CeH suffered from an episodic, moderate to severe intense headache with a mean duration of 4.1 hours/episode. Referred pain from the neck was reported by all 18 participants. Headache or neck pain was not present at the start of the test procedure.

3.3
Pain processing characteristics (Table 4) CSI: No significant differences were seen between groups for the ordi-

Lifestyle and psychosocial characteristics (Table 5)
Lifestyle characteristics: The recommended level of weekly physical activity (i.e., "high" in Table 5  of participants in the CeH group compared to 0% of participants in control group.
No significant differences between the groups were seen concerning the DASS-depression and DASS-anxiety.

3.5
Relations between headache, pain processing, lifestyle, and psychosocial characteristics (Table 6) Following Table 6 summarizes significant relations between: Lifestyle and headache characteristics, psychosocial and headache characteristics, pain processing and headache characteristics in the CeH group, and between pain processing and lifestyle characteristics, psychosocial, and pain processing characteristics in both groups.

DISCUSSION
The aim of this study was to analyze relations between biopsychoso-  (Bezov et al., 2011;Chua et al., 2011). Indications for CS in the CeH group were further supported by the lower extra-cephalic PPTs compared to the control group, and by 77.8% of participants with CeH reporting at least mild symptoms of CS on the CSI (Neblett et al., 2015). It was already reported that higher scores on the CSI predict higher pain-related disability in patients with musculoskeletal disorders in a primary care setting (Tanaka et al., 2019). This finding was also reflected by the significant relation (p = .006, Fisher's exact test) between more reported symptoms of CS and passive responses (such as rest, analgesic intake or a combination of both) to a headacheattack by the CeH group in our study. Such passive self-management was previously reported by patients with chronic CeH in the Akershus study. The latter revealed that patients might have low expectations towards traditional medicine, and that CeH is falsely expected to be self-manageable (Kristoffersen et al., 2013). The general practitioner or neurologist will only be consulted in case of persistent pain or functional limitations. Such behavior might contribute to the chronification of CeH.
A footnote concerning the CSI is its interpretation. Results on the CSI do not refute or confirm CS, but merely are a representation of general distress (Kregel et al., 2018). This consideration is clinically relevant since exclusively relying on the CSI to confirm CS might increase the risk of false-positive outcomes, which influences therapy approach.

4.2.1
Worse sleep quality, more stress and impact of headache on quality of life in the CeH group In the CeH group, the reported sleep quality and headache-related quality of life were worse, and the level of stress higher compared to in the control group. The combination of predominant pain, along with multiple comorbid features such as lowered extra-cephalic TA B L E 5 Lifestyle and psychosocial characteristics in the CeH group (n = 18) and control group (n = 18)
PPTs, stress, sleep problems, and a diminished quality of life in the CeH group might be indicative for an already disturbed pain process in participants with episodic CeH, and supports previous studies reporting that CS is not an exclusive feature of chronic headache (Bernstein & Burstein, 2012;Fumal & Schoenen, 2008;Staffe et al., 2019). Such disturbed pain processing is influenced by top-down lifestyle and psychosocial risk factors which means that these factors could maintain, and even drive a pain process (Staffe et al., 2019).

Conditioned pain modulation (CPM) and temporal summation
should however be assessed in episodic CeH to determine a nociceptive profile (Yarnitsky, 2015). Such profile, which can be either pro-or anti-nociceptive, might provide essential information for future patient-centered interventions (Lumley et al., 2011).

4.3
Are headache, pain processing, lifestyle, and psychosocial characteristics related in both groups?

4.3.1
Lifestyle and psychosocial characteristics relate to pain processing in the CeH group Thirty multidimensional quality indicators for headache care were previously agreed on (Peters et al., 2012). Although lifestyle and psychosocial factors were identified as such indicators, no study addresses these factors in patients with episodic CeH. In the CeH group, relations were seen between: (1) worse sleep quality, more stress, and more signs and reported symptoms of CS on the one hand and (2) between a lower level of physical activity, lower quality of life, more stress, and signs of peripheral sensitization on the other hand. Presence of both negative psychosocial characteristics such as stress and more symptoms of CS has already been proposed by physiotherapists to contribute to nonresponsiveness to treatment in patients with CeH (Liebert et al., 2013).
Further, negative psychosocial variables like depression, anxiety, and distress are among the most robust general predictors for transition from acute to chronic pain, for example tension-type headache, neck and back pain (Fumal & Schoenen, 2008;Lumley et al., 2011;Mills et al., 2019). Questioning the patient's self-reported symptom burden such as headache-frequency is another quality indicator (Mills et al., 2019). Headache-frequency is often used to evaluate therapy success within the context of chronification. This outcome measure is additionally used to estimate the degree of sensitization (Adams & Turk, 2015;Buchgreitz et al., 2006;Luedtke et al., 2016). Higher headache-frequency might lead to an increased pressure pain sensitivity, and with time CS could develop (Buchgreitz et al., 2006). CeH in our study is, based on the range of frequencies (10 to 15.8 days/month), merely an episodic headache. However, potential catalysts for chronification were detected. Both long sitting duration at work and anxiety were independently related to a higher headachefrequency in the CeH group. Most research on relations between headache-frequency and CS concerns primary headaches (e.g., tension-type headache, migraine). However, the novel key findings in our work pointing in the direction of disturbed pain processing, and previous preliminary work suggesting autonomic dysregulation, can be supportive arguments to justify future studies analyzing central nervous system involvement in episodic CeH (Mingels & Granitzer, 2020).
Nevertheless, more research is needed before our results can be transferred to the clinical practice. Based on recent work it is advised to desensitize the central nervous system in patients with CS by designing an individually tailored multimodal treatment plan comprising pain neuroscience education, cognition targeted exercise therapy, sleep management, stress management, and/or dietary intervention (Arendt-Nielsen et al., 2018;Nijs et al., 2014Nijs et al., , 2019Van Wilgen et al., 2014).

Limitations and suggestions
In this study, several statistical analyses were used requiring some caution when interpreting the results. No Bonferroni correction was applied to analyze simple relations (simple linear regression) among biopsychosocial characteristics themselves, and between these characteristics and headache characteristics. Such correction is not demanded since we examined the effect of one independent variable on the outcome of interest. Thus, only one hypothesis was tested at a time (i.e., no increase in the false positive error rate) (Andrade, 2019).
However, caution is needed to interpret and generalize results. Several variables were selected based on priori hypotheses and entered in the regression model, leading to many hypotheses being tested. The backward stepwise approach was used to downsize the model. The VIF was used in case two independent variables were related to the dependent variable.
The rather small sample size (n = 36) will tend to overestimate an effect. Post hoc power calculations for most relevant outcomes were: TA B L E 6 Summary of significant relations between headache, pain processing, lifestyle, and psychosocial characteristics in the CeH group (n = 18) and control group (n = 18)

Relation CeH group Control group
Lifestyle (independent) and headache characteristics (outcome) Sitting-time at work related to headache-frequency Screen-time related to headache-intensity Estimate . a Analysis of the relation between the nominal (severe and extreme vs. moderate) version of DASS-anxiety vs. headache-frequency. Higher headachefrequency was reported if the level of anxiety was severe to extreme compared to moderate. b Reporting more symptoms (i.e., moderate and hard) of CS was related to higher scores on HIT-6. c 33% of participants with poor sleep quality reported moderate to hard symptoms of CS, compared to 22% with optimal sleep quality reporting subclinical symptoms. d Being highly physical active was related to higher right suboccipital PPTs compared to being moderate physical active. e Shorter (i.e., 7-14 hours/week) screen-time was related to higher right suboccipital PPTs compared to longer (i.e., >21 hours/week) screen-time; Δ = right tibialis anterior PPTs were higher with a screen-time of 7-14 hours compared to >21 hours; ΔΔ = being low physical active was related to lower PPTs on the left tibialis anterior compared to being moderate and high physical active. f Higher CSI-scores (i.e., moderate and hard) were related to higher DASS-stress scores (i.e., moderate to extreme). Based on a contingency table, distributions of associations between HIT-6 and CSI differed significantly (Fisher's exact test p < .001) with 38.9% of participants with moderate symptoms of CS reporting headache had a severe impact on quality of life. g Distributions of associations between the CSI and PSQI differed; 44% of participants with optimal sleep quality reported no symptoms of CS.
A statistical correct interpretation for results on, for example, the CSI is that the null hypothesis will be erroneously rejected in 54% (=β probability) of the cases, were this study is to be repeated a large number of times. However, we only performed our study once, and no probability can be assigned to a singular, observed result. Thus, we currently have no method for deciding whether this one case was a false-negative or a true-negative finding (Levine & Ensom, 2001).
Psychosocial characteristics in participants with episodic CeH should be further explored. More research on maladaptive perceptions is needed. Such perceptions influence a patient's health behavior and contribute to the maintenance of chronic pain (Simpson et al., 2018).
Additionally, more attention should be given to positive characteristics that may be protective against chronic pain.
In our study, 33% of participants with episodic CeH and poor sleep quality reported higher degrees of symptoms of CS. Since chronic exposure to insufficient sleep may increase vulnerability to chronic pain by altering processes of nociceptive habituation and sensitization, we recommend to assess additional CPM and temporal summation in patients with episodic CeH (Staffe et al., 2019). Although a decreased CPM and enhanced temporal summation are predictive for pain, to our knowledge, only one study assessed CPM in patients with chronic CeH (Chua et al., 2011). Having a less efficient CPM when being pain-free at baseline, suggests that upon a pain-generating event, a patient is at higher risk to develop pain than patients showing an efficient CPM at baseline (Yarnitsky, 2015).
Next, the inconsistency between low extra-and cephalic PPTs, and mainly mild to moderate reported symptoms of CS (cfr. CSI) in the CeH group might raise questions. A first consideration concerning this result is that the CSI provides no direct measure of CS, but results rather represent general distress (Neblett et al., 2015). Our results are in line with Kregel et al. (2018). They reported a weaker relation between results on the CSI and measurements of PPT and CPM, compared to a stronger relation between results on the CSI and measurements of pain intensity, quality of life, pain disability, current pain status, and pain catastrophizing (Kregel et al., 2018). Therefore, it seems advised not to use the CSI as a sole screening instrument, but complementary to a comprehensive screening protocol.

CONCLUSION
Exploring multidimensional factors in CeH demonstrated relations between (1) worse sleep quality, more stress, and more signs and reported symptoms of CS, and between (2) lower level of physical activity, lower quality of life, more stress, and signs of peripheral sensitization. These key findings are novel and fill a gap in current scientific literature on relevant quality indicators for headache care. However, future research is needed to determine a pain profile and to evaluate a patient-centered intervention based on such profile.

ACKNOWLEDGMENTS
This study was supported by the Special Research Fund (BOF) of Hasselt University (BOF19OWB28).

CONFLICT OF INTEREST
The authors declare that they have no conflict of interest.