Treatment of whiplash-associated disorders-part I : noninvasive interventions

CRD summary The authors recommended mobilisation for treating pain and compromised range of motion in patients with acute whiplash-related disorders, but recommended further research into other treatments. Limitations in the review methods and questionable methods of meta-analysis mean some results should be treated with caution. It should be noted that the overall recommendation is based on only 2 trials.

W hiplash-associated disorders (WAD) represent a signifi- cant public health problem and socioeconomic burden throughout the industrialized world.In a benchmark review of the scientific literature and expert opinions, the Quebec Task Force (QTF) defined WAD as "an acceleration-deceleration mechanism of energy transfer to the neck ... [which] may result in bony or soft tissue injuries (whiplash), which may in turn lead to a variety of clinical manifestations" (1).Whiplash is a common injury, with an incidence of approximately 3.8 cases per 1000 population per year (2).While the prognosis for the majority of patients is good, with most studies reporting permanent disability in only 6% to 18% of patients (3), whiplash is a potentially debilitating and costly injury.The economic costs associated with WAD, including medical care, disability, sick leave and lost work productivity, total approximately $3.9 billion annually in the United States; this figure rises to more than $29 billion when litigation costs are considered (4,5).
A variety of interventions have been used to treat patients suffering from WAD, yet many of these interventions have not undergone sufficient clinical trials to confirm their effectiveness (6).The QTF itself concluded that the scientific evidence published before 1993 regarding whiplash was "sparse and generally of unacceptable quality" (1) and the QTF had to rely on consensus opinion for the majority of its mandated treatment recommendations.Yet, in the years since the review of the literature by the QTF, many new scientific and nonscientific studies regarding noninvasive, medical and surgical interventions for WAD have been published.The objective of this review is to identify and evaluate the literature on treatment of acute and chronic whiplash injury that has been published since the completion of the literature review by the QTF in January 1993, and to provide recommendations for clinical practice and future research.The present review, the first in a two-part series, evaluates noninvasive interventions for WAD.In part II (pages 33-40), medical and surgical interventions are reviewed.

Study identification and selection
The MEDLINE and CINAHL databases and the Cochrane Central Register of Controlled Trials (CENTRAL) were searched for studies on the treatment and rehabilitation of WAD published between 1993 and 2003.The MEDLINE search key words included "whiplash injury" and "therapy" or "rehabilitation" or "drug therapy" or "radiotherapy", and the limits were English, human and subjects older than 18 years of age.The CINAHL search key words were "whiplash injury" and "rehabilitation" or "therapy" or "diet therapy" or "drug therapy" or "surgery", with the limits of English and age greater than 18.The CENTRAL was searched using the term "whiplash injury" and the limits of English and age over 18 were applied.
Articles identified through the database searches were included in the review if the purpose of the article was to study the effect of a specific, clearly defined treatment protocol on improvement of WAD secondary to motor vehicle collision.Randomized controlled trials (RCTs) and epidemiological studies, including cohorts, casecontrol studies and case series, were included in this review.An RCT is an experiment in which subjects in a population are randomly allocated into groups to receive or not receive a therapeutic intervention and the subjects are prospectively studied to measure the outcome of interest.A cohort study is a nonrandomized epidemiological study that involves identifying two groups of subjects, one that received the treatment and one that did not, and following the groups to measure the outcome of interest.A case-control study is a retrospective epidemiological study which involves identifying subjects who have the outcome of interest (cases) and subjects without the same outcome (controls), and then reviewing the subjects to determine which ones had the treatment of interest.A case series is an epidemiological report on a series of patients with a treatment of interest.No control group is reported.Meta-analyses, reviews, abstracts, letters and case reports of single patients were read but excluded.Studies were not excluded on the basis of treatment protocol, outcome measure or quality assessment.
The MEDLINE search identified 88 articles, of which 20 with met the inclusion criteria.Eighty-five articles were found through CINAHL, including six that were initially identified through the MEDLINE search.Four of the remaining articles met the inclusion criteria.The CENTRAL search yielded 43 articles, of which 11 were previously identified and two met the inclusion criteria.Finally, three studies that were cited in review articles and the included articles but not identified through the three databases were also retrieved, bringing the total number of clinical trials to 29.
After all the clinical trials on the treatment of WAD were identified, three categories of Studies were selected for meta-analyses if four criteria were met.First, each study had to meet the definition of an RCT.Second, the studies had to assess the difference between one treatment and no treatment, one treatment and a placebo or sham treatment, or between two treatments.Dose-escalating studies were not included.Third, each study had to report the results of at least one common type of outcome measure for the calculation of a pooled effect size that incorporated the data from at least two studies.Fourth, the duration of WAD injury was used to determine appropriate pooling of subject data for meta-analysis.Acute WAD was defined as any injury of less than three months duration, while chronic WAD was injury of more than three months.Studies that did not meet the inclusion criteria for meta-analysis were categorized, reviewed and summarized.

Data abstraction
The review process consisted of four parts: abstraction of data from each study regarding methodology, outcome measures, results and final conclusions; assessment of quality of the included RCTs; meta-analysis where inclusion criteria for meta-analysis were met; and summary of the findings and evidence in support of each treatment modality.
A single reviewer (AC) abstracted the data from each of the included RCTs using a predetermined data abstraction form.The information sought included sample population, subject inclusion and exclusion criteria, follow-up time period and outcome measures.The outcome measures were categorized as pain measures, physical measures, and function/coping measures.The pain measures included the Visual Analog Scale out of 10 or 100, the Pain Disability Index, the McGill Pain Questionnaire and subjective reports of pain.The physical measures included cervical range of motion (flexion, extension, left and right lateral flexion, and left and right rotation), kinesthetic sensibility and head posture.The function/coping measures included the Self-Efficacy Score, Vernon-Mior Score, sick-leave profile, self-reported psychological distress and self-reported ability to complete activities of daily living.
Two independent evaluators (AC and SB) evaluated the quality of the RCTs according to the Physiotherapy Evidence Database (PEDro) standardized rating scale, a validated tool used for the assessment of the quality of RCTs (7).The PEDro scale (Table 1) consists of a list of 10 equally weighted criteria for quality assessment of RCTs.Raters were blinded to each other's results until all studies were assessed.Any discrepancies in the assessment of the articles were resolved by discussion between the reviewers until consensus was reached.

Statistical analysis
The RCT results were analyzed to compare activation-based interventions with their control treatments.Data on the outcomes of each trial were pooled to arrive at an overall estimate of the effectiveness of the procedure.Analyses were based on the data provided at the end of the follow-up period.Subgroup analyses were attempted to determine the effect of each treatment modality on various outcomes of interest.For continuous data, including mean scores and mean change in scores, the results were presented as weighted mean differences (WMD), where the difference between the treatment group and the control group was weighted by the inverse of the variance.For dichotomous outcomes, results were presented as an OR or RR.Fixed effects models were used where statistical homogeneity was demonstrated (P>0.05).
Evidence for and against all treatment protocols was summarized in tables.The summary tables identify treatments by category and duration of WAD injury.The findings of all RCTs within a given treatment category are reported.For treatment categories for which no RCTs have been published, the results of all reported epidemiological studies are summarized.

RCTs
A total of eight RCTs examined noninvasive interventions (Table 2).Of these, three studied exercise alone (8)(9)(10), two studied exercise in conjunction with multimodal intervention (11,12), and two compared mobilization with immobilization (13,14) and one studied pulsed magnetic field treatments (15) (Table 3).The median methodology quality score was five, while the range was three to six (Table 2).The criteria of random allocation, baseline homogeneity of subjects, and between-group statistical comparison were commonly fulfilled.One study failed to demonstrate both baseline homogeneity of subjects and between-group statistical comparison (8).However, intention to treat analysis was not completed in any of the eight trials.
Rosenfeld et al ( 9) randomized patients to four groups: treatment initiated within 96 h versus treatment delayed for two weeks and active treatment (neck exercises 10 times per waking hour) versus standard treatment (active movements up to three times daily and optional use of a soft collar for comfort and immobility).Pain and range of motion was assessed initially and at six months.The study revealed that active treatment resulted in significant improvements in pain (P<0.001) but not range of motion.In addition, when active treatment was provided, it was better when it was provided early; when standard treatment was provided, it was better when it was provided late.However, other studies regarding physiotherapy exercises failed to demonstrate significant differences between treatment and control groups.Soderlund et al (10) compared the outcome of patients who completed additional exercises to improve kinesthetic sensibility and neck muscle coordination to patients who did not, and found no significant difference between the groups for improvement in pain, physical parameters or function at the six-month follow-up.Fitz-Ritson (8) compared pain disability among patients undergoing chiropractic manipulation and phasic movements about the eyes, head, neck and arms to patients receiving chiropractic care alone, but failed to provide a between-group analysis and offered no conclusions on the comparative efficacy of these treatments.

Continued on next page
Two RCTs assessed the effectiveness of exercise in conjunction with multimodal interventions.Provinciali et al (12) found that patients who received multimodal treatment consisting of relaxation training, cervical spine mobilization, fixation exercises and psychological support showed significantly greater improvement in pain (P<0.001),self-assessment of outcome (P<0.001) and delay in return to work (P<0.001) at 180 days compared with patients receiving physical interventions alone.However, Soderlund and Lindberg (11) found no significant differences in the outcome measures of pain intensity, pain disability, cervical range of motion, and head posture when they compared patients whose treatment plan included functional behavioural analysis with patients whose treatment plan did not.
Two noninvasive intervention studies compared mobilization to immobilization (13,14).The studies reported that patients who did not use a soft collar for immobilization fared significantly better than those who did on outcome measures of pain, cervical range of motion (13), neck stiffness, memory and concentration (14).
One RCT (15) reported that pulsed magnetic field treatment was an effective form of treatment for WAD.Patients receiving this treatment had significantly less pain (P<0.03) and significantly greater range of motion (P<0.05)compared with controls (8).

Pooled analysis of RCTs
Pooled analysis was considered for the eight RCTs.Three of the RCTs were excluded from pooled analysis on the basis of failure to report standard deviations (8,9,12).Pooled analysis was thus possible for five studies and a total of 850 patient data points (10,11,(13)(14)(15).
Categorically, active treatment was not found to be superior to control treatment on the outcome measure of degree of pain as measured by a variety of pain scales (WMD -1.10, 95% CI -1.18 to -1.02) (10,11,14,15) (Figure 1).However, beginning on the first day of treatment and global improvement were evaluated at concentration.There was also a and a 5-day prescription for a nonsteroidal 2 weeks and 6 months.significant reduction of symptoms from anti-inflammatory drug.Patients were randomly the time of intake to 24 weeks after the assigned to "act-as-usual" group (no sick treatment period in both groups. leave or collar) or an immobilized group (14 days of sick leave and collar use).

Pulsed magnetic field treatments
Thiule and Walzl, Ninety-two patients with whiplash injury of Neck, head, shoulder and arm pain and Patients receiving the magnetic field 2002, Austria (15) undefined period of time were randomized cervical flexion, extension and rotation were treatment had significantly less pain than to two groups.The treatment group assessed initially and after treatment.controls (P<0.03)and significantly received diclofenac, tizanidine and pulsed greater range of motion (P<0.05).
magnetic field treatment twice daily for two weeks, while the control group received the medications alone.

TENS Transcutaneous electrical nerve stimulation
pooled analysis for presence of pain at four anatomical areas as measured by Bonk et al (13) indicated that mobilization was superior to soft collar use (OR 0.56, 95% CI 0.31 to 1.01) (Figure 2).As well, pooled analysis for the outcome measure of cervical range of motion across five studies (10,11,(13)(14)(15) revealed that subjects undergoing noninvasive interventions had significantly greater improvement in cervical range of motion than did subjects in the control groups (WMD 4.70; 95% CI 4.34 to 5.07) (Figure 3).Analysis of other physical outcome measures, including cervicothoracic posture and kinesthetic sensibility, demonstrated no significant difference (P=0.28) between control and treatment groups when the results of the 261 patients from two studies were pooled (10,11) (Figure 4).Finally, pooled analysis was also completed for noninvasive treatments on the outcome measures of selfefficacy and other measures of function (10,14); however, there was no significant difference between the control and treatment groups (P=0.72 and P=0.92, respectively) (Figures 5  and 6).

Nonrandomized studies
Ten nonrandomized studies assessed the impact of noninvasive treatment on outcome in whiplash patients.The populations and methods, outcome measures and results are summarized in Table 4. Soderlund and Lindberg ( 16) conducted a multiple baseline design study of three patients with WAD of at least four months duration who underwent psychological and behavioural  functional analyses and physiotherapy.Compared with the patients' baseline measures, all measures of pain intensity and cervicothoracic posture were significantly improved.Neck range of motion and self-efficacy scores also improved in two of three patients on completion of the multimodal intervention.
Heikkila and Astrom ( 17) assessed eight patients with chronic whiplash injury after completion of a six-week multidisciplinary rehabilitation program involving physical and occupational therapy, psychology and social work.Active head repositioning was significantly more precise after the rehabilitation program for cervical rotation and extension; however, no significant improvement was found on pain intensity as measured by the Visual Analog Scale.
In another study (18), patients with chronic WAD with temporomandibular disorder and patients with temporomandibular disorders alone underwent eight weeks of muscle exercises, counselling and splint stabilization.Compared with the control group, WAD patients had significantly more headaches, tender muscles, somatic complaints and psychological distress.As well, WAD patients showed improvement in the proportion of tender muscles, while the control group showed improvements on all outcome measures (18).
Two case series (19,20) also documented the effectiveness of multimodal interventions with exercise on the outcome of WAD patients.Sterner et al (19) followed 90 patients with whiplash injury of undefined duration who completed a program consisting of hydrotherapy, body awareness therapy, ergonomics, pharmacology and pain education.At the sixmonth follow-up, pain intensity in the neck and upper back were significantly decreased (P=0.018).Vendrig et al (20) assessed 26 patients with WAD of six months duration or longer who completed a four-week treatment program to restore muscle strength, muscle endurance and aerobic fitness, and abolish inappropriate pain behaviour.When reassessed at six months, the patients demonstrated significant improvement on the measures of pain intensity, disability, somatic complaints and psychological symptoms.
Gennis et al (21) conducted a cohort study to assess the effectiveness of activation-based treatment by comparing the outcomes of WAD patients assigned to wear a soft collar following initial presentation to the emergency department.There was no significant difference between the groups on the measures of degree of pain, recovery, improvement or deterioration.Goodman and Frew (22) assessed the effectiveness of exercise alone for the treatment of whiplash of varying duration by following 10 patients who underwent 18 strength-training sessions over six weeks.Patients demonstrated improvement in cervical range of motion and isometric strength, but the statistical significance of the improvement was not reported.
Three case series (23)(24)(25) assessed the impact of manipulation on various outcome measures in WAD patients.In one     (24), however, found significant improvement in neck pain and cervical flexion and extension among 57 patients with WAD of varying duration who underwent subluxation-based chiropractic care to the point of maximal improvement.As well, Woodward and colleagues (25) found that a significant proportion of patients (P<0.001) with WAD of at least three months had categorical improvement based on reports of pain, analgesics usage, and function following spinal manipulation, proprioceptive neuromuscular facilitation and cryotherapy.
Since the QTF report, two RCTs have been published that provide evidence that exercise does not improve range of motion in acutely injured WAD patients (9,10).However, these RCTs also give conflicting evidence regarding the utility of exercise for the treatment of pain in acute WAD.Other studies exist regarding the independent contribution of exercise, but the methodologies render the findings noncontributory.The QTF also found weak cumulative evidence to restrict the prescription of cervical collars and rest as interventions for WAD, as well as weak evidence to support mobilization.Since 1993, two RCTs (13,14) have been published which substantiate the QTF findings.These studies indicate that mobilization combined with a lack of soft collar use is an effective intervention for pain and cervical range of motion in the acutely injured WAD patient.A third RCT found no difference in the outcome of mobilized and immobilized patients with WAD (21); however, due to the poor methodological quality of this trial, the validity of the results is questionable and the evidence is subsequently interpreted among the non-RCT studies.
One study was cited by the QTF regarding the utility of pulsed magnetic field treatment; however, due to the lack of statistically significant differences between the treatment groups, no recommendations for or against pulsed magnetic field treatments were made.This review identified one RCT supporting pulsed electromagnetic field treatment for improvement in pain and range of motion.However, the authors failed to specify the duration of the injury of the patients in the study, and thus, the implications for clinical practice are unclear.
Studies regarding the efficacy of a single chiropractic manipulation in the treatment of WAD were discussed by the QTF, but due to the design of the studies, no recommendations were made.Since 1993, three non-RCTs (23)(24)(25) demonstrating improvements in the measures of pain, range of motion and function in WAD of varied duration have been published.However, because of the lack of methodological rigour of these non-RCTs, only limited evidence exists in support of chiropractic manipulation, and future research is recommended.
Four RCTs that evaluated multimodal intervention with exercise were identified by the QTF.While the QTF criticized each study for failing to evaluate the independent effect of exercise, it nonetheless concluded that the cumulative evidence at that time suggested that active exercises as part of a multimodal intervention can be beneficial.In the multimodal studies reviewed here, conflicting evidence was identified in two RCTs (11,12) and four non-RCTs (16)(17)(18)(19) on the outcome measure of pain.
Overall, mobilization appears to be the most effective noninvasive form of intervention for the treatment of both pain and cervical range of motion in the acutely injured WAD patient.As well, substantiated evidence exists to suggest that exercise does not improve range of motion in acute WAD.For all other noninvasive interventions, evidence regarding their effectiveness ranges from limited to conflicting.However, since the completion of the literature review by the QTF in 1993, multiple studies have also been published which assess the utility of medical and surgical interventions.Part II (pages 33-40) of this series provides a review of these interventions and offers further recommendations for clinical practice and future research.
interventions were developed: noninvasive interventions, medical interventions and surgical interventions.A total of 18 studies on noninvasive interventions were identified, including eight RCTs and 10 non-RCTs.Eleven studies on medical-and surgical-based interventions were included in the review (see part II, pages 33-40).Noninvasive interventions were subcategorized as physiotherapy exercise alone, multimodal treatment including physiotherapy exercise, patient mobilization, strength training, pulsed magnetic field treatment and chiropractic manipulation.
Noninvasive interventions for WAD Pain Res Manage Vol 10 No 1 Spring 2005 23

Figure 1 )
Figure 1) Active treatment versus control treatment: degree of pain.WMD Weighted mean difference.Soderlund (1) refers to reference 10 and Soderlund (2) refers to reference 11

Figure 2 )Figure 3 )
Figure 2) Active treatment versus control treatment: presence of pain.WMD Weighted mean difference

TABLE 1
9. The results of between-group statistical comparisons are No/Yes reported for at least one key outcome 10.The study provides both point measures and measures No/Yes of variability for at least one key outcome Data from reference 7

TABLE 2
AB Assessor blinding; BC Between-group comparison; BS Baseline similarity of subjects; CA Concealed allocation; ITT Intention-to-treat analysis; OM Outcomes measures of 85% of subjects or more; PVM Point and variability measures; RA Random allocation; SB Subject blinding; TB Therapist blinding

TABLE 3
Experimental group treatment consisted of learning basic physical and psychological skills, application and generalization, and maintenance in accordance with a functionalbehavioural analysis.
Active treatment versus control treatment: other measures of function.WMD Weighted mean difference Active treatment versus control treatment: other physical parameters.WMD Weighted mean difference.Soderlund (1) refers to reference 10 and Soderlund (2) refers to reference 11 Active treatment versus control treatment: self-efficacy.WMD Weighted mean difference.Soderlund (1) refers to reference 10

TABLE 4
Continued on next page study (23), 23 patients with whiplash injury of three to six weeks' duration underwent cervical spinal manipulation.Patients demonstrated improvement in cervical range of motion and neck and arm muscle strength, but the statistical significance was not reported.McCoy and McCoy (21)lizationGennis et al, 1996, Cohort -Adults with whiplash injury following Pain intensity and categorical pain (more, No significant difference between the United States(21)an automobile crash who were treated at an same, less or none) was evaluated initially groups was found for degree of pain, urban emergency department were assigned and at 6 weeks.recovery,improvement or deterioration.towear a soft collar (92 patients) or no collar (104 patients).