Long-term recovery from post-traumatic neck pain: protocol for a prospective cohort study in the emergency department

Abstract

Introduction

Post-traumatic neck pain is common, representing a substantial human and societal burden. About 15%–25% of individuals involved in an accident causing whiplash continue to experience moderate-to-severe symptoms and functional impairment 1 year post-trauma. Factors such as age, high pain intensity, hypersensitivity to pain and early post-traumatic hyperarousal are associated with persistent neck pain. However, multiple questions remain unanswered regarding how best to improve early care. As such, research on recovery patterns (including indicators for health economic burden) and their predictors is still needed, including biomarkers for pre-traumatic and peri-traumatic stress, and the value of early prediction tools.

Methods and analysis

This prospective cohort study will include 100 participants (≥18 years) suffering from post-traumatic neck pain sustained within 72 hours of an accident. At baseline (a combination of inclusion and 1 week assessment), eligible participants will undergo a thorough evaluation, including assessment of descriptive characteristics, self-reported variables (eg, pain, disability, sleep quality and post-traumatic stress), biomarkers (eg, heart rate variability (HRV) and hair cortisol) and clinical tests (eg, cervical range of motion). Follow-up will be conducted at 3, 6 and 12 months post-trauma. Further, register data (eg, data on labour market attachment) will be added for the period. Among other methods, a receiver operating characteristic (ROC) curve and multivariable regression analyses will be used to evaluate performance and associations of the prediction tools and their associations with measures of HRV.

The sample size calculation is based on previous studies, estimating that 15% of participants will develop moderate-to-severe ongoing symptoms. Using a conservative estimate, 64 participants are needed to achieve a statistical power of 90% with an expected area under the curve of 0.80. Accounting for a 25% loss to follow-up, 80 participants are required. For regression analysis, 100 participants will be included. The prediction tool will be validated using ROC analysis, sensitivity and specificity. Logistic regression models will be performed with and without biomarkers and pain sensitivity. Health economic costs will be compared across groups. Multivariable regression will examine the link between HRV and post-traumatic stress disorder, adjusting for confounders and a moderation analysis will assess hair cortisol as a potential moderator.

Ethics and dissemination

The study is approved by the Regional Committee on Health Research Ethics of Southern Denmark (S-20230037). Due to the acute nature of recruitment, the study design does not allow for a 24-hour reflection period; however, this approach has been approved by the Committee.

Study results will be published in peer-reviewed journals and disseminated through non-scientific outlets, including patient and professional publications, press releases and social media. If effective, workshops for clinicians will be organised. Results will be published regardless of outcome, with coauthorships following ICMJE guidelines.

Trial registration number

NCT06176209.

STRENGTH AND LIMITATIONS OF THIS STUDY.

• Heart rate variability is used as a biomarker of autonomic hyperarousal in patients with neck pain following accidents.

• This multisite study recruits participants to assess precrash stress as a vulnerability factor for the development of chronic neck pain.

• Despite the multisite nature of the study, not all individuals with post-traumatic neck pain present to the emergency department, which may affect the representativeness of the sample.

• Comparing health economic costs across groups will provide an insight into the financial burden of post-traumatic neck pain.

• Given that recruitment is limited to three emergency departments in Denmark, the representativeness of the sample to the broader Danish population and internationally cannot necessarily be assured.

INTRODUCTION

Post-traumatic neck pain (often referred to as whiplash associated disorders) sustained after accidents is common,¹ representing a substantial human and societal burden that can cause significant long-term pain and disability.2,4 The worldwide prevalence of post-traumatic neck pain after traffic accidents is estimated to range from 16 to 200 cases per 100,000.⁵ In addition, post-traumatic neck pain following other types of accidents is also suspected to be common.

Previous estimates have indicated that there are 6000 new cases of post-traumatic neck pain annually in Denmark,⁶ though international research indicates that the figure may be more than double.7,9 Up to half of those injured continue to experience symptoms,1,3 and 15%–25% report moderate-to-severe symptoms and functional impairment a year after the accident, leaving them at risk of leaving the labour market.^{10 11} Post-traumatic neck pain is often accompanied by symptoms such as headaches, fatigue, concentration problems and sensitivity to light and sound.34 11,15 A number of studies also point to frequent psychological symptoms after injuries like stress, depression and anxiety,1114 16,18 highlighting significant heterogeneity in symptoms. In addition to the human costs, the economic and societal burden is substantial. While no up-to-date or comparable data exist for the economic burden of post-traumatic neck pain, a study covering the incidence across 10 European countries estimated total claims cost at approximately $500 million as of 2008.⁵ Taking indirect and accumulated costs into consideration, the costs are even higher.¹⁹ However, these figures need to be interpreted with caution, since incidence rates and compensation claim systems differ considerably across countries. Due to the human suffering and societal costs, it is important to create evidence-based knowledge about early recovery indicators post-trauma, as well as to develop and validate early screening tools to predict risk of poor recovery, in order to provide early targeted care to prevent ongoing symptoms.

Previous studies have demonstrated various risk factors associated with persistent post-traumatic neck pain, including age, early pain intensity and early hypersensitivity to pain.1014 15 20,23 Stress responses immediately after an accident have also been shown to play a significant factor in the development of chronic post-traumatic neck pain. For instance, post-traumatic hyperarousal symptoms have been associated with increased sensitivity to pain, catastrophic thinking and misinterpretation of common bodily symptoms,^{24 25} while early post-traumatic stress symptoms have been reported to predict poorer recovery over time.^{16 26}

To the best of our knowledge, only two clinical prediction tools exist for evaluating patients with symptoms after whiplash injury: the Australian WhipPredict^{21 22} and the Danish Whiplash Group Risk Assessment Score (DWGRAS).¹⁵ WhipPredict by Ritchie et al^{22 27} is a prediction rule to identify patients at risk of developing post-traumatic neck pain based on the patient’s age, Neck Disability Index (NDI) score and hyperarousal symptoms (five subscale items from the Post-traumatic Diagnostic Scale). WhipPredict allows for classification into low, medium and high risk for chronicity.^{22 27} DWGRAS is a 19-point tool based on the number of non-painful neurological symptoms, the severity of neck pain and/or headache and the reduction in active range of motion of the cervical spine.¹⁵ Based on this, patients are stratified into seven different risk strata depending on the total score, prognosis worsening as scores increase.¹⁵ However, their validation has so far been limited, also across different contexts, raising questions about their clinical utility, especially in acute scenarios. Therefore, further testing and comparison of these two screening tools, also in a European and an (sub)acute context, is needed.

In addition, and to potentially improve precision of the prediction tools, there is a need to assess whether physiological stress mechanisms are linked to lack of recovery. Preferably, stress and hyperarousal should be assessed using objective measures such as heart rate variability (HRV) measured in the acute phase, within 72 hours after the traumatic event,²⁸ while pretrauma stress can be indicated by hair cortisol.²⁹ HRV refers to the variability between successive heartbeat intervals, reflecting the balance between sympathetic and parasympathetic nervous systems activity.³⁰ HRV is a non-invasive and easily obtainable measure.³¹ HRV is impacted by stress,³² and current neurobiological evidence supports its use in the objective assessment of psychological health and stress.³² Low HRV has been linked to autonomic dysregulation (ie, stress response) and development of post-traumatic stress symptoms.33,39 Conversely, a high HRV is associated with cardiovascular recovery and improved self-regulation during stressful situations.⁴⁰ To the best of our knowledge, only one study has examined HRV as a predictor of post-traumatic stress disorder (PTSD) following severe traffic trauma, finding that HRV measured 2 days after an accident correctly identified 72% of those who subsequently developed PTSD after the accident.^{32 41} However, no studies have assessed HRV in relation to recovery from post-traumatic pain. Pretrauma stress levels may also be of importance⁴² and can be determined from hair samples, which is non-invasive, easily conducted and generally well tolerated by participants.²⁹ Hair cortisol levels obtained from hair samples are retrospective in nature (ie, capturing cortisol levels prior to the accident) and have been shown to be reliable.⁴³ Using hair cortisol as a measure of pretrauma stress allows us to explore whether it signifies a stress response either to the accident or to pre-existing stress. Therefore, this study also aims to explore whether the addition of HRV measurements can increase the precision of the presented tools and assess whether pretrauma stress impacts recovery. Taken together, these considerations provide a rationale for the current study.

Aims

The objectives of this prospective cohort study are to:

1. Explore recovery over time (ie, 3, 6 and 12 months) and validate and compare two existing prediction and screening tools for post-traumatic neck pain in a Danish setting.

   We hypothesise that the screening tools will have similar predictive probabilities in a Danish context.

2. Explore whether the additional HRV measurements can improve the predictive value of the existing screening and prediction tools.

   We hypothesise that¹ the addition of HRV to the existing screening and prediction tools in the acute phase (ie, <72 hours) after an accident can contribute to better and earlier identification of high-risk patients who continue to experience moderate-to-high degree of impairment 12 months after the accident, as measured using the NDI, that is, NDI>32%.^{22 27 44}

3. Evaluate the role of low HRV in the development of PTSD symptoms 3 months after the accident.

   We hypothesise that low HRV is a predictor of the development of PTSD symptoms 3 months after the accident.

4. Evaluate the associations between low HRV, PTSD symptoms and pain sensitivity.

   We hypothesise that the change in pain sensitivity 3 month after an accident is associated with low HRV and the development of PTSD symptoms.

5. Evaluate the role of pretrauma stress (ie, hair cortisol) as a vulnerability factor (moderator) for lack of recovery over time.

   We hypothesise that pretrauma stress, as measured by hair cortisol, serves as a moderator of lack of recovery over time.

6. Examine whether the health economic costs for the group that do not recover are greater after 12 months compared with those who do recover, as measured using NDI score <10% vs >32%.^{22 27 44}

   We hypothesise that the health economic costs for the group that do not recover are statistically significantly larger compared with the group that do recover.

METHODS AND ANALYSIS

Study design

Participants who meet the eligibility criteria will be consecutively recruited in this prospective cohort study for early screening using prognostic tools combined with measurement of selected biomarkers (HRV, resting heart rate, respiration and sleep), quantitative sensory testing (QST), clinical tests (neck mobility and neurological assessment), register data and self-report questionnaires. The study will be reported in accordance with Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)⁴⁵ and Tripod⁴⁶ statements for cohort and clinical prediction studies, respectively.

Participants and setting

Possibly eligible participants will be identified in two ways:

1. Participants presenting with neck pain after an accident will be recruited from three emergency departments (ED) in Denmark: the ED at the Næstved-Slagelse and Ringsted Hospitals, the Hospital of Southern Jutland, and Kolding Hospital. Participants will be recruited on site by study personnel at the end of their visit to the ED.

2. Potential participants who are not identified in the ED will be identified through screening of patient records at the respective hospitals 1–2 days after the accident and will be contacted by phone.

All eligible participants will receive both written and verbal study information before giving verbal and written informed consent. Participants will be given ample time to review the consent form and are encouraged to bring a close friend or family member for support. All information will be given in a private, calm and quiet room at the respective hospital departments. Subsequently, participants will be invited to attend an appointment at their respective ED 1 week later. Figure 1 illustrates the flow of participants through the study. All included participants will receive travel reimbursement for their participation in the study in accordance with the per-kilometre rate set by the Danish government. There was no patient involvement in any phase of this project.

Figure 1. Flow chart of participants.

To be considered eligible for inclusion, participants must meet the following criteria:

Inclusion criteria: 18 years of age or older and post-traumatic neck pain of musculoskeletal nature acquired within the past 72 hours in connection with an accident. For the purposes of this study, ‘accident’ will be defined broadly as any event resulting in neck pain, including motor vehicle collisions, sports injuries and falls. Participants must be able to read and speak Danish.

Exclusion criteria: previous pain after similar accidents, existing chronic back pain at the time of the accident, fractures or slippage of vertebrae after the accident, suspected or known spinal pathology, including confirmed fracture or slippage of vertebrae at the time of the accident, previous back or neck surgery, spinal cord injury, severe psychiatric history (eg, schizophrenia and depression) or existing rheumatological or neurological disorders.

Data collection

The study will use a combination of different measurements: self-reported descriptive information, questionnaire data, biomarkers, clinical tests and registry data. Data will be collected at different time points. For a full list of all variables and time points, please see table 1. Data gathered at inclusion and at the visit after 1 week will be utilised as baseline measurements. For the additional follow-ups, participants will be contacted by phone at 3, 6 and 12 months. At 3 months, an additional appointment for onsite testing will be arranged. At 6 and 12 months, the Firstbeat device will be sent to participants by mail, and participants will be asked to repeat the previous questionnaires online.

Table 1. Summary of self-reported information and physical assessments.

Variable	<72 hours	1 week	3 months	6 months	12 months
Descriptive information:
Baseline characteristics	X
Use of medication and treatment (eg, physiotherapy, etc)	X	X	X	X	X
Biomarkers
HRV (firstbeat), respiration frequency, resting heart rate, sleep and hair cortisol*	X		X	X	X
Clinical tests
Neck rotation (CROM)	X	X	X
Pain sensitivity (QST)		X	X
Neurological evaluation of upper and lower limb	X	X	X
Self-reported variables
Acute self-reported stress (NRS)51	X
Expected improvement (0%–100%)	X	X	X	X	X
Pain (SF-MPQ)49	X	X	X	X	X
Level of function (NDI)47 48		X	X	X	X
Stress, anxiety and depression (DASS-21)50		X	X	X	X
PTSD (PDS, PCL-5)53 54		X	X	X	X
Paincatastrophisingg (PCS)55		X	X	X	X
Quality of life (EQ-5D-5L)60		X	X	X	X
Global rating of change (PGIC)56		X	X	X	X
Self-reported sleep (KSS)57		X	X	X	X
Comorbidity58		X	X	X	X
Physical activity59		X	X	X	X
Adverse events			X
Register data†
DREAM					X
The National Patient Register					X
The Register of Pharmaceutical Sales					X

^*Hair cortisol will only be collected at the first visit.

^†All register data will be collected after 12 month follow-up has ended for all participants.

CROM, Cervical Range of Motion Instrument; DASS-21, Depression Anxiety Stress Scale 21; DREAM, The Register-based Evaluation of the Extent of Marginalisation in Denmark; GLA:D, Good Life with osteoArthritis in Denmark; HRV, heart rate variability; KSS, Karolinska Sleepiness Scale; NDI, Neck Disability Index; NRS, Numerical Rating Scale; PCL-5, Posttraumatic Stress Disorder Checklist; PCS, Pain Catastrophizing Scale; PDS, Stress Diagnostic Scale; PGIC, Patient Global Impression of Change; PTSD, post-traumatic stress disorder; QST, Quantitative Sensory Testing; SF-MPQ, Short-form McGill Pain Questionnaire.

Self-reported questionnaires

Participants will be asked to complete questionnaires addressing various factors, including quality of life, sleep, comorbidities, PTSD symptoms and physical activity (see table 1). These questionnaires will undergo pilot testing at all study sites prior to administration. The estimated completion time is 30 min. Questionnaires will be administered via tablet during visits to the emergency department and online during follow-up at 1 week, 3 months, 6 months and 12 months. All data will be collected through the Research Electronic Data Capture (REDCap) platform. In addition, self-reported demographic information, including height, age, weight, educational level and employment status, will be gathered at baseline.

Outcome

Primary outcome

The primary outcome measure in this project will be the NDI.^{47 48} The NDI is a validated, reliable and responsive^{47 48} self-report measure designed to assess neck pain-related disability.^{49 50} The NDI is scored from 0 to 100, with a higher score suggesting greater disability. In this study, participants will be classified as recovered if their NDI score is less than 10% at the 12 month follow-up, while NDI scores above 32% will indicate moderate-to-severe persistent symptoms.

Secondary outcomes

Secondary self-reported outcomes include the following: Self-reported pain Numerical Rating Scale;⁵¹ Pain mannequin from the Short-form McGill Pain Questionnaire;⁴⁹ Depression Anxiety Stress Scale 21;⁵⁰ Post-traumatic Stress Diagnostic Scale;⁵² Post-Traumatic Stress Disorder Checklist for DSM-5;^{53 54} Pain Catastrophising Scale;⁵⁵ Patient Global Impression of Change⁵⁶ using an 11-point scale that ranges from −5 (vastly worse) to +5 scale (completely recovered); Karolinska Sleepiness Scale;⁵⁷ Self-reported Comorbidity;⁵⁸ Physical activity adopted from Good Life with osteoArthritis in Denmark⁵⁹ and Quality of life (EQ-5D-5L).⁶⁰

Clinical assessment

During the initial visit, a specially trained physiotherapist will perform a clinical evaluation of the participants’ active cervical range of motion using a digital goniometer.⁶¹ Any pain, restriction in range of motion or discomfort will be noted. A neurological assessment of the upper and lower extremities will be conducted, encompassing the evaluation of motor function, sensory function and reflexes. Motor function will be assessed through a muscle strength test graded on a 0–5 scale.⁶² For sensory function, a dermatomal assessment⁶² will be applied. Reflexes will be evaluated using deep tendon reflexes and judged on a 0–4 scale.⁶³ The neurological evaluation will serve as a screening protocol for adverse events and will not be reported in the results. Clinical assessments, including range of motion measurements, will be repeated during each visit 1 week and 3 months after the accident.

As part of the clinical assessment, QST will be performed at baseline and at 3 months. QST will be performed to objectively assess pain sensitivity, capturing changes in pain processing that may not be reflected by self-reported pain intensity. QST will consist of pressure pain threshold followed by cold pressor pain and pressure pain threshold. Pressure pain threshold will be performed bilaterally on the upper anterolateral aspect of the tibialis anterior and intermediate part of trapezius halfway between C7 and acromion64,66 using a handheld algometer (Somedic, Hörby, Sweden) with a 1 cm² probe mounted. Immediately after the pressure pain threshold test, participants will be asked to submerge their non-dominant hand in 0°C–2°C cold water for a maximum of 2 min or until pain becomes intolerable.⁶⁷ Immediately thereafter, the pressure pain threshold will be repeated. Both the cold pressor test and pressure pain threshold have been found to be reliable and valid in chronic pain populations.⁶⁸ The result of QST (change in PPT) will be used in the statistical analysis.

Biomarkers

HRV, resting heart rate, respiration frequency and sleep patterns will serve as biomarkers for stress in this study. These will be measured using the Firstbeat Bodyguard 2 tracker (Firstbeat Technologies Ltd, Helsinki, Finland), which records continuous data over a period of 4 days. HRV measurements will be taken at baseline, 3, 6 and 12 months, and will be used to evaluate stress and recovery levels by calculating the root mean square of successive differences (RMSSD) between normal heartbeats. The RMSSD is computed by first determining the time difference between each consecutive heartbeat in milliseconds, then squaring these differences, averaging them and taking the square root of the average.^{32 69} The Firstbeat Bodyguard 2 (Firstbeat Technologies Ltd, Helsinki, Finland) tracker has been shown to provide valid and reliable measurements.^{70 71}

Hair samples will be collected within the first week of participant inclusion. These samples contain cortisol, providing a retrospective index of integrated cortisol secretion over several months, with approximately 1 cm of hair corresponding to 1 month of secretion.⁷² For this study, hair samples will reflect a 3 month period, necessitating a minimum length of 3 cm. As hair samples can only be obtained from the scalp, participants are required to have at least 3 cm of head hair.

Adverse events

Adverse events will be reported in accordance with regional Ethics Committee regulations and systematically recorded at all follow-up visits by inquiring about potential events through open-ended questions to ensure comprehensive documentation. An adverse event is defined as the worsening of pre-existing symptoms, such as pain or disability. If an adverse event leads to hospitalisation, extended inpatient care, surgery or is life-threatening, results in death, or causes permanent disability or damage, it will be classified as a serious adverse event.⁷³

Health economics

Registry data will be extracted covering the period from 1 year prior to 1 year after the accident. The data will include information from the National Patient Register, the Health Insurance Register, the Register-based Evaluation of the Extent of Marginalisation in Denmark (DREAM) and the Register of Pharmaceutical Sales. These data will be used both to describe the cohort and to validate the predictive tool, serving as an indicator of recovery over time. All registry data will be collected only after the completion of the 12 month follow-up period for all participants.

Sample size calculation and statistical analysis

Based on previous cohort studies, at least 25% of the participants are expected to develop moderate-to-severe ongoing symptoms.^{10 14 15 21 22} A more conservative estimate of 15% is used as the starting point for our sample size calculation (easyROC web-tool V.1.3.1) for a ROC analysis. At an expected area under the curve of 0.80 and an allocation ratio of 7, a sample of 64 patients will be sufficient to give a statistical strength of 90% and yield a significant result (alpha 0.05). An expected loss to follow-up of up to 25% at 12 months is anticipated. Therefore, 80 participants will be required for the analysis of the primary outcome. To retain the statistical power for the regression model (described below), a total of 100 participants will be included. Distribution of normality will be examined using histograms and qq-plots. Parametric and non-parametric analysis will be applied as appropriate.

The prediction tool will be validated by analysing its ability to predict those who recover from the traffic accident (NDI<10%) and those who develop moderate-to-severe symptoms (NDI>32%) at 12 months follow-up. The tool’s precision will be examined using ROC analysis, and sensitivity and specificity will be calculated by various cut-off criteria. Participants with intermediate scores (NDI 10%–32%) will be included as controls in both analyses (ie, recovery and moderate-to-severe) and will therefore not be excluded from the validation. K-fold cross-validation will be performed to assess the stability and internal generalisability of the predictive models. Additionally, logistic regression analyses will be conducted, both with and without biomarkers (eg, hair cortisol, HRV) and pain sensitivity, to determine the optimal prediction model. Health economic costs will be compared across groups, with labour market attachment, medication usage and healthcare service utilisation used as proxies for recovery at 12 months. Registry data will only be collected after all participants have completed the 12 month follow-up period.

Multivariable regression analysis will be used to investigate the association between HRV and PTSD, adjusted for relevant confounders. Furthermore, a moderation analysis will be conducted investigating pretrauma stress (hair cortisol) as a moderator of the associations between the prediction models and recovery. Relevant confounders will be identified using directed acyclic graph⁷⁴ specifically the dagitty tool for R package.⁷⁵ Missing data will be managed by applying multiple imputation strategies. All data will be analysed using Stata (V.17, StataCorp, Texas, USA) and R (V.4.3.1 (Beagle Scouts)).

Perspectives and implications

We expect the results of this study to be of interest to clinicians, researchers, legal and insurance professionals and policymakers. The results may contribute to both national and international knowledge about the early identification of patients at high risk of chronic neck pain and the role of biomarkers. To the best of our knowledge, this study is the first European effort to investigate the role of biomarkers in the development of accident-induced neck pain. More specifically, this study marks the first attempt to use objective measures of acute and pretrauma stress as predictors, and potential moderators, of chronic neck pain development. It is also the first direct comparison of healthcare costs between patients who recover and those who do not. If confirmed, the findings could enable healthcare systems worldwide to allocate resources more efficiently and potentially reduce the human and societal burden of post-traumatic neck pain.

Patient and Public Involvement

There was no patient involvement in any phase of this project.

Ethics and Dissemination

The study is approved by the Regional Committee on Health Research Ethics of Southern Denmark (S-20230037). The study will comply with the Declaration of Helsinki,⁷⁶ the General Data Protection Regulation (GDPR)⁷⁷ and the Data Protection Act. Access to registry data will be requested from Statistics Denmark.

In general, no significant long-term side effects from the data collection process are expected. However, in some cases, the QST procedure may result in short-term pain flare-up. Further, given the context of this study, the study design does not allow for the minimum 24-hour reflection period for participants. This procedure has been used in previous ED studies and has been approved by the Regional Committee on Health Research Ethics for Southern Denmark (S-20230037).

The findings from this study will be published in peer-reviewed scientific journals and disseminated through non-scientific outlets such as patient and professional publications, press releases and social media. If proven effective, practical workshops for interested clinicians will be organised to train them in the screening method. The study’s results will be published regardless of whether they are positive, negative or inconclusive. Coauthorships will follow the guidelines set by the International Committee of Medical Journal Editors (ICMJE).⁷⁸