Determinants of quality of life in degenerative cervical myelopathy: a systematic review

Abstract

Background

Degenerative cervical myelopathy (DCM) is the most common cause of chronic, progressive spinal cord impairment worldwide. Patients experience substantial pain, functional neurological decline and disability. Health-related quality of life (HRQoL) appears to be particularly poor, even when compared to other chronic diseases. However, the determinants of HRQoL are poorly understood. The objective was to perform a systematic review of the determinants of quality of life of people with DCM.

Methods

A systematic search was conducted in MEDLINE and Embase following PRISMA 2020 guidelines (PROSPERO CRD42018115675). Full-text papers in English, exclusively studying DCM, published before 26 March 2020 were eligible for inclusion and were assessed using the Newcastle-Ottawa Scale and the Cochrane Risk of Bias 2 (RoB 2) tool. Study sample characteristics, patient demographics, cohort type, HRQoL instrument utilised, HRQoL score, and relationships of HRQoL with other variables were qualitatively synthesised.

Results

A total of 1176 papers were identified; 77 papers and 13,572 patients were included in the final analysis. A total of 96% of papers studied surgical cohorts and 86% utilised the 36-Item Short Form Survey (SF-36) as a measure of HRQoL. HRQoL determinants were grouped into nine themes. The most common determinant to be assessed was surgical technique (38/77, 49%) and patient satisfaction and experience of pain (10/77, 13%). HRQoL appeared to improve after surgery. Pain was a negative predictor of HRQoL.

Conclusion

Current data on the determinants of HRQoL in DCM are limited, contradictory and heterogeneous. Limitations of this systematic review include lack of distinction between DCM subtypes and heterogenous findings amongst the papers in which HRQoL is measured postoperatively or post-diagnosis. This highlights the need for greater standardisation in DCM research to allow further synthesis. Studies of greater precision are necessary to account for HRQoL being complex, multi-factorial and both time and context dependent.

Keywords:

• Cervical cord
• myelopathy
• spondylosis
• stenosis
• disc herniation
• ossification posterior longitudinal ligament
• degeneration
• disability
• recovery
• quality of life

Introduction

Degenerative cervical myelopathy (DCM) is the most common cause of spinal cord impairment worldwide, with a prevalence estimated to be as high as 5% of over 40s.^1–3 The pathophysiology comprises a range of degenerative changes in the cervical spine, including ligament hypertrophy, calcification and ossification, osteophytosis and vertebral disc herniation, that precipitate narrowing of the spinal canal, compression of the spinal cord and functional neurological decline.^1,2

A wide range of symptoms are common in patients with DCM: neck pain and stiffness, limb or body pain, weakness or stiffness, limb or body numbness, paraesthesia or sensory disturbance,^4,5 deterioration of manual dexterity, balance and coordination and autonomic dysfunction, such as poor control of the bladder and bowels.⁶ Diagnosis is challenging given the heterogeneity of symptoms.^1,4 The burden of disease is high in DCM: only a very small proportion of individuals currently make a full recovery; most individuals are left with lifelong disabilities and high rates of dependence and unemployment.^7,8

Health-related quality of life (HRQoL) is a subjective measure of how well a person functions in their life and his or her perceived wellbeing in physical, mental, and social domains of health.⁹ A number of instruments have been used to measure HRQoL in DCM clinical research.^10,11 Most instruments are generic and can be applied to any disease, whilst others are disease-specific.

The HRQoL of both DCM patients¹² and DCM carers¹³ is particularly poor when compared to other chronic diseases, such as diabetes, cancer and depression¹². Whilst often considered a consequence of the significant physical disability, trial data show that this is only weakly correlated.¹⁴ In fact, the factors contributing to DCM HRQoL are currently unknown. Their better characterisation would enable targeted treatments or support. This was selected as an international DCM research priority by the AO Spine and James Lind Alliance RECODE-DCM priority setting partnership.¹⁵ The emerging recognition of prevalent affective disorders,¹⁶ employment difficulties,⁸ strained relationships with carers¹³ and chronic pain may be significant.¹⁷

The objective of this systematic review was therefore to synthesise and summarise the current evidence on the determinants of quality of life of people with DCM.

Methods

Protocol and registration

A systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.¹⁸ The protocol was registered on PROSPERO (CRD42018115675) before commencement of the review. No amendments have been made since registration.

Information sources, eligibility criteria and paper selection

A search was conducted in Embase and Ovid MEDLINE on 12^th December 2018. The search was re-run on 26 March 2020. All English-language articles published before 26 March 2020 were eligible for inclusion. The full search strategy is described in Appendix 1. Two reviewers (Y. H. and O. D. M.) independently screened the titles and abstracts of the search results. Disagreements on inclusion of papers were discussed by all authors until mutual agreement was reached. Included papers were limited to English language. Papers with both randomised and non-randomised study designs were included. Case reports, opinion articles, editorials, conference papers, proposals, corrections and unpublished literature were excluded due to being poor quality of evidence.¹⁹ Animal studies and papers without validated HRQoL measures were excluded on the basis of having poor relevance to our review question. Papers reporting on mixed myelopathy and radiculopathy patient cohorts were excluded due to the focus of this review being on myelopathy. Existing systematic reviews and meta-analyses were not included. Reference lists of included articles and trial registries were not searched.

Full texts were searched and if available, screened for eligibility according to the following inclusion criteria:

• Primary research paper

• DCM is the primary focus of the paper

• Research which discusses determinants of HRQoL

• Patient quality of life is a focus of the paper

• English language

• Full-text available

Assessment of risk of bias

Risk of bias assessment was performed using the Newcastle-Ottawa Scale (NOS)²⁰ for cohort and case-control studies (Supplementary Table 2). A higher NOS score indicates less bias. Cochrane Risk of Bias 2 (RoB 2) tool²¹ was used to classify randomised control trials as 'Low', 'High' risk of bias, or ‘express some concerns' (Supplementary Table 3).

Risk of bias was independently completed in duplicate by two authors (Y. H. and O. D. M.).

Data collection and qualitative analysis

The following datapoints were extracted directly from each paper by one reviewer (Y. H.):

1. Author names

2. Publication year

3. Study design

4. Level of evidence, scored using the Oxford Centre for Evidence-based Medicine (CEBM) criteria¹⁹ where level 1 is a higher level of evidence compared to level 2

5. Conflicts of interest

6. Study sample characteristics: single/multiple centre, number of patients, country of patients or corresponding author, start and end dates of study

7. Patient demographics: age, gender, ethnicity, any other relevant demographics

8. Cohort type: surgical/non-surgical, aetiology of myelopathy

9. Name of validated HRQoL instrument utilised

10. HRQoL scores

11. Statistical methods used

12. Predictive factor assessed (e.g. age, comorbidities)

13. Factors associated with patient quality of life and statistical significance

Template data collection forms, data extracted from included studies and data used for all analyses have been made available (Supplementary Table 1).

HRQoL tools

Validated HRQoL instruments were defined as those that have been tested on an independent cohort from that on which they were developed and been shown to reliably distinguish between different degrees of quality of life. Validated HRQoL instruments used in the 77 included papers were:

• Short Form 36 (SF-36) is a generic HRQoL questionnaire composed of 36 questions which examines 8 domains related to HRQoL: physical functioning, physical role, physical pain, general health perceptions, vitality, social functioning, emotional health and mental health.²² An algorithm is used to calculate scores for each domain. These in turn are used to calculate composite scores of physical and mental well-being: a physical component score (PCS) and a mental component score (MCS).

• Short Form 12 (SF-12) consists of a subset of 12 items from the SF-36, examining the same 8 domains of HRQoL, but with fewer questions, making it more applicable in time-constrained circumstances.²³

• Short Form 6 (SF-6D)²⁴ is a preference-based measure derived from 11 questions of SF-36.²⁵ It expresses health in terms of 6 dimensions and describes 18,000 different health states.

• EuroQol-5 Dimension (EQ-5D) is a generic preference-based instrument which measures HRQoL in 5 dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression.²⁶

• World Health Organization Quality of Life (WHOQOL-BREF) questionnaire is an abbreviated version of WHOQOL-100 which is a generic, cross-culturally applicable questionnaire that examines QoL through 4 domains: physical health, psychological, social relationships and environmental health.²⁷

• Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire (JOACMEQ) is a DCM-specific outcome measure. It is a self-administered questionnaire comprising 24 questions and 5 domains: cervical spine function, upper extremity function, lower extremity function, bladder function and quality of life.²⁸

Statistical analysis

Formal meta-analysis was not conducted due to the heterogeneity between papers. A qualitative synthesis is presented

Results

A total of 1176 papers were identified by our search; 1080 remained after de-duplication. Following screening, 162 underwent full text screening and 77 papers were included in the final analysis (Figure 1). The majority (65/77, 84%) were cohort studies (prospective and retrospective) with cohort sizes ranging from 10 to 757. A total of 74 papers (74/77, 96%) studied surgical cohorts.

Figure 1. PRISMA flow diagram.

The most commonly utilised HRQoL instrument was the SF-36 (Table 1), which was utilised in 66 papers (66/77, 86%): 59 papers (59/77, 77%) presented both PCS and MCS scores, whilst 7 papers (7/77, 9%) reported only PCS scores. EQ-5D was reported in 10 papers (10/77, 13%), whilst 6 papers (6/77, 8%) reported SF-12. A total of 4 papers (4/77, 5%) reported JOACMEQ scores. SF-6D, PHQ-9 and WHOQoL-BREF scores were each reported by 1 paper (1/77, 1%). Some papers utilised more than one HRQoL instrument.

Table 1. Validated HRQoL instruments and number of included papers which utilised them.

Quality of Life Instruments used in DCM Literature	Number of papers
SF-36 (PCS and MCS)	59
EQ-5D	10
SF-36 (PCS only)	7
SF-12	6
JOACMEQ	4
SF-6D	1
PHQ-9	1
WHOQOL-BREF	1

Potential HRQoL determinants were grouped into 9 themes (Table 2). Themes were developed, piloted and iteratively refined to group identified papers into the minimum number of distinct categories, as mutually agreed by all authors. Themes were not mutually exclusive since some papers assessed multiple distinct determinants of HRQoL. The most common determinant to be assessed was surgical technique (38/77, 49%) followed by patient satisfaction and experience of pain (10/77, 13%).

Table 2. Potential determinants of DCM HRQoL. Interpretation of the quality of evidence using the GRADE criteria is as follows: High quality (+ + + +) denotes further research is very unlikely to change our confidence in the estimate of effect. Moderate quality (+ + + −) denotes further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality (+ + − −) denotes further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality (+ − − −) denotes any estimate of effect is very uncertain.^103,104

Theme	Sub-theme	Relationship examined (HRQoL & X)	Total (significant, NS) Significant (+, -)	CEBM Level of Evidence	Conclusion (quality of evidence using the GRADE criteria)^30
1. Surgical management	Comparing pre- and post-operative	Postoperative HRQoL (PCS and MCS) compared to pre-operative HRQoL (PCS and MCS)	14 (11,3) → (^7^,^29–38, ^39–41) 11 (11, 0) → (^7^,^29–38, 0)	2B to 4	Significantly higher (+ + + −)
		Postoperative HRQoL (PCS only) compared to pre-operative HRQoL (PCS only)	5 (5,0) → (^82–86, 0) 5 (5,0) → (^82–86, 0)	2B	Significantly higher (+ + + +)
	Comparing surgery and conservative management	HRQoL following surgery compared to conservative management	1 (0,1) → (0,^42)	2B	No significant difference (+ − − −)
	Comparing two surgical techniques	HRQoL following different surgical approaches	15 (7,8) → (^14^,^43–48^,^83^, ^49–55)	1B to 4	Uncertain (+ + − −)
		Postoperative HRQoL following cervical microendoscopic laminotomy (CMEL) vs conventional expansive laminoplasty (ELAP)	2, (0,2) → (0,^53^,^55)	3B	No significant difference (+ − − −)
		Postoperative HRQoL following French door laminoplasty vs open-door laminoplasty	1 (1,0) →(^46, 0) 1 (1,0) →(^46, 0)	1B	Significantly higher after French door (+ − − −)
		Postoperative HRQoL following anterior vs posterior surgery	5 (4,1) →(^43–45^,^47^,^50) 4 (4, 0) →(^43–45^,^47, 0)	2B	Significantly higher after anterior approach (+ + − −)
		Postoperative HRQoL following anterior vs combined anterior and posterior approaches	1 (0,1) →(0,^49)	2B	No significant difference (+ − − −)
		Postoperative HRQoL following arthroplasty vs anterior discectomy and fusion	1 (1,0) →(^48, 0) 1 (1,0) →(^48, 0)	2B	Significantly higher in arthroplasty (+ − − −)
		Postoperative HRQoL following laminoplasty vs laminectomy and fusion	3 (0,3) →(0,^51^,^52^,^55)	2B to 3B	No significant difference (+ + − −)
		Postoperative HRQoL following Double-door Laminoplasty vs Selective Laminoplasty	1 (0,1) → (0,^54)	4	No significant difference (+ − − −)
2. Additional peri-operative management		HRQoL with and without additional peri-operative treatment	3 (0, 3) → (0,^87–89)	1B to 2B	No significant difference (+ − − −)
3. Patient experience after surgery	Satisfaction	Level of satisfaction and HRQoL (MCS and PCS)	1 (1,0) → (^90,0) 1 (1,0) → (^90,0)	2B to 4	Significant relationship (+ + − −)
		Level of satisfaction and HRQoL (PCS only)	2 (2, 0) → (^56^,^91,0) 2 (2, 0) → (^56^,^91,0)	2B	Significant Relationship (+ + − −)
	Pain	Level of pain and HRQoL	7 (5, 2) → (^40^,^57–60, ^61^,^92) 5 (0, 5) → (0, ^40^,^57–60)	2B to 4	Significant relationship (+ + + −)
	Postoperative complication	Postoperative complications and HRQoL	1 (0,1) → (0,^93)	2B	No significant relationship (+ − − −)
4. Cervical alignment, balance and range of motion	–	Structural spinal deformities and HRQoL (PCS only)	9 (7,2) → (^43^,^61–66^,^64^,^94) 7 (7,0) → (^43^,^61–66, 0)	2B to 4	Significant relationship (+ + + −)
5. Radiological characteristics of the spinal cord	Cord compression	Cord compression and HRQoL	2 (1, 1) → (^96^,^95) 1 (0,1) → (0,^96)	2B	Uncertain (+ − − −)
	Diffusion Tensor Imaging	Fractional anisotropy and HRQoL	1 (0,1) → (0,^68)	2B	No significant relationship (+ − − −)
6. Functional disability and severity of DCM		Disability due to DCM and HRQoL (PCS only)	5 (5,0) → (^56^,^69–72,0) 5 (0,5) → (0, ^56^,^69–72)	2B to 4	Significant relationship (+ + + −)
		Disability due to DCM and HRQoL (PCS and MCS)	4 (3,1) → (^14^,^73^,^74, ^97) 3 (0,3) → (0, ^14^,^73^,^74)	2B to 4	Significant relationship (+ + + −)
7. Demographics	Age	Age and HRQoL	3 (2, 2) → (^14^,^42, ^75^,^98) 2 (0,2) → (0, ^14^,^42)	2B	Uncertain (+ − − −)
	Race	Race and HRQoL	2 (1,1) →(^14^,^99)	2B	Uncertain (+ − − −)
	Geographic variation	Continent and HRQoL	2 (1,1) →(^100, ^7)	2B	Uncertain (+ + − −)
8. Co-morbidities		Presence of comorbidities and HRQoL	8 (7,1) → (^14^,^76–81, ^101) 7 (0,7) →(0, ^14^,^76–81)	2B	Significant relationship (+ + + −)
9. Biochemistry and molecular genetics	Polymorphism in brain derived neurotrophic factor (BDNF) gene	Val66Met and HRQoL	1 (1,0) →(^102, 0) 1 (0,1) →(0,^102)	2B	Significant relationship (+ − − −)

Owing to heterogeneity in choice of outcome tools and outcome reporting, quantitative synthesis of data was not possible. Instead, results are presented qualitatively.

Surgical management

A total of 19 papers of evidence level 2B to 4 and similar risk of bias scores (NOS 6-8) compared pre- and postoperative HRQoL. Eleven papers, consisting mainly of prospective cohort studies, reported postoperative improvement in both mental and physical components of HRQoL in a total of 1807 patients.^7,29–38. Four retrospective cohort studies and one prospective cohort study reported statistically significant improvement in the physical component but not mental component of HRQoL.^82–86 In contrast, 3 papers, including an international cohort study of 479 patients, reported no significant improvement in HRQoL following surgery.^39–41 Moreover, 1 prospective cohort study of 62 patients reported no consistent surgery-associated improvement in HRQoL compared to non-operative management.⁴²

In addition, 15 papers of evidence level 2B to 4 compared two surgical techniques, including laminectomy and fusion versus laminoplasty, French door laminoplasty versus open door laminoplasty, anterior versus posterior surgery and arthroplasty versus anterior cervical discectomy and fusion (ACDF). A total of 8 cohort studies all with a NOS risk of bias score of 8, encompassing a total of 1268 patients, found that surgical approach did not affect HRQoL.^49–55,83 One paper used a principal component analysis to show that surgical approach and number of operated levels contributed to SF-36 PCS scores.¹⁴ In addition, 6 papers, 5 of which has NOS risk of bias scores from 7-8 and 1 of which is a randomised control trial of level 1B evidence, identified a significant difference in HRQoL between different surgical approaches: higher HRQoL following French door laminoplasty compared to open-door laminoplasty,⁴⁶ an anterior approach compared to an posterior approach (level 2B evidence)^43–45,47 and following arthroplasty compared to anterior discectomy and fusion (level 2B evidence).⁴⁸ Hitchon et al. (2019) provided level 2B evidence that there were differences in the subcomponents of SF-36 that improved following anterior and posterior surgery; for example, in the anterior group general health index did not significantly increase postoperatively, whilst in the posterior group energy and fatigue scores did not change significantly postoperatively.⁴⁷

Additional management in the peri-operative period

In a cohort study of 203 patients of level 2B evidence, no significant difference in pre- and postoperative SF-36 was found in those who did and those who did not receive pre-operative epidural anaesthesia, physiotherapy and opioids.⁸⁷ Two small randomised controlled trials of level 1B evidence identified no significant improvement in short-term peri-operative HRQoL following postoperative collar fixation⁸⁸ and postoperative dexamethasone⁸⁹ designed to mitigate against dysphagia and airway compromise.

Patient experience after surgery: satisfaction, pain and complications

Three cohort studies of level 2B to 4 evidence and NOS risk of bias score 6, including a total of 313 patients, each reported a significant positive correlation between patient satisfaction and physical domains of HRQoL.^56,90,91 Kimura et al. (2014) reported significantly greater mental domain scores of HRQoL, indicating better state of mental wellbeing, for those with greater satisfaction.⁹⁰ A total of 5 cohort studies of level 2B to 4 evidence and NOS risk of bias score 6, with a total sample size of 619, reported that pain is associated with decreased SF-36 and EQ-5D in both mental and physical domain scores, however the mental component to a greater extent.^40,57–60 In contrast, Thakar and Rajshekhar (2012) reported that neither the SF-36 MCS nor the PCS scores correlated significantly with reported pain levels in their cohort study of 60 patients, which has a NOS risk of bias score of 6. Moreover, Fujiwara et al. (2018) did not find a significant correlation between pain and JOACMEQ score in their observational cohort study of 57 patients, which has the same risk of bias score.^61,92 One retrospective cohort study of level 2B evidence studying 203 patients identified postoperative complications such as cerebrospinal fluid leak, postoperative radiculopathy and excessive bleeding as possible determinants of immediate HRQoL, however reported that these complications had no significant influence on HRQoL at 2 years post-surgery.⁹³

Cervical alignment, balance and range of motion

A total of 7 cohort studies of level 2B to 4 evidence with NOS risk of bias scores ranging from 6 to 8, studying a total of 484 patients, reported that deformities in cervical alignment, balance and range of motion negatively impacts the physical component of SF-36 but not the mental component.^43,61–66 However, two papers with the same range of risk of NOS bias scores reported no significant correlation between spinal deformity and imbalance and any measure of HRQoL.^67,94

Radiological characteristics of the spinal cord

In their cohort study of level 2B evidence with a NOS risk of bias score of 6 studying 134 patients, Karpova et al. (2014) reported that the degree of cord compression did not correlate significantly with HRQoL despite a correlation with physical symptoms such as Hoffman’s sign, hyperreflexia and sphincter dysfunction.⁹⁵ However, Arvin et al. (2013) reported, in their blinded cohort study of level 2B evidence and the same risk of bias score studying 57 patients, a significant negative correlation between pre-operative maximal cord compromise measured on mid-sagittal MRI and postoperative SF-36 MCS. Despite postoperative improvements in physical function such as walking and balance scores, patients with high indices of cord compromise and cord compression had poor perception of well-being and low MCS scores.⁹⁶ This study also reported significantly poorer postoperative SF-36 physical scores in patients with focal T1 signal change as opposed to diffuse T2 signal change.⁹⁶ Another paper reported that fractional anisotropy demonstrates a strong positive correlation with measures of physical function such as mJOA, however does not predict postoperative SF-36.⁶⁸

Functional disability and severity of DCM

A total of 5 cohort studies of level 2B to 4 evidence and NOS risk of bias scores ranging from 6-8 on a total sample size of 1771, reported a significant negative correlation between physical disability and the PCS component of HRQoL,^56,69–72 whilst 3 cohort studies of level 2B to 4 evidence and NOS risk of bias score 8 on a total of 776 patients reported a significant correlation of HRQoL with both PCS and MCS components of HRQoL.^14,73,74 One observational cohort study on 75 patients, utilised a modified Italian version of mJOA and reported no relationship between disability and patient-reported HRQoL.⁹⁷

Demographics: age, race and geographic variation

Two studies of level 2B evidence with sample sizes of 62 and 154, reported a significant correlation between increasing age and poorer HRQoL, especially PCS. ^14,42 In contrast, a prospective cohort study of the same evidence level and the same NOS risk of bias score of 8, studying 479 participants found that old age results in reduced functional status but not decreased HRQoL.⁹⁸ Isogai et al. (2018) studied elderly participants and found that being over 80 years does not affect surgical outcome.⁷⁵ Nagoshi et al. (2016) found no significant difference in SF-36 between different races,⁹⁹ whilst Badhiwala et al. (2018) used a principal component analysis to show that race impacts the MCS component of SF-36.¹⁴ Fehlings et al. (2015) compared SF-36 in 479 patients from North America, Latin America, Asia Pacific and Europe in a prospective cohort study of level 2B evidence and found no differences in the HRQoL scores postoperatively despite variation in surgical practices.⁷ However, in a more recent prospective cohort study of level 2B evidence studying 757 patients,¹⁰⁰ Fehlings et al. (2018) reported significant geographic differences in postoperative changes in the PCS and MCS of SF-36.¹⁰⁰ At 2 years, improvements in PCS were highest in patients from Asia Pacific (p = 0.0109) whilst improvements in MCS were highest in Latin America (p < 0.0001) compared to other regions.

Co-morbidities

A total of 7 international papers of level 2B evidence with risk of bias scores ranging from 6-8 reported that there was a negative correlation between the presence of comorbidities, including renal disease, rheumatoid arthritis, osteoarthritis, obesity, diabetes, Parkinson’s disease, depression and bipolar disorder, and HRQoL in DCM patients.^14,76–81 In contrast, a retrospective matched-pair cohort study of equal evidence level and risk of bias score of 8 reported no significant relationship between diabetes and DCM HRQoL in 58 subjects.¹⁰¹

Biochemistry and molecular genetics

Abode-Iyamah et al. (2016) reported a statistically significant correlation between the presence of the Val66Met polymorphism in brain derived neurotrophic factor (BDNF) gene and lower SF-36 PCS scores in 10 DCM patients.¹⁰² Patients with the Val66Met variant also had poorer levels of daily activity and more severe symptoms compared to those without the variant.

Discussion

Summary of findings

Data on the determinants of HRQoL in DCM are limited, contradictory and heterogeneous. HRQoL typically improves after surgery. Adjuvant treatments and complications in the peri-operative period are suggested to be unimportant to HRQoL, whilst levels of pain and patient satisfaction appear important. The extent of spinal deformity appears important, however the extent of radiological compression uncertain. Functional disability appears to be a possible contributor to HRQoL. Evidence on the effect of race and geography is inconsistent. Increasing comorbidity correlates strongly with poorer HRQoL. A single genetic study identified a strong relationship between one genetic variant and HRQoL.

The majority of papers assessing HRQoL in relation to surgery suggested a postoperative improvement in HRQoL, especially in the physical domain scores, supporting arguments that surgical intervention is cost-effective.³³ However, postoperative changes in PCS and MCS scores do not always correlate. For example papers of level 2B evidence by Hitchon et al. (2018) and Blizzard et al. (2017) report improvements in the physical domain scores of SF-36 but not mental domain scores.^83,85 The reason for this is unknown. One possible explanation may be because improvement in neurological function do not necessarily equate to improvement in mental health.³⁶ Surgery has been shown to be effective at halting disease progression, but commonly functional recovery remains incomplete.^1,7,35 Zhou et al. (2014) reported that initially the mental component of HRQoL did not improve.³⁶ Thus, whilst PCS may reflect largely objective functional scales, such as the mJOA, MCS may also capture patient perception of their recovery and health state. This illustrates the importance of measuring both PCS and MCS.

Another possible reason for the discrepancy between PCS and MCS is patient perception of wellbeing and satisfaction. A number of papers have shown that irrespective of physical recovery, HRQoL is determined by subjective perception of surgical success.^56,90,91 In other words, even the physical component of quality of life scores may be influenced to a greater degree by patient satisfaction than by functional recovery.

Furthermore, pain is an affective and physical experience closely related to patient satisfaction. A total of 5 out of the 7 papers studying pain conclude that presence of pain is associated with worsening in both the physical and mental HRQoL domain scores, the mental component to a greater extent.^40,57–60 Thus, even if physical function such as mobility and dexterity improve after surgery, if there is persisting pain, the HRQoL score, especially the mental components, appears to remain very low. This is supported by level 2B evidence from Kimura et al. (2015), who reported a significant negative correlation between axial pain intensity and all HRQoL measures at 6 months follow-up.⁴⁰ In our recent cross-sectional survey of people with DCM, we identified pain as the number one recovery priority of people with DCM.¹⁷

However, two prospective studies of level 2B evidence did not report a significant relationship between pain and HRQoL.^61,92 The discrepancy could be due to the small number of subjects (60 and 57, respectively), different geographical location and different cultural practices of expressing pain as both these studies recruited patients from single countries (India and Japan respectively). Pain rather than decline in sensory, motor and visceral functions is a common reason to initially seek DCM treatment.¹ Thus, pain may be the biggest determinant of well-being, having a substantial effect on satisfaction and HRQoL. Prioritising pain as the critical symptom to be treated may therefore improve HRQoL in DCM patients. Currently no peri-operative treatments have been found effective in decreasing long-term pain. More research and trials in this area will be of benefit.

Evidence suggests that the degree of spinal deformity is important, as it determines regional and global neck stability and function, however, it only appears related to the physical component of quality of life. This could be due to the fact that surgery restores some physical functioning, but this may not be sufficient to leave patients satisfied. On the contrary, two papers found that cervical alignment does not affect HRQoL. This could be explained by the fact that the primary goal of surgery in myelopathy is to decompress and reduce neurological symptoms and sufficient decompression can still be achieved with residual kyphosis.⁶⁵ Restoration of regional and global cervical balance may contribute to regaining of physical function and activities of daily living, however restoration of alignment on its own seems to have minimal impact on HRQoL, if spinal cord compression is not corrected. However, radiological evidence of spinal compression does not seem to correlate with HRQoL.

Limitations and future directions

First, there is a substantial issue of contradictory evidence, mostly of similar quality. For example, 16 papers report HRQoL improves significantly following surgery whilst 3 papers report it does not; 7 papers suggest surgical technique affects HRQoL whilst 8 papers show no evidence of relationship. It is crucial to elucidate the cause of this contradiction and how we can mitigate against it. The heterogeneity of evidence may be due to different sampling: DCM is a powerful umbrella concept for a range of degenerative pathology in the cervical spine with a final common pathway of cord compression however DCM is not a single condition and consists of a number of subtypes, such as ossification of the posterior longitudinal ligament and cervical disc herniation. There may be value in considering different subcategories of DCM separately as they may have different HRQoL profiles.

Amongst our search result there is also a heterogeneity in the demographics of the subjects being studied e.g. age, race, culture, and severity of DCM, as well as the tool used to measure HRQoL. Thus, the lack of standardisation makes interpretation of data extremely difficult and prevented meta-analysis. This highlights the urgent need for a universally agreed definition of DCM, a core measurement set, a core outcome set and standardisation of all relevant factors in DCM research to improve research efficiency and synthesis, which is currently being undertaken by the RECODE-DCM project.¹⁵

Furthermore, HRQoL is time-dependent and there is marked variation amongst the papers when HRQoL is measured postoperatively or post-diagnosis. For example in Gerling et al. (2017) postoperative complications affected HRQoL in the short term but not after 2 years.⁹³ The determinants of HRQoL are likely to differ at different time points e.g. at diagnosis, pre-op, 1-week postoperatively, 1-year postoperatively and 10 years postoperatively. Further studies are required to examine these factors in detail, especially the long-term effect of surgery on HRQoL.

Moreover, many of the papers report correlative evidence, which cannot be interpreted as causative, leaving the ultimate drivers of HRQoL undefined. For example, papers that show HRQoL improvement after surgery cannot exclude the effect of confounding variables such as improvements in pain or satisfaction, and even these confounding variables themselves likely interact e.g. reduced satisfaction due to pain. Moreover, the relationship between these factors is likely to be dynamic and evolve over time.

Given these limitations, current research does not provide sufficient data to fully elucidate the HRQoL determinants in DCM. We anticipate that the DCM HRQoL is affected by wider factors such as culture and socioeconomics as well as disease-specific factors which vary with time and context. Studies of greater precision are needed in recognition of the fact that HRQoL is highly multi-factorial, complex and may be different for different DCM subtypes. Future work will thus require multiple simultaneous streams. First, future work must consider what factors may be universally important determinants of HRQoL for people with any chronic disease, independent of the disease e.g. socioeconomic and cultural factors, whilst also considering how the specifics of DCM may interact with and modify these general factors. Second, future work must identify determinants specific to DCM HRQoL, and furthermore those factors most important to specific subtypes of DCM, in different contexts and at different points in time.

Conclusion

Current evidence on the determinants of HRQoL in DCM is inconsistent. In order to provide targeted and effective support for the patients and their carers, better characterisation of determinants of HRQoL in DCM is essential. This will only be possible with greater standardisation in DCM research.

Disclaimer

This report is independent research arising from a Clinician Scientist Award, CS-2015-15-023, supported by the National Institute for Health Research. The views expressed in this publication are those of the authors and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health and Social Care.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

M. R. K. gratefully acknowledges support by the Cambridge NIHR Brain Injury MedTech Cooperative and is funded by a NIHR Clinician Scientist Award CS-2015-15-023. B. M. D. is supported by a Royal College of Surgeons Research Fellowship and an NIHR Clinical Doctoral Research Fellowship.

Appendix 1

Embase Search Strategy

Initial date: 12/12/18

Re-search: 26/03/20

1. quality of life/

2. exp health status/

3. "activities of daily living"/

4. life qualit$.mp.

5. exp self-concept/

6. health level.mp.

7. level of health.mp.

8. wellness.mp.

9. well being.mp.

10. (activities of daily life or daily living activities).mp. [mp = protocol supplementary concept, rare disease supplementary concept, title, original title, abstract, name of substance word, subject heading word, unique identifier] {No Related Terms}

11. functional ability.mp.

12. good health.mp.

13. healthiness.mp.

14. patient reported outcomes.mp.

15. social adjustment/

16. physical limitations.mp.

17. psychiatric status.mp.

18. pain measurement/

19. functional assessment.mp.

20. sense of coherence.mp.

21. (sf36 or cesd or bdi).mp.

22. or/1-21

23. exp Cervical Vertebrae/or exp Cervical Cord/or cervical.mp. or (phrenic nucleus or accessory nucleus).mp. or (("Japanese Orthop?edic Association" adj2 score*) or (joa adj2 score*)).mp.

24. myelopath*.mp. or exp Spinal Cord Diseases/or (spinal cord adj3 (diseas* or disorder*)).mp. or myeloradiculopath*.mp. or spondylomyelopath*.mp. or spondylomyeloradiculopath*.mp. or (Spinal Cord adj3 Compress*).mp. or exp Spinal Cord Compression/

25. 23 and 24

26. exp "Ossification of Posterior Longitudinal Ligament"/

27. 25 or 26

28. exp Atlanto-Occipital Joint/or exp Arteriovenous Fistula/or exp Radiotherapy/or exp Vitamin B 12/or exp Radiation/or exp Radiation Injuries/or exp Re-Irradiation/or exp Craniospinal Irradiation/or exp Whole-Body Irradiation/or exp Motor Neuron Disease/or exp Amyotrophic Lateral Sclerosis/or exp Neoplasm Metastasis/or exp Hemangioma/or exp neoplasm/or exp metastasis/or exp Nervous System Malformations/or exp "autoimmune diseases of the nervous system"/or exp "congenital, hereditary, and neonatal diseases and abnormalities"/or exp virus diseases/

29. 27 not 28

30. 22 and 29

31. limit 30 to english language

Appendix 2

MEDLINE Search Strategy

Initial date: 12/12/18

Re-search: 26/03/20

1. exp "quality of life"/

2. exp health status/or good health.mp. or healthiness.mp.

3. *daily life activity/or (activities of daily life or daily living activities).mp.

4. life qualit$.mp.

5. (level of health or health level).mp.

6. exp self-concept/

7. exp *wellbeing/or wellness.mp.

8. exp functional status/

9. exp functional status assessment/

10. (sf36 or cesd or bdi).mp.

11. exp patient-reported outcome/

12. social adaptation/

13. exp physical disability/co, dm, ep, et, rh, si, su, th [Complication, Disease Management, Epidemiology, Etiology, Rehabilitation, Side Effect, Surgery, Therapy]

14. physical limitations.mp.

15. exp mental health/ep [Epidemiology]

16. exp *pain assessment/or exp *pain measurement/or exp *chronic pain/or exp *pain/

17. "sense of coherence"/

18. or/1-17

19. exp cervical vertebra/or exp cervical spine/or exp cervical spinal cord/or (phrenic nucleus or accessory nucleus).mp. or (("Japanese Orthop?edic Association" adj2 score*) or (joa adj2 score*)).mp.

20. (myelopath* or spondylomyeloradiculopath*).mp. or *spondylosis/or exp *cervical spondylosis/or *spinal cord disease/or *myeloradiculopathy/or *spinal cord compression/or *ligament calcinosis/or exp *cervical myelopathy/or degenerative cervical myelopathy.mp.

21. 19 and 20

22. atlantooccipital joint/or exp arteriovenous fistula/or exp radiotherapy/or exp radiation/or exp cyanocobalamin/or exp re-irradiation/or exp irradiation/or exp craniospinal irradiation/or exp whole body radiation/or exp motor neuron disease/or exp amyotrophic lateral sclerosis/or exp metastasis/or exp neoplasm/or exp hemangioma/or exp nervous system malformation/or exp autoimmune disease/or exp virus infection/or (congenital, hereditary, and neonatal diseases and abnormalities).mp. or acute spinal injury.mp. or acute trauma.mp.

23. 21 not 22

24. 18 and 23

25. limit 24 to english language