To what degree patient‐reported symptoms of central sensitization, kinesiophobia, disability, sleep, and life quality associated with 24‐h heart rate variability and actigraphy measurements?

Chronic musculoskeletal pain is associated with decreased parasympathetic and increased sympathetic activity in the autonomic nervous system. The objective of this study was to determine the associations between objective measures of heart rate variability (a measure of autonomic nervous system function), actigraphy (a measure of activity and sleep quality), respiration rates, and subjective patient‐reported outcome measures (PROMs) of central sensitization, kinesiophobia, disability, the effect of pain on sleep, and life quality.

[9][10][11] Patient-reported outcome measures (PROMs) are standardized, validated, and subjective self-administrated questionnaires. 12,13Numerous PROMs are used to evaluate various factors related to CMP. 14,15 Some widely used and validated CMP-related PROMs assess central sensitization, 16 kinesiophobia (fear of movement), 17 low back pain-related disability, 18 quality of life, 19 and effect of pain on sleep. 20Research evidence of associations between CMP-related PROMs and function changes of the ANS are limited to three studies with 10 or fewer minutes of heart rate variability (HRV) measures.][23] Heart rate variability refers to the temporal variation of beat-to-beat intervals between heartbeats. 24 growing body of literature suggests that organized variability in the heart rate pattern is a reasonable index of physical and emotional health. 25,26HRV is a commonly used method for assessing the balance between the sympathetic and parasympathetic parts of the ANS. 24Increased sympathetic nervous system activity is associated with "fight-or-flight" and decreased HRV.Conversely, increased parasympathetic nervous system activity is associated with the "rest and digest" activity of ANS and increased HRV. 27In addition to the cardiovascular system, the ANS is part of the regulation system of wakefulness and sleep. 28CMP and sleep disturbance are highly correlated, with more severe pain being associated with more severe impairment in sleep quality. 29,30tudies have revealed that subjects with CMP often have difficulty with sleep initiation and maintenance during the night. 31Actigraphy is a commonly used method to assess sleep quality parameters.It uses an accelerationbased method, often with wrist-worn devices, for measuring movement, which can help estimate sleep-wake patterns, sleep continuity versus fragmentation, and general sleep quality. 32Actigraphy has shown over 90% sensitivity in detecting subjects' sleeping state compared to polysomnography, which is considered a gold standard method of assessing sleep. 32 faster breathing rate is identified in subjects with CMP. 33,34][37] However, despite strong evidence of a treatment effect of therapeutic breathing methods, respiration rate during the 24 h related to the most commonly studied CMP symptoms are not directly studied previously.
The objective of this work was to study the associations between PROMs of central sensitization, disability, kinesiophobia, the effect of pain on sleep, and quality of life and 24-h ambulatory HRV and actigraphy measurements during wakefulness and sleep.In addition, we studied association between PROMS and respiration rate during wakefulness and sleep.To the best of our knowledge, these objectives have not been previously studied simultaneously with HRV and actigraphy measurements in 24-h measurement.

Ethical approval and consent to participate
Ethical approval for the study was obtained from the Research Ethics Committee of the Northern Savo Hospital District with identification number 1106/13.02.00/2018.Written informed consent was received from all subjects before the data collection.The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines were adhered to in this study. 38

Data collection and subjects
The subjects were recruited from a cross-cultural validation study of the Central Sensitization Inventory (CSI).This study was carried out partly simultaneously in a single chiropractic clinic in Helsinki Finland from May 2019 to March 2020. 39The subjects completed an online demographic form including age, gender, height, and weight.Body mass index was calculated in the data analysis phase from subject-reported height and weight data.All subjects included in this study met the following inclusion criteria: (a) Age between 18 and 65 years and (b) Proficient in written and spoken Finnish language.Exclusion criteria were as follows: (a) History of a malignant tumor; (b) History of diagnosed trauma potentially negatively affecting the central nervous system (including whiplash or mild traumatic brain injury); (c) History of diagnosed disease negatively affecting the central nervous system (including multiple sclerosis, Alzheimer's disease, Parkinson's disease, and dementia); (d) chronic obstructive pulmonary disease; (e) bundle-branch block or chronic arrhythmias; (f) spinal surgery in the last 12 months; (g) a cardiac pacemaker; and (h) not completing online form of demographic data.
For each subject in this study, a collection of physiological measurements of actigraphy and HRV were carried out between December 2019 and March 2020 and between August 2020 and November 2020.The break in data collection was due to the COVID-19 outbreak in Finland.From a total of 229 subjects recruited in the CSI validation study, 39 those with CSI scores ≤30 (low CSI subgroup) and CSI scores ≥40 (high CSI subgroup) were invited to participate in this study.Group scores were based on previously established clinically relevant severity levels of CSI, where the score of ≤30 clinically translates as mild and ≥40 as severe. 40The recruitment process was stopped when the required 20 subjects per group were recruited.After data collection began, two additional subjects with low CSI scores were excluded due to the poor quality of HRV data, which left 18 subjects in the low CSI subgroup and 20 subjects in the high CSI subgroup.Subjects also completed an online form of pain history and PROMs on the same day, or the following day, as the physiological 24-h measurements were carried out.A flowchart of subjects is presented in Figure 1.

Pain history
All subjects completed a structured web-based pain history assessment with binary questions (yes/no), including the presence of chronic low back pain (CLBP), referral to a leg (if yes to CLBP), the experience of other ongoing chronic musculoskeletal pain, and presence of chronic headaches.The subjects were further divided into three pain history groups (a) pain-free control group (no CLBP, pain intensity 0, no other chronic musculoskeletal pain, and no chronic headache), (b) pain in a single body area (CLBP group with or without leg referral or other chronic musculoskeletal pain or chronic headache), and (c) multisite chronic pain (two or more of the following: CLBP with or without radiculopathy, other chronic musculoskeletal pain and/or chronic headache).CLBP is the most common CMP diagnoses 41 and is defined as pain present for more than 3 months and more than 3 days per week. 42Generally, subjects with CMP tend to have PROMs scores indicating more severe related CMP symptomology, but not without exceptions, 13 which also have been confirmed in previously published studies of this same cohort. 39,43,44Moreover, subjects with multisite pain distribution tend to have more comorbid biopsychosocial health issues. 45In this study, we concentrated not only on associations between HRV/actigraphy measurements and PROMs, but also included pain status to inform the pain history of subjects.

PROMs
The Central Sensitization Inventory (CSI) was developed as a screening tool for symptoms related to central sensitization. 16It is considered the leading PROM for assessing CS-related symptomology. 46The CSI is a two-part questionnaire.Part A includes 25 questions about CSrelated symptomology, with a total score range of "0" to "100."Items are rated on the Likert scale: 0 = never, 1 = rarely, 2 = sometimes, 3 = often, and 4 = always.A ≥40 cutoff score has been proposed for reliable discrimination of subjects whose presenting symptomology is likely related to central sensitization. 16,40Part B contains questions about previously diagnosed Central Sensitization syndromes and related disorders in the form of "No/ Yes, and year diagnosed."CSI part B is only for additional information and is not scored.It includes binary questions (yes/no) and year of previous diagnoses such as fibromyalgia, neck injury, restless legs syndrome, temporomandibular joint disorder, or migraine/tension headaches. 16In this study, we extracted the number of subjects who reported previous fibromyalgia diagnoses from CSI part B for further analysis.A Finnish version of the CSI, which has been previously translated and cross-culturally validated in a Finnish population. 39he Tampa Scale of Kinesiophobia (TSK) evaluates kinesiophobia (fear of movement).The TSK is a 17-item questionnaire used to assess subjective kinesiophobia on a Likert scale: 1 = strongly disagree, 2 = disagree, 3 = agree, and 4 = strongly agree.The range of scores is from 17 to 68.Higher scores indicate a more severe level of kinesiophobia. 47A Finnish version of the TSK, which has previously been translated into Finnish and validated in the Finnish population, was used in this study. 48he Roland-Morris Disability Questionnaire (RMDQ) is a 24-item measure designed to evaluate the perceived level of disability related to chronic low back pain. 49For each item, disability in performing specific daily activities is indicated by "yes" or "no."The RMDQ is scored by adding up the number of items checked "yes."Total scores range from 0 to 24, with higher scores indicating a higher level of disability related to low back pain. 50I G U R E 1 Flowchart.The Pain and Sleep Questionnaire 3-Item Index (PSQ-3) assesses the impact of pain on sleep during the past week.51 It is measured on a numerical 11-point rating scale from 0 to 10. Zero indicates "never" and 10 indicates "always."Thus, the final score range is from 0 to 30.A Finnish version of the PSQ-3, which has been previously translated and cross-culturally validated in a Finnish population, was used in this study.43 The EuroQol (EQ-5D-5L) assesses health-related quality of life in the five dimensions.52 The dimensions are mobility, self-care, usual activities, pain/discomfort, and anxiety/depression.Each dimension has five response levels: 0 = no problems, 1 = slight problems, 2 = moderate problems, 3 = severe problems, and 4 = unable to /extreme problems. A econd part of the EQ-5D-5L is the EQ visual analog scale (EQ VAS).52 Because there is currently no Finnish standard value set available, a value set from Denmark was used to calculate the index value as recommended by the EuroQol EQ-5D-5L User Guide.53

Twenty-four-hour physiological measurements
Measurements always began on Tuesday afternoon and ended ∼24 h later on Wednesday.The physiological measurement equipment was setup by a trained colleague at the clinic and the equipment was returned via mail.Simultaneous electrocardiography (ECG) and acceleration-based actigraphy data were recorded for 24 h.The ECG was recorded with a Bittium Faros 180 Holter device (Bittium Oyj, Oulu, Finland) with a 250 Hz sampling rate using three wet gel electrodes (BlueSensor VLC, Ambu A/S, Ballerup, Denmark) placed under the left and right collarbones and below the left rib cage.Simultaneously, actigraphy was measured with an ActiGraph GT9X link research-grade activity bracelet (ActiGraph LLC., Pensacola, FL) with a 30 Hz sampling rate.

Heart rate variability (HRV) measures
Heart rate variability analyses were carried out using Kubios HRV Premium 3.5 software (Kubios Oy, Kuopio, Finland).The software automatically detects RR intervals (time intervals between successive ECG R-waves), and corrects missed, extra, and misaligned (including ectopic) beats using a validated algorithm. 54,55urthermore, periods of noisy measurement data were automatically identified and excluded from HRV analysis.Finally, very low-frequency components were removed from the HRV data since the baseline drift of ambulatory HRV data is not directly related to the short-term regulation of heart rate by the sympathetic and parasympathetic branches of the ANS. 56All HRV analyses were carried out by a trained data analyst (SK), who also visually verified that only good-quality sinus rhythm data were analyzed.
Heart rate variability was assessed using the mean RR interval (mean RR), the standard deviation of normalto-normal beat intervals (SDNN), the root mean square of successive RR interval differences (RMSSD), and the ratio of the standard deviations SD2 and SD1 from the Poincaré plot (SD2/SD1).In addition, an estimate of the respiratory rate derived from the ECG data was obtained from the Kubios HRV software.The respiratory rate estimate is based on analyzing the respiration-induced changes in the ECG R-wave amplitude and RR interval time series. 57The descriptions of the HRV parameters are given in Table 1.
Sleep-time HRV analysis was conducted using two approaches.First, HRV variables were assessed for the entire duration of sleep, commencing at bedtime and concluding upon awakening, to evaluate sleep-time HRV across different study groups.Second, recognizing the inherent interindividual variability in HRV parameters, 58 a 15-min baseline HRV was established for each subject, starting from the time they went to bed.Individual HRV changes from this baseline were subsequently evaluated during the initial 4 h of sleep at 15-min intervals.Essentially, the first 4 h of sleep were partitioned into nonoverlapping 15-min segments, HRV variables were computed for each segment, and changes in HRV from the baseline were analyzed in relation to time.This latter analysis aimed to identify potential group differences in the initiation of sleep.

Actigraphy measures
The actigraphy data were analyzed with Actilife 6.0 analysis software (ActiGraph LLC., Pensacola, FL).Actilife uses the Cole-Kripke algorithm 59 for sleep scoring.In this work, the sleep-awake patterns were analyzed with a 60-s window.The sleep quality parameters were evaluated using the Actilife implementation of the Tudor-Locke method. 60Sleep quality was assessed through total sleep time (TST), sleep efficiency (SE), number of awakenings (NOA), and wake after sleep onset (WASO).The sleep quality parameters are described in more detail in Table 1.In addition, actigraphy data were used to determine the daytime activity levels.Activity levels were evaluated with cut points described by Freedson et al. 61 The cut points were 0-99 counts for sedentary, 100-1951 for light, 1952-5724 for moderate, 5725-9498 for vigorous, and 9499 and above for very vigorous activity.

Diary
In addition to the physiological measures, subjects kept an activity diary during the 24-h measurement period.They were asked to document their prescribed medications, daily activities and estimated time of sleep onset, and estimated rising time in the precision of 30 min.The beginning and end of sleep were extracted for each subject based on their HRV and actigraphy data and diary notes.

Statistical methods
Statistical analysis of demographics and subjectreported data was performed using the SPSS version 27 (IBM SPSS Statistics for Windows, Version 27.0.IBM Corp, Armonk, NY).Statistical significance was defined as p < 0.05.Data were shown as N (%) or mean (95% confidence interval lower and upper bound or standard deviation).Normal or non-normal data distribution was evaluated by Shapiro-Wilks tests and histograms.Group comparisons for non-normally distributed data were calculated by Mann-Whitney U-test.Categorical variables were compared by Pearson Chi-square (χ 2 ) tests.Physiological measurements were analyzed by comparing the HRV, sleep quality parameters, and activity levels between the low CSI (≤30) and high CSI (≥40) subgroups of scores.The statistical differences were evaluated by one-way-ANOVA using a built-in function anova1 on MATLAB (version R2022a, MathWorks inc., Natick, MA).Before ANOVA, a one-sample Kolmogorov-Smirnov normality test was applied for each parameter using the kstest function with the default 0.05 significance level on MATLAB.There are no previous studies with similar 24-h HRV measurement protocol study association with PROMs used in this study.Hence, group sample sizes were not based on sample size calculation.

R E SU LT S Demographic characteristics of the sample
Demographic and subject-reported symptoms are presented in Table 2.There were no differences in age, gender, height, weight, and BMI between the low and high CSI groups.However, significant differences were found between the two groups in all pain parameters and subject-reported symptoms on the PROMs.

Wakefulness HRV and activity
Wakefulness HRV and activity results for the study groups are presented in Table 3.No significant differences were found in wakefulness HRV parameters, respiration rate, and activity levels between the two groups.

Sleep-time HRV and sleep quality
Night-time HRV and sleep quality were assessed for the entire night, starting from the detected beginning of sleep at bedtime and ending at the detected wake-up In addition to the whole night HRV analysis, we analyzed the first 4 h of sleep in 15-min windows to see how the HRV changed at the beginning of the sleep.This time trend analysis was carried out because subjects with CMP may have challenges in sleep initiation. 31Since the magnitude of HRV at rest is highly interindividual, 58 the HRV parameter time trends are reported as changes from the first 15-min window, that is, as differences to T A B L E 3 Group comparison of wakefulness HRV and activity levels (N = 38).Note: Data presented as N (%) or mean (95% confidence interval lower and upper bound).The standard deviation of normal-to-normal beat intervals (SDNN), the root mean square of successive RR interval differences (RMSSD), and the ratio of the standard deviations SD2 and SD1 from the Poincaré plot (SD2/SD1).

Variable
T A B L E 2 Group comparison of baseline data (N = 38).

DI SC US SION
Subjects were initially divided into low CSI and high CSI subgroups.Only very few significant associations were found between the subgroups in measures of 24-h HRV, 24-h actigraphy, and subjective symptoms of central sensitization, kinesiophobia, low back painrelated disability, pain-related sleep disturbance, and quality of life.However, there was overall little trend toward increased sympathetic nervous systems activity and poorer sleep quality in the higher score CSI subgroup.Clinically, this was the most pertinent finding because the associations between subjective central sensitization and HRV had not been studied before.Previously, higher scores of CSI have shown weak or no associations between objective other measures of nociceptive sensitivity of pain threshold, heat pain threshold, conditioned pain modulation, and temporal summation. 62he recent high-quality study demonstrated a similar lack of significant HRV findings with pain intensity. 63ence, our findings challenge the use of HRV measurements as an objective outcome measurement in future clinical trials related to CMP conditions, because there is only little association with the subjective core outcome measures of pain intensity, disability, and quality of life. 64t is known that demographic factors, such as sex, age, and body mass index have major effects on HRV results. 65In our cohort, there were no significant intergroup differences in demographics or activity levels during waking hours or baseline values of sleep quality between the two groups, which may have affected our findings.
The most marked finding was a stronger decrease in heart rate and an increase in HRV parameters during the first 2 h of sleep in subjects who reported lower levels of central sensitization.This indicates higher parasympathetic recovery, which has previously been linked to better sleep quality and shorter sleep latency. 66However, the overall trend showed smaller differences in HRV and there were only little subgroup differences in sleep quality measured by actigraphy.
We also investigated the association between respiration rates and subjective symptoms between the low and high CSI groups.No significant differences in respiratory rates were found between the groups, either during wakefulness or during sleep.These negative findings challenge the common understanding that CMP and contributing factors are meaningfully associated with faster mean respiration rates. 33,34eart rate variability measurement methods vary greatly in studies involving subjects with CMP. 8 Only three previous studies have directly assessed the association of PROMs to HRV in CMP subjects.][23] A previous meta-analysis of HRV studies comparing chronic pain subjects with pain-free controls showed evident differences in HRV measures. 8The results of this meta-analysis were heavily influenced by studies that OUTCOME MEASURES AND AUTONOMIC NERVOUS included subjects with fibromyalgia, which was not the case in our study.Only 10% of subjects in the high subjective symptoms group, and none of the subjects in the low symptom group, reported a previous fibromyalgia diagnosis in CSI part B. This difference may partially explain our more marginal results compared to this metaanalysis.It should also be noted that the meta-analysis compared chronic pain subjects with healthy controls, which was not the case in this study.

ST R E NGT H S
This study had several strengths, including (a) strict inclusion and exclusion criteria; (b) well-defined study groups; (c) a state-of-the-art 24-h HRV measurement protocol, including simultaneous actigraphy; and (d) reliable differentiation between periods of sleep and wakefulness.

L I M I TAT ION S
Most of the study subjects were females (84%), which limits the generalization of results to male populations.HRV measurements differ greatly between individuals and across the studies, leading to the unavoidable situation where the variability of the results is large. 58This trend was observable also in our results, and hence, it is likely that our study cohort was too small for the adequately powered study.However, previous studies with similar cohort sizes have shown The changes in HRV parameters and respiration rate during the first 4 h of sleep for the two study groups.Data presented as mean (bold lines) and 95% confidence intervals (dash lines).
meaningful ANS function differences between subjects with CMP and pain-free controls on HRV 8 and actigraphy measures. 67,68Another limitation is that we did not incorporate medication use as a factor in our analysis.
Almost half of the study participants (18/38) reported regular use of one or more medications.Medication use was more common in group 2 (14/20) compared to group 1 (4/18).However, as there were no significant differences between the groups in the daytime HRV, it appears that the medications used did not have a significant effect on HRV.

CONC LUSION
Almost all HRV and actigraphy parameters and subjective measures of central sensitization, disability, kinesiophobia, the effect of pain on sleep and quality of life showed only little association during wakefulness and sleep.Overall, there were small and nonsignificant trend for increased sympathetic nervous system activity and poorer sleep quality in the high central sensitization subgroup.
Heart rate variability and actigraphy measures.
T A B L E 1Abbreviations: ANS, autonomic nervous system; PNS, Parasympathetic nervous system; SNS, sympathetic nervous system.andrisingtime.Night-time HRV and sleep quality for the study groups are presented in Table4.There were no statistically significant differences in night-time HRV or sleep quality variables between the study groups.Though nonsignificant, the mean RR was somewhat longer (lower HR), sleep efficiency was about 2% higher, and total sleep time was about 36 min longer for group 1 compared to group 2.
T A B L E 4Note: