**The Impact of Older Age on Functional Recovery and Quality of Life Outcomes after Surgical Decompression for Degenerative Cervical Myelopathy: Results from an Ambispective, Propensity-Matched Analysis from the CSM-NA and CSM-I International, Multi-Center Studies**

### **Jamie R. F. Wilson 1,2, Jetan H. Badhiwala 1,2, Fan Jiang 1,2, Je**ff**erson R. Wilson 1,3, Branko Kopjar 4, Alexander R. Vaccaro <sup>5</sup> and Michael G. Fehlings 1,2,\* on behalf of Investigators from the AOSpine North America and CSM-International Studies**


Received: 20 August 2019; Accepted: 13 October 2019; Published: 17 October 2019

**Abstract:** Background: The effect on functional and quality of life (QOL) outcomes of surgery in elderly degenerative cervical myelopathy (DCM) patients has not been definitively established. Objective: To evaluate the effect of older age on the functional and QOL outcomes after surgery in an international, multi-center cohort of patients with DCM. Methods: 107 patients aged over 70 years old (mean 75.6 ± 4.4 years) were enrolled in the AOSpine CSM-North America and International studies. A propensity-matched cohort of 107 patients was generated from the remaining 650 adults aged <70 years old (mean 56.3 ± 9.6 years), matched to gender, complexity of surgery, co-morbidities, and baseline functional impairment (modified Japanese Orthopedic Association scale (mJOA). Functional, disability, and QOL outcomes were compared at baseline and at two years post-operatively, along with peri-operative adverse events. Results: Both cohorts were equivalently matched. At two years, both cohorts showed significant functional improvement from the baseline but the magnitude was greater in the younger cohort (mJOA 3.8 (3.2–4.4) vs. 2.6 (2.0–3.3); *p* = 0.007). This difference between groups was also observed in the SF-36 physical component summary (PCS) and mental component summary (MCS) outcomes (*p* = <0.001, *p* = 0.007), but not present in the neck disability index (NDI) scores (*p* = 0.094). Adverse events were non-significantly higher in the elderly cohort (22.4% vs. 15%; *p* = 0.161). Conclusions: Elderly patients showed an improvement in functional and QOL outcomes after surgery for DCM, but the magnitude of improvement was less when compared to the matched younger adult cohort. An age over 70 was not associated with an increased risk of adverse events.

**Keywords:** degenerative cervical myelopathy; elderly; old age; outcomes; complications; mJOA; SF-36

### **1. Introduction**

Degenerative cervical myelopathy (DCM) is a family of non-traumatic spinal cord injuries that contribute to spinal cord compression and the progressive onset of neurological deficits [1,2]. DCM encompasses a number of related conditions such as cervical spondylotic myelopathy (CSM), ossification of the posterior longitudinal ligament (OPLL), ossification of the ligamentum flavum, degenerative disc disease, and various congenital malformations that cause stenosis or instability leading to eventual spinal cord dysfunction [2]. It is now the commonest form of spinal cord dysfunction in adults [3], and has become an increasingly important area of focus for spine surgeons and clinician scientists in recent years [4–9]. This renewed focus has primarily been driven by a greater understanding of the benefit of timely intervention [5,7,10,11], and international clinical practice guidelines for the management of DCM were published in 2017 [7].

The aging population, juxtaposed with developments in medical technologies, has led to the "epidemiological transition"; the shift away from traditional causes of disease and mortality (infectious, nutritional deficiency, parasitic) toward chronic and degenerative diseases [3]. The incidence of DCM increases with age [3]; cervical spondylosis affects half of individuals aged 50–59 years, and nearly all individuals over the age of 70 [12]. The natural history of DCM is one of progressive neurological deficit, with the potential irreversible loss of dexterity, quadriplegia, and sphincter dysfunction [3,12,13]. DCM is set to become a major cause of chronic disability and global disease burden, if the current demographic expansion continues [14].

Decompressive surgery has been shown to not only halt neurological deterioration in DCM, but can produce a significant recovery in the neurological impairment that individuals experience [7,15]. The extent of functional and quality of life (QOL) improvement after surgery is influenced by the duration of symptoms, severity of functional impairment at presentation, the presence of co-morbidities, tobacco smoking status as well as age at presentation [8,11,16–20]. Elderly patients often present with a longer duration of symptoms, and more severe functional deficit compared to younger adults with DCM [11,12,18,19]. In addition, elderly patients have a higher prevalence of degenerative conditions, co-morbidities, an increased risk of osteoporosis, and higher peri-operative morbidity and mortality when compared to younger patients [3,12,21].

Age is a common surrogate for frailty or physiological reserve. The effect that frailty alone has on the effect of spine surgery has been studied through various means, and seems most relevant in the practice of spinal deformity where the procedures are often invasive and of long duration [22–28]. However, the significance of age on the effect of surgical outcomes for DCM is unclear. Available evidence would suggest that increasing age is associated with a worse functional outcome after surgical decompression [29–32]. However, many articles based on retrospective case series have also presented evidence demonstrating no significant differences in outcomes in terms of the modified Japanese Orthopedic Association score (mJOA), Nurick, and SF-36 scores when directly compared to standardized, younger patient cohorts [33–35]. Older patients are also more likely to undergo posterior surgery with a higher number of operated cervical levels when compared to younger adults [15,16,19].

The objectives of the current study were to (1) explore the functional and QOL impairment of adults over 70 years old with DCM compared to younger adults; (2) define how the variable of age over 70 alone can contribute to functional outcomes after DCM surgery, when accounting for other common age-related variables; and (3) define how elderly age can affect the QOL outcomes when compared to an equivalent adult cohort. We hypothesized that (1) adults over 70 (when compared to a matched cohort of younger adults with the same functional impairment) would exhibit less quality of life improvement; (2) when the effects of age-related co-morbidities and surgical factors were adjusted for, age over 70 would remain a significant risk factor for worse functional improvement after surgery; and (3) younger adults would have a more sustained improvement in QOL outcomes when compared to the elderly cohort.

### **2. Methods**

### *2.1. Subjects*

The AOSpine CSM-NA study recruited 278 patients with symptomatic DCM and correlating magnetic resonance imaging findings over a duration of two years from 12 North American centers [15]. The AOSpine CSM-International study recruited 479 symptomatic DCM patients from six Asian, five European, three Latin American, and two North American sites over a duration of four years [17]. Both studies were prospective observational studies, and patients were eligible for enrolment if they met the following inclusion criteria: (1) symptomatic DCM with one or more clinical signs of myelopathy; (2) imaging evidence compatible with spinal cord compression; (3) aged 18 years or older; and (4) no previous history of cervical spine procedures. Exclusion criteria included patients with malignancy or metastatic disease, active systemic infection, trauma, rheumatoid arthritis (or other inflammatory disease such as ankylosing spondylitis), symptomatic tandem lumbar stenosis, or asymptomatic DCM patients. All centers obtained approval from their respective local ethical boards prior to commencement of the study.

### *2.2. Baseline Characteristics*

All patients enrolled in both studies had demographic data recorded prior to surgery (age, race, socioeconomic status, tobacco smoking status, body mass index (BMI), presence of co-morbidities) together with a focused myelopathy history including duration of symptoms, clinical signs, and etiology of DCM. All patients underwent surgical treatment, with the surgical approach and number of spinal levels operated left to the discretion of the treating surgeon. Peri-operative demographics were recorded including the surgical procedure of choice (anterior discectomy, corpectomy, laminoplasty, laminectomy with or without instrumented fusion). Post-operative complications up to 24 months after surgery were recorded.

### *2.3. Outcomes*

Functional and QOL assessments were performed prior to surgery at baseline, then six, 12, and 24 months after surgery. Functional status was assessed by the mJOA scale; a standardized assessment of neurological and functional impairment that was administered by investigators [4,5,36]. Disability and QOL assessments were self-reported outcome measures in the form of the Neck Disability Index (NDI—specific to cervical degenerative pathologies) and the Short Form-36 version 2 (SF-36—a generic health-related QOL measurement). The SF-36 was further separated into the SF-36 Physical Component Summary (PCS) and Mental Component Summary (MCS) in an effort to distinguish between the patient-reported perception of physical health compared to mental and emotional well-being.

### *2.4. Statistical Analysis*

All statistical analyses were performed using Stata 15 (Stata Corp, College Station, TX, USA) and R version 3.5.1 (R Foundation for Statistical Computing, Vienna, Austria) with an a priori specified significance level of *p* = 0.05 (two-tailed). Descriptive statistics are listed as mean and standard deviation (SD) for continuous variables and count and percentage for categorical variables.

### 2.4.1. Propensity Score Matching

The propensity score matching algorithm was developed and executed using the 'MatchIt' package for R statistical software by one of the coauthors (J.H.B.). Variables to be included as covariates in the generation of propensity scores were defined a priori by author consensus of clinical relevance.

Propensity scores were calculated as the probability of age ≥ 70 years versus age < 70 years using the logit method with the baseline mJOA score, duration of DCM symptoms, cardiac disease, diabetes, smoking (current), psychiatric disease, surgical approach, and number of operated levels as

covariates (independent variables). Propensity score matching was performed in a one-to-one ratio using the 'optimal matching' technique to minimize the average absolute distance across all matched pairs. This resulted in two study groups (age < 70 vs. age ≥ 70) adjusted for the baseline covariates specified above. The baseline characteristics were compared between the study groups by the t-test for continuous variables and chi-square test for categorical variables.

### 2.4.2. Analysis of Outcomes

Two year outcomes for functional status (mJOA score), disability (NDI), and health-related quality of life (HRQOL; SF-36 PCS, and MCS) were compared between the study groups using an analysis of covariance (ANCOVA) adjusting for baseline score. Effect sizes for each outcome measure were summarized by β coefficients (mean difference) and associated 95% confidence intervals (CIs).

### **3. Results**

Both young and older cohorts were matched sufficiently in terms of sex, duration of symptoms, smoking history, co-morbidities, number of levels, surgical approach, and baseline functional impairment (Table 1). The mean age of the younger cohort was 56.3 ± 9.6 when compared to 75.6 ± 4.4 in the older cohort (*p* < 0.001). The younger cohort demonstrated a significantly worse baseline SF-36 Physical Component Score (PCS, 30.7 ± 8.2 vs. 33.5 ± 8.8 (*p* = 0.019)) and Mental Component Score (MCS, 38.4 ± 12.8 vs. 43.0 ± 13.1, (*p* = 0.011)) when compared to the older cohort.


**Table 1.** Baseline demographics of the elderly cohort and younger cohort after propensity matching for gender, duration of symptoms, smoking status, co-morbidities, surgical factors, and baseline functional impairment.

mJOA = modified Japanese Orthopedic Association scale, NDI = Neck Disability Index, SF-36 PCS = SF-36 Physical Component Score, SF-36 MCS = SF-36 Mental Component Score. \* *p* value < 0.05.

Both cohorts demonstrated an improvement in functional impairment at two years, as defined by the mJOA scale (*p* = 0.001, see Table 2). Significant improvements were also seen in the NDI, PCS, and MCS scores, with the exception of the elderly MCS score (*p* = 0.077). The functional outcomes in the younger cohort were of a greater magnitude when compared to the older group (mean difference 3.8 ± 3.0 versus 2.6 ± 3.3; *p* = 0.007; see Table 3). This difference was also present in the QOL

measurements (PCS and MCS, *p* < 0.001, *p* = 0.007). The change in NDI scores between the groups at two years showed no significant difference (*p* = 0.094).


**Table 2.** Functional, disability, and quality of life assessment scores of both cohorts at the two-year interval.

mJOA = modified Japanese Orthopedic Association scale, NDI = Neck Disability Index, SF-36 PCS = SF-36 Physical Component Score, SF-36 MCS = SF-36 Mental Component Score, MD = Mean Difference.

**Table 3.** Delta values (change in the scores from baseline) at two years for younger and older cohorts for all outcomes.


mJOA = modified Japanese Orthopedic Association scale, NDI = Neck Disability Index, SF-36 PCS = SF-36 Physical Component Score, SF-36 MCS = SF-36 Mental Component Score, MD = Mean Difference.

The total number of all-cause adverse events over two years, including peri-operative complications and worsening functional impairment, was lower in the younger cohort (*n* = 16; 15%) when compared to the older cohort (*n* = 24; 22.4%; *p* = 0.161), but this was not statistically significant (see Table 4). Post-operative infection and hardware failure were equivalent between the younger and older cohorts (*p* = 0.701 and *p* = 0.313, respectively). The incidence of post-operative dysphagia was significantly higher in the older cohort when compared to the younger cohort (*n* = 6 compared to *n* = 0; *p* = 0.013). The incidence of the worsening of myelopathy symptoms was higher in the older cohort when compared to the younger cohort, which was just shy of statistical significance (*n* = 17 (15.9%) vs. *n* = 8 (7.5%); *p* = 0.055).

**Table 4.** List of complications of elderly and younger adult cohorts over the two-year follow up.


\* *p* value < 0.05.

### **4. Discussion**

Renewed focus on functional and quality of life outcomes after surgery for DCM in adults has come from rigorous scrutiny of the results of several large-scale studies in the last decade [2,4,5,7,10,17,37–42]. The traditional doctrine of the use of surgery to 'arrest clinical progression' has been replaced with a more modern approach with decompressive surgery used as a means to provide both functional and quality of life improvement [5,7]. The assessment of quality of life metrics has become particularly important in the examination of patients with mild functional impairment (or asymptomatic cord compression), and is the focus of a number of studies [5,7,43]. Careful evaluation has discovered a number of predictors of clinical outcomes from surgery for DCM, including baseline severity of impairment, duration of symptoms, presence of systemic or psychological co-morbidities, obesity, and of course age [11,18,21]. With simple univariate analysis of outcome measures, many studies have reported worse functional and QOL outcomes with increasing age after surgery for DCM [29–32]. However, elderly patients often present with increased duration and worse severity of symptoms, with an increased incidence of co-morbidities [11,18,29]. Many elderly patients also present a high, sometimes unacceptable, surgical risk profile that makes decisions regarding surgical management less appropriate. DCM is a disease that can cause a significant amount of potentially reversible or preventable neurological disability in the elderly population, and therefore concerted efforts should be made to improve decision-making strategies for this patient group.

The results of the current study demonstrate a number of important conclusions, some of which have not been described in previous literature. Firstly, even when matched for co-morbidities, duration of symptoms, sex, smoking status and severity of functional impairment, younger adults report significantly worse effects on their baseline physical and mental quality of life scores compared to the older aged cohort. This could be explained through 2 potential mechanisms. Firstly, as the younger cohort were of working age, the psychological impact of DCM may be greater if employment security is perceived to be threatened. Secondly, the elderly patients demonstrate psychological resilience or adaptations/support that mean the symptoms of DCM have a less pronounced effect on their quality of life. This is an interesting concept, and is an important consideration for clinicians managing elderly DCM patients.

All patients showed improved functional impairment scores after 2 years (see Table 2). Taking the older cohort in isolation, this is good evidence that surgery is an effective modality to produce functional improvements in elderly DCM patients. This is an important stand-alone conclusion from this study. When the change from baseline scores at 2 years is calculated, it becomes evident that the functional improvement seen after surgery is of a greater magnitude in younger patients compared to the older cohort. This conclusion is similar to those described from previous univariate analyses in prior studies [11,18,29,32,44]. Suggested mechanisms for this discrepancy include the fact that elderly patients often present with increased duration of symptoms at diagnosis, often have difficulty accessing specialist assessment and imaging, and also have less neurological plasticity or reserve compared to younger adults [11,12,18,45,46]. The level of disability as measured by the NDI was improved significantly by surgery for all patients, but the order of magnitude was similar for both age groups (*p* = 0.094).

All patients demonstrated improved SF-36 PCS and MCS from baseline at the 2-year interval, with the exception of the MCS scores in the elderly group. Both PCS and MCS metrics echoed the functional improvement in that the younger cohort had a significantly greater degree of improvement from baseline compared to the older cohort (*p* < 0.001, *p* = 0.007). This provides good evidence that surgery leads to increased physical and mental perception of quality of life outcomes in all patients undergoing surgery for DCM. It appears that older patients' mental perception of quality of life remains stable throughout the treatment period, despite having a worse functional improvement, and less magnitude of quality of life improvement overall. This study is the first to report this difference in quality of life measures between younger and older aged cohorts and has potentially large implications on pre-surgical assessment and counselling in older patients with DCM. However, age-related effects on health perception have been found to influence the SF-36 score, and the contribution from the

perception of physical effects often misses a number of important determinants of overall quality of life in older persons [47,48]. For these reasons, it has been suggested that the focus of quality of life determinants in the elderly should be weighted toward mental components rather than the physical [48]. This also raises certain questions about the efficacy of functional assessment measures for DCM in the elderly, but this was beyond the scope of the current study.

All forms of adverse events over two years were higher in the elderly cohort when compared to the younger adults, but not significantly so (*p* = 0.161, see Table 4). The risk of infection, hardware failure, CSF leak, and need for revision surgery were all equivalent between groups. The incidence of dysphagia after the anterior approach was significantly greater in the elderly cohort, which has been well described [49]. The incidence of worsening functional impairment after surgery was non-significantly higher in the elderly cohort, which is also consistent with previously described studies and may be a reflection of the disease process in DCM, rather than directly related to complications from surgery [11,40,50]. These findings suggest that, contrary to popular opinion, surgery for DCM in the elderly does not carry a significantly higher risk of adverse events when compared to younger adults that are matched for co-morbidities and complexity of surgery. This is good evidence to demonstrate that age alone does not necessarily confer an increased risk of adverse events in DCM surgery, but that the other risk factors associated with older age (increased number of levels of pathology, posterior surgery, increased co-morbidities, etc.) ostensibly play more of a role in the determinant of peri-operative risk. This concept of frailty, and its association with the assessment of surgical risk, is emerging as an important tool for surgical decision making and has been a recent focus of interest in pathologies such as spinal cord injury and adult deformity surgery [51–53]. However, the impact of frailty on the outcomes from DCM appears to be less well defined.

There are limitations of the current study. The results were from the pooled analysis of two harmonized datasets, and although the data came from prospective, multi-center sources, the original studies were not designed or powered to measure the effect of old age on the outcomes after surgery for DCM. Although both cohorts were well matched (Table 1), there could exist significant heterogeneity between the groups or hidden confounders such as drug histories that may have affected the results. The use of propensity-matching methodology helps to reduce the over-fitting seen in mixed effects or regression models, but does have obvious effects in the sample size. Therefore, some aspects of the results that show trends and not significant differences may indicate that the results are underpowered in some areas.

### **5. Conclusions**

To consider our previous hypotheses, elderly patients, when compared to younger adults (matched to functional impairment and age-related risk factors), exhibited better SF-36 PCS and MCS prior to surgery for DCM. All patients in this study demonstrated improved functional impairment two years after surgery, but the magnitude of improvement seen was greater in the younger cohort, even when baseline functional impairment and age-related risk factors were adjusted for. Elderly patients also showed improved QOL physical and mental component scores after surgery, but the extent of the increase in the physical component was reduced when compared to younger adults. Aside from the incidence of post-operative dysphagia, older age alone was not associated with a higher incidence of adverse events.

The authors believe that the results of this study provide good evidence that surgery for DCM in the elderly is effective in terms of both functional and QOL outcomes. Perhaps most importantly, these results demonstrate that the elderly DCM age group should have different expectations with regard to the extent of functional and QOL outcomes after surgery. Patients over the age of 70 with a diagnosis of DCM are likely to require specialist considerations and should be counseled appropriately with adjusted expectations. The exact degree to which each modifiable risk factor contributes to perioperative risk, and the components that affect functional and QOL outcomes, remain to be determined and should be an important focus of further research into the effects of aging on

surgery for DCM. Developing a prediction model using age (or measures of frailty) and related covariate adjustment would significantly improve the calculation of the risk profile in DCM patients undergoing assessment for decompressive surgery.

**Author Contributions:** Conceptualization, J.R.F.W., J.H.B., J.R.W., B.K., and M.G.F.; Data curation, J.R.F.W. and B.K.; Formal analysis, J.R.F.W., J.H.B., J.R.W., and B.K.; Investigation, J.H.B., F.J., J.R.W., A.R.V., and M.G.F.; Methodology, J.R.F.W. and J.H.B.; Resources, M.G.F.; Supervision, M.G.F.; Writing—original draft, J.R.F.W., J.H.B., and F.J.; Writing—review & editing, J.R.F.W., J.H.B., F.J., J.R.W., B.K., A.R.V., and M.G.F.

**Funding:** The original collection of data in the CSM-NA and CSM-I studies was supported by AOSpine. J.H.B. is supported by a Fellowship Award from the Canadian Institutes of Health Research (CIHR).

**Acknowledgments:** M.G.F. would like to acknowledge the support from the Gerry and Tootsie Halbert Chair in Neural Repair and Regeneration and the DeZwirek Family Foundation.

**Conflicts of Interest:** The authors declare no relevant conflicts of interest.

### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Degenerative Cervical Myelopathy in Higher-Aged Patients: How Do They Benefit from Surgery?**

**Oliver Gembruch 1,\*, Ramazan Jabbarli 1, Ali Rashidi 1, Mehdi Chihi 1, Nicolai El Hindy 1,2, Axel Wetter 3, Bernd-Otto Hütter 1, Ulrich Sure 1, Philipp Dammann <sup>1</sup> and Neriman Özkan <sup>1</sup>**


Received: 20 October 2019; Accepted: 24 December 2019; Published: 26 December 2019

**Abstract:** Background: Degenerative cervical myelopathy (DCM) is the most common reason for spinal cord disease in elderly patients. This study analyzes the preoperative status and postoperative outcome of higher-aged patients in comparison to young and elderly patients in order to determine the benefit to those patients from DCM surgery. Methods: A retrospective analysis of the clinical data, radiological findings, and operative reports of 411 patients treated surgically between 2007 and 2016 suffering from DCM was performed. The preoperative and postoperative neurological functions were evaluated using the modified Japanese Orthopedic Association Score (mJOA Score), the postoperative mJOA Score improvement, the neurological recovery rate (NRR) of the mJOA Score, and the minimum clinically important difference (MCID). The Charlson Comorbidity Index (CCI) was used to evaluate the impact of comorbidities on the preoperative and postoperative mJOA Score. The comparisons were performed between the following age groups: G1: ≤50 years, G2: 51–70 years, and G3: >70 years. Results: The preoperative and postoperative mJOA Score was significantly lower in G3 than in G2 and G1 (*p* < 0.0001). However, the mean mJOA Score's improvement did not differ significantly (*p* = 0.81) between those groups six months after surgery (G1: 1.99 ± 1.04, G2: 2.01 ± 1.04, G: 2.00 ± 0.91). Furthermore, the MCID showed a significant improvement in every age-group. The CCI was evaluated for each age-group, showing a statistically significant group effect (*p* < 0.0001). Analysis of variance revealed a significant group effect on the delay (weeks) between symptom onset and surgery (*p* = 0.003). The duration of the stay at the hospital did differ significantly between the age groups (*p* < 0.0001). Conclusion: Preoperative and postoperative mJOA Scores, but not the extent of postoperative improvement, are affected by the patients' age. Therefore, patients should be considered for DCM surgery regardless of their age.

**Keywords:** degenerative cervical myelopathy; cervical canal stenosis; cervical spine surgery; higher-aged patients; neurological outcome; mJOA Score; MCID

### **1. Introduction**

Degenerative cervical myelopathy (DCM) is a slowly ongoing degenerative disease of the cervical spine caused by progressive narrowing of the cervical canal and compression of the spinal cord. The DCM is age-dependent and is the most common degenerative disease of the cervical spine in

elderly patients with a progressive degeneration of the intervertebral discs, joints, and ligaments. In the 4th decade of life, 30% of the population shows cervical spine degeneration and in the 6th decade about 90% suffer from cervical spine degeneration [1]. The United Nations analyzed demographic changes in the global population and showed that in 2015, 12.6% of the population worldwide was aged ≥60 years. By 2030, those people will represent 16.5% of the worldwide population and in 2050, 21.5% will be aged ≥60 years [2]. The German Census Bureau data indicates that persons above the age of 60 years are the fastest growing of all of the age groups. By 2030, 34.6% of the German population will be more than 60 years old and 37.6% of the German population will be more than 60 years old by 2050 [3]. Therefore, the frequency of patients presenting with DCM will increase. The question of surgical treatment of higher-aged patients suffering from DCM is becoming more and more important.

Previous systematic reviews have shown that non-surgical treatment in moderate to severe DCM is not recommended because of the inferior outcomes compared to surgery [4–6]. Fehlings et al. presented clinical practice guidelines for the management of DCM [6,7]. The cost-effectiveness of surgical treatment has also been proven in industrialized countries [8].

Surgery remains the gold standard for the treatment of DCM, but there is still no consensus among experts about therapy for higher-aged patients with DCM. There is a reluctance from surgeons to perform surgery in old patients because age is an independent factor increasing morbidity and is associated with additional comorbid medical conditions [9].

However, surgical outcomes in elderly patients with DCM are controversial. A lower surgical outcome (based on the Japanese Orthopedic Association Score (JOA Score), the neurological recovery rate (NRR), and the JOA Score improvement) in old patients has been reported in several studies [10–12], while other authors described no significant differences in neurological improvement between old and young patients [13–15].

Nevertheless, the analysis of the surgical benefit of higher-aged patients with DCM remains underrepresented in literature with respect to demographic changes.

The aim of this study was to analyze the preoperative function and the postoperative functional outcome in patients undergoing DCM surgery using the modified JOA Score (mJOA Score), the mean mJOA Score improvement, NRR, and the minimum clinically important difference (MCID) based on different age groups, with a particular focus on elderly patients.

### **2. Materials and Methods**

### *2.1. Study Population*

A retrospective analysis of the clinical and radiological data and operative reports of patients suffering from DCM was performed.

Data from 968 patients suffering from cervical degenerative disorders who were treated surgically in our hospital between 2007 and 2016 were analyzed applying the following exclusion criteria: 1. cervical degenerative disorders others than DCM; 2. congenital abnormalities of the cervical spine; 3. metastatic or rheumatoid diseases; 4. fractures of unknown age; 5. instability of the cervical spine; or 6. traumatic spinal cord injury.

Posterior fixation was only used in four DCM patients. Those patients were excluded from the study cohort due to the limited amount of cases. Subgroup analysis was not possible, and exclusion of those patients did not influence statistical analysis.

Therefore, 411 patients (263 males, 148 females; mean age: 62.6 ± 12.1 years; range: 31–96 years) with a spinal stenosis (249 patients, 60.6%) or a herniated disk (162 patients, 39.4%) were studied.

Patients were operated on with a ventral or posterior approach using anterior cervical discectomy and fusion (ACDF), laminoplasty, or decompressive laminectomy. Posterior fusion was not performed in DCM surgery.

The patients were divided into three groups depending on their age: G1: ≤50 years of age, 74 patients; G2: 51–70 years of age, 204 patients; and G3: >70 years of age, 133 patients (Figure 1). The following data were collected for all patients: age, sex, comorbidities (using the Charlson Comorbidity Index (CCI) [16], symptom presentation until surgery, preoperative status (mJOA Score) (Table 1), and the neurological outcome (mJOA Score, mean mJOA Score improvement, NRR and the MCID) (Table 2).

**Figure 1.** Cervical Myelopathy in a patient of G1 (**A**), G2 (**B**) and G3 (**C**).



**Table 2.** Modified Japanese Orthopedic Association Score (mJOA Score), neurological recovery rate (NRR) and the mean mJOA Score improvement according to the age groups (G1–G3) postoperative, three and six months after surgery.


### *2.2. Assessment of Clinical Outcome*

The severity of the DCM was evaluated before and after surgery according to mJOA Score proposed by the Japanese Orthopedic Association for cervical myelopathy [17]. The postoperative mJOA Score was assessed during the stay at the hospital, as well as three and six months after surgery. The NRR was calculated using the formula suggested by Hirabayashi and Satomi [NRR (%) = (postoperative mJOA Score−preoperative mJOA Score)/(full score (18)−preoperative mJOA Score) × 100] [18].

The mJOA Score improvement was also evaluated (postoperative mJOA Score–preoperative mJOA Score) to analyze the postoperative outcome. Comorbidities were analyzed using the CCI. A total of 27 patients (6.6%) did not attend the three-month follow-up examination and 95 patients (23.1%) were lost to follow-up six months after surgery.

Additionally, the MCID was evaluated after the operation. It is defined as the minimum change in a measurement that a patient would identify as being beneficial [19]. For DCM, the MCID is defined as follows: 1 point for patients with mild DCM (mJOA Score ≥ 15), 2 points for patients with moderate DCM (mJOA Score of 12–14), and 3 points for patients with severe DCM (mJOA Score < 12). A "poor" outcome was therefore defined as a postoperative change of mJOA Score less than the MICD.

### *2.3. Statistical Analysis*

Data were analyzed using SPSS 23.0 (Statistical Package for the Social Sciences, SPSS Inc., Chicago, IL, USA). Metric data were described by mean and standard deviation and nominal data by frequency and valid percent. Data were checked for possible deviations from the assumption of normal distribution using the Shapiro-Wilk test. The mJOA Scores were assessed preoperatively, postoperatively, three months after surgery, and six months after surgery and were compared by Friedman-Tests for non-normally distributed data. The Man-Whitney-U-Test and the Kruskal-Wallis-Test were used to evaluate significant differences between the groups. The Wilcoxon signed-rank test was used to test between pairs of repeated measurement. Analysis of variance (ANOVA) was also used to detect statistical differences between the age groups. Pearson Chi2 statistics were applied to compute two-sided asymptotic statistical significance. In addition, a contingency coefficient served as a measure for the symmetry of the association. McNemar-Test was used within the age groups to determine significant changes in MCID improvement within a period of six months after surgery.

### *2.4. Ethics*

The study has been carried out in accordance with The Code of Ethics of the World Medical Association (The Declaration of Helsinki) and was approved by the Institutional Review Board (Medical Faculty, University of Duisburg-Essen, Registration number: 16-6270-BO).

### **3. Results**

### *3.1. Symptom Presentation*

First symptoms in group G1 were cervicobrachial neuralgia (48.6%), followed by sensory deficits (24.3%). These results were similar to those of group G2 (43.1% and 18.6%, respectively). However, the number of patients with ataxia was higher with increasing age, particularly in G3. Here, only 30.1% of the patients suffered from cervicobrachial neuralgia as the first symptom, but 40.6% complained about ataxia (Table 3).

**Table 3.** Analyzing the first presenting symptom in relation to the age groups.


### *3.2. Surgical Treatment*

Surgical treatment included ACDF (243 patients, 59.22%; G1: 64; G2: 130; G3: 49), laminoplasty (117 patients, 28.40%: G1: 9; G2: 61; G3: 47), and decompressive laminectomy without posterior fusion (51 patients, 17.92%; G1: 1; G2: 13; G3: 37) (Table 4).

**Table 4.** Surgical treatment and complications according to the age groups; ACDF: Anterior cervical discectomy and fusion.


ACDF was chosen for patients with a ventral one or two-level narrowness caused by a spinal canal stenosis or a herniated disk. In patients with multilevel spinal canal stenosis, posterior decompression was favored, while laminoplasty was performed in patients with a predominantly dorsal multilevel narrowness.

However, according to the surgical approach, the number of operated levels and the surgical treatment showed no significant MCIDs six months after the operation (Table 5).

**Table 5.** The minimum clinically important difference (MCID) six months after surgery according to the surgical approach, the number of operated levels, and the surgical treatment; anterior cervical discectomy and fusion: ACDF.


### *3.3. Preoperative and Postoperative mJOA Score*

The mean preoperative mJOA Score in G1 was 14.99 ± 2.17, the mean postoperative mJOA Score was 15.78 ± 2.22, while the mean mJOA Score three months after surgery was 16.58 ± 1.90 and the score six months after surgery was 17.25 ± 1.36.

The mean preoperative mJOA score in G2 was 14.57 ± 2.27, the mean postoperative mJOA score was 15.32 ± 2.47, while the mean mJOA score three months after surgery was 16.30 ± 1.92 and the score six months after surgery was 16.99 ± 1.39.

The mean preoperative mJOA Score in G3 was 13.57 ± 2.51, the mean postoperative mJOA Score was 14.23 ± 2.56, while mean the mJOA Score three months after surgery was 15.45 ± 2.02 and the mJOA Score six months after surgery was 16.32 ± 1.70.

The postoperative mJOA Score improved significantly (*p* < 0.001, respectively) in every age group, but higher-aged patients showed a lower postoperative mJOA Score compared to younger and elderly patients (Table 2 and Figure 2).

**Figure 2.** Modified Japanese Orthopedic Association Score (mJOA Score) according to the age-dependent groups (G1–G3).

Additionally, postoperative mJOA Scores increased significantly independently of the surgical procedure (Table 6).

**Table 6.** Modified Japanese Orthopedic Association Score (mJOA Score), neurological recovery rate (NRR) and the mean mJOA Score improvement according to the surgical treatment; anterior cervical discectomy and fusion: ACDF.


### *3.4. Neurological Recovery Rate*

The median NRR in patients according to G1 was 37.6% postoperative, 65.2% three months after surgery, and 83.4% six months after surgery in patients G1.

The median NRR in patients according to G2 was 32.9% postoperative, 60.3% three months after surgery, and 75.8% six months after surgery in patients G2.

The median NRR in patients according to G3 was 19.5% postoperative, 46.2% three months after surgery, and 61.9% six months after surgery.

NRR improved significantly in every age group at the postoperative, three and six months post-surgery follow-up examinations (*p* < 0.001, respectively). Additionally, patients belonging to G1 or G2 presented a significantly better recovery rate than patients representing G3, due to the higher preoperative mJOA Scores (*p* < 0.001) (Table 2 and Figure 3).

**Figure 3.** The neurological recovery rate (NRR) based on the age-dependent groups.

Additionally, postoperative NRR increased significantly independently of the surgical procedure used (Table 6).

### *3.5. Mean mJOA Score Improvement*

The mean mJOA Score improvement in G1 was 0.76 ± 0.79 postoperative, 1.58 ± 0.90 three months after surgery, and 1.95 ± 1.04 six months after surgery.

The mean mJOA Score improvement in G2 was 0.74 ± 0.97 postoperative, 1.60 ± 0.89 three months after surgery, and 2.01 ± 1.04 six months after surgery.

The mean mJOA Score improvement in G3 was 0.66 ± 1.02, 1.58 ± 1.13 three months after surgery, and 2.00 ± 0.91 six months after surgery.

The mean mJOA Score improvement was significantly better in every age group at the postoperative, three and six months post-surgery follow-up examinations (*p* < 0.001, respectively). However, the mean mJOA Score improvement did not differ significantly between the age groups (*p* = 0.81) (Table 2 and Figure 4).

**Figure 4.** mJOA Score improvement based on the age-dependent groups.

Additionally, postoperative mJOA Score improvement increased significantly independently of the surgical procedure used (Table 6).

### *3.6. Minimum Clinically Important Di*ff*erence*

In our study, we were able to show favorable MCID immediately postoperatively in 52.7% of the patients of G1, in 49.0% of the patients of G2, and in 34.6% of the patients of G3. Furthermore, 98.4% (G1), 93.3% (G2), and 85.7% (G3) of the patients achieved favorable MCID six months after surgery. MCID was significantly within the different age groups. Unfortunately, the MCID achievements were significantly reduced in elderly patients compared to younger patients (Table 2).

Furthermore, the MCID six months postoperative revealed no significant differences regarding the different surgical approaches (ventral vs. dorsal), the number of operated levels (monosegmental, bisegemental, and multisegmental), and the different surgical treatment (ACDF, laminoplasty and laminectomy) (Table 5).

### *3.7. Comorbidities and the Charlson Comorbidity Index*

Comorbidities of the patients were collected and grouped as 1. cardiovascular (e.g., arterial hypertension, coronary heart disease myocardial infarction), 2. pulmonal (e.g., chronic obstructive lung disease, pneumonia, asthma bronchiale), 3. neurological (e.g., transitory ischemic attack, stroke, polyneuropathia), 4. oncological (e.g., lung cancer, breast cancer or prostate cancer), 5. endocrine disease (e.g., thyreoditis, diabetes mellitus), and 6. surgical (abdominal, cardial, pulmonal or orthopedic).

The CCI was evaluated for each age group, showing a statistically significant group effect (*p* < 0.001) according to ANOVA. Additionally, CCI showed a significant association with the preoperative and postoperative mJOA Score (*p* < 0.001) (Table 1).

### *3.8. Duration of Myelopathic Symptoms Prior to Surgery*

In our study, the duration lasting from the first symptom presentation until DCM surgery was much longer in G3 (45.9 ± 66.0 weeks; range: 1–350) than in G1 (22.2 ± 22.08 weeks; range: 1–122) and G2 (33.3 ± 43.6 weeks; range: 1–300). ANOVA revealed a significant group effect on the delay (weeks) between symptom onset and surgery (*p* = 0.003) (Table 1).

### *3.9. Duration of Hospitalization*

Stay at the hospital in G1 was around 8.6 ± 3.8 days, 9.6 ± 4.0 days in G2, and 10.5 ± 4.7 days in G3. The duration of the stay at the hospital did differ significantly (*p* < 0.001) between the age-groups according to the ANOVA results (Table 1).

### *3.10. Surgical and Non-Surgical Complications*

Complications were analyzed according to the age groups. Surgical and non-surgical complications were evaluated. Surgical complications were defined as 1. postoperative bleeding, 2. poor wound healing, 3. cerebrospinal fluid leakage, and 4. acute myelon compression. Non-surgical complications were defined as 1. pneumonia, 2. heart attack, and 3. stroke.

Non-surgical complications were highest in G3 with 6.0%, whereas those complications were similar within G1 and G2 (1.4% vs. 1.5%).

Surgical complications were reported to increase with age, affecting 1.4% (G1), 2.5% (G2), and 6.8% (G3) of the patients (Table 4).

### **4. Discussion**

The demographic changes of the population especially in western countries [3] have led to an increase of age-dependent diseases. Therefore, understanding the treatment of elderly patients has become more relevant. At present, there are no guidelines for management of degenerative spine diseases in elderly individuals. Moreover, only a few articles have addressed this topic so far, returning partially discrepant results in functional outcomes [10–15].

Tetreault et al. showed in their systematic review that patients with a more severe DCM expressed by a lower mJOA Score and patients with a longer duration of the symptoms are more likely to have a worse surgical result. They concluded that both severe and chronic compression of the spinal cord may lead to irreversible damage due to demyelination and necrosis of the grey matter. They were also able to show that age is a potential predictor when analyzing the postoperative outcome [20]. Holly et al. also analyzed age, the duration of symptoms, and preoperative neurological function as predictors of the neurological outcome and they could show similar results in their review [21]. In our study, the time between the first symptom presentation and DCM surgery was much longer in G3 than in G1 or G2. This effect might be caused by the fact that patients in G3 suffer from more comorbidities than younger patients and that an age-dependent decrease in the daily condition might be seen as normal or might not be recognized early in older patients and therefore, the time until diagnosis is prolonged. The reduced physical condition is also expressed by the prolonged hospitalization in G3 compared to G1 and G2 and the significantly lower CCI. The CCI also showed a strong correlation with the preoperative and postoperative mJOA Scores. Additionally, the first symptom of G1 patients was cervicobrachial neuralgia, whereas the first symptom of G3 patients was ataxia followed by cervicobrachial neuralgia. This could highlight the difficulties of distinguishing DCM from other age-related diseases.

The evaluation of the preoperative mJOA Score showed significantly lower scores in G3 compared to the preoperative mJOA scores of G1 and G2. The postoperative mJOA Score and the mJOA Score three and six months after surgery showed similar results.

In our study, the significantly lower preoperative and postoperative mJOA Scores of G3 patients as compared to G2 and G1 patients was strongly associated with the lower CCI of those patients. Mean preoperative and postoperative mJOA Scores may be lower in G2 and especially in G3 due to physical weakness caused by age and known comorbidities such as cerebral vascular disorders, hip and knee osteoarthritis, entrapment of peripheral neuropathy (carpal or cubital tunnel syndrome), diabetic neuropathy, benign prostatic hypertrophy, or urinary stress incontinence [22]. Machino et al. [23] also concluded that the preoperative JOA Score might be influenced by those comorbidities.

Nagashima et al. evaluated the neurological outcomes of 37 patients over 80 years of age and compared them with that of a younger population. The NRR was lower in the elderly population, but JOA Scoring improved in a way that life style was positively influenced [10].

In the present study, there was a significant improvement according to the NNR after surgery in all age groups. Interestingly, NRR was significantly lower in G3 compared to G1 and G2 despite similar mean mJOA Score improvements.

Nevertheless, there are limitations concerning the validity of the NRR despite its popularity. The results of the NRR are strongly influenced by the preoperative mJOA Scores. For example, if patients have a low preoperative mJOA Score, then NRR is lower than in patients with a higher preoperative mJOA Score even though the mean mJOA Score improvements are the same [23].

Due to the limitations of the NRR evaluating the neurological outcome in patients with a lower preoperative mJOA Score, the mean mJOA Score improvement or the MCID might be more valuable for comparing the neurological improvements between those patients.

In our study, the mean mJOA Score improvements were similar in every age-dependent group (Table 3), although the preoperative mJOA Score was significantly lower in G3 than in G1 or G2.

The results of the mean mJOA Score improvement are in line with the findings of Machino et al. [23]. They analyzed 520 patients with CSM treated by laminoplasty and divided their patients into nonelderly (<65 years), young-old (65–75 years) and old-old (>75 years). Elderly group patients showed significantly lower recovery rates of JOA Scores compared with the nonelderly group, but mean JOA Score improvements showed no difference among these groups. Preoperative JOA Scores were also significantly lowered similarly to our patients.

Madhavan et al. also performed a meta-analysis of old DCM patients evaluating the postoperative outcome and the operative risks. They found, like in our evaluation, a significantly lower preoperative JOA Score and a lower postoperative JOA Score associated with a lower NRR. But postoperative and long-term improvements in old patients have been remarkable in terms of improvements in mobility and independence, leading to reduced nursing care being required. The incidence of postoperative complications did not show a significant difference [24].

Additionally, MCID was favorable in the majority of elderly patients (85.7%) six months after surgery. This means that those patients showed acceptable clinical improvement after surgery despite their age.

In summary, mean mJOA Score improvements did not differ significantly among the age-dependent groups, but clinical improvement after surgery according to the mJOA Score was much better in old patients compared to younger patients. This improvement led to an improvement in mobility and independence, hence requiring reduced nursing care. This was shown by Yoshida et al., who were able to show in a study with 76 patients older than 75 years of age that the nursing care requirements based on JOA Score and functional independence measure scoring was reduced [11]. Furthermore, a different surgical approach, number of operated levels, and surgical treatment revealed no significant MCID differences six months post-surgery.

### *Limitations*

The present study has several limitations. First, this is a retrospective, non-randomized study with the associated inherent bias. The analyzed data were collected from documented electronic records, operative reports, radiological data, and reports of the patients. Secondly, the mJOA Scoring system might be influenced in the elderly group by several comorbidities such as hip and knee osteoarthritis, cerebrovascular diseases, diabetic neuropathy, or prostate hypertrophy. Additionally, the NRR system also has some limitations. Lower preoperative scores indicate lower NRR, although the mean mJOA Score improvement was the same. Furthermore, the follow-up period of six months is relatively short, caused by the retrospective nature of the study and the resulting losses in follow-up. The evaluation of the postoperative outcome might be too early directly after surgery, but we were still able to show an improvement of the neurological status over that short period.

However, future prospective studies with a longer follow-up are needed to evaluate the neurological long-term outcome in elderly patients. Nevertheless, we could show in a large population of elderly patients that surgery is still useful due to clinical improvements of the symptoms and the resulting lower need for daily care.

### **5. Conclusions**

The preoperative and postoperative mJOA Scores are significant lower in older patients compared to younger individuals, but the mean mJOA Score improvement is similar. The lower mJOA Score in those patients correlates with a lower CCI. With respect to the postoperative mJOA Score improvement and MCID, older patients still benefit from surgery. Therefore, surgical treatment of DCM is a valuable option for those patients.

**Author Contributions:** Conceptualization: O.G., A.R. and N.Ö.; Data curation: O.G. and A.R.; methodology: O.G., A.R. and N.Ö.; Formal analysis: O.G., M.C. and B.-O.H. Visualization: O.G., R.J., N. EH., A.W., U.S., P.D. and N.Ö.; Writing—original draft: O.G., A.R., R.J., B.-O.H. and N.Ö.; revising it critically for important intellectual content: M.C., N. EH., A.W., U.S. and P.D.; Final approval of the version to be published: O.G., R.J., A.R., M.C., N. EH., A.W., B.-O.H., U.S., P.D. and N.Ö. All authors have read and agreed to the published version of the manuscript.

**Funding:** An IFORES grant (D/107-40960) to Oliver Gembruch from the University of Duisburg-Essen supported the research. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

**Acknowledgments:** We acknowledge support by the Open Access Publication Fund of the University of Duisburg-Essen.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
