Biopsychosocial sequelae and recovery trajectories from whiplash injury following a motor vehicle collision

Abstract

BACKGROUND CONTEXT:

Five out of 10 injured in a motor vehicle collision (MVC) will develop persistent pain and disability. It is unclear if prolonged symptoms are related to peritraumatic pain/disability, psychological distress, muscle fat, lower extremity weakness.

PURPOSE:

To test if widespread muscle fat infiltration (MFI) was (1) unique to those with poor recovery, (2) present in the peritraumatic stage, (3) related to known risk factors.

STUDY DESIGN/SETTING:

A cohort study, single-center Academic Hospital.

PATIENT SAMPLES:

A total of 97 men and women (age 18–65) presenting to an urban academic emergency medicine department following MVC, but not requiring inpatient hospitalization.

PRIMARY OUTCOME MEASURE:

Neck disability at 12-months.

METHODS:

Participants underwent magnetic resonance imaging (MRI) to quantify neck and lower extremity MFI, completed questionnaires on pain/disability and psychological distress (< 1- week, 2-weeks, 3-, and 12-months) and underwent maximum volitional torque testing of their lower extremities (2-weeks, 3-, and 12-months). Percentage score on the Neck Disability Index at 12-months was used for a model of (1) Recovered (0%–8%), (2) Mild (10%–28%), and (3) Moderate/Severe (≥ 30%). This model was adjusted for BMI and age.

RESULTS:

Significant differences for neck MFI were revealed, with the Recovered group having significantly lower neck MFI than the Mild and Moderate/Severe groups at all time points. The Mild group had significantly more leg MFI at 12-months (p=.02) than the Recovered group. There were no other significant differences at any other time point. Lower extremity torques revealed no group differences. The Traumatic Injury Distress Scale (TIDS) and MFI of the neck at 1-week postinjury significantly predicted NDI score at 12-months.

CONCLUSIONS:

Higher neck MFI and distress may represent a risk factor though it is unclear whether this is a pre-existing phenotype or result of the trauma.

TRIAL REGISTRATION:

ClinicalTrials.gov Identifier: NCT02157038.

Introduction

Whiplash-associated-disorders (WAD) from a motor vehicle collision (MVC) affect ~ 4-million Americans each year [1]. While recovery is rapid for ~ 50% of those injured, ~ 50% will continue to report persistent interference in daily life [2,3], ranging from self-reported mild symptoms to severe disability. The precise reasons ~50% fail to fully recover remain unclear but anxiety, depression, distress, including reports of higher pain intensity, have demonstrated prognostic value [4].

Recent work has expanded the biopsychosocial model of recovery [5–8], by reconceptualizing WAD as a multisystem tissue- (biological) and stress-based (psychological and socioenvironmental) interaction [8]. Central to this is that an MVC is both potentially injurious and distressing, and that there may be no single or simple threshold of a “good-or-bad” reaction to an MVC applied across all people and all types of injury events. It seems intuitive that collision severity is related to a poor clinical outcome. However, vehicular variability in the seat and head restraint responses [9] and even wider variability between individuals in their physical and psychological responses to an MVC undermine the utility of such a generic relationship when applied to an individual person exposed to a specific MVC [10,11].

Emerging evidence also suggests the presence and number of advanced imaging findings in the acute and chronic stages of WAD differentiate between those recovering rapidly from those recovering more slowly. These include (1) larger number of pre-existing degenerative pathologies and lower muscle attenuation on peritraumatic computed tomography (CT) scans [12,13], (2) larger amounts of neck MFI on magnetic resonance imaging (MRI), [14–20] (3) functional MRI (fMRI) markers of altered brain network modulation [21], and (4) reductions in myelin integrity in spinal cord white matter [22–24].

To highlight and confirm the heterogeneity of whiplash recovery, we tested a three-group model of recovery (recovered, mild persistent disability, moderate/severe disability). We hypothesized larger MFI in the neck and lower extremities, lower extremity weakness, and higher levels of acute and persistent distress would be present to a larger extent in the groups with persistent and severe disability.

Methods

This study was a primary analysis of data investigating the psychosocial and neuromuscular mechanisms underlying poor recovery following MVC (Clinicaltrials.gov Identifier: NCT02157038). All analyses were performed using imaging, sociodemographic, and longitudinal outcome data from a prospective cohort of men and women enrolled following MVC.

Participants were eligible if reporting MVC-related neck pain (≥ 4 on a numeric pain rating scale) and were within the Quebec Task Force Classification [25] category of WAD Grade II (tenderness to palpation, movement restriction with no radicular symptoms), and within 1-week of injury. Exclusion criteria included participant age younger than 18 or older than 65, spinal fracture from the MVC, prior spinal surgeries, history of one or more MVCs or previous diagnoses of cervical or lumbar radiculopathy, history of neurological disorders (eg, Multiple Sclerosis, previous stroke, myelopathy), inflammatory/autoimmune diseases (eg, Hepatitis, Arthritis, Ankylosing Spondylitis, Crohn’s disease, Fibromyalgia) or metabolic disorders (eg, Diabetes, hyper- and hypothyroidism) and institutional contraindications to undergoing a MRI. We also excluded patients with injuries requiring surgical consult or inpatient hospitalization.

The Institutional Review Board of Northwestern University, Feinberg School of Medicine granted approval and all methods (both in the enrollment of the MVC cohort and all subsequent sample processing and analyses) were in accordance with the approval. All enrolled participants underwent serial MRI examination at < 1 week, 2-weeks, 3-, and 12-months postinjury to quantify MFI in select neck and leg muscles and maximal volitional plantarflexor torques at 2-weeks, 3- and 12-months postinjury. Participants also completed a suite of questionnaires at each time point.

Self-reported neck-pain-related disability (obtained at each time point)

Self-reported neck-pain-related disability (ie, how neck pain affects ability to manage everyday life) was measured using the 10-item Neck Disability Index (NDI) [26]. Percentage scores ≥ 30% have been reported to indicate moderate/severe neck-related disability [18,19].

Self-reported pain intensity (obtained at each time point)

The Numeric Pain Rating Scale was used as a unidimensional measure of pain intensity in which the respondent selects a whole number (0 (no pain) – 10 (extreme pain) integers) [27]. Higher initial pain (> 5.5/10) intensity has been associated with worse outcomes [4].

Measures of psychological distress (obtained at each time point)

The Tampa scale of Kinesiophobia (TSK) - is a 17-item self-report measure of fear of reinjury due to movement (kinesiophobia) and has been shown to be a robust tool for traumatic neck pain [28,29].

The Impact of Events Scale (IES) - is a 15-item questionnaire measuring a stress response related to a particular event [30]. The IES has been validated in studies investigating emotional responses to acute trauma, [31] including SCI [32].

The Traumatic Injuries Distress Scale (TIDS) - is a 12-item self-report tool designed to capture the experiences of acute traumatic distress. It has demonstrated sound measurement properties, including a stable 3-factor structure (uncontrolled pain, negative affect, intrusion/hyperarousal), good internal consistency, and evidence of both concurrent and predictive validity. Each item is scored on a frequency-based scale of 0 (never or not at all) to 2 (often or all of the time) for a total scale range of 0 (no distress) to 24 (severe distress) [33,34].

MRI muscle fat analysis

Muscle fat from MRI has been described previously [19,35]. All post-MVC MRI data were collected with a 3.0T Prisma scanner (Siemens, Erlangen, Germany). A localizer scan and a T2-weighted sagittal turbo spin echo sequence was performed to determine the location of the high-resolution axial 3D fat-water images of the cervical spine. Fat-water images were acquired using a dual-echo gradient-echo sequence (2-point Dixon, TR = 7.05 ms, TE₁ = 2.46 ms, TE₂ = 3.69 ms, flip angle = 12°, bandwidth = 510 Hz/pixel, FOV = 190×320 mm², slab oversampling of 20% with 40 partitions to prevent aliasing in the inferior-superior direction, in-plane resolution = 0.7×0.7 mm², slice thickness = 3.0 mm, 44 slices, number of averages = 6, acquisition time = 4 minutes 5 seconds). The scanner outputs the in- and opposed-phased data as well as the water and fat images. A 64-channel head/neck coil was used as a receiver coil to improve signal-to-noise. This scan covered the cephalad portion of C3 through the caudal portion of the C7 vertebral end plates.

A similar gradient echo sequence (slice thickness increased to 5.0 mm and FOV increased to 60 slices to fully cover the anatomy) was used for the lower extremity muscles using a 16-channel body array surface coil. The left and right lower extremities were acquired in a single acquisition. Localizer scans were used to obtain the fat/water data in the axial plane perpendicular to the bone.

Muscle water-fat quantification

Using a custom MATLAB script (The MathWorks, Inc, Natick, MA, USA), regions of interest were manually drawn within the fascial borders of the multifidus and semispinalis cervicis muscles from C4 to C7 on the fat and water images. For both legs, the dorsiflexors, gastrocnemius, soleus, posterior tibialis, and peroneus muscle groups were segmented across 10 slices. The software obtains the mean fat and water signal intensities within each region of interest. The MFI (%) was calculated as the mean pixel intensity of fat-only (Fat) and the mean pixel intensity of water-only (Water) images using the following equation:

$$\text{MFI (\%) = Fat/(Fat + Water) }\times \text{ 100}$$

Total MFI (2 sides × 2 muscles) of all the cervical levels (between C4 and C7) for each participant was calculated. For the legs, the total MFI for each muscle was defined as a mean across all slices, and an average of total MFI for all left and right leg muscles (2 sides × 5 muscles) for each participant was calculated.

Plantarflexor torque

Maximum Plantarflexor torques were measured for each subject’s bilateral ankle strength. The participant was seated in a comfortable position with hips flexed to 75 degrees, knees flexed to 20 degrees, and the foot / ankle positioned in neutral and affixed to an isokinetic dynamometer (Biodex Rehabilitation System v3, Shirley, NY, USA). Five trials of maximal voluntary plantarflexion were performed for each participant. Each participant was asked to maximally contract their plantarflexors against the static footplate (isometric MVT) and both verbal encouragement and a biofeedback screen were provided to fully optimize the 3 to 4 second maximal isometric force generation [22,36–38]. The maximum MVT value of five trials was used for analysis, and MVTs of both right and left lower extremities were assessed and included.

Statistical analysis

All data were analyzed with IBM SPSS Version 28 software (Windows, Armonk, NY, USA: IBM Corp) and STATA version 16.1 software (StataCorp, College Station, TX, USA).

The outcome was defined as: Recovered (0%–8%), Mild (10%–28%), and Moderate/severe (≥ 30%) disability based on 12-month follow-up scores. The longitudinal data were analyzed using linear mixed modeling as it does not rely on assumptions of equal variance and correlations among repeated measurements. Separate, random intercept linear mixed models for the model was used to estimate the effects for each outcome variable: MFI of the Neck, MFI of the right and left lower extremities, MVT of the right and left plantarflexors, and overall responses on the patient reported outcomes of pain and psychological distress. Group and time were modeled as fixed effects and BMI and age were included as covariates as they could influence neck and leg MFI [39,40]. The group-time interaction at each timepoint was the hypothesis of interest and was examined using pairwise comparisons of the estimated marginal means. Significance for all statistical analyses was set a priori using 2-tailed alpha = 0.05.

Univariable associations between single-domain predictor candidates and the NDI% score at 1 year were assessed using correlation analysis (point biserial correlations for dichotomous variables, Spearman rho correlations for ordinal variables, and Pearson product-moment correlations for continuous variables). Multivariable relationships between the predictor candidates captured at baseline and the NDI at 1 year following injury were evaluated using stepwise multiple linear regression analysis with predictor candidates significantly (p<.10) correlated with NDI at 1 year postinjury being eligible for entrance into the model. A stepwise fashion of both entering and removing factors with a significance value of less than 0.05 for model entry and greater than 0.10 for removal was used to identify the most parsimonious subset of predictor variables. Identifying a multiple linear regression model that significantly fit the data (p<.05) and consisted of variables that each significantly contributed to the model across all participants (p<.05) was the overall goal.

Sample size

Sample sizes of 27 in the severe group and a total of 63 in the mild and recovered groups (total of 100 with an expected 10% dropout resulting in 90 with complete follow-up) have 80% power to detect a difference in mean MFI of 0.66 standard deviations, assuming a two-tailed test and a Type I error rate of 0.05. Previous results for neck MFI data [18] indicated a standard deviation of 3.6% with a mean at 3 months of 17.0% fat in the mild/recovered group, and 25.1% fat in the moderate/severe group (difference = 8.1%). Assuming continued progression of neck MFI in the severe group, the study was powered to detect differences between the moderate/severe group and the mild/recovered groups at 12 months of 2.4% (0.66×3.6), which is smaller than the previously observed difference of 8.1%. However, power may be decreased due to the multivariable analyses performed when covariates are included in the model. The effect of adjustment is difficult to predict a priori since one or both of the estimated mean and the residual variation may be affected by adjustment. Pilot data on plantarflexor muscle fat indicated a standard deviation of 0.5% with 7.0% fat in the mild/recovered group, and a mean (SD) of 14.5% (2.1%) in the moderate/severe group with varying duration of symptoms (difference = 7.5%). The study was powered to detect differences between the moderate/severe group and the mild/recovered groups at 12 months of 1.4% (0.66×2.1), which is smaller than previously observed difference of 7.5%.

Results

Of the 143 eligible participants recruited and enrolled, 46 did not attend their initial appointment. Accordingly, 97 participants consented and enrolled in the full cohort. All 97 participants attended and completed all questionnaires at time point 1, 2, and 3. Nineteen were lost to follow up between time points 3 (3-months) and 4 (12-months). Seventy-eight (78) completed all four time points, which could introduce bias by attrition. Fig. 1 details a participant flowchart.

Fig. 1.

Participant flowchart.

Three-group recovery model

The descriptive statistics of baseline demographic, history and questionnaire response data are summarized in Table 1.

Table 1.

Baseline demographics for 3 groups. ^*(Recovered with NDI baseline of 0–8%; Mild with NDI baseline of 10%–28%; Moderate/Severe with NDI baseline of ≥ 30%)

	Recovered (n=37)		Mild (n=44)	Moderate/Severe (n=16)
Age, y		34.1 ± 10.0	34.4± 11.3	36.9 ± 13.4
Sex assigned at birth (% male)		56.8%	88.6%	75%
BMI, kg/m2		25.3 ± 4.0	24.5 ± 4.5	27.1 ± 5.6
Days from MVC		5.0 ± 1.3	5.3 ± 1.5	5.7 ± 1.3
NPRS at 1 week		4.7 ± 2.4	4.9 ± 2.2	6.1 ± 1.5
NDI at 1 week		30.4 ± 19.0	37.3 ± 14.7	47.1 ± 8.6
TIDS at 1 week		13.22 ± 10.9	15.1 ± 7.9	20.2 ± 8.1
MFI of the Neck at 1 week		17.4 ± 4.0	20.6 ± 5.4	22.1 ± 8.4
MFI of the Leg at 1 week		10.4 ± 2.4	10.9 ± 2.9	10.5 ± 3.7
MVT of plantarflexors on right at 2 weeks		126.4 ± 23.0*	117.2 ± 28.5	120.1 ± 33.3
MVT of plantarflexors on left at 2 weeks		125.9 ± 26.0*	116.8 ± 28.2	126.0 ± 43.4

Abbreviations: MVC, motor vehicle collision; MFI, muscle fatty infiltration; NDI, Neck Disability Index; TIDS, Traumatic Injuries Distress Scale.

^*Values are mean ± SD unless otherwise indicated.

^*recovered MVT right n = 30 and left n = 29; mild right n = 41 left n = 37; mod/severe right n = 16 left n = 15.

Table 2 details the primary outcomes of MFI of the neck and legs at 1-week, 2-weeks, 3- and 12-monts post-MVC for 3 groups.

Table 2.

Primary outcomes of muscle fatty infiltration of the neck and leg at 1 week, 2 weeks, 3-months and 12-months post-MVC for 3 groups (recovered, mild, moderate/severe)

	Non-Recovered		Mild Recovered		Recovered
	Mean (SD)	Mean Change from Week 1 (95% CI)	Mean (SD)	Mean Change from Week 1 (95% CI)	Mean (SD)	Mean Change from Baseline	Adjusted Between-Group Difference (Mild Recovered – Recovered)	P Value	Adjusted Between-Group Difference (Non-Recovered – Recovered)	P Value	Adjusted Between-Group Difference (Non-Recovered – Mild Recovered)	P Value
MFI Neck
1 week	22.14 (8.40)		20.55 (5.41)		17.41 (4.05)		3.09 (1.16, 5.02)	0.02*	3.55 (0.85, 6.25)	0.01*	0.46 (−2.18, 3.10)	0.73
2 weeks	20.89 (7.77)	−1.04 (−2.04, −0.04)	20.62 (5.27)	0.06 (−0.56, 0.68)	16.96 (4.26)	−0.46 (−0.94, 0.03)	3.61 (1.68, 5.54)	<0.01*	3.00 (0.31, 5.68)	0.03*	−0.62 (−3.24, 2.01)	0.65
3 months	21.73 (6.84)	−0.15 (−1.29, 0.99)	20.85 (5.52)	0.29 (−.47, 1.05)	17.50 (4.54)	0.09 (−0.44, 0.61)	3.30 (1.36, 5.23)	<0.01*	3.29 (0.60, 5.98)	0.02*	−0.00 (−2.63, 2.62)	0.99
12 months	21.20 (7.59)	−1.14 (−2.62, 0.34)	20.27 (6.57)	0.62 (−0.39, 1.63)	17.06 (5.11)	−0.44 (−1.49, 0.61)	3.97 (1.99, 5.94)	<0.01*	2.84 (0.12, 5.57)	0.04*	−1.12 (−3.77, 1.52)	0.41
MFI Leg
1 week	10.50 (3.70)		10.87 (2.95)		10.45 (2.41)		0.67 (−0.19, 1.52)	0.13	−0.83 (−2.0, 0.33)	0.16	−1.50 (−2.64, −0.36)	0.01*
2 weeks	10.43 (3.59)	−0.07 (−0.28, 0.15)	10.88 (2.97)	0.01 (−0.19, 0.22)	10.29 (2.24)	−0.16 (−0.42, 0.09)	0.84 (−0.02, 1.70)	0.06	−0.74 (−1.90, 0.42)	0.21	−1.58 (−2.72, −0.43)	0.01*
3 months	10.58 (3.35)	0.08 (−0.31, 0.48)	10.79 (3.05)	−0.08 (−0.33, 0.17)	10.38 (2.50)	−0.06 (−0.38, 0.25)	0.65 (−0.21, 1.51)	0.14	−0.69 (−1.85, 0.47)	0.25	−1.34 (−2.48, −0.19)	0.02*
12 months	10.09 (2.91)	0.09 (−0.50, 0.68)	10.54 (3.00)	−0.06 (−0.45, 0.34)	9.27 (1.75)	−0.46 (−0.87, −0.05)	1.06 (0.18, 1.93)	0.02*	−0.26 (−1.45, 0.92)	0.67	−1.32 (−2.48, −0.16)	0.03*

^*Denote statistical significance p<.05

MFI of the neck and legs

When participants were stratified into 3 groups based on their NDI scores and including age and BMI as covariates, there were significant differences at each followup for MFI of the neck. The recovered group had significantly less neck MFI than both the mild and moderate/severe groups. Differences were not significant between the Moderate/Severe and Mild groups. When assessing the outcome variable MFI of the leg with age and BMI as covariates, the mild group had significantly more leg MFI at 1 year (p=.02) than the recovered group; however, there were no other significant differences between these two groups at other time-points. There were significant differences at all 4-time points between the moderate/severe and mild groups, with the latter having more leg MFI at each time point.

Plantarflexor torque

There were no significant group-by-time interactions at 2 weeks, 3-months, or 12-months postinjury for MVT on the right and the left.

Multivariable model

Table 3 provides results from the univariable associations between single-domain predictor candidates and the NDI score at one year. Peri-traumatic MFI (p=.015), TIDS (p=.007 and sex (p=.014) were significant predictors and correlated with NDI and were thus entered into the regression model.

Table 3.

Univariate predictors of poor outcome and correlation with NDI

Variable	Correlation Coefficient	P Value
Age	0.076	0.456
BMI	0.140	0.171
Gender	−0.212	0.014*
MFI of the Neck at 1 week	0.253	0.015*
Impact Score at 1 week	0.133	0.193
PDS Total SX at 1 week	0.137	0.180
HADS Anxiety at 1 week	0.128	0.212
PMI Active at 1 week	−0.019	0.855
TIDS at 1 week	0.274	0.007*

^*Entered into the regression model.

Table 4 reveals the results of the multiple linear regression that significantly fit the 12-month recovery status data (p<.05), consisting of two variables that significantly contributed to the model (TIDS score and neck MFI at 1-week postinjury). Fifteen percent (15%) of the overall variance in 12-month NDI score was explained by this model.

Table 4.

Results of the final model (based on 3-group model)

Variables Retained in Final Model	Unstandardized Beta Coefficient (95% CI)	Significance of Beta Coefficient	Adjusted R2	Significance of Model Fit
Model 1
TIDS at one week post injury	0.39 (0.09 to 0.69)	0.01	0.07	0.01
Model 2
TIDS at one week post injury	0.42 (0.14 to 0.71)	< 0.01	.15	<0.01
MFI of the Neck at one week post injury	0.68 (0.20 to 1.16)	<0.01

Discussion

This study provides further empirical evidence that MFI in the cervical region and psychological distress is featured in those with poor recovery following whiplash injury.

In considering a three-group model of recovery [41], significant group differences in neck MFI were observed at all time points. This is consistent with previous work across three countries and different research groups [14,17–19,22,42], suggesting the presence and expression of neck MFI could reflect a risk factor that is either a pre-existing phenotype [43] or a rapid local response to the injurious event. In keeping with colleagues [44,45], significant differences were not observed in leg MFI or plantarflexor torques. These findings should be interpreted with caution and any suggestion [46] that whiplash from an injurious MVC does not result in damage to the central nervous system is premature. Such a definitive stance is likely detrimental to those few patients who may have may have sustained an injury involving the brain, [21,47,48] spinal cord white matter pathways [23] with concomitant lower extremity weakness and muscle wasting [22,24,37]. Such injuries are likely to be radiologically occult with blunt imaging tools where reproducibility data is nonexistent, [46] but may be revealed with available higher-resolution acquisition techniques and quantitative measures [21–24].

Of interest, this US-based study uniquely featured ethnically diverse participants (African American [n = 49], Caucasian [n = 57], Asian [n = 7], Hispanic [n = 15]) presenting for care in an urban level-1 trauma certified emergency department. Whereas previous work enrolled and included participants primarily of Caucasian European descent [14,17–19,22,42,45]. Our results emphasize the need for understanding the ethnographic influences on skeletal muscle adiposity and its presence as a risk factor for poor health [49] following injurious soft-tissue traumas or other diseases affecting the human condition [50].

Furthermore, it is acknowledged that whiplash [18,19,41] (and other conditions [eg, low back pain, [51,52] mild traumatic brain injury, [53] extremity injury [54]]) outcomes consistently involve three recovery groups, when using the NDI or another patient-reported, body- or region-specific scale to predict outcomes [54]. Accordingly, now may be a good time for the wider network of investigators studying, and stakeholders managing, adverse trauma outcomes to reconceptualize and define these injurious events as multisystem tissue- (biological) and stress-based (psychological and socioenvironmental) spectrum of interactions [4,8,55–58]. A central theme to this reconceptualization is that no one experiences their trauma and recovery in a vacuum; that is their recovery is influenced by existing vulnerabilities or resiliencies that are further impacted by the socioenvironmental context within which the person lives and functions [8,50,59].

The lack of findings in our study of higher MFI content in the legs or reductions in plantarflexor torque in those with poor recovery supports that the typical whiplash injury mechanism would more likely affect structures in/around the neck (a local injury (or neck strain/whiplash) with local physiological changes) than soft-aqueous skeletal muscles distal to the neck (eg, a potential systemic injury with distal consequences). However, it could also highlight the challenges (which are not unique to our team) of recruiting acutely injured participants who may have sustained a more severe injury with neurologic signs without fracture (we excluded participants with QTF Grades III or IV) [25] or evidence of a more severe injury requiring further workup. It is also recognized that negative findings on clinically warranted peritraumatic imaging would likely be met with relief and satisfaction that there is “nothing seriously wrong” [13],

However, it has been our clinical and research experience that (1) persistent symptoms, (2) numerous encounters with varying health practitioners, and (3) varying responses to a number of available procedures, these patients, enquire, seek, and enroll in proof-of-concept, [24] or cross-sectional studies [22]. These small studies are inspired by clinical observations and questions that foster interdisciplinary collaborations, which often reveal new insights: widespread MFI, [24] altered myelin in white matter pathways [22,24], brain network modulation [21], and lower extremity weakness [22,24] in some with poor recovery. It also remains plausible that those with poor-recovery have a larger number of degenerative pathologies [13] and higher neck MFI before the MVC, [12] suggesting their presence could be considered a risk-factor (some of which are modifiable) and this needs further investigation. What’s more is the possibility that ethnic, gender, socioeconomic, or other intersectional identity differences might influence differences in body fat or skeletal muscle mass with (or without) trauma. This too needs further investigation.

It is acknowledged WAD reflects a spectrum of clinical presentations, suggesting it is not only a medical condition but one that is influenced by non-injury-related factors [60], or other comorbidities [43]. Our findings align with this position, and it is our contention that combining available measures to quantify biological and psychosocial factors advances our mechanistic understanding of why some, but not others, develop chronic WAD-related disability. Our predictive model involving a wide array of biopsychosocial variables explained 15% of the variance, which (while significant), strongly suggests more work is needed.

While we have not evaluated the positive predictive value of our combined measures, our results suggest that post-traumatic stress and anxiety symptoms are good candidates for use in, and easily translated to, clinical practice. Semi-automated MFI measures on advanced cervical spine imaging remain predictive in research-based models but it is not appropriate to suggest that early assessment strategies in clinical practice include brain, spinal cord, muscle MRI for all patients. Future work is required to improve the accuracy of prediction models and to implement their use in clinical practice where a growing dataset can be used to inform assessment, prediction, and management options. Those future prediction models should leverage and include available and evolving imaging-based methods [61,62] obtained through rapid acquisition, measured with deep learning automatic segmentation techniques [63,64], and combined with available (cost- and time-effective) tools. Such findings could result in optimizing the use of necessary imaging, revolutionizing the cost- and time-effective clinical assessment to predict WAD outcomes accurately, consistently, and confidently…and thus, inform management.