Musculoskeletal morbidity following spinal cord injury: A longitudinal cohort study of privately-insured beneficiaries

Abstract

Background:

People living with spinal cord injuries (SCIs) experience motor, sensory and autonomic impairments that cause musculoskeletal disorders following the injury and that progress throughout lifetime. The range and severity of issues are largely dependent on level and completeness of the injury and preserved function.

Objective:

High risk of developing musculoskeletal morbidities among individuals after sustaining a traumatic SCI is well known in the clinical setting, however, there is a severe lack of evidence in literature. The objective of this study was to compare the incidence of and adjusted hazards for musculoskeletal morbidities among adults with and without SCIs.

Methods:

Privately-insured beneficiaries were included if they had an ICD-9-CM diagnostic code for SCI (n=9,081). Adults without SCI were also included (n=1,474,232). Incidence estimates of common musculoskeletal morbidities (e.g., osteoporosis, sarcopenia, osteoarthritis, fractures, etc.) were compared at 5-years of enrollment. Survival models were used to quantify unadjusted and adjusted hazard ratios for incident musculoskeletal morbidities.

Results:

Adults living with traumatic SCIs had a higher incidence of any musculoskeletal morbidities (82.4% vs. 47.5%) as compared to adults without SCI, and differences were to a clinically meaningful extent. Survival models demonstrated that adults with SCI had a greater fully-adjusted hazard for any musculoskeletal morbidity (Hazard Ratio [HR]: 2.41; 95%CI: 2.30, 2.52), and all musculoskeletal disorders, and ranged from HR: 1.26 (1.14, 1.39) for rheumatoid arthritis to HR: 7.02 (6.58, 7.49) for pathologic fracture.

Conclusions:

Adults with SCIs have a significantly higher incidence of and risk for common musculoskeletal morbidities, as compared to adults without SCIs. Efforts are needed to facilitate the development of improved clinical screening algorithms and early interventions to reduce risk of musculoskeletal disease onset/progression in this higher risk population.

Introduction

Spinal cord injury (SCI) results in loss of motor, sensory and autonomic functions below the level of the lesion with a range of completeness and incompleteness of neurological injury. Cervical injuries result in tetraplegia and thoracic/lumbar injuries result in paraplegia. It is well known clinically that there are multiple sequelae to the musculoskeletal system directly due to or related to SCI level, completeness and duration of injury. Loss of motor and sensation function directly result to loss of muscle mass and muscle atrophy, as well as low bone mineral density.^1-8 Primary metabolic dysfunction related to autonomic dysfunction in SCI also results to heightened bone loss in the acute phase and is persistent throughout life which increases risk for fractures;^9-15 and can likewise result to abnormal ossifications and bony growths.^16-18 Many of the musculoskeletal issues arise from repetitive stress to muscles, ligaments, tendons and joints particularly in the upper extremities,^19-28 but also occurs in the lower extremities in those who have ability to stand or walk.^29,30 Decreased mobility and spasticity after SCI lead to progressive joint contractures.^31-36 Furthermore, people with SCI are not exempt from developing osteoarthritis, although patterns of osteoarthritis in people with SCI are most likely different from the norm since stress on various joints are different for wheelchair users or for abnormal biomechanics from compensated standing and gait.^28,37-40 Impaired sensation may also contribute to unintended trauma to both upper and lower extremities that adds further risk to musculoskeletal injury. All these disorders can be large contributors to pain and morbidity in the SCI population, and therefore must always be strongly considered, included in evaluation and closely monitored.^19,41-45

However, there have been very few nationwide studies to examine the risk for developing musculoskeletal disorders in the SCI population as compared to a non-SCI population. The purpose of this study was to examine the 4-year incidence of common musculoskeletal morbidities in a large sample of privately-insured adults with SCI, as compared to adults without.

Methods

Data Source

This is a retrospective cohort study of adults with SCI whose SCI diagnosis could have existed across any patient care setting. This study used a national, private insurance claims database, Clinformatics DataMart Database (OptumInsight, Eden Prairie, MN). This is a de-identified administrative claims database of over 80 million adults and children with commercial insurance representing those on a single, large U.S. private payer who had both medical and pharmacy coverage throughout the enrollment. Enrolled beneficiaries’ emergency department, outpatient, and inpatient encounters are captured. This study was deemed exempt by the University of Michigan Institutional Review Board at the researchers’ institution.

Sample Selection

All individuals 18 years of age and older at the time of their enrollment which could start from 2007 to 2017 were potentially eligible for this analysis. We excluded individuals with less than 12 months of continuous enrollment to require sufficient baseline claim history.

Identification of Patients with a Spinal Cord Injury

All members with a diagnosis of SCI were identified using International Classification of Diseases, Ninth revision, Clinical Modification (ICD-9-CM) (Supplementary Table S1). Members that had SCI prior to 2007 were excluded due to poorer coverage of diagnosis codes during 2001 to 2006 in the database. Members without a diagnosis code when they were 18 years or older during enrollment were excluded. To allow adequate longitudinal follow up for all patients with SCI, only those that had four or more continuous years of enrollment following their first SCI diagnosis date within the study period were included.

A comparison cohort of controls without SCI were also identified using the same aforementioned inclusion criteria. Additional exclusion criteria for identifying the control cohort included removal of any individual with other physically disabling neurological disorders (i.e., non-SCI paraplegia, non-SCI quadriplegia, non-SCI hemiplegia, cerebral palsy, spina bifida, and multiple sclerosis). Among remaining members without SCI, we obtained a 20% simple random sample of general population controls, using a fixed randomization seed. We further examined that no unintentional bias was introduced due to random sampling by conducting post-hoc effect size calculations between the full general population control cohort and the 20% sample on baseline covariates such as demographics and prevalent comorbidities. We considered an unbiased random sample if post-hoc effect sizes indicated no meaningful differences.

Musculoskeletal Morbidities

Physician-diagnosed musculoskeletal health disorders were identified based on a single encounter that included at least one of pertinent ICD-9 or ICD-10 codes (see Supplementary Table S1). All musculoskeletal morbidities were chosen based on established categories through the AHRQ indicators of clinical classification software (CCS).⁴⁶ CCS is a software tool that aggregates ICD-9-CM diagnoses codes into higher levels of clinical classifications. The decision to follow the AHRQ clinical definitions was made a priori to provide uniformity. The primary outcome was time in days to any incident musculoskeletal morbidity following enrollment on the plan. Secondary outcomes were component incident musculoskeletal morbidity, including: (1) osteoarthritis; (2) osteoporosis; (3) pathologic fracture; (4) disorders of muscle, joint, ligament, tendons, and connective tissue (e.g., upper extremity tendonitis; synovitis and tenosynovitis; other disorders of synovium and tendon [e.g., synovial hypertrophy]; bursitis; enthesopathies-lower limb [e.g., hip tendonitis, etc.]; other enthesopathies [e.g., lateral epicondylitis]; other and unspecified soft tissue disorders, not elsewhere classified [e.g., panniculitis]; calcification and ossification of muscle; (5) sarcopenia and weakness; (6) myalgia; and (7) rheumatoid arthritis, myositis, and musculoskeletal infections.

Covariates

Explanatory covariates included age, sex, race, educational attainment, household net worth, and a modified Elixhauser comorbidity index. The Elixhauser comorbidity index was modified to remove one conditions that would be correlated with incident musculoskeletal morbidity: rheumatoid arthritis. Therefore, the revised index only considers 30 comorbidities (Supplemental Table S2).

Statistical Analysis

Bivariate analyses of baseline demographic characteristics between patients with SCI and controls were examined. For categorical variables, column percentages were compared between both groups using effect size calculations with Cohen’s h. The Cohen’s h effect size calculation was used since, due to large sample sizes, being statistically overpowered would not provide clinically meaningful differences in proportions between groups. For continuous variables, means and standard deviations as well as medians with upper and lower bounds on interquartile ranges were calculated. Cohen’s d standardized mean differences were calculated for continuous variables to ascertain clinically meaningful differences between groups.

To capture full comorbidity history within the study period, all patients with sufficient continuous enrollment within the study period of five years were retained to enable sufficient follow-up. For the SCI cohort, we used a one year look back period prior to first SCI diagnosis date to capture comorbidity history and to examine if any prevalent musculoskeletal outcomes existed.

To examine disease-free survival of individuals with SCI compared to controls, those patients that had no evidence of composite musculoskeletal morbidity in each group were plotted using Kaplan-Meier product limit survival curves for a three-year period. To establish incidence in claims, we used a one-year lookback period from the index date in each group to obtain evidence of any service utilization with a diagnosis of any musculoskeletal morbidity. These patients were excluded from the product-limit survival curves and other subsequent analyses.

To estimate the unadjusted and adjusted hazard of the composite and each musculoskeletal morbidity, a series of survival models were developed. For each musculoskeletal morbidity, all patients who had evidence of the specific musculoskeletal morbidity were excluded from the model. For example, if osteoporosis was being considered as the incident outcome, all patients with prevalent osteoporosis in the one year prior to the index date would be excluded from the model. Therefore, sample sizes of patients included for each outcome varied based on evidence of prevalent disease in the one year prior to the index date. Survival models were then used to quantify unadjusted and adjusted hazard ratios for each incident musculoskeletal morbidity. Appropriate survival models were based on distributional assumptions that included testing Weibull, lognormal, exponential, gamma, logistic, loglog, and Normal distribution with respect to the follow-up in days by minimizing critical model fit statistics. Critical assessment of Akaiki Information Criterion (AIC) was used as a basis for minimization amongst all candidate distributions. Use of the parametric Weibull regression for incident musculoskeletal outcome was applied stepwise. To examine the effects of incremental adjustment on the exposure variable (SCI), a series of models for each musculoskeletal outcome was evaluated. All patients were right censored if they did not experience the outcome in the follow-up period or disenrolled from the plan. Both unadjusted and all adjusted hazard ratios and 95% confidence intervals for the exposure to SCI were calculated.

All analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC). Statistical testing was two-tailed with a significance level of 0.05 and effect sizes used a 0.2 meaningful difference cutoff.

Results

In this study, we examined common musculoskeletal disorders among individuals after sustaining a traumatic SCI (n=9,081), as compared to adults without SCIs (n=1,474,232). The median time in the plan for eligible enrollees was 10.3 (25th Percentile: 8.0; 75th Percentile: 13.0) and 7.6 (25th Percentile: 6.0; 75th Percentile: 10.3) years for patients with SCI vs controls respectively.

Our primary findings revealed that adults living with SCI had a significantly higher 4-year incidence of any musculoskeletal morbidity (82.4% vs. 47.5%) as compared to adults without SCI, and differences were to a clinically meaningful extent. Moreover, adults with SCI had higher 4-year incidence of all musculoskeletal outcomes, including osteoarthritis (40.2% vs. 18.6%), osteoporosis (24.0% vs. 6.2%), pathologic fracture (20.2% vs. 1.4%), disorders of muscle, joint, ligaments, tendons and connective tissues (77.0% vs. 44.2%), sarcopenia and weakness (31.1% vs. 6.2%), myalgia (15.8% vs. 8.4%), and rheumatoid arthritis, myositis, and musculoskeletal infections (4.7% vs. 2.0%), as compared to adults without SCI (all P<.01 and SMD≥0.2) (Table 2).

Table 2.

Incidence of any and all musculoskeletal morbidities among adults with and without SCI with one-year clean enrollment period.

Spinal Cord Injury	† No Outcome at Baseline
	Case/Denominator	Control/Denominator
Any Musculoskeletal	2518/3057 (82.4%)*	511657/1077396 (47.5%)
Osteoarthritis	2659/6620 (40.2%)*	249195/1342834 (18.6%)
Osteoporosis	1760/7338 (24.0%)*	88662/1421643 (6.2%)
Pathologic fracture	1543/7648 (20.2%)*	20965/1469029 (1.4%)
Disorders of muscle, joint, ligaments, tendons, and connective tissues	3375/4382 (77.0%)*	514011/1163784 (44.2%)
Sarcopenia and weakness	2540/8177 (31.1%)*	90152/1456285 (6.2%)
Myalgia	1310/8311 (15.8%)*	119512/1422838 (8.4%)
Rheumatoid arthritis, myositis, and musculoskeletal infections	414/8724 (4.7%)*	29099/1455696 (2.0%)

^*P<.01 and standard mean difference (SMD) ≥0.2

^†Denominators for both cases and controls reflect a one-year clean period during their enrollment for the specific condition. For instance, among cases (SCI), there exist 6,620 patients whose first year of enrollment had no evidence of osteoarthritis; therefore, inferred incident osteoarthritis could be estimated for this subset of the full SCI cohort. As a result, all patient cohorts’ denominators dynamically change conditional on the incident outcome being measured to ensure a clean period in the first year of enrollment.

A Kaplan-Meier curve for the unadjusted disease-free survival for any musculoskeletal morbidity in adults with SCI and controls is provided in Figure 1. For the secondary analyses, unadjusted survival models demonstrated a robust hazard ratio (HR) for each of the musculoskeletal morbidities among adults with SCI, and ranged from HR: 1.92 (1.82, 2.03) for myalgia to HR: 13.28 (12.42, 14.20) for pathologic fracture (all p<0.001).

Figure 1.

Disease-free survival and Kaplan-Meier product-limit survival curves (3-year) for adults with SCI (blue) and without SCI (red), for any musculoskeletal morbidity.

Finally, fully adjusted survival models demonstrated that adults with SCI had a greater hazard for any musculoskeletal morbidity (HR: 2.41; 95%CI: 2.30-2.52) (Supplemental Table S3), and all musculoskeletal disorders, and ranged from HR: 1.26 (1.14, 1.39) for rheumatoid arthritis, myositis, and musculoskeletal infections to HR: 7.02 (6.58, 7.49) for pathologic fracture (Table 3).

Table 3.

Survival models with parametric Weibull regression was completed stepwise for each incident musculoskeletal outcome to examine the effects of incremental adjustment on the exposure variable (SCI).

	Model 1	Model 2	Model 3	Model 4
Any Musculoskeletal	2.96 (2.83, 3.09)***	2.76 (2.63, 2.88)***	2.42 (2.32, 2.54)***	2.41 (2.30, 2.52)***
Osteoarthritis	2.46 (2.37, 2.56)***	1.58 (1.52, 1.65)***	1.31 (1.26, 1.36)***	1.31 (1.26, 1.36)***
Osteoporosis	3.89 (3.69, 4.09)***	2.41 (2.29, 2.54)***	2.20 (2.09, 2.32)***	2.19 (2.08, 2.30)***
Pathological fracture	13.28 (12.42, 14.20)***	8.58 (8.04, 9.15)***	7.09 (6.65, 7.56)***	7.02 (6.58, 7.49)***
Disorders of muscle, joint, ligaments, tendons, and connective tissues	2.68 (2.58, 2.78)***	2.25 (2.17, 2.33)***	1.94 (1.87, 2.02)***	1.93 (1.86, 2.01)***
Sarcopenia and weakness	5.80 (5.56, 6.05)***	3.53 (3.39, 3.68)***	2.68 (2.57, 2.80)***	2.66 (2.55, 2.77)***
Myalgia	1.92 (1.82, 2.03)***	1.82 (1.72, 1.93)***	1.47 (1.39, 1.55)***	1.46 (1.38, 1.55)***
Rheumatoid arthritis, myositis, and musculoskeletal infections	2.30 (2.08, 2.54)***	1.62 (1.46, 1.79)***	1.25 (1.13, 1.38)***	1.26 (1.14, 1.39)***

Model 1: Unadjusted

Model 2: Model 1 + Demographic variables (age, sex, race, geographic region).

Model 3: Model 1 + Model 2 + Modified Elixhauser Comorbidity Index

Model 4: Model 1 + Model 2 + Model 3 + Education + Income

^*As with incidence estimates (Table 2), all survival models used cases (SCI) and control cohorts consistent with Table 2, which required a one-year clean period with no evidence of the condition being measured

^***P-value < 0.001

Discussion

The principal finding of this study was that adults living with traumatic SCI had a higher incidence of any and all musculoskeletal morbidities than adults without SCI. Moreover, both unadjusted and adjusted survival models revealed a significant association between traumatic SCI and hazards for developing subsequent musculoskeletal morbidities, as compared to adults without SCIs. These findings indicate a need for clinical awareness regarding the musculoskeletal disorders experienced and risks among adults with SCI, as well as enhancing clinical screening strategies, and developing efficient referral resources for coordinated care may help reduce the burden of bone, muscle, joint, ligament, tendon, connective tissue health disorders in this high risk population.

Significant compromise arising from musculoskeletal disorders in people with SCI mostly occur in the chronic phase of traumatic SCI. Due to ongoing clinical research, early identification of problems and treatment related to organ systems affected in the acute and subacute phases have improved rates of survival and life expectancy in SCI patients.^47-49 However, this has allowed the more chronic effects of SCI to manifest in various body systems. Factors that may contribute to this include change in metabolism, repetitive stress, altered and compensatory biomechanics, and the close interplay with aging and breakdown. It is particularly evident in the musculoskeletal system and its reciprocity with the rest of the organ systems and how this impacts the individual with SCI overall – physically, functionally, psychologically, etc. For instance, due to quadriplegia or paraplegia, people with SCI are at higher risk of obesity and cardiovascular disease because caloric consumption typically far exceeds caloric expenditure.^50-52 Pain related to chronic upper extremity or lower extremity conditions like tendonitis and osteoarthritis can further limit activity and function, and are expected to facilitate decline in cardiovascular health and quality of life.⁴⁸ Despite the established literature pertaining to the musculoskeletal morbidities related to aging in the non-SCI population, the extent to which these conditions arise after SCI are less well-understood. Previous studies have largely been limited to small, cross-sectional clinical or prospective studies and there is a lack of information pertaining to the natural history and trajectories of musculoskeletal morbidities among adults with SCI. For example, in a very recent small cross-sectional study using dual-energy X-ray absorptiometry, Frotzler et al.⁵³ demonstrated that compared with able-bodied women and men, individuals with chronic motor complete SCI showed considerably lower bone densitometry values and a higher historical fracture rate. Moreover, another study demonstrated that adults with chronic tetraplegia or paraplegia stemming from SCI are known to have high risk for combined low muscle mass and increased relative adiposity (i.e., “sarcopenic obesity) following injury.⁵⁴ Secondary health complications following traumatic SCI for musculoskeletal and other organ systems are extremely common and pervasive,^55-58 and thus enhanced clinical follow-up of people with SCI following discharge into the community may increase access to rehabilitation opportunities for those most in need.

Preventive measures, early diagnosis and treatment, reduction of disease progression and morbidity for these musculoskeletal conditions are imperative and cannot be overemphasized.^{19,41-44,59-61} Ultimately, the primary goal is to reduce the consequences of these musculoskeletal disorders impact on overall health, wellbeing, function and quality of life in people with SCI.^{2,5,7,41-43,62} Moreover, behavioral interventions including adapted physical activity including loading and weight bearing exercise is critical for maintaining health and integrity of muscle and bone after SCI. The 2018 Physical Activity Guidelines (PAG) for Americans provides recommendations on amount and intensity of physical activity for the general population to decrease risk for musculoskeletal morbidity. The PAG recommendations for individuals living with chronic disease or disability are similar, with suggestions to adjust to the individual’s ability.⁶³

Strengths and Limitations

This study was able to identify the high incidence of numerous musculoskeletal disorders in the SCI population based on diagnostic codes and highlights a major problem that has not been given adequate attention. Increased awareness facilitates appropriate investigation and provision of care for these conditions. Findings and outcomes of this study can be used to encourage and motivate prospective and more in-depth research in the future. Limitations of this study include the inability to analyze specific processes of musculoskeletal disorders over a period of time and inability to determine cause and effect of these disorders. Findings of this study are not guaranteed to be truly representative of the problem and judgment is required when applying generalized results to the individual level. In addition, there can be systematic differences in the sources where data were collected; there is no data available on confounding factors. Moreover, we were unable to determine the severity of SCI through claims-based data. However, we suspect that our sample may be more reflective of a healthier, higher functioning segment of the traumatic SCI population, because they had to be enrolled in private insurance, either by purchasing their own insurance, or by being covered through employment or marriage to someone who had private insurance. We also did not examine differences between upper and lower extremity MSI disorders, and did not provide incidence estimates across age categories or genders. Thus, future work is needed to explore differences in health outcomes and disparities in healthcare utilization among different age categories, genders, and across racial/ethnic minorities. Finally, we cannot rule out time-varying confounding since baseline measurements of all covariates were included in our final models. Thus, whether having SCI “causes” an elevated risk for musculoskeletal morbidity, or if changes in other health parameters (e.g., diabetes, a known predictor of bone and muscle quality) themselves, are a cause of poor musculoskeletal health, is an interesting topic. Thus, we were unable to determine if other competing risks or unmeasured confounding (i.e., other risk factors [e.g., lack of physical activity, loss of functional independence, etc.]) may have influenced the observed findings.

Conclusion

In conclusion, adults with SCI have an elevated risk of developing a variety of musculoskeletal morbidities compared to the general adult population of privately insured beneficiaries without SCI. Individuals with SCI frequently utilize healthcare services as part of their routine clinical care. Therefore, increasing clinical awareness of the musculoskeletal disorders experienced and risks among adults with SCI, improving clinical screening strategies, and developing efficient referral resources for coordinated care may help reduce the burden of bone, muscle, joint, ligament, tendon, connective tissue health disorders in this high need population.

Supplementary Material

Supplemental File 1

Supplemental Table S1. Diagnostic codes for SCI, and all musculoskeletal morbidities using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and the International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes.

Supplemental File 2

Supplemental Table S2. Modified Elixhauser Index.

Supplemental File 3

Supplemental Table S3. Fully adjusted survival model for any musculoskeletal morbidity.

Table 1.

Descriptive characteristics among adults with SCI (case) or without SCI (control).

Spinal Cord Injury	Case	Control
Overall	9,081 (100%)	1,474,232 (100%)
Full Enrollment Length
Mean (SD)	10.8 (3.4)	8.5 (3.2)
Median (Q1-Q3)	10.3 (8.0-13.0)	7.6 (6.0-10.3)
Years Post Eligibility Start Date †
Mean (SD)	6.0 (1.7)	5.3 (1.5)
Median (Q1-Q3)	5.6 (4.7-7.0)	4.7 (4.2-5.8)
Age Group
18-44	1384 (15.2%)	542106 (36.8%)
45-64	2547 (28.0%)	512676 (34.8%)
65 or Older	5150 (56.7%)	419450 (28.5%)
Gender
Female	5252 (57.8%)	774282 (52.5%)
Male	3829 (42.2%)	699950 (47.5%)
Race
Asian	263 (2.9%)	56134 (3.8%)
Black	637 (7.0%)	117545 (8.0%)
Hispanic	718 (7.9%)	129689 (8.8%)
Unknown	1758 (19.4%)	285255 (19.3%)
White	5705 (62.8%)	885609 (60.1%)
Education
<High School Diploma	58 (0.6%)	8418 (0.6%)
High School Diploma	2386 (26.3%)	354512 (24.0%)
<Bachelor Degree	4999 (55.0%)	784246 (53.2%)
Bachelor Degree	1451 (16.0%)	284927 (19.3%)
Unknown/Missing	187 (2.1%)	42129 (2.9%)
Net Worth
Unknown	1554 (17.1%)	250582 (17.0%)
<$25K	1479 (16.3%)	222644 (15.1%)
$25K-$149K	1622 (17.9%)	258445 (17.5%)
$150K-$249K	886 (9.8%)	151485 (10.3%)
$250K-$499K	1376 (15.2%)	245106 (16.6%)
≥$500K	2164 (23.8%)	345970 (23.5%)

^†All adults with SCI have their Index Date set the same as start of eligibility date (start of 2007, year when turned 18, or enrollment start date, whichever was the latest)