Antonio Romeo, Carla Vanti, Valerio Boldrini, Martina Ruggeri, Andrew A. Guccione, Paolo Pillastrini, Lucia Bertozzi

Transcription

1 Review Cervical Radiculopathy: Effectiveness of Adding Traction to Physical Therapy A Systematic Review and Meta-Analysis of Randomized Controlled Trials Antonio Romeo, Carla Vanti, Valerio Boldrini, Martina Ruggeri, Andrew A. Guccione, Paolo Pillastrini, Lucia Bertozzi Background. Cervical radiculopathy (CR) is a common cervical spine disorder. Cervical traction (CT) is a frequently recommended treatment for patients with CR. Purpose. The purpose of this study was to conduct a review and meta-analysis of randomized controlled trials (RCTs) on the effect of CT combined with other physical therapy procedures versus physical therapy procedures alone on pain and disability. Data Sources. Data were obtained from COCHRANE Controlled Trials Register, PubMed, CINAHL, Scopus, ISI Web of Science, and PEDro, from their inception to July Study Selection. All RCTs on symptomatic adults with CR, without any restriction regarding publication time or language, were considered. Data Extraction. Two reviewers selected the studies, conducted the quality assessment, and extracted the results. Meta-analysis employed a random-effects model. The evidence was assessed using GRADE criteria. Data Synthesis. Five studies met the inclusion criteria. Mechanical traction had a significant effect on pain at short- and intermediate-terms (g = 0.85 [95% CI = 1.63 to 0.06] and g = 1.17 [95% CI = 2.25 to 0.10], respectively) and significant effects on disability at intermediate term (g= 1.05; 95% CI = 1.81 to 0.28). Manual traction had significant effects on pain at short- term (g = 0.85; 95% CI = 1.39 to 0.30). Limitations. The most important limitation of the present work is the lack of homogeneity in CR diagnostic criteria among the included studies. Conclusions. In light of these results, the current literature lends some support to the use of the mechanical and manual traction for CR in addition to other physical therapy procedures for pain reduction, but yielding lesser effects on function/disability. A. Romeo, PT, OMT, Department of Biomedical and Neuromotor Sciences (DIBINEM), Alma Mater Studiorum, University of Bologna. C. Vanti, PT, MSc, OMT, Department of Biomedical and Neuromotor Sciences (DIBINEM), Alma Mater Studiorum, University of Bologna. V. Boldrini, PT, OMT, Department of Biomedical and Neurological Sciences, University of Bologna. M. Ruggeri, PT, Department of Biomedical and Neurological Sciences University of Bologna A.A. Guccione, PT, PhD, DPT, Department of Rehabilitation Science, College of Health and Human Services, George Mason University, Fairfax, Virginia. P. Pillastrini, PT, MSc, Department of Biomedical and Neuromotor Sciences (DIBINEM), Alma Mater Studiorum, University of Bologna, Bologna, Italy. Address all correspondence to paolo. pillastrini@unibo.it. L. Bertozzi, PT, MSc, Academic Clinical Coordinator, School of Physical Therapy, Alma Mater Studiorum, University of Bologna. [Romeo A, Vanti C, Boldrini V, et al. Cervical radiculopathy: effectiveness of adding traction to physical therapy a systematic review and meta-analysis of randomized controlled trials. Phys Ther. 2018;98: ] 2018 American Physical Therapy Association Published Ahead of Print: January 5, 2018 Accepted: January 1, 2018 Submitted: February 2, 2018 Post a comment for this article at: April 2018 Volume 98 Number 4 Physical Therapy 231

2 Cervical radiculopathy (CR) is an affliction of a cervical spinal nerve root, most commonly caused by a cervical disk herniation or other space-occupying lesion (eg, osteophytic encroachment associated with cervical spondylosis), that may result in nerve root impingement, inflammation, or both. 1,2 CR as a direct result of cervical trauma or metastases is infrequent. 3 The most common level of nerve root compression is C7, followed by C6; compression of C5 and C8 roots is less frequent. 1,2,4 The clinical diagnosis of CR is based on the presence of neck pain referred to the arm, associated with upper extremity paresthesia or numbness and accompanied by signs of nerve root compression. 5 Diagnostic imaging (eg, magnetic resonance imaging) and electrophysiological testing (eg, nerve conduction velocity, electromyography) are commonly used to confirm the clinical diagnosis of CR. There are no widely accepted, well-defined clinical criteria for the diagnosis of CR. Terms such as cervicobrachial pain, cervicobrachial syndrome, CR, and neck and arm pain are often used as synonyms. 5 It has been suggested that radiculopathy should be differentiated from radicular pain because radiculopathy describes a neurological condition characterized by motor and/ or sensory dysfunction that is caused by a conduction block along a spinal nerve or its root. However, radicular pain and radiculopathy can occur together. 6 This review considered CR as an overarching classification. The prevalence of CR is reported as 3.5 per 1000 people 7 and the annual incidence from 83 cases 7 to 210 cases 1 per 100,000 people, with a peak from 50 to 59 years. The rate of recurrence has been reported as 32% over a period of 4.9 years. 1 A recent systematic literature review of the natural course of CR concluded that substantial improvements tend to occur within the first 4 to 6 months after onset; the time to complete recovery ranges from 24 to 36 months; and recovery is generally maintained over 2 or 3 years. 8 Combined neck and arm pain seem to be much more disabling than either symptom alone. Chronic symptoms lasting more than 6 months have a negative impact on overall mental health, significantly greater than in cases of less symptom duration. 9 Management of CR can be surgical or conservative. There is low-quality evidence that surgery may provide pain relief faster than physical therapy or hard-collar immobilization in CR, but there is little or no difference in the long-term. 10 Surgery may be associated with a 4% adverse events rate. 11 Therefore, conservative management is the initial treatment of choice for most patients with CR, even though patients should be referred to a spine surgeon in the presence of intractable radicular symptoms unresponsive to nonoperative management over a 6-week period, motor weakness persisting for more than 6 weeks, progressive neurologic deficit at any point after symptom onset, signs or symptoms of myelopathy, and spinal instability or deformity. 12 Therefore, studies of conservative management are critical to understanding the first line of intervention for this condition. Conservative management includes invasive and noninvasive techniques. Relative to invasive treatments, there is some evidence on the effectiveness of injection therapy 13 and acupuncture. 14 Noninvasive therapy includes different procedures performed by physical therapists, such as cervical traction, postural education, exercise, and manual therapy applied to the cervical and thoracic spine. These different procedures are often used in a multimodal approach to intervention. 15,16 In recent years, 2 systematic reviews 17,18 evaluated the effectiveness of noninvasive treatments in CR. According to their results, there was inconclusive evidence for the effectiveness of noninvasive management, and no intervention seemed to be superior or consistently more effective than other ones. Nonetheless, cervical traction (CT) is still frequently recommended for patients with CR, and it is often used as an adjunct modality in outpatient rehabilitation. 19 It may be applied mechanically or manually, intermittently or continuously. 20 The physiological effects of CT may include separation of vertebral bodies, movement of facet joints, expansion of intervertebral foramen, and stretching of soft tissues. 19,20 To date, no systematic review has evaluated the effectiveness of cervical traction combined with other physical therapy procedures in CR. For this reason, we recognized the need of a systematic review and meta-analysis of randomized controlled trials (RCT) to investigate the effects of manual and mechanical cervical traction combined with other physical therapy procedures versus nontraction-related physical therapy procedures alone on pain and disability in adults with CR. Methods Data Sources and Searches Our literature search was aimed at identifying all available studies that evaluated in adults with CR the effects of manual and mechanical cervical traction combined with other physical therapy procedures versus physical therapy procedures alone on pain and disability. Records were identified by searching multiple literature databases, including Cochrane Controlled Trials Register, MEDLINE, CINAHL, Scopus, ISI Web of Science, and PEDro, from their inception to July Several keyword combinations were used together: For localization: cervical, neck For dysfunction and symptoms: radicular syndrome, radiculopath*, nerve root pain, neuropath*, nervous system disease, cervicobrachial pain, nerve root injuries, radicular pain, neck pain, cervical pain, arm pain, nerve pain, herniated disc, hernia, neurologic manifestation For treatment: traction, physiotherapy, physical therapy, manual therapy Additional records were searched through other sources to complement the database findings; manual search of reference lists of relevant literature reviews and indexes of peer reviewed journals were used. Two review authors (V.B., M.R.) independently conducted study selection and data extraction. All results, 232 Physical Therapy Volume 98 Number 4 April 2018

3 including title, abstract, and references, were imported to Endnote Web (Endnote basic; EndNote/Clarivate Analytics, Philadelphia, Pennsylvania) to search for and delete duplicates. Then titles, abstracts, and full texts were independently screened and thoroughly analyzed, to verify their correspondence with the inclusion criteria. Two experts (A.R., C.V.) were consulted in the case of persisting disagreement. Study Selection Types of studies. We included published RCTs without any restrictions on publication date or language. Quasi- RCT and nonrandomized controlled trials were excluded. Among RCTs, only trials with nontraction physical therapy procedures alone as a control or a comparison group were considered for inclusion in the study. Types of participants. The participants had to be adults with symptoms, aged 18 years or older (regardless of sex). Participants were required to be diagnosed with CR as a result of at least one of the following: 1) diagnosis based on magnetic resonance imaging, electromyography, or nerve conduction velocity testing, 2) positive results on at least 3 out of 4 tests, according to clinical prediction rules, 21 or 3) symptoms associated with CR, such as pain radiating to the arm, with or without motor or sensitive dysfunction. We excluded trials on patients affected solely by neck pain or cervical myelopathy, tumor, rheumatic disease, or central neurological disorder (eg, multiple sclerosis). Types of interventions. We included RCTs comparing the effects of cervical traction, both manual and mechanical, either constant or intermittent, combined with other physical therapy procedures (massage, exercise, education, manual therapy, physical modalities [eg, ultrasound, transcutaneous electrical stimulation]) versus these other physical therapist interventions alone. The position of traction needed to be similar in the different studies, that is, a slight neck flexion (about 15 ) or the most painless position. We excluded RCTs comparing traction versus placebo or usual care or RCTs investigating the addition of traction to conservative treatments different from physical therapist interventions. Types of outcome measures. We included RCTs that used at least one of the following : pain intensity measured by the Visual Analog Scale (VAS) or the Numeric Rating Scale (NRS); function/disability measured by validated scales, such as the Neck Disability Index (NDI), the Patient- Specific Functional Scale (PSFS), and the Self-Perceived Disability on Visual Analog Scale. Outcome measures had to be collected at least at baseline and another follow-up, at short term (up to 3 months after the baseline), mid term (from 3 to 6 months), or long term (from 6 to 12 months). A summary of the characteristics of the selected studies can be found in Table 1. Data Extraction and Quality Two authors (V.B., A.R.) independently evaluated the selected articles based on the PEDro score. Reviewers were not blinded to information regarding the authors, the journal of origin, or for each reviewed article. Disagreements were resolved by a third reviewer (P.P.). Papers were not excluded based on the PEDro score. A PEDro score is composed of 11 questions addressing the main methodological characteristics of an RCT. The maximum possible score is 10, given that the first question (external validity) does not contribute to the final score. A PEDro score is a valid measure of methodological quality of clinical trials. 22,23 Data Synthesis and Analysis Data were synthesized using a meta-analytic method based on a random effects model due to the significant heterogeneity of the included studies and due to this method accounting for both within-study and between-study variance. This approach weights studies by the inverse of the variance and incorporates heterogeneity into the model. 24 All effect sizes were computed using the Hedges g statistic. Hedges g is an inferential measure that assesses the magnitude or size of differences between 2 groups, not the statistical significance of the difference, which is influenced by sample size. We used Hedges g because it incorporates a small sample bias correction. 25 ProMeta v.2.0 software (Internovi by Scarpellini Daniele s.a.s., Cesena [FC], Italy; now owned by Idostatistics) 26 was used for the statistical analyses. Standardized mean differences (SMDs) with 95% confidence intervals (95% CIs) were calculated for continuous data. The SMD was used because different measures were adopted by each study to address the same clinical outcome. To interpret effect size calculated with SMD, we used the method described by Cohen 27 as a guide to identify small (0.20), medium (0.50), or large (0.80) effects. Calculation of effect size was based first on the best possible data (ie, final means, standard deviations, and sample sizes of intervention and control groups). Selected studies for which these or other crucial data were not directly reported or obtainable by contacting authors were not included in the review. The Q and I-square statistics were used to assess heterogeneity among studies. The Q statistic has low power as a comprehensive test of heterogeneity, 28 especially when the number of studies is small (ie, most meta- analyses). Conversely, the Q statistic has too much power as a test of heterogeneity if the number of studies is large. 29 A significant Q value indicates a lack of homogeneity of findings of studies. Following the approach of Higgins and Thompson, 29 heterogeneity was qualified as low (25% 50%), moderate (50% 75%), or high (75%). The potential publication bias was assessed using the Egger t test. A qualitative analysis to evaluate the overall quality of the evidence was planned 30 and independently conducted by 2 authors (L.B., P.P.) using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach. 31 GRADE is a systematic and explicit approach to making judgments about quality of evidence and strength of recommendations. Using GRADE, we April 2018 Volume 98 Number 4 Physical Therapy 233

4 Table 1. Characteristics of Selected Studies a Study Inclusion Criteria Exclusion Criteria Groups Treatment Outcome Measures Results Jellad et al 33 (2009), Tunisia Moustafa and Diab 34 (2014), Egypt CR (of < 3 mo) with herniated or degenerated disk confirmed by CT and/or MRI Clinical examination CR secondary to herniated disk confirmed by MRI or CT Wainner test item cluster Numbness and pain in arm Repeated movements test Physical examination Cervical surgery Red flags Shoulder disease, carpal tunnel syndrome Recent rehabilitation for current CR Worsening of pain or intolerance of manual cervical traction test Inability to tolerate cervical flexion or extension position Complete loss of sensation along involved nerve root Cervical or thoracic surgery Red flags Pregnancy Central nervous system pathologies Intractable pain 39 participants (3 groups of 13) 216 participants (3 groups of 72) 4 wk (12 sessions) 1. Standard b + intermittent manual traction in supine position 20 tractions: 20 s with 10-s rest interval Head posture: most pain-free position Force applied: ~6 kg 2. Standard b + mechanical traction (pulley) in supine position 2 tractions: 25 min with 10-min rest interval Head posture: most pain-free position Force applied: gradually increased from 5 to 12 kg 3. Standard b 4 wk (12 sessions) 1. Multimodal c + mechanical ventroflexion traction in supine position 20 min: 50 s of traction with 10-s rest interval Head posture: 24 of ventroflexion Force applied: from 10% of body weight to kg from first to last sessions 2. Multimodal c + mechanical H-reflex traction 20 min: 50 s of traction with 10-s rest interval Head posture: selected according to H-reflex findings At end of treatment (4 wk) and at 1, 3, and 6 mo after treatment Neck pain (VAS) Radicular pain (VAS) Self-perceived disability (VAS) Analgesic consumption At end of treatment (4 wk) and after 12 mo Neck pain Arm pain Disability (NDI) Latency and amplitude of flexor carpi radialis H-reflex Significant improvements in groups A and B for neck pain and radicular pain in relation to group C at short and long term; significant improvements in groups A and B for disability in relation to group C at short term; slightly better results in group B than in group A The gain in consumption of analgesics in the 3 groups was significant Group B better than groups A and C in all results No significant differences between groups A and C in all results Force applied: from 10% of body weight to kg from first to last sessions 3. Multimodal c (Continued) 234 Physical Therapy Volume 98 Number 4 April 2018

5 Table 1. Continued Study Inclusion Criteria Exclusion Criteria Groups Treatment Outcome Measures Results Fritz et al 35 (2014), United States Aydin and Yazicioǧlu 36 (2012), Turkey Young et al 37 (2009), United States Report of neck pain with symptoms (pain or numbness) extending distal to the acromioclavicular joint or caudal to the superior border of the scapula NDI of 10 or more CR secondary to herniated disk confirmed by MRI Provocation test (Spurling test, compression test) Unilateral upper extremity pain, paresthesia, or numbness confirmed by MRI or CT Wainner test item cluster (3/4 positive) Repeated movements test Deep-tendon reflex (biceps, triceps, brachioradial), myotomal assessment, grip strength Thoracic or cervical surgery Recent motor vehicle accident Red flags Central nervous system pathologies CR from spondylosis Shoulder-arm-hand pathology or surgery Red flags Local or general arthritis Central nervous system signs Pregnancy Upper extremity functional limitations Red flags Steroid medications Central nervous system symptoms Thoracic or cervical surgery 86 participants in 3 groups (28 in exercise group, 31 in mechanical traction + exercise group, 27 in over-door traction + exercise group) 27 participants in 2 groups (13 in traction group, 14 in control group) 81 participants in 2 groups (45 in experimental group, 36 in control group) 4 wk (10 sessions) 1. Exercise d 2. Exercise d + mechanical traction in supine position (Sauders) 15 min: 60 s of traction with 20-s rest interval Head posture: 15 of ventroflexion Force applied: from 5.44 kg to patient s tolerance or symptom response 3. Exercise d + over-door traction (Chattanooga) 15 min on days between physical therapy sessions Head posture: patient seated facing door Force applied: from 3.63 or 5.44 kg to 9.07 kg 3 wk (15 sessions) 1. Standard e physical therapy + manual intermittent cervical traction 20 min: 7 s of traction with 5-s rest interval Head posture: most pain-free position Force applied: gradually increased from 5 to 12 kg 2. Standard e physical therapy 4 wk (average of 7 sessions) 1. Manual therapy and exercise + mechanical intermittent cervical traction 15 min: 50 s of traction with 10-s rest interval Head posture: ~15 of flexion Force applied: gradually increased from 10% of patient s body weight to kg from first to last sessions At 4 wk and at 6 and 12 mo after treatment Disability (NDI) Neck pain (VAS) Arm pain (VAS) At end of treatment (3 wk) Neck pain MGS (dynamometer) At 2 and 4 wk after treatment Numeric Pain Rating Scale NDI Patient-Specific Functional Scale Fear-Avoidance Beliefs Questionnaire Global Rating of Change Scale Pain diagram (symptom distribution) Satisfaction Grip strength Significant improvement in disability for mechanical traction group compared to exercise group and over-door traction group at 6 and 12 mo Significant improvement in pain for mechanical traction group compared to exercise group and over-door traction group at 4 wk and at 6 and 12 mo Significant improvement in neck pain and MGS for both groups but significantly greater for traction group No significant difference between groups for any primary or secondary outcome measure at 2 or 4 wk (Continued) April 2018 Volume 98 Number 4 Physical Therapy 235

6 Table 1. Continued Study Inclusion Criteria Exclusion Criteria Groups Treatment Outcome Measures Results Young et al 37 (2009), United States (cont d) 2. Manual therapy and exercise + cervical traction (sham) 15 min: 50 s of traction with 10-s rest interval Figure 1. Flow diagram of the included studies. Head posture: ~15 of flexion Force applied: 2.27 kg a CR = cervical radiculopathy, CT = computed tomography, MGS = maximum grip strength, MRI = magnetic resonance imaging, NDI = Neck Disability Index, VAS = visual analog scale. b Ultrasound, infrared, massage, cervical mobilization, and muscle strengthening and stretching. c Infrared, interferential current, massage, muscle strengthening, and thoracic manipulation. d Neck and scapular deep-muscle strengthening. e Ultrasound, hot packs, transcutaneous electrical nerve stimulation, and muscle strengthening and stretching. rated the evidence not by individual study but across studies for specific clinical. Domains that may decrease the quality of the evidence are: study design, risk of bias, inconsistency of results, indirectness (not generalizable), and imprecision (insufficient data). The quality of the evidence was adjusted by a level based on the performance of the studies against the 5 domains. 31,32 Levels of evidence were defined as follows: High-quality evidence. Further research is very unlikely to change our confidence in the estimate of effect. At least 75% of the RCTs with no limitations of study design have consistent findings, direct and precise data, and no known or suspected publication biases. All domains are met. Moderate-quality evidence. Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. One of the domains is not met. Low-quality evidence. Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Two of the domains are not met. Very-low quality evidence. Any estimate of effect is very uncertain. Three of the domains are not met. No evidence. No RCTs were identified that addressed this outcome. 236 Physical Therapy Volume 98 Number 4 April 2018

7 Figure 2. Meta-analysis for mechanical cervical traction (pain) (A), meta-analysis for mechanical cervical traction (disability) (B), meta-analysis for manual cervical traction (pain) (C). ES = effect size, sig. = significance, W = weight. April 2018 Volume 98 Number 4 Physical Therapy 237

8 Table 2. Physiotherapy Evidence Database (PEDro) Scores PEDro Score Jellad et al 33 Moustafa and Diab 34 Fritz et al 35 Aydin and Yazicioǧlu 36 Young et al 37 1) Eligibility criteria were specified ) Subjects were randomly allocated to groups ) Allocation was concealed ) The groups were similar at baseline regarding the most important prognostic indicators. Results We identified 22,019 studies through database searching. No additional eligible studies were identified through other sources. After duplicates were removed and titles and abstracts of all remaining unique articles were screened, there were 11 full-text articles that needed to be assessed to verify their eligibility for inclusion in the present study. Ultimately, 6 of them were excluded for various reasons (Fig. 1), resulting in 5 studies selected for this review All 5 studies were eligible for quantitative synthesis by pooling their data for meta-analysis. Three studies 34,35,37 analyzed the effect of the mechanical cervical traction combined with physical therapy procedures versus physical therapy procedures alone, and 1 study 36 evaluated the effect of the manual cervical traction combined with other physical therapy procedures versus physical therapy procedures alone. One study evaluated the effects of both manual and mechanical traction. 36 All studies evaluated pain, and 4 of 5 studies 33 35,37 also evaluated disability. All studies had short-term follow-ups for both, and 2 of them 34,35 also had mid- and long-term follow-ups. Overall, the 5 included studies from Africa (Egypt, Tunisia), America (United States), and Asia (Turkey) were published from 2009 to The number of patients who were enrolled and completed baseline assessments was 449 (range = ), with a mean sample size of 90 participants. The mean age of the study participants was approximately 44 years. All studies used intermittent traction, with the number of sessions ranging from 7 to 15, and duration of application from 15 to 50 minutes. The starting position was a pain-free position, 33,36 signfree position, 34 or 15 degrees of flexion position. 35,37 The starting traction force ranged from 5 to 6 kg 33,35,36 and 9 kg, 34,37 and was incrementally adjusted based on patient tolerance and symptom response. All selected trials were judged by the reviewers to be clinically homogeneous, and meta-analysis was performed. For these studies, effect sizes and associated 95% CIs for the individual trials ) There was blinding of all subjects ) There was blinding of all therapists who administered the therapy ) There was blinding of all assessors who measured at least 1 key outcome. 8) Measures of at least 1 key outcome were obtained from more than 85% of the subjects initially allocated to groups. 9) All subjects for whom outcome measures were available received the treatment or control condition as allocated or, where this was not the case, data for at least 1 key outcome was analyzed by intention to treat. 10) The results of between-group statistical comparisons are reported for at least 1 key outcome. 11) The study provides both point measures and measures of variability for at least 1 key outcome Total were calculated and were presented in a forest plot grouped according to treatment, followed by and follow-ups (Fig. 2). Quality Assessment The methodological quality of the studies was assessed by the PEDro scale. Two evaluators independently judged all of the studies included in the review according to the scale criteria. Then, they compared their evaluation with the published PEDro scores, when available, and reached an agreement in the other cases by a meeting consultation with a third author (P.P.). Four studies reached the minimum score (6/10) to be considered of good quality. The range was from 5/10 to 7/10, with an average higher than this threshold (mean score = 6.6). The 3 worst scored quality criteria on which all the studies failed to obtain a positive score were related to blinding, specifically the blinding of: 1) the physical therapists, 2) the patients toward the treatment, which was not possible given the type of treatment performed, and 3) the blinding of assessors. Concealed allocation of patients to groups 238 Physical Therapy Volume 98 Number 4 April 2018

9 Table 3. Mechanical and Manual Cervical Traction Type of Cervical Traction Mechanical Manual Follow-Up Short-term (0 3 mo of Midterm (3 6 mo of Long-term (6 12 mo of Short-term (0 3 mo of Midterm (3 6 mo of Long-term (6 12 mo of Short-term (0 3 mo of No. of Studies and data analysis according to the intention-to-treat method were found in 4 of 5 studies. The best favorable scored criteria for all studies were those related to the statistical analysis of the results, the randomization of subjects in the groups, and the initial comparability of the most important prognostic factors (Tab. 2). Outcomes of Intervention Table 3 shows the follow-up study findings for pain and disability with respect to the pooled effect size for intervention, 95% CI values, assessment of heterogeneity across studies (Q and I-square statistics), and Egger t test for potential publication bias. Forest plots for each outcome are shown in No. of Participants Pooled Effect Size, Reported as Hedges g (95% CI) ( 1.63 to 0.06) ( 2.25 to 0.10 ) ( 3.40 to 0.79) ( 3.50 to 0.51) ( 1.81 to 0.28) ( 7.79 to 2.06) ( 1.39 to 0.30) Figure 2. Forest plots depict the effect size calculated for each study by outcome as well as the overall effect size obtained for the outcome across studies at each time interval. Forest plots also indicate whether the effects obtained in each study across studies favor the control group or the intervention group. Mechanical Cervical Traction Combined With Other Physical Therapist Interventions Versus Physical Therapist Interventions Alone Pain. Four studies 33 35,37 assessed this outcome in the short term (up to 3 months after the baseline) and 2 studies assessed this outcome in the mid term (from 3 to 6 months) and P Pain Q for Heterogeneity P I 2 (%) z for Publication Bias Not Not Disability long term (from 6 to 12 months). 33,35 In the 4 studies that assessed pain at short term, the overall effect size of mechanical cervical traction was large and significant (g = 0.85), with a range from 1.63 to In the studies that assessed pain in mid term after the intervention, the overall effect size of mechanical traction was large and significant (g = 1.17), with a range from 2.25 to At long-term follow-up, the overall effect size of mechanical traction was large but not significant (g = 1.30), with a range from 3.40 to A high heterogeneity of findings appeared for the 3 : short-term pain, mid-term pain, and long-term pain (P <.05). A significant and negative Egger t test was found for the first of the P Not Not Not Not Pain Not Not Not Not April 2018 Volume 98 Number 4 Physical Therapy 239

10 3, although it was not possible to apply this test to the other due to the low number of studies. Disability. Four studies 33 35,37 assessed this outcome in the short term, and 2 studies assessed this outcome in the mid term and long term. 33,35 In the 4 studies that assessed disability at short term, the overall effect size of mechanical cervical traction was large and not significant (g = 1.50), with a range from 3.50 to In the studies that assessed disability from 3 to 6 months after the intervention, the overall effect size of mechanical traction was large and significant (g = 1.05), with a range from 1.81 to At the longterm follow-up, the overall effect size of mechanical traction was large and not significant (g = 2.87), with a range from 7.79 to A high heterogeneity of findings appeared for the 3 : short-term disability, mid-term disability, and long-term disability (P <.05). A significant and negative Egger t test was found for the first of the 3, although it was not possible to apply this test to the other due to the low number of studies. In summary, using GRADE criteria, this review found moderate-quality evidence that mechanical cervical traction combined with other physical therapy procedures is more effective in improving pain and disability than physical therapy procedures alone at short-term, mid-term, and long-term follow-ups (eappendix 1, available at academic.oup.com/ptj). Manual Cervical Traction Combined With Other Physical Therapist Interventions Versus Physical Therapist Interventions Alone Pain. Two studies 33,36 of manual cervical traction assessed this outcome in the short term. The overall effect size of manual cervical traction was large and significant (g = 0.85), ranging from 1.39 to A moderate heterogeneity of findings appeared for this outcome (P <.05). An Egger t test was not applied due to the small number of studies. In summary, using GRADE criteria, there is low-quality evidence that manual cervical traction combined with other physical therapy procedures is more effective in improving pain than physical therapy procedures alone at shortterm follow-up (eappendix 2, available at Discussion This review aimed to investigate the effectiveness of the traction added to physical therapy procedures for pain and disability in CR. Our meta-analyses demonstrated significant results of mechanical traction on the reduction of pain at the immediately posttreatment and 6-month follow-ups, based on moderate evidence according to GRADE criteria. With respect to manual traction, significant improvements in pain and disability were reported immediately posttreatment, although this finding is supported only by low-quality evidence according to GRADE criteria. We also found that the effects of mechanical traction on disability appeared to be large at different time intervals, but significant only at mid-term follow-up. We did not find a rationale to explain the differences among these results obtained at various follow-ups. The results of our meta-analyses albeit based on a literature judged as having low-to-moderate quality according to GRADE criteria despite the PEDro scores of individual studies suggest that both mechanical and manual traction may be effective treatment for pain when added to various other physical therapy procedures in CR. Nevertheless, it is beyond the scope of the present review to determine whether mechanical or manual traction induces significantly different effects from each other. Physical therapy procedures that are most often combined with traction include physical modalities, passive mobilization, massage, stretching, and active exercises following a multimodal approach. Typically, implementation of a multimodal approach in the clinic proceeds first with passive modalities (eg, electrical modalities, soft tissue techniques, thoracic spine mobilization/ manipulation, stretching, traction) to prepare the neuromusculosketal tissues for more active intervention such as exercise. Such considerations of typical clinical practice in developing research protocols might enhance the ability to translate research findings into clinical protocols. Judging the literature as a whole, only low-to-moderate quality evidence is currently available to support the effectiveness of traction, both at short term and long term, for pain and disability. 17,18,38 However, the current literature does not indicate which physical therapy procedures are most effective in CR, either alone or when added to traction, despite the rigor of the studies we analyzed as indicated by PEDro score. Furthermore, we are unable to ascertain what impact the incorporation of newer procedures such as dry needling or acupuncture might have in future systematic reviews and meta-analyses. To date, this is the first systematic review aimed to investigate the effectiveness of the traction added to physical therapy procedures for pain and disability in CR. Previous reviews investigated the effectiveness of noninvasive treatment for CR 17,18 and compared the effects of active and sham traction. Salt et al 17 found that there was no statistically significant difference between the effectiveness of active and placebo traction in terms of pain, function, or disability, either at postintervention or at various follow ups. Thoomes et al 18 confirmed these results, concluding that a low level of evidence supported the lack of difference in effects between active and placebo traction. On the contrary, we found significant results of manual and mechanical traction, when added to other physical therapy procedures, on the reduction of pain at short- and mid-term. The reason for our different results may depend on 4 additional studies we included, of which were published after the most recent review by Thoomes et al. 18 There is some variability in the number of treatment sessions used in studies on the effectiveness of physical therapist intervention for CR, including traction procedures, with number of sessions ranging from 7 to 15. The duration of each treatment session also varies, 240 Physical Therapy Volume 98 Number 4 April 2018

11 from 15 to 50 minutes. We can observe that some studies with a higher number of sessions and longer traction times 33-34,36 show better results for pain and disability, suggesting a potential dose-response relationship. The results of our meta-analysis are relevant for clinical practice due to its external (i.e., ecological) validity because the treatments and the outcome measures included in our review are very similar to those more commonly used in a clinical context. Specifically, a mean number of 11 sessions, 15,39 41 a mean duration of the application force of 25 minutes, the use of traction in addition to other forms of physical therapist interventions, and the evaluation of pain and disability are typical in the treatment of patients with CR. Our search was extensive and used many databases without restrictions that can reduce the collected titles. Furthermore, the selection and qualitative assessment were made by 2 independent assessors. The most important limitation of the present work is the lack of homogeneity in CR diagnostic criteria among the included studies. For example, in 1 study, the eligibility criteria were a chief report of neck pain with symptoms (pain or numbness) extending distal to the acromioclavicular joint or caudal to the superior border of the scapula. 35 In another one, the inclusion criteria were the presence of a disk herniation confirmed by imaging, dermatomal numbness, continuous or intermittent pain or discomfort persisting for more than 3 months, radiation of arm pain in the dermatomal areas, diminished deep tendon reflexes in the affected arm, and no spinal deformity. In addition, symptoms had to be modified depending on different movements or positions. Finally, inclusion criteria included a cluster of tests identified by Wainner et al. 21,34 Therefore, the diagnosis of CR in some cases was intended to convey a clinical presentation as simply pain radiating to the arm, yet in other cases as a deficit of nerve conduction. Moreover, there was a lack of homogeneity of the underlying pathogeneses. These discrepancies could be considered as random errors and might influence the effect estimates. No included study adopted continuous traction, and therefore we cannot draw any conclusion on this method of traction application. Our study is unable to draw conclusions regarding standard duration and positioning of traction, which are likely key determinants of clinical. Another limit of this review is related to small sample sizes in 2 studies with results favoring traction. Publication bias also is a potential limitation of any review. A strong publication bias, however, is unlikely in this review because studies in all languages and for any year of publication as well as studies from all countries were included. Using only clinical trials may have influenced the potential publication bias, but this approach allowed us to derive our conclusions using more rigorous studies according to PEDro scoring. In light of our results, the current literature lends support to the use of the traction in addition to other physical therapy procedures for pain reduction, with less substantial impact on function and disability. Further studies should investigate the most effective traction method and dosage, the subgroups of patients with CR, or the pain stage (acute, subacute, or chronic) most benefited by this intervention and the physical therapy procedures that yield the best results when combined with traction. Author Contributions Concept/idea/research design: A. Romeo, C. Vanti, P. Pillastrini, L. Bertozzi Writing: A. Romeo, C. Vanti, V. Boldrini, A.A. Guccione, L. Bertozzi Data collection: A. Romeo, V. Boldrini, M. Ruggeri Data analysis: A. Romeo, M. Ruggeri, A.A. Guccione, L. Bertozzi Project management: A. Romeo, C. Vanti Consultation (including review of manuscript before submitting): A. Romeo, C. Vanti, A.A. Guccione, P. Pillastrini Disclosures All authors completed the ICJME Form for Disclosure of Potential Conflicts of Interest and reported no disclosures relevant to this work. DOI: /ptj/pzy001 References 1 Radhakrishnan K, Litchy WJ, O Fallon WM, Kurland LT. Epidemiology of cervical radiculopathy: a population-based study from Rochester, Minnesota, 1976 through Brain. 1994;117: Nobuhiro T, Yoshinori F, Howard S, Yoshikazu I. The anatomic relation among the nerve roots, intervertebral foramina, and intervertebral discs of the cervical spine. Spine. 2000;25: Shelerud RA, Paynter KS. Rarer causes of radiculopathy: spinal tumors, infections, and other unusual causes. Phys Med Rehabil Clin North Am. 2002;13: Kim HJ, Nemani VM, Piyaskulkaew C, Vargas SR. Cervical radiculopathy: incidence and treatment of 1,420 consecutive cases. Asian Spine J. 2016;10: Thoomes EJ, Scholten-Peeters GG, De- Boer AJ, Olsthoorn RA. Lack of uniform diagnostic criteria for cervical radiculopathy in conservative intervention studies: a systematic review. Eur Spine. 2012;21: Bogduk N. On the definitions and physiology of back pain, referred pain, and radicular pain. Pain. 2009;147: Salemi G, Savettieri G, Meneghini F, et al. Prevalence of cervical spondylotic radiculopathy: a door-to-door survey in a Sicilian municipality. Acta Neurol Scand. 1996;93: Wong JJ, Cote P, Quesnele JJ, Stern PJ. The course and prognostic factors of symptomatic cervical disc herniation with radiculopathy: a systematic review of the literature. Spine J. 2014;14: Daffner S, Hilibrand A, Hanscom BS, Brislin BT. Impact of neck and arm pain on overall health status. Spine. 2003;28: Nikolaidis I, Fouyas IP, Sandercock PA, Statham PF. Surgery for cervical radiculopathy or myelopathy. Cochrane Database Syst Rev Jan 20;(1):CD Carragee EJ, Hurwitz EL, Cheng I, et al. Treatment of neck pain: injections and surgical interventions results of the Bone and Joint Decade Task Force on Neck Pain and Its Associated Disorders. Spine (Phila Pa 1976). 2008;33(4 suppl):s153 S Albert TJ, Murrell SE. Surgical management of cervical radiculopathy. J Am Acad Orthop Surg. 1999;7: Diwan S, Manchikanti L, Benyamin RM, et al. Effectiveness of cervical epidural injections in the management of chronic neck and upper extremity pain. Pain Physician. 2012;15(4):E405-E Trinh KV, Graham N, Gross AR, et al. Acupuncture for neck disorders. Cochrane Database Syst Rev. 2006;(3):CD Cleland JA, Whitman JM, Fritz JM, Palmer JA. Manual physical therapy, cervical traction, and strengthening exercises in patients with cervical radiculopathy: a case series. J Orthop Sports Phys Ther. 2005;35: April 2018 Volume 98 Number 4 Physical Therapy 241

12 16 Waldrop MA. Diagnosis and treatment of cervical radiculopathy using a clinical prediction rule and a multimodal intervention approach: a case series. J Orthop Sports Phys Ther. 2006;36: Salt E, Wright C, Kelly S, Dean A. A systematic literature review on the effectiveness of non-invasive therapy for cervicobrachial pain. Man Ther. 2011;16: Thoomes EJ, Scholten-Peeters W, Koes B, Falla D. The effectiveness of conservative treatment for patients with cervical radiculopathy: a systematic review. Clin J Pain. 2013;29: Michlovitz SL, Nolan TP. Modalities for Therapeutic Intervention. 4th ed. Philadelphia, PA: FA Davis Co; 2005; Kisner C, Colby LA. Therapeutic Exercise: Foundations and Techniques. 3rd ed. Philadelphia, PA: FA Davis Company; Wainner RS, Fritz JM, Irrgang JJ, Delitto A. Reliability and diagnostic accuracy of the clinical examination and patient self-report measures for cervical radiculopathy. Spine. 2003;28: Maher CG, Sherrington C, Herbert RD, Moseley AM. Reliability of the PEDro scale for rating quality of randomized controlled trials. Phys Ther. 2003;83: De Morton NA. The PEDro scale is a valid measure of the methodological quality of clinical trials: a demographic study. Aust J Physiother. 2009;55: DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials.1986;7: Hedges LV, Olkin I. Statistical Methods for Meta-Analysis. Orlando, FL: Academic Press; ProMeta software. com. Accessed December 30, Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates; Gavaghan DJ, Moore RA, McQuay HJ. An evaluation of homogeneity tests in meta-analyses in pain using simulations of individual patient data. Pain. 2000;85: Higgins J, Thompson SG. Quantifying heterogeneity in a meta analysis. Stat Med. 2002;21: Furlan AD, Pennick V, Bombardier C, et al updated method guidelines for systematic reviews in the Cochrane Back Review Group. Spine (Phila Pa 1976). 2009;34: Atkins D, Best D, Briss PA, et al. Grading quality of evidence and strength of recommendations. BMJ. 2004;328: BMJ Clinical Evidence. What is GRADE? static/ebm/learn/ html. Accessed December 30, Jellad A, Ben Salah Z, Boudokhane S, Migaou H. The value of intermittent cervical traction in recent cervical radiculopathy. Phys Rehabil Med. 2009;52: Moustafa IM, Diab AA. Multimodal treatment program comparing 2 different traction approaches for patients with discogenic cervical radiculopathy: a randomized controlled trial. J Chiropr Med. 2014;13: Fritz JM, Thackeray A, Brennan GP, Childs JD. Exercise only, exercise with mechanical traction, or exercise with over-door traction for patients with cervical radiculopathy, with or without consideration of status on a previously described subgrouping rule: a randomized clinical trial. J Orthop Sports Phys Ther. 2014;44: Albayrak Aydin N, Yazicioǧlu K. Cervical intermittent traction: does it really work in cervical radiculopathy due to herniated disc? Turkish Journal of Physical Medicine and Rehabilitation. 2012;58: Young IA, Michener LA, Cleland JA, Aguilera AJ. Manual therapy, exercise, and traction for patients with cervical radiculopathy: a randomized clinical trial. Phys Ther. 2009;89: Eubanks J. Cervical radiculopathy: nonoperative management of neck pain and radicular symptoms. Am Fam Physician. 2010;81: Langevin P, Desmeules F, Lamothe M, Robitaille S, Roy JS. Comparison of 2 manual therapy and exercise protocols for cervical radiculopathy: a randomized clinical trial evaluating shortterm effects. J Orthop Sports Phys Ther. 2015;45: Radpasand M. Use of a multimodal conservative management protocol for the treatment of a patient with cervical radiculopathy. J Chiropr Med. 2011;10: Savva C, Giakas G. The effect of cervical traction combined with neural mobilization on pain and disability in cervical radiculopathy: a case report. Man Ther. 2013;18: Physical Therapy Volume 98 Number 4 April 2018