Donald R. Murphy, DC, Eric L. Hurwitz, DC, PhD, Amy Gregory, DC, Ronald Clary, DC
Department of Community Health,
Brown University Medical School,
Providence RI, USA.
OBJECTIVE: The objective of this study was to describe the clinical outcomes of patients with cervical radiculopathy (CR) treated nonsurgically after a diagnosis-based decision rule.
METHODS: A prospective observational cohort study on consecutive patients with CR was performed. Data on 35 consecutive patients were collected at baseline, at the end of the active treatment, and at a minimum of 3 months after cessation of treatment. Disability was measured using the Bournemouth Disability Questionnaire. Pain intensity was measured using the Numerical Pain Rating Scale. Patients were also asked to self-rate their improvement.
RESULTS: Complete outcome data were available for 31 of the 35 patients. Twenty-seven patients were reached for long-term follow-up. The mean number of months from last treatment to follow-up was 8.2 months. Seventeen patients (49%) reported their improvement as “excellent” and another 14 (40%) did so as “good.” The mean patient-rated improvement was 88.2%. The mean percentage of improvement in the Bournemouth Disability Questionnaire score was 78%. The mean percentage of improvement in the Numerical Pain Rating Scale score was 72%. Twenty-four of 31 (77.4%) patients had a clinically significant improvement from baseline to the end of treatment, and 25 of 27 (92.6%) had a clinically significant improvement from baseline to long-term follow-up.
CONCLUSIONS: The management strategy that we studied yielded favorable outcomes in this patient sample and appears to be a safe option for patients with CR. However, the absence of randomization and a control group limits interpretation with regard to clinical effectiveness. Randomized clinical trials are necessary to distinguish treatment effects from the natural history of CR.
From the FULL TEXT Article
Although it was not the purpose of this study to assess the results of manipulation alone, this is the first prospective observational study on patients with a documented cause of CR (ie, LCS and/or CDH) who were treated with a protocol in which manipulation was a central part. Schliesser et al  looked retrospectively into 39 patients treated with flexion-distraction–type manipulation. Only 22 of those patients had complete data to be analyzed. They found a mean improvement of 41.4 mm on a 100-mm visual analogue scale. Unfortunately, effect of treatment on disability was not measured. BenEliyahu  published a case series of 27 patients with herniated disks, 11 of which involved the cervical spine. However, in only 6 of the 27 patients was there specific nerve root encroachment shown on MRI. Twenty-two of the 27 patients showed “good” clinical outcomes.
The current study supports the notion that nonsurgical measures are a viable alternative to surgery in patients with CR. Specifically, it suggests that the approach taken here may be beneficial in this patient population. Other studies have shown similar findings. [8, 11, 12] Heckmann et al  retrospectively reviewed 60 patients, 65% of whom were treated nonsurgically and 35% of whom were treated with surgery. Although they did not use rigorous outcome measures with regard to pain and disability, they found that 89.7% of the nonsurgically treated patients did not feel disabled at an average of 5.5 years later, whereas only 66.7% of the surgically treated patients did not feel disabled. Persson et al,  in a prospective randomized controlled trial of 81 patients, compared surgical intervention with conservative treatment. At 4 months, there was greater improvement in pain, sensory disturbance, and muscle strength in the surgically treated group; however, at 1 year, there was essentially no difference. Importantly, the equality of outcome between the groups at 1 year was attributed to gradual improvement in the nonsurgical group as well as recurrence of pain in many surgical patients. This may suggest that rehabilitation after surgery with a focus on prevention of recurrence may be important in surgically treated patients.
Interestingly, in the study by Heckmann et al,  the authors noted that whereas upper-extremity symptoms generally improved or resolved, neck pain improved in only 36% of the conservatively managed group and in 20% of the surgically managed group. Improvement in neck pain vs improvement in extremity symptoms was not specifically measured in the present study, but it was our observation that there was a direct correlation between improvement in neck and extremity symptoms; patients who improved were not generally left with significant residual neck pain. This may relate to the idea inherent in the management protocol used in this study; that is, the recognition that pain in patients with cervical radicular syndromes is often multifactorial. As a result, an attempt to detect and treat pain-generating tissues other than just the nerve root was central to the clinical approach. It may be that the neck pain itself in these patients arises from a joint, disk, or muscle, which must be treated for effective resolution of the neck pain component of the clinical picture. Additional work in this area is needed to further assess this issue.
It is difficult to compare the results presented here with those of other studies on nonsurgical management of patients with CR owing to differences in design, outcome measurement, and selection criteria. For example, in the study by Saal et al,  patients who eventually went on to require surgery were eliminated from the data, whereas in this study, the data from the single patient who required surgery were included in the outcome calculations. Also, in the studies by Persson et al,  patients with “disk bulge,” a finding believed to be of questionable clinical significance,  were included.
Close to 90% of patients described their improvement as either “excellent” or “good.” Clinically significant improvement was seen in disability both at final follow-up and over the long term. Pain intensity improved by 62% at final follow-up and by 72% over the long term. Importantly, no major complication was seen in any patient secondary to treatment, with minor transitory discomfort occurring in just less than half of the patients. Only 9 of these cases were secondary to manipulation. Only 1 patient went on to require surgery.
The scores for self-rated percentage of improvement (75.4 vs 88.2), percentage of improvement in disability (53.0 vs 78.0), and percentage of improvement in pain intensity (62.0 vs 72.0) were all greater at long-term follow-up than at the end of treatment. This would suggest that patients continued to improve after being discharged from care. The reason for this may relate to the emphasis in the management strategy on cervical stabilization exercise, which is designed to train optimal motor control ; neural mobilization exercise, which is designed to improve and maintain nerve root mobility ; and education to increase patient confidence, decrease or limit fear-avoidance beliefs, and improve posture and movement patterns.  However, the study design does not allow for firm conclusions to be drawn. Nonetheless, these data support the theory that the management strategy used in this study promotes continued improvement even after cessation of active treatment.
The management strategy reported involved a systematic approach to diagnosis that took into consideration nonmusculoskeletal causes or serious illnesses, the specific pain-generating entities that may be present, and perpetuating factors that may be involved. With regard to pain-generating tissues, some controversy exists as to whether these can be identified in most patients with spinal pain; however, some degree of reliability and/or validity has been shown for joint pain, [18, 19] nerve root pain,  and myofascial trigger points [38, 39] in the cervical spine. With regard to perpetuating factors in patients with CR, far less is known, and the methods applied in this study are based on clinical experience  as well as on studies that have evaluated motor control dysfunction in patients with cervical pain  and pain hypersensitivity  and nonorganic and fear-avoidance factors [43, 44] in spine pain patients in general. The treatment procedures used were based on those factors that were deemed by the treating chiropractic physician to be most important in each individual patient. Therefore, the management strategy was individualized. Because all patients had at least one pain-generating entity in common (ie, CR), there was significant overlap in treatment. Some form of neural mobilization was used in all patients. Also, painful joint dysfunction was found based on palpation in all patients; thus, all patients were treated with manipulation. Although each patient was given some form of basic cervical stabilization exercises, only those in whom test results for poor cervical motor control were positive were provided detailed cervical stabilization training. 
The mean number of visits that was deemed by the treating clinicians to be required was approximately 12. We feel that this accurately represents the optimal number of treatments for patients with CR. This is supported by the fact that patients appeared to continue to improve even after dismissal from care. Were the number of treatments provided here not adequate, it would be expected that the patients would likely deteriorate once treatment was withdrawn. It should be emphasized that the number of treatments deemed to be required in each patient was made on an individual basis, with some requiring as few as 4 visits and others as many as 24. This study does not support the use of a uniform number of visits automatically provided for all patients without regard for individual patient needs.
Some controversy exists over the use of manipulation in patients with CR. It has been stated that this treatment is contraindicated in the presence of a herniated disk.  Saal et al,  in their study on nonoperative management of patients with CR, made it a point to state that “Forceful joint manipulation was not used.” Haas et al  listed disk herniation and CR as “contraindications to cervical manipulation.” Wolff and Levine,  citing Atchison et al,  stated that HVLA techniques are contraindicated in patients with central CDH. However, of the 35 patients in the current study, 8 had the finding of central CDH, 4 of whom were treated with some form of HVLA technique. No complication, even short-term increase in pain, was seen in any of these patients. The data presented here would suggest that manipulation, when applied by properly trained and experienced practitioners, is potentially a safe option for patients with CR. The absence of a control group in this study does not weaken this conclusion because it is felt that, although such absence limits the interpretation of a study with regard to outcome, it does not limit interpretation with regard to safety.  However, because of this study’s sample size, rare complications are not likely to be detected. Larger studies are needed to confirm the safety of the treatments used in this study.
Eighteen patients in this study were treated with some form of HVLA manipulative technique and 13 were treated with some form of ME technique, which does not involve a short and quick movement but rather uses slow maneuvers that are aided by breathing and eye movements.  There was no difference in outcome between patients treated with HVLA and those treated with ME. This would appear to support other studies that have found no difference in outcome between patients treated with HVLA manipulation and those treated with a non-HVLA technique. [50, 51] However, the patients in the present study were not randomized to receive these treatments and the sample size is too small to draw firm conclusions.
There are several important limitations to this study. It is not a randomized controlled trial; thus, interpretation of the results must be made with caution. The natural history of CR is generally thought to be favorable ; however, data on untreated patients with this disorder are lacking. Because of this paucity of natural history data and because the current study did not involve an untreated or placebo control group, there is no way to compare the outcomes reported here with those in the untreated situation.
This study was pragmatic in nature (ie, it assessed a treatment approach in which individual patients were treated with one or more methods). Thus, it does not allow one to determine the extent to which any particular method contributed to the clinical outcome in each patient. However, one of the factors that make this study useful is that it measures the outcome of a treatment protocol that is used in everyday practice (ie, it is a real world study). Practitioners who have the appropriate training can apply this protocol in the office setting. Also, it allows for the assessment of an approach that individualized the treatment for each patient.
The long-term follow-up was done by having a research assistant read the BDQ over the phone and record the answers on the questionnaire. The validity of using the BDQ in this way is unknown. Because the follow-up phone calls were made by a research assistant and not by the treating practitioner, the possibility of halo effect may be diminished. Electromyogram was not used in every case to confirm the diagnosis of CR. However, as clinical historical and examination factors have been found to be reliable and accurate in diagnosing CR,  particularly when imaging such as MRI is available,  it was not felt that an electromyogram was necessary in confirming the diagnosis.
It would appear from this study and the other studies discussed here that nonsurgical approaches that are relatively safe and possibly effective are available to patients with CR. Randomized controlled trials are called for to further assess the efficacy of this approach compared with untreated control or other nonsurgical treatments.