The role of functional, radiological and self-reported measures in predicting clinical outcome in spondylotic cervical radiculopathy

Abstract

BackgroundCervical radiculopathy (CR) results in significant disability and pain and is commonly treated conservatively with satisfactory clinical outcomes. However, a considerable number of patients require surgery to prevent irreversible neurological damage or when pain is unremitting. Both conservative and surgical treatments are characterised by a plethora of strategies and interventions, making comparisons between treatment groups a difficult task.The ability to predict the likely outcomes of treatment is important to clinicians involved in treatment and serves as an important basis for health policy, resource management and core knowledge essential for patients to have an informed consent when being offered both conservative and surgical interventions. Generic and condition specific measures in clinical conditions may have a role in predicting clinical outcome after treatment. Common measures employed in the examination and evaluation of cervical radiculopathy patients include self reported measures of pain and disability, measures of impairment and function, neurological and radiological evidence of the disease. Unfortunately, few studies exist that provide knowledge in understanding factors associated with the presentation of cervical radiculopathy and the ensuing long-term clinical outcome.The objective of this thesis was to address the limited understanding of the impairment measures and prognostic outcomes associated with cervical radiculopathy in an Indian population for both surgical and conservative treatment strategies. In these aspects, this thesis was original.AimThe principal aim of this thesis was to evaluate the role of clinical (pain, disability), functional (posture, cervical ROM), radiological (radiographic segmental curvature and segmental movements in the sagittal plane), socio-demographic and lifestyle factors (age, gender, BMI, work characteristics, physical activity, smoking, life-style changes, duration of symptoms, co-morbidities and number of previous episodes) in predicting clinical outcome (pain and disability) at one year in a cervical radiculopathy cohort which was treated conservatively or surgically. To support and substantiate the primary aim, associations, relationships, differences between the outcome variables at different measurement times, and effects of treatment were determined. Furthermore, a series of studies were undertaken examining key assessments of cervical spine impairment to further support the primary clinical outcome study.MethodsThe main element of the thesis derived clinical data from a sample of convenience of 163 patients (109 patients in the conservative group and 54 in the surgical group). All measurements were done at baseline, that is, prior to any form of intervention, and then at pre-determined intervals until a one year follow-up period in both treatment groups. After baseline assessments, the conservative group underwent medical treatment and physiotherapy whereas the surgery group underwent surgery, and post-surgery, were given ergonomic advice and exercises. An intention to treat analysis approach was adopted for dropouts (14% at 12 months).A large series of other studies were undertaken prior to and in parallel with the main clinical trial (Agarwal et al. 2005 a,b,c; Agarwal et al. 2006). Emphasis was placed on tester reliability for each measure used in this study and consequently methodological studies establishing reliability and validity of measures were carried out and published. Cervical range of motion, a key measure of impairment and neck functions, frequently used by physiotherapists as an assessment tool, was analysed to determine the effects of age, gender and repeat measurements over time in asymptomatic individuals (n = 219) as well as differences in range of motion between asymptomatics and patients with cervical radiculopathy. During the initial course of this research, it became evident that the pain and disability questionnaire (Neck Pain and Disability scale) in the English language was not applicable to a section of the patient population. This was therefore translated into Hindi (the national language of India).The outcome variables with their respective measuring instruments were neck pain and arm pain (101 Numerical Pain Rating Scale), pain and disability (NPAD), depression (NPAD factor 3 scores), posture (head neck angle), cervical range of movement (Spin-T goniometer), radiographic sagittal segmental curvature (Posterior Tangent Method) and radiographic sagittal segmental motion (Penning’s method). Demographics and lifestyle factors consisted of either continuous or dichotomous variables.A range of parametric and non-parametric tests analysed the correlations and differences between outcome variables at different times of measurements as well as determined treatment efficacy. The Clinical Prediction Rule (CPR) analysis was used to determine the group of predictor variables which could result in a successful outcome in a CR cohort, following conservative or surgical intervention. The NPAD (English and Hindi) was the outcome criterion for the CPR analysis of this study with a score of <22 (minimal or no pain and disability) used as a responder threshold.ResultsThe Hindi version of the NPAD was tested as valid and reliable. The maximum typical error values between repeat measurements of radiographic segmental curvature (levels C2-7), radiographic segmental flexion-extension motion (levels C2- 7) and composite active cervical ROM did not exceed 3°. However, when composite cervical ROM was tested for reliability of repeat measurements at 3 months intervals (total 6 months from baseline), typical error values were slightly higher (not exceeding 5°). Similarly, typical error values for Head-neck angle were 4° for same day measurements, but higher, reaching up to maximum 8°, for repeat measurements at 3 weeks.The effects of age, gender and clinical condition on composite active cervical ROM showed that within a normal population, for both genders, a systematic change in cervical ROM was noted with the rate of range of motion loss varying between 3° to 5° per decade from age 20 to 80 years. Differences in composite active cervical ROM between a cervical radiculopathy group and matched controls measurements suggested that flexion and extension range of motion were more likely reduced in the CR cohort. Similarly, patients with CR had systematically decreased radiographic sagittal flexion-extension motion of the cervical spine (C2- 7) compared to an asymptomatic cohort.At baseline, the two treatment groups were comparable for age, gender, BMI, marital status, duration of symptoms (in weeks) and co-morbidities, radiographic segmental curvature (C2-7) as well as head-neck angle measurements whilst the surgery group patients showed more severity with higher levels of pain and disability and neurological deficits. Simultaneously, the surgery group also demonstrated more radiological segmental flexion-extension motion and composite active cervical ROM when compared to the conservative group.Outcome at 12 months showed a statistically significant improvement in both groups for neck pain, arm pain and disability measures. The surgery group with higher baseline scores demonstrated a larger reduction in pain scores than the conservative group. For repeat measurements following intervention, the conservative group demonstrated a systematic pattern suggesting improvements for radiographic sagittal segmental curvature and flexion-extension motion values and composite range of motion in all directions. However, this consistency for the same variables was not so in the surgical group during repeat measurements following surgery. Head – neck angle was neither sensitive nor responsive over the treatment period for either group. Both treatment groups showed improvement from baseline to final measurements for all neurological scores. Hundred percent improvement at final follow-up was not achieved for any neurological sign in either group.Significant bivariate correlations were consistent in establishing a negative relationship between radiographical sagittal segmental curvature values at symptomatic levels and arm pain scores, between composite active cervical ROM and pain and disability measures and between radiological segmental flexionextension motion at the diagnosed symptomatic level/s and final neck pain scores. This implies that reduced radiographic curvature and flexion-extension motion as well as composite range of motion are correlated with increased pain and disability or vice-versa.Further, to test the principal hypothesis, the combination of baseline factors that predicted good clinical outcomes at different time points in Indian CR patients, treated conservatively and surgically were:Conservative (3 months): Age < 40 years, BMI <24.4, No recent lifestyle changes, Duration of symptoms ≥ 33 weeks, NPAD factor 1 < 18, Neck flexion < 40°, Number of previous episodes <2. Three of the seven predictors generated an 8.54 fold increase [likelihood ratio 95% Confidence Interval (CI) 4.17 to 17.48] for the individual to be a responder. With three predictors, post-test probability of success increased to 82% from a pre-test probability of success of 35%.Conservative (12 months): Education level ≥ post graduate, PTM C2-7 ≥ 11°, PTM (symptomatic level) ≥ 2.5°, Rotation (right) > 55°. Three of the four predictors generated a 9.8 fold increase (likelihood ratio 95% CI 3.3 to 29.8) for the individual to be a responder, with a post-test probability of success at 93%.Surgery (3 months): Age < 40 years, Head neck angle < 40°, NPAD scores < 55. Two of the three predictors generated a positive likelihood ratio of 10.15, which is a 10 fold increase (likelihood ratio 95% CI 3.4 to 30.7) for the individual to be a responder, increasing the post-test probability of success from 31% to 82%.Surgery (12 months): Duration of symptoms ≥ 33 weeks, NPAD scores < 55, Neck flexion < 40°, Rotation (right) > 55°. Two of the four predictors generated a 16-fold increase (likelihood ratio 95% CI 2.3 to 112) for the individual to be a responder and the post test probability of success increased to 94% from a pre-test probability of success at 50%.ConclusionIn conclusion, this study was able to provide original and long-term assessments of cervical spine movement characteristics as well as translating the primary outcome into Hindi. The clinical finding was able to identify predictor clusters that provide level IV evidence of predicting outcomes at different timelines for cervical radiculopathy for both conservative and surgically treated patients. The use of these predictors in future studies may help in decision making for the appropriate type of treatment and expected outcome in CR patients.