Greater adipose tissue distribution and diminished spinal musculoskeletal density in adults with cerebral palsy

Abstract

Objective

The purposes of this study were to examine differences in adipose tissue distribution, lumbar vertebral bone mineral density (BMD), and muscle attenuation in adults with and without cerebral palsy (CP), and to determine the associations between morphological characteristics.

Design

Cross-sectional, retrospective analyses of archived computed tomography (CT) scans.

Setting

Clinical treatment and rehabilitation center for persons with CP.

Participants

Adults with CP with a mean ± SD age of 38.8 ± 14.4 years; body mass: 61.3 ± 17.1 kg; Gross Motor Function Classification level of I-V, and a matched cohort of neuro-typical adults. Of the 41 adults with CP included in the study, 10 were not matchable due to low body masses.

Interventions

Not applicable

Main Outcome Measure(s)

Computed tomography scans were assessed for visceral and subcutaneous adipose tissue (VAT and SAT areas), psoas major area and attenuation in Hounsfield units (HU), and cortical and trabecular BMDs.

Results

Adults with CP had lower cortical (β=−63.41 HU, p<0.001) and trabecular (β=−42.24 HU, p<0.001) BMDs, as well as psoas major areas (β=−374.51 mm², p<0.001) and attenuation (β=−9.21 HU, p<0.001), after controlling for age, sex, and body mass. Adults with CP had greater VAT (β=3914.81 mm², p<0.001) and SAT (β=4615.68 mm², p<0.001). Muscle attenuation was significantly correlated with trabecular (r=0.51, p=0.002) and cortical (r=0.46, p<0.01) BMD; whereas VAT was negatively associated with cortical BMD (β=−0.037 HU/cm²; r²=0.13; p=0.03).

Conclusions

Adults with CP had lower BMDs, smaller psoas major area, greater intermuscular adipose tissue, and greater trunk adiposity than neuro-typical adults. VAT and cortical BMD were inversely associated.

Cerebral palsy (CP) is caused by a malformation or lesion to the developing brain which affects motor control centers, and causes alterations in growth, development, and overall health and function throughout the lifespan. Once established, the brain insult or structural malformation does not progress with time, but individuals with CP can develop secondary conditions which may interfere with important aspects of quality of life, such as independence, activity, and participation.¹ As a result, increasing clinical and research attention has begun to highlight the unique problems facing adolescents with CP as they transition to adulthood, as well as those specific to aging in CP. Various secondary conditions may predispose young to middle-aged adults with CP to sustain morphologic and functional decline similar to those seen in late-middle age and elderly adults without CP.² Though the underlying mechanisms of this accelerated aging is not well-established, research has demonstrated that individuals with CP have significantly lower fitness,³ and are therefore at risk for functional and cardiometabolic health declines throughout adulthood.⁴ Moreover, individuals with CP have diminished muscle size and strength⁵ and lower bone mineral density (BMD)⁶-collectively highlighting the risk for frailty in this population.

Due to substantial deficits in overall lean tissue,^7-10 normal BMIs in CP may disguise excess adiposity in visceral or other ectopic adipose depots (e.g., liver, muscle, bone marrow, etc.).¹¹ There is some evidence to suggest greater general adiposity in children with CP;¹² however, this has yet to be well-studied in the adult CP population. In 2009, Johnson and colleagues¹³ provided some of the first evidence that young children with quadriplegic CP had significantly greater intermuscular adiposity than matched, typical neuro-developing children, and that the extent of infiltration was robustly associated with objectively measured physical inactivity. Early declines in function among individuals with CP may occur as a result of accelerated-aging; however, it is equally plausible that the hallmark neuropathic symptoms (e.g., chronic spasticity, fatigue, antagonist co-activation, low muscle tone, etc.) merely increase the likelihood of exaggerated sedentary behavior from early childhood.^2,14,15

Along with the neuromuscular deficits and altered morphology in persons with CP, the exaggerated sedentary lifestyles often found in this population have prompted a comparison model of disability to persons with spinal cord injury¹⁶-a population with significant muscle wasting, increased adipose tissue deposition, and elevated risk for cardiometabolic abnormalities.¹⁷ Although a plausible comparison, virtually no research has addressed the long-term cardiometabolic consequences of CP, or the interrelationships between adipose tissue partitioning, altered muscle composition and atrophy, and BMD in this population. Therefore, the purposes of this study were to examine the differences in trunk adipose tissue distribution (i.e., subcutaneous and visceral adipose tissues (SAT and VAT)), lumbar trabecular and cortical BMDs, and muscle attenuation (i.e., an indicator of intermuscular adipose tissue) in adults with and without CP as well as between ambulatory and non-ambulatory persons with CP, and to determine the associations between these morphological characteristics after adjusting for age, sex, and body mass.

Method

Participants

Abdominal and thoracic computed tomography (CT) scans and patient records were obtained from a convenience sample of 41 adults (aged 18-65 years) with CP (age: 38.8 ± 14.4 years; body mass: 61.3 ± 17.1 kg), and a cohort of 311 neuro-typical available clinical patients, matched for sex, age, and body mass. All CT scans took place in the same University Health System, between the years 2007-2013. Of the 41 subjects with CP, 10 subjects were not matchable due to very low body mass, and thus were included only as a comparison subset to the main CP cohort. The Gross Motor Function Classification System (GMFCS) was recorded for each patient with CP, and ranged from level I-V. GMFCS assesses mobility status with a five level ordinal grading scale. Specifically, the GMFCS is used clinically to describe mobility status of individuals with CP, on the basis of self-initiated movement and with emphasis on sitting, walking, and wheeled mobility. Distinctions between levels are also based on the need for assistive technology, including hand-held mobility devices (walkers, crutches, etc.), and/or wheeled mobility. Individuals at “Level I” can generally walk without significant restrictions, but may experience limitations in advanced motor-related skills. Conversely, individuals at “Level V” are usually very restricted in the ability to function, even with external assistive technology. Studies of the GMFCS to classify mobility status in the adult population with CP show it to be reliable (interclass correlation coefficient (ICC) = 0.93) with excellent interrater reliability (quadratic kappa value of 0.978).¹⁸ All patients were examined by the same physician investigator to classify GMFCS level, which were obtained during chart reviews, and the breakdown for the different levels was as follows: GMFCS (I) n=4; (II) n=5; (III) n=9; (IV) n=12; and (V) n=11. The study was approved by the University of Michigan Institutional Review Board.

Anatomical and morphomic data

Patient CT scans were processed and analyzed using a proprietary, semi-automated procedure developed using MATLAB® software (MathWorks Inc., Natick, MA), as previously described.^19,20 Trained CT processors were required to confirm the locations of vertebral levels, identify key anatomic landmarks for semi-automated visceral fascial determination, and draw contours of psoas major muscles at the L4 level. The algorithm applied these inputs to automatically generate anatomic and morphomic measurements. Trabecular bone density was measured as the mean density within a circle that was half the size of the vertebral body. Cortical bone density was measured as the mean of the level of half-max of the bone signal peak at every angle within a 60 degree wedge. Psoas major muscle attenuation was measured as the mean of voxel attenuation in Hounsfield units (HU) within the psoas contours. Psoas major area was measured as the total area inside the psoas contours, and lean psoas as area multiplied by normalized psoas density (mapping (−85,85) to (0,1)). VAT area was measured as the total area inside the abdominal fascia meeting fat density thresholds, and SAT area as the total area between abdominal fascia and skin meeting fat density thresholds.

Statistical Analysis

All statistical analyses were conducted using R statistical package (R Foundation for Statistical Computing, Vienna, Austria), and SAS 9.3 (SAS Institute, Cary, NC). Anatomic and morphomic characteristics were examined between adults with CP and matched, neuro-typical adults, and are provided as means and standard deviations. Unadjusted differences between groups were tested using two independent samples t-test. Adjusted differences in each characteristic were tested using multiple linear regression after creating a dummy variable for group, and controlling for age, sex, and body mass. Among the adults with CP, we also examined correlation coefficients to assess the association between bone/muscle densities and age, body mass, VAT area, and SAT areas. Due to the concern of potential outliers in measuring densities, we report Spearman’s rank correlation coefficients (ρ), which is a nonparametric measure of statistical dependence. In using this method, we were able to also examine both linear and non-linear associations. Moreover, among only the adults with CP, we used simple linear regression to examine differences between all characteristics in the 10 unmatched versus 31 matched adults. Normality of the residuals was tested using a Shapiro-Wilks test and homogeneity of the variance of the residuals was tested with standard regression diagnostics. A minimum criterion alpha level of p ≤ 0.05 was used to determine statistical significance.

Results

Unadjusted anatomic and morphomic data are presented in Table 1, as means, standard deviations, t-statistics, and p-values for adults with and without CP. Adults with CP had significantly lower BMDs, smaller psoas major muscle areas and lower muscle attenuation, but larger VAT areas.

Table 1. Summary of demographics and morphomics at L4 by group.

	Cerebral Palsy (Mean±SD) N=31	Matched Controls (Mean±SD) N=311	t-Statistics (p-Values)
Age, year	39.41±15.71	34.31±14.55	−2.10 (0.04)
Weight, kg	66.55±14.51	69.67±12.67	1.46 (0.15)
Cortical BMD, HU	282.8±81.37	340.7±73.5	4.33 (<0.01)
Trabecular BMD, HU	187.3±63.18	231.7±57.4	4.25 (<0.01)
SAT area, mm2	19948.4±14080.8	19987.8±11689.4	0.02 (0.99)
VAT area, mm2	9127.1±8109.1	6313.9±5376.6	−2.51 (0.01)
Psoas area, mm2	2145.8±989.2	2470.9±812.7	2.11 (0.03)
Psoas density, Hu	54.54±13.48	61.47±8.41	2.91 (0.01)
Lean psoas, Hu*mm2	889.2±449.1	1064.9±362.3	2.55 (0.01)

After controlling for age, sex, and body mass, adults with CP had significantly lower cortical (β=−63.41 HU, p<0.001) and trabecular (β=−42.24 HU, p<0.001) BMDs, as well as smaller psoas major areas (β=−374.51 mm², p<0.001) and lower attenuation (β=−9.21 HU, p<0.001), as compared to matched adults without CP (Table 1). Moreover, adults with CP had significantly greater adiposity in both the visceral (β=3914.81 mm², p<0.001) and subcutaneous (β=4615.68 mm², p<0.001) depots (see Figure 1).

Figure 1.

Computed tomography image at vertebral level L4, depicting trunk adiposity distribution and muscle size in: (a) a 53 year old, neuro-typical male (65 kg body mass), and (b) a 54 year old male with CP (66 kg body mass).

Compared to the 31 matched subjects with CP, the 10 unmatched comparison-subset of adults with CP had significantly lower body mass (45.49±8.33 kg vs. 66.34±16.05 kg; p<0.001) and smaller psoas major area (1359.58±600.31 mm² vs. 2049.51±1074.84 mm²; p=0.03), but less visceral (5035.45±2888.77 mm² vs. 10047.45±8536.60 mm²; p=0.01) and subcutaneous adiposity (12796.58±7639.65 mm² vs. 21344.50±14929.41 mm² p=0.04). Among all adults with CP, muscle attenuation was significantly correlated with trabecular (r=0.51, p=0.002) and cortical (r=0.46, p=0.006) BMD (Figure 2); whereas, visceral adipose tissue was negatively associated to cortical BMD (β=-0.037 HU/cm²; r²=0.13; p=0.03). Adults with GMFCS IV-V had lower BMDs, smaller muscle areas, and lower muscle attenuation as compared to GMFCS I-III; however, these differences were not significant (Table 2; all p>0.05).

Figure 2.

Scatter plot depicting the correlation between muscle attenuation and bone mineral density among adults with cerebral palsy.

Table 2. Summary of demographics and morphomics in adults with CP, by grouped GMFCS.

	GMFCS I-III (Mean±SD) N=18	GMFCS IV-V (Mean±SD) N=23	Diff (95% CL)
Age, year	36.43±15.80	40.32±13.48	3.88 (−5.5-13.2)
Weight, kg	70.50±17.77	55.36±13.92	−15.14 (−25.2-5.1)*
Cortical BMD, HU	289.68±80.84	277.26±83.46	−12.41 (−68.5-43.7)
Trabecular BMD, HU	199.66±63.94	177.33±62.34	−22.33 (−65.3-20.7)
SAT area, mm2	23659.27±17013.05	16856.01±10609.90	−6803.3 (−16694.7-3088.2)
VAT area, mm2	10522.31±8850.61	7964.43±7491.30	−2557.9 (−8357.1-3241.4)
Psoas area, mm2	2249.48±1022.16	2053.56±979.08	−195.9 (−895.4-503.6)
Psoas density, Hu	57.87±10.84	51.58±15.15	−6.29 (−15.6-3.0)
Lean psoas, Hu*mm2	945.71±447.97	838.97±456.96	−106.70 (−423.4-209.9)

^*p<0.05

Discussion

The primary findings of this investigation are that adults with CP have significantly lower trabecular and cortical BMDs, increased muscle attenuation, and greater visceral and subcutaneous adipose tissue areas, as compared to matched-, neuro-typical adults. Each of these differences was significant, even after controlling for age, sex, and body mass. This is an extremely important finding as it provides some of the first evidence regarding differences in trunk adipose tissue partitioning in adults with CP, as well as the potential concomitant risk of premature aging and musculoskeletal fragility in this population. It is well known that children with moderate to severe CP are chronically immobile,²¹ have significant feeding issues, and are at greater risk for malnutrition, which are collectively thought to be the drivers of osteoporosis, musculoskeletal deformation and stunted growth.^22,23 Traditional clinical recommendations to quell these prognoses have included calorie-dense nutritional regimens, such as oral nutritional support or enteral tube feeding, with the hope of preserving body mass and growth trajectories, as well as to prolong life expectancy. While these recommendations are thought to be of particular relevance for children with oropharyngeal dysfunction, what remains to be fully understood are the health implications of such interventions through the lifespan.

The current findings highlight a critical concern pertaining to the disproportionate ratio of trunk adipose tissue to muscle in adults with CP, as compared to neuro-typical adults matched for age, sex and body mass. Interestingly, among the adults with CP, muscle attenuation was significantly correlated with trabecular and cortical BMD; whereas VAT was negatively associated to cortical BMD. These findings are in agreement with another recent study which demonstrated, at the tissue level, that neuro-typical women with greater central adiposity had inferior bone quality and stiffness and markedly lower bone formation.²⁴ Likewise, we have recently shown that increased VAT is robustly associated with lower BMDs and muscle attenuation in young and middle-aged adults without CP, even after adjusting for age, sex, and BMI.²⁰ Thus, in combination with other research to indicate that obesity may be associated with an increased risk of fracture ^25-27, these collective findings support the need for additional research to identify the mechanisms linking greater adiposity to adverse effects on bone density and muscular fragility. Indeed, the accumulation of lipids in visceral and ectopic depots is a well-established, lipotoxic process that occurs as a result of disequilibrium between energy intake and incomplete lipid oxidation, and is associated with a host of cardiometabolic abnormalities.^28-30 However, mounting evidence has demonstrated the equally-pathophysiologic consequences of abnormal lipid partitioning on bone and skeletal muscle integrity and function.^31-34

Sedentary aging adults are at increased risk for weakness and sarcopenic obesity,³⁵ which are in-turn the primary drivers of musculoskeletal fragility,^25,36 cardiometabolic abnormalities³⁷ and early all-cause mortality.³⁸ Thus, considering the highly sedentary lifestyles of most patients with CP,^14,15,39 these current findings provide strong support that this population is at increased risk for accelerated morphologic aging and cardiometabolic abnormalities. Perhaps just as importantly, whereas standard BMI cutoffs are known to have excellent specificity to screen for adiposity in adults with functional or mobility impairments, it lacks the sensitivity to detect excessive adiposity in non-obese persons.⁴⁰ Therefore, since body mass is still used in the majority of clinical settings to categorize growth and “nutritional status” in persons with CP, it is quite plausible that many patients classified as normal weight or even underweight, could have much greater visceral adiposity than their neuro-typical counterparts. In our cohort of 41 patients with CP, the average body masses were approximately 62kg and 61kg for women and men respectively. Due to considerable orthopedic and pathologic factors that prevent accurate height measurement in CP (e.g. scoliosis, joint contractures, spasticity, etc.), we were not able to calculate and report BMIs. Despite this limitation, our findings lend support for the notion that body mass (and thus BMI) is not a sufficient indicator of adiposity in the CP population, as there were significantly greater visceral, subcutaneous, and intermuscular adipose tissue stores than neuro-typical adults matched for body mass. This is congruent with our previous work,⁴¹ and that of others⁴² which demonstrates that indicators of central adiposity (e.g., waist circumference or WHR) are potentially more sensitive than BMI or body mass for risk stratification of cardiometabolic health in adults with CP. Clinical screening for risk must be altered in the population with CP to include direct measures or proxy indicators of trunk adiposity, as well as clinically-relevant serum indicators of musculoskeletal and cardiometabolic health.

Limitations

An additional limitation of this investigation is the use of a retrospective, cross-sectional design from existing imaging and patient clinical records. Despite our attempts to adjust for the most important covariates, we were not able to control the selection of potential model predictors, and/or distinguish an actual causal link between exposures and outcomes. For example, physical activity is a known, robust predictor of musculoskeletal health in adults.⁴³ Thus, whether a mechanism inherent to CP “causes” an increased risk for frailty and accumulation of adiposity, or if chronic physical inactivity, itself, is merely the driver of progressive physical dysfunction, musculoskeletal deterioration, and increased adiposity deposition in this population, is an intriguing and complex topic. Although degree of mobility impairment is typically considered the strongest predictor of onset and severity of secondary conditions among persons with CP,⁴⁴ we found no statistical differences in BMD, muscle size and attenuation, or adiposity between patients with GMFCS I-III versus those with IV-Vs. Moreover, it is possible that the indication for the scan, itself, may have confounded or exaggerated the results, and thus, future prospective research is needed to highlight the longitudinal alterations in muscle, bone and adipose tissue distribution in a larger, more heterogeneous sample of adults with CP, as well as to better understand the role of modifiable risk factors such as physical activity and healthy nutrition to attenuate health risk.

Conclusions

Adults with CP have significantly lower cortical and trabecular BMDs, smaller psoas major areas, greater intermuscular adipose tissue, and larger visceral and subcutaneous adiposity areas than matched, neuro-typical adults. Moreover, muscle attenuation and BMD was positively correlated in CP, and visceral adiposity and cortical BMD were inversely associated. Screening for frailty and cardiometabolic abnormalities should thus include direct measures of muscle and bone quality, as well as clinical cardiometabolic health in CP. Future studies are also needed to identify early preventive and treatment options to simultaneously target musculoskeletal preservation and healthy body composition among persons with CP.

Abbreviations

CP

cerebral palsy

BMI

body mass index

SAT

subcutaneous adipose tissue

VAT

visceral adipose tissue

BMD

bone mineral density

CT

computed tomography

GMFCS

Gross Motor Function Classification System

HU

Hounsfield units