記住我
Click on image for details.
ORIGINAL ARTICLE
Year : 2022 | Volume
: 70
| Issue : 8 | Page : 166-174
AIIMS Cervical Myelopathy Score – A New Comprehensive and Objective Patient-Reported Modification of mJOA: Results of an Internal Validation Study
Kanwaljeet Garg1, Vaibhav Vij1, Shashwat Mishra1, Deepti Vibha2, Manmohan Singh1, P Sarat Chandra1, Shashank S Kale1
1 Department of Neurosurgery, All India Institute of Medical Sciences, New Delhi, India
2 Department of Neurology, All India Institute of Medical Sciences, New Delhi, India
Correspondence Address:
Shashwat Mishra
720, CNC, AIIMS, New Delhi - 110 029
India
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/0028-3886.360920
Background: Several scoring systems have been developed for assessment of patients with compressive cervical myelopathy. However, all of these have some shortcomings. We proposed a new modification of the modified Japanese Orthopedic Association (mJOA) score—the AIIMS cervical myelopathy score (ACMS).
Objective: The aim of this study was to compare the ACMS with mJOA score and Nurick score.
Methods: We prospectively studied patients with cervical compressive myelopathy. The new ACMS, mJOA, Nurick proposed by Benzel, and Nurick scores were recorded preoperatively and at three months postoperatively in patients.
Results: Sixty-two patients completed the 3-month follow-up and were included in the final analysis. The mean preoperative and postoperative Nurick, mJOA, and ACMS scores were 3.3 and 3.0, 12.3 and 13.8, and 15.1 and 17.7, respectively. High correlation (Pearson's r > 0.8, 95% CI: 0.94 to 0.97, P < 0.005) was observed between ACMS and mJOA scores for all the individual components of both scales, both in pre- and postoperative assessments. A negative correlation was observed between the occupational ability scores ACMS and the Nurick scale (r = −0.76, 95% CI: −0.83 to − 0.68). No correlation was found with cord/canal ratio on magnetic resonance imaging (MRI) with any of the three scoring systems (preoperative, postoperative, or recovery rates).
Conclusions: The ACMS score showed a good correlation with the mJOA score for evaluation of functional disability in the setting of cervical myelopathy. The patients could themselves report the scores using the ACMS scoring chart. The occupational component of the ACMS also correlated well with the Nurick score.
Keywords: Cervical spine, Japanese orthopedic association score, myelopathy, Nurick grade, outcome
Key Message: We proposed a modification of the mJOA score: the AIIMS cervical myelopathy score (ACMS). The ACMS score offers a composite alternative to both mJOA and Nurick scores in the evaluation of a patient's preoperative and postoperative clinical status.
Cervical spine spondylosis is the leading cause of myelopathy in the middle-aged population.[1],[2],[18],[19] The symptoms of myelopathy result from compression of various spinal cord tracts and are related to their degree of compression. Patients with cervical spondylosis can present with axial neck pain or with symptoms and signs of either radiculopathy or myelopathy. The diagnosis of cervical spondylotic myelopathy (CSM) is primarily clinical with supporting radiology. Furthermore, it is important to stratify patients based on symptom progression in the early stage of the disease because relatively emergent interventions would be indicated in those with rapid progression.[20]
Assessment of a patient's deficits and clinical status is essential for execution of proper management strategies including surgical plan.[21] Many measurement tools are used by clinicians for patients with CSM to quantify the disease severity, assist in decision-making, and to evaluate the outcome of surgical intervention. Assessment becomes more objective with the use of assessment scales. Many of these scales combine assessment of deficits related to the different domains of neural function: motor, sensory, and autonomic. Moreover, the surgical decisions are made objectively rather than subjectively with the use of these assessment tools. It is also possible for the surgeon to objectively ascertain the benefit received from surgery.
Several scoring systems have been developed for this purpose and the two of the most popular ones are the mJOA (Benzel's[3] modified Japanese Orthopedic Association) and Nurick grades. Though both mJOA and Nurick scores assess separate domains of functionality, none is complete and adequate in itself, and a new score is still required that needs to be more comprehensive.[4]
The mJOA score has four sections with a total of 22 items. Total score ranges from 0 to 18. The lesser is the dysfunction on individual parameters, the higher is the score. However, the mJOA score lacks clearly definable elements of the symptomatology associated with cervical myelopathy in certain domains of daily functioning. It uses nebulous terms such as “mild” and “severe” for grading sensory dysfunction but has not defined them further to aid the clinician in categorizing the patient's symptoms accurately. This ambiguity can sometimes lead to confusion between observers and interfere in the accurate and objective assessment of functional impairment.
Nurick score, which was used initially to evaluate functional outcomes in cervical myelopathy patients relates ambulatory status of a patient with employment. The reasoning behind this scheme was that any change in the ambulatory status of a patient would translate into a change in the employment status of a person. However, subsequent studies found that the improvement in function of the lower limb was discordant with the patient's ability to regain employment.[5]
Keeping these drawbacks of the mJOA scores in mind, a new modification of the mJOA score was proposed by us (KG, SM and SSK): the AIIMS cervical myelopathy score (ACMS) [Table 1]a and [Table 1]b. It introduces more objectivity into the scoring system by defining specific aspects of the sensory and sphincter dysfunction. Furthermore, to make it reportable by the patient, we incorporated those features of neurological dysfunction which are easily noted by the patient in their daily activities. Additionally, there is a separate section to evaluate perturbation of occupational activities/ability for self-care.
Thus, we have cam up with a comprehensive scoring system by which we tried to incorporate major aspects of functional disability which was expected in patients of CSM.[22] ACMS builds upon the strengths of existing and popular scoring systems and tries to overcome their shortcomings.
The aim of this study was to compare the newly developed AIIMS modification of mJOA (ACMS) with mJOA score and Nurick score by studying the correlation between the various heads of the assessment scores and patient outcomes following surgical intervention for cervical myelopathy.
Materials and Methods
Study design
We conducted a prospective, single-center, observational study after receiving approval from the institute's ethics committee.
Patient population
Inclusion criteria
Patients with clinical features of cervical myelopathy due to extramedullary compression (cervical disc prolapse, ossified posterior longitudinal ligament [OPLL]) who underwent surgical intervention were included in the study.
Exclusion criteria
Patients who refused to participate in the study or had other coexisting diseases which could produce motor paresis, sensory deficits, gait abnormalities, and autonomic symptoms (e.g., poliomyelitis, Parkinsonism More Details) were excluded from the study. Similarly, patients who had previously undergone any surgical procedures in the cervical spine were also excluded.
Data collection
Patients underwent detailed clinical examination and laboratory investigations. They were operated by either anterior or posterior approach depending upon the magnetic resonance imaging (MRI) and clinical findings.
The new ACMS, mJOA proposed by Benzel[3] and Nurick grade were recorded preoperatively and at three months (±1 week) postoperatively in patients by the same interviewer to minimize interobserver variation. The mJOA and Nurick scores were recorded by the treating clinician while the scores on the ACMS scale were registered by the patient themselves. The patients also evaluated their improvement on a visual analogue scale ranging from 1 to 100 following surgery.
The T1 hypointensity and T2 hyperintensity observed on sagittal MRI were recorded as absent or present. Furthermore, the vertical extent of involvement was divided into three categories as follows:
None: showing no T1/T2 changes;
Single: showing T1/T2 changes limited to one segmental level;
More than one: showing T1/T2 changes at more than one segmental level.
One segment was defined as the length of the spine of the spinal cord seen opposite to the disc space up to the adjacent halves of the superior and inferior vertebral bodies.
Statistical analysis
Though well-established criteria for sample size calculation are lacking for validation of patient reported outcome scales, we estimated an appropriate sample size of 60 based on the thumb rule of maintaining a subject-to-item ratio of more than 5. The data from the patients was recorded into a Microsoft Excel spreadsheet. For comparison between raw scores on the mJOA and ACMS scales, values were normalized to a 100-point scale (minimum-maximum normalization) according to the following formula.
Normalized score = (Patient's score − Minimum possible score)/(Maximum possible score − Minimum possible score) ×100
For example, if a patient scored 13 on the mJOA scale then his normalized score would be
(13 − 0 (minimum possible score))/(18 (maximum possible score) – 0 (minimum possible score)) × 100 = 13/18 × 1 00 = 72.2.
Normalized recovery scores were calculated in a similar manner for each subcategory and the aggregate score
Normalized recovery score = (Postoperative score − Preoperative score)/(Possible range of score) × 100
The recovery rate was calculated as per the following formula:
Recovery rate = (Postoperative score − Preoperative score)/(Maximum possible score − Preoperative score) × 100. Since Nurick grade appreciates with increasing disability, unlike mJOA and ACMS, the differences were reversed in case of Nurick grade.
The extent of agreement between the ACMS scale and mJOA was compared using correlation analysis because the sub-categories in these scales were similar.
Correlation coefficients (Pearson or Spearman rank correlation, depending upon the distribution of the pertinent data) were calculated between the following variables:
Raw ACMS and mJOA scoresNormalized ACMS and mJOA scoresNormalized recovery scores between the subcategory of occupational ability/self-care and observed improvement in subcategory scores pertaining to upper and lower extremity motor/sensory disability, as applicable.Extent of improvement of subjective analogue scale (1–100) and the normalized ACMS and mJOA recovery scores.Paired t-tests compared the extent of change in the mJOA and ACMS scores against the perceived improvement on a subjective scale. For this analysis, the subjective improvement was subcategorized as worse/no improvement, up to 50% improvement, 51%–75% improvement, and 76%–100% improvement.
The change in ACMS and mJOA scores was also studied in relation to the extent of myelopathy-specific MRI changes. A linear regression model of the predicted ACMS score was developed using raw mJOA scores as the independent variable. Cronbach's alpha, which is a measure of the internal consistency of the test items, was calculated to check the level of agreement between the answers provided by the patients with respect to the different functional domains tested.
Data were analyzed using R Programming (R: A language and environment for statistical computing; R Foundation for Statistical Computing, Vienna, Austria).
ResultsThe study included patients who were operated with a diagnosis of extramedullary cervical spine pathologies with myelopathy at our institute from March 2019 to April 2020. Our study included 66 patients. Three (4.5%) patients were lost to follow-up in this study. One patient passed away (due to respiratory distress; had preop severe cervical myelopathy). Sixty-two patients completed the 3-month follow-up and were included in the final analysis.
[Table 2] describes the basic demographic details of the included patients and the distribution of pathologies among the included patients.
The mean preoperative and postoperative Nurick, mJOA, and ACMS were 3.3 and 3.0, 12.3 and 13.8, and 15.1 and 17.7, respectively [Table 3]. [Figure 1](a–c) shows the bar plots of the change in Nurick grade, mJOA, and ACMS following surgery. It was evident that the patients did not show as much improvement in Nurick grade as was evident in the mJOA and the ACMS scales. This was further supported by the scatterplot between the mJOA and the ACMS scales for all values (both preoperative and postoperative) [Figure 1]d. [Table 3] shows the change in Nurick, mJOA, and ACMS scores following surgery. The recovery rate for the different scales shows that Nurick grade was poor at detecting improvement [Table 3] and [Figure 2]a.
Table 3: Scores on various assessment scales and respective recovery rate (n=62 patients)
Figure 1: (a–c) Distribution between mJOA, ACMS, and Nurick score in the pre- and postoperative period. (d) Scatter plot with regression line between the ACMS and mJOA scores
Figure 2: (a) Boxplot comparing the recovery rate across the three scales. (b) Scatter plot comparing the perceived subjective improvement following the change in the ACMS and mJOA scores. (c) Scatter plot comparing the change in occupational ability with the change in other domains of functional ability. (d) Comparison of change in normalized mJOA and ACMS scores within the grades of subjective improvement following surgeryComponent-wise correlation between mJOA and ACMS scores
High correlation (Pearson's r > 0.8, P < 0.005) was observed between ACMS and mJOA scores for all individual components of both scales, both in pre- and postoperative assessments [Table 4]. Expectedly, a negative correlation was observed between the aggregate ACMS scores and the Nurick grade (r = −0.76)
Table 4: Correlation between individual components of the various scoresCorrelation of scores with MRI findings
No correlation was found with cord/canal ratio in MRI with any of the three scoring systems (preoperative, postoperative, or recovery rates). Only preoperative MRIs were available. No consistent relationship was found between the presence of T1 hypointensity and the myelopathy scores.
Presence of T2 hyperintensity in preoperative MRI was found to affect preoperative Nurick (P = 0.02), mJOA (P = 0.01), and ACMS scores (P = 0.01) significantly.
Similarly, the extent of T2 hyperintensity had a positive correlation with preoperative Nurick score (ρ = 0.24) but not significant (P = 0.06), whereas it had a statistically significant negative correlation with mJOA (Spearman correlation coefficient, ρ = −0.34, P = 0.01) and ACMS (Spearman's correlation coefficient, ρ = −0.3, P = 0.01) scores [Figure 3]a.
Figure 3: (a) Normalized mJOA and ACMS scores before and after surgery compared across the extent of T2 signal changes seen on preoperative cervical spine MRI. (b) Change in normalized mJOA and ACMS scores in relation to the preoperative extent of signal changes. (c) Overall clinical improvement or worsening among the study subjectsFurthermore, patients with T2 hyperintensity confined to a single level on preoperative MRIs showed statistically significant improvement in their postoperative scores on both ACMS and mJOA scales. [Table 5], [Figure 3]b
Table 5: Change in normalized ACMS and mJOA scores following surgery in relation to the extent of T2 hyperintensityCorrelation of subjective improvement with changes in normalized scores
Compared to the mJOA scores, the ACMS scores show a higher agreement with perceived improvement by the patients, though the difference between the two scores is not statistically significant [Figure 3]c.
Improvement in the motor and sensory scores for upper extremity and the motor scores for lower extremity were found to correlate positively with improvement in the occupational/self-care scores following surgery [Figure 2]c.
We also compared the normalized ACMS and mJOA scores with the patient's subjective perception of improvement or worsening following surgery. The change in mJOA and ACMS scores following surgery were highly correlated and consistent [Figure 2]b. The normalized ACMS scores showed a higher sensitivity to worsening or dramatic improvement following surgery [Figure 2]d, which was statistically significant (P < 0.05)
Calculation of predicted total ACMS score from raw mJOA score and internal consistency of ACMS scale
We obtained the following formula for estimation of the ACMS score from mJOA using the linear regression model: ACMS score = −2.23 + 1.4 × mJOA score (Adjusted R2 = 0.9188)
Cronbach's alpha was calculated as 0.82 (CI: 0.75 − 0.86), indicating that there was a satisfactory level of agreement between the answers provided by the patients with respect to the different functional domains tested.
DiscussionCSM is a progressive degeneration of the spine leading to its functional impairment. Successful management of myelopathy depends upon timely detection of its signs by the treating clinician and identification of its typical symptoms, which are reported by the patient.
Why do we need assessment scales?
A simple and universal applicable tool is essential for the objective assessment of severity of disease, quantification of disease progression, and identification of the need for surgical intervention. Moreover, uniformity in reporting is essential for unambiguous communication among the treating team, where all members may not have the opportunity to interview and physically examine the patient individually.
The diagnosis of CSM may be delayed by as much as 6 years on an average from presentation due to the inconspicuous nature of the initial symptoms.[6],[7],[23],[24] Furthermore, comorbidities associated with old age like diabetes (being associated with peripheral neuropathy) or vitamin B12 deficiency can cause peripheral neuropathy, which imitates the sensory findings of CSM like proprioceptive loss and sensory impairment in the glove and stocking distribution. Knowing this is important because early diagnosis and early management have good outcomes in patients with CSM.[8],[9],[10],[11],[25]
Nurick and mJOA scales
In order to assess and quantify this functional disability in patients with CSM, it was Nurick who initially developed a 6-grade ordinal scale.[12] The Nurick grade includes only ambulatory status as a marker of employment. However, if we consider an example of a shopkeeper who has a desk job, his employment could be more dependent on manual dexterity rather than ambulatory ability. So, in this case, a decline in lower extremity motor function would not result in a change in employment and upper extremity disability would be more predictive of occupational impairment. Additionally, his occupational impairment would be mitigated even with improvement confined to the upper limb. This could explain the discordance of results between ACMS and Nurick grade in our study. The observable improvement in ACMS score is more conspicuous as compared to Nurick grade.
Besides this discrepancy, other aspects of myelopathic involvement like bladder problems or upper limb deficits were not assessed using Nurick grade.
These shortcomings of Nurick grade lead to development of a more comprehensive JOA score which was later modified by Benzel et al.[3] in order to make it more compatible with western cultural milieu (populations in western countries uses cutlery instead of chopsticks). It is a clinician-based measure and includes upper and lower extremity motor function, hand sensation, and micturition.
Available literature on Nurick and mJOA
Although Nurick and mJOA scores have different structure and method of evaluation, a good correlation between both mJOA score and Nurick grade has already been established in a study by Vitzthum et al.[13] It was a retrospective study which included 43 patients of CSM who underwent anterior decompression. Both preoperative and postoperative scores of Nurick grade and mJOA showed a good correlation. However, the outcome following surgery showed a wide variation in terms of both the scoring systems. While 81% of patients showed improvement in their JOA score, only 33% of patients showed improvement in their Nurick grade. The authors reasoned that the Nurick grade did not assess the upper limb motor and sensory functions which strongly affected the employment status and quality of life of patients, and which could explain the discordance between the two scores. This could have also resulted in the statistically significant differences in the mean postoperative recovery rates observed through the two scales (23% as per Nurick grade and 37% with JOA scoring system). Thus, use of multiple scoring systems was advised by the authors for assessing outcomes in postoperative patients.
Later, Kato et al.[14] compared mJOA score with JOA score in his study of a total of 92 CSM patients. He showed a strong correlation between the two scores with a correlation coefficient (Spearman's rho) of 0.87 and of 0.75 between their recovery rates. Based on the results, mJOA was advocated as the standard scale for CSM grading in the Western population.[15]
Advantages of AIIMS cervical myelopathy scale
One of the major discrepancies between the Nurick and the mJOA scales at follow-up evaluation was that despite any improvement in the mJOA scores, people were regaining employment with a change in occupation. Thus, a score is still required which would include employment in addition to mJOA score and would be a better predictor of outcome from the patient's perspective. Thus, we came up with a comprehensive scoring system that reflected major domains of functional outcome in a patient.
Despite their excellent applicability, users realized few shortcomings of these traditional assessment scales. The assessment through mJOA requires the presence of a healthcare worker well versed with the symptoms of CSM. Some subcategories of functional assessment evaluated by mJOA contain ambiguous terms like”mild or severe” which are open to varied interpretation even among experienced clinicians. Thus, a need was felt for a new score with very objective and exacting criteria which could be interpretable by patients while eliminating the ambiguity in quantification of the various domains of functional assessment.
Since the questionnaire is meant for patients, vernacular translation was also required. The educational background of our subjects was varied and did not affect the answers to the questionnaire, although a small number of patients in the various subcategories of highest educational qualification did not permit a rigorous statistical analysis of this relationship.
This questionnaire is suitable for distribution over online platforms and through telemedicine as it can be filled up and returned by patients without substantial assistance from healthcare personnel. This may prove to be especially useful during the extraordinary circumstances imposed by the COVID-19 pandemic when we were able to evaluate some patients, who were not a part of this study, distantly without the need for physical interview.
The simplistic assumption of the Nurick grade connecting functional improvement with regaining employment does not take into consideration a wide variety of scenarios where such correlations may not hold.
Certain culture-specific daily practices like squatting and sitting cross-legged affect the outcome of intervention from the patient's perspective and cannot be ignored. However, incorporation of culture specific elements of physical function within the scoring systems renders them unsuitable for comparison with global studies.[16] The mJOA system is a very popular and widely implemented method of functional assessment, and drastic departures from its structure would make any new scale difficult to adapt and popularize. Moreover, comparison of scores with global experience may become poorly interpretable. Thus, the new score has been devised with the considerations that the criteria for evaluation are globally generalizable and unaffected by the cultural background of the patient.
In our study, we established a good correlation between ACMS and mJOA (Pearson r > 0.85 in all the shared categories of evaluation). We also devised a formula for conversion between mJOA and ACMS scales. Thus, past studies done using mJOA can be retrospectively compared and interpreted with this scale without difficulty.
Despite this correlation, a wide variation was noted in the outcomes following surgery. While per Nurick grade only 26% of patients improved, with mJOA and ACMS 50% and 58% of patients, respectively, showed improvement. A most likely explanation for this discrepancy could be because of certain different functional domains that are assessed by Nurick grade and mJOA/ACMS scores. Any improvement in either upper limb sensation or its motor component of the patient was not reflected in Nurick score which deals only with ambulation/lower limb status. While on the other hand these changes were apparent in both mJOA score and ACMS and were also found to affect employment.
Correlation of post-surgical clinical outcome with preoperative MRI signal changes (T2 hyperintensity and T1 hypointensity was established by Suri et al.[17] in his prospective study of 146 patients. We also explored the relationship between presence and extent of T1/T2-weighted signal changes on the MRI and the degree of improvement following intervention. Interestingly, it was observed that the absence of T2 hyperintensity or extensive signal changes involving multiple spinal cord segments did not result in statistically significant improvement following surgery. The absence of T2 hyperintensity was associated with high functional score even preoperatively. Thus, the scope for further improvement was diminished. On the contrary, extensive signal changes were related to advanced disease and irreversible injury to the spinal cord. Thus, though these patients benefited from surgery, the improvement was not as conspicuous as noted in patients with T2 hyperintensity limited to a single level.
ACMS has corroborated most closely with subjective improvement (65%) and was found to be most representative among the three scores in assessing outcome following surgery in patients with CSM.
With this newly described ACMS score, we tried to address the deficiencies of the previous scores as mentioned above and made the new score simple, patient reportable, and convertible with respect to the well-established mJOA score. Thus, ACMS was found to be clinically and radiologically appropriate for assessment and follow-up of patients with cervical myelopathy.
Limitations
Our study had several limitations. Our article described a new scale, and it was tested at our center. It needs to be externally validated before it can be recommended for wider use. Correlation with mJOA and Nurick can be better established once it is studied in a larger cohort, as our study had a relatively smaller patient population.
ConclusionsThe ACMS score showed a good correlation with the mJOA score for evaluation of functional disability in context of cervical myelopathy and patients could themselves judge the scoring using the ACMS scoring chart without the need of an experienced interpreter. The occupational/self-care component of the ACMS also correlated well with Nurick score. The ACMS score may offer a composite alternative to both mJOA and Nurick scores in the evaluation of a patient's preoperative and postoperative clinical status. Further studies are needed to confirm the findings of this study.
Ethics approval
Taken from the institute review board.
Consent to participate
Patient consent was taken.
Consent for publication
Patient consent for publication was taken.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References
留言 (0)