Classification of obesity, cardiometabolic risk, and metabolic syndrome in adults with spinal cord injury

Abstract

Objective: To describe and compare (1) classification of obesity using clinical proxies of body composition that are easily accessible in the outpatient clinic setting, (2) cardiometabolic risk using existing screening tools and staging systems, and (3) the presence of metabolic syndrome (MetS) using four commonly-used definitions in adults with spinal cord injury (SCI).

Design: Retrospective chart review

Setting: Outpatient Veterans Affairs (VA) SCI Annual Evaluation Clinic

Participants: Patients who attended an annual evaluation appointment with demographic, anthropometric, and biochemical data documented in their medical records as part of routine medical care.

Outcome measures: Obesity classification (body mass index, waist circumference, ideal body weight percentage), cardiometabolic risk scores (Framingham Risk Score, Cardiometabolic Disease Staging System, Edmonton Obesity Staging System), and MetS classification (using four commonly-used definitions) were described and compared.

Results: Of the 155 veterans included in this analysis, 93% were considered “at risk” by at least one of the measurements studied. However, there was considerable variation between the different screening tools. The κ-agreement between various definitions of MetS ranged from fair to moderate.

Conclusion: Screening tools that were developed for the non-SCI population produced variable assessments of risk when applied to veterans with SCI. Due to the fair to moderate inter-rater agreement between MetS definitions, it is unknown which definition is superior to identify MetS in the SCI population. An SCI-specific screening tool is needed to accurately classify obesity, cardiometabolic risk, and MetS in order to provide timely education and intervention.

Introduction

Current literature suggests an increased prevalence of obesity, cardiometabolic risk, and metabolic syndrome (MetS) among individuals with spinal cord injuries (SCI), potentially due to decreased physical activity, lower metabolism, and decreased muscle mass.^1–8 However, evidence-based screening tools designed specifically for the SCI population to identify and classify those at risk currently do not exist.^9,10

Obesity, defined by excess body fat accumulation, is associated with diabetes, all-cause mortality, and cardiovascular disease (CVD)-mortality.^11,12 Identification of obesity can lead to appropriate interventions and improved health outcomes.¹³ Methods used to assess or estimate body fat percentage include air displacement plethysmography, bioelectrical impedance analysis (BIA), and imaging methods such as computerized tomography (CT), magnetic resonance imaging (MRI), and dual-energy X-ray absorptiometry (DXA).¹⁰ Although many of these methods are used in SCI-related research, they are often not available in an outpatient clinical setting.^3,10,14–20

When accurate body composition measurement methods are not accessible in the outpatient clinical setting, clinical proxies of body composition are often used to classify obesity. Body mass index (BMI) is an anthropometric measure of weight adjusted for height (kg/m²) often used to classify obesity.²¹ However, since BMI underestimates adiposity in individuals with SCI, it is not sensitive for detecting obesity status and is a poor predictor of CVD in this population.^{1,3,19,22–25} Adjusted BMI classification schemes have been proposed for the SCI population to account for the changes in body composition after injury; however, they have not yet been validated in a large population.^10,24,26 Similarly, it has been proposed that waist circumference (WC) may be an indicator of obesity-related comorbidities, but there are currently no SCI-specific cut-off points that have been validated.^27,28 In place of BMI, it has been suggested to calculate and adjust ideal body weight (IBW) for individuals with SCI, but these approaches are dated, not validated, and are not accompanied by a classification scheme.^29,30 A standard classification of obesity in the SCI population remains unknown.¹⁰

In addition to solely relying on obesity classification, risk clustering has been used to determine cardiometabolic or CVD risk.^23,30–34 Based on the concept of risk clustering, several screening tools and staging systems have been developed to assist providers with clinical decision making using easily-accessible data. Among these screening tools are the Framingham Coronary Heart Disease Risk Score (FRS),^35,36 Cardiometabolic Disease Staging System (CMDS),^37,38 and Edmonton Obesity Staging System (EOSS).^39,40

The FRS is one of the most widely-used CVD risk prediction scores.⁴¹ It involves a sex-specific scoring system to identify 10-year CVD risk.³⁵ Although it has been applied in the SCI population,^28,42 it has not been validated in a large cohort in this population. The CMDS was developed to evaluate the stage and severity of the cardiometabolic disease to predict the risk of diabetes, CVD mortality, and all-cause mortality. This 5-stage system was validated using two large national cohorts of the non-SCI population,³⁸ but to our knowledge has not been studied or validated in the SCI population. The EOSS is a 5-stage system that incorporates morbidity and functional limitations associated with excess weight to predict all-cause mortality.⁴⁰ While the EOSS has been studied in the non-SCI population,^43–47 it has not been studied nor validated in the SCI population, to our knowledge.

Additionally, MetS classification is used to define the risk of CVD and diabetes by identifying a cluster of traits.¹¹ Common algorithms for classifying MetS have been established by the World Health Organization (WHO);⁴⁸ National Heart, Lung, and Blood Institute and American Heart Association (NHLBI/AHA);⁴⁹ National Cholesterol Education Program (NCEP);⁵⁰ and the International Diabetes Federation (IDF).⁵¹ Previous studies investigating the prevalence of MetS in SCI populations have yielded equivocal results and have demonstrated the limitations of the existing definitions due to the physiologic changes associated with SCI.^8,9,52–56

These tools and classification schemes were developed in the non-injured population, and it is unknown whether they can be extrapolated to consistently assess risk among those with SCI. Therefore, the objectives of this study were to describe and compare (1) classification of obesity using clinical proxies of body composition that are easily accessible in the outpatient clinic setting, (2) cardiometabolic risk using existing screening tools and staging systems, and (3) the presence of MetS using four commonly-used definitions.

Methods

A retrospective chart review was conducted of veterans who attended an annual evaluation appointment in the outpatient clinic of the Spinal Cord Injuries/Disorders unit at the Edward Hines, Jr. Veterans Affairs (VA) Hospital in Hines, Illinois from January 1, 2017 through December 31, 2017. Inclusion criteria included age 18–89 years and a diagnosis of an SCI. Patients with a diagnosis of multiple sclerosis (MS), amyotrophic lateral sclerosis (ALS), or a non-SCI spinal cord disorder were excluded. Pregnant females, patients with limb amputations, and patients who had experienced the SCI ≤ 6 weeks prior to the appointment were excluded due to expected changes in body composition in addition to SCI-related changes.

Demographic, anthropometric, and biochemical data were obtained and documented in the computerized patient record system (CPRS) by trained nursing staff during a single visit as part of routine medical care for an annual evaluation. As part of standard care in the clinic, patients were instructed to fast before the appointment, standard venipuncture procedures were followed, and blood was analyzed in a CLIA-certified laboratory. As part of routine care, WC was obtained by trained clinic staff using a flexible measuring tape while the patient was either standing, sitting, or in the supine position, depending on the patient’s level of injury and practicality of positioning. The patient records were retrospectively reviewed by a single author in January 2018 after the appointments were completed.

As described below, these data were then used to determine obesity classification, cardiometabolic risk scores, MetS classification, and additional descriptive frequencies.

Obesity classification

BMI was calculated for each patient. The prevalence of obesity was determined using the cut-off point of ≥30 kg/m² according to the WHO ²¹ definition, as well as proposed SCI-specific cut-off points of ≥25.3 kg/m² per Ayas et al.²⁶ and >22 kg/m² per Laughton et al.²⁴

The WC cut-off point of ≥102 cm²⁷ and the SCI-specific cut-off of ≥ 94 cm per Ravensbergen et al.²⁸ were used as proxy indicators of excess visceral adiposity and risk for obesity-related comorbidities.

IBW was calculated for each patient using the Hamwi^57,58 method based on height in inches and weight in pounds. For men, Hamwi IBW is calculated as 106 pounds plus 6 pounds for every inch over 60 in. of height. For example, if a patient was 67 in. tall, the following equation would be used to calculate IBW:

IBW = 106 + [6 × (67 – 60)] = 148 pounds

Using recommendations specific to the SCI population from the Academy of Nutrition and Dietetics Evidence Analysis Library,²⁹ this estimation was further adjusted by subtracting 10–15% of body weight or 15–20 pounds for individuals with tetraplegia or by subtracting 5–10% of body weight or 10–15 pounds for individuals with paraplegia. For example, if a veteran with paraplegia was 67 in. tall and had a calculated IBW of 148 pounds per the Hamwi method, two additional adjusted IBW measurements were calculated:

IBW adjusted by percentage low range: 148 − (148 × 0.1) = 133.2 pounds

IBW adjusted by percentage high range: 148 − (148 × 0.05) = 140.6 pounds

IBW adjusted by pounds low range: 148 − 15 = 133 pounds

IBW adjusted by pounds high range: 148 − 10 = 138 pounds

This individual would have an adjusted IBW range of 133.2–140.6 pounds (adjusted by percentage of body weight) or 133–138 pounds (adjusted by pounds).

Each individual’s percent IBW (%IBW) was calculated for each estimation by dividing actual body weight by calculated IBW.

Cardiometabolic risk scores

Demographic, anthropometric, and biochemical data were entered into existing screening tools (FRS,³⁵ CMDS,³⁸ EOSS⁴⁰) to determine cardiometabolic risk. A FRS >20% was considered high risk of developing CVD. A CMDS or EOSS stage >0 was considered at risk. Full criteria for each of the screening tools can be found in Supplemental Materials.

Metabolic syndrome classification

MetS was calculated using criteria as defined by IDF,⁵¹ NCEP,⁵⁰ NHLBI/AHA,⁴⁹ and WHO⁴⁸ (Supplemental Materials).

Additional descriptive frequencies

As part of routine care in the outpatient annual evaluation clinic, nursing staff documented and asked each veteran, “How would you rate your overall health?” Each veteran was given the option to reply with “excellent,” “good,” “fair,” or “poor.” These responses were collected in this chart review. Medications were also recorded from the active medications section of the electronic health record. Medications were categorized as “yes” if there was the presence of at least one active medication in a given category or “no” if the medication was not listed as active for each of the following categories: hypertensive medications, diabetes medications, or lipid medications.

Statistical analysis

Descriptive statistics were used to summarize demographic, anthropometric, and biochemical data. To compare the agreement between MetS definitions, inter-rater reliability was measured for each combination of the four MetS definitions using Kappa (κ) values and categorized using divisions published by Landis et al. (κ < 0.00: poor, 0.00–0.20: slight, 0.21–0.40: fair, 0.41–0.60: moderate, 0.61–0.80: substantial, 0.81–1.00: almost perfect).⁵⁹ Simple linear regression was used to determine the difference between the adjusted IBW calculated by subtracting pounds versus percentage of body weight. Visualization method was used to determine the difference of adjustment in the category of SCI by examining regression plots graphically. Analyses were performed using IBM® SPSS® Statistics Version 25. All tests were two-tailed, and P < 0.05 was considered statistically significant.

Results

Table 1 displays each of the screening tools and measurements used in this study, along with the strengths and limitations of each. Of the 197 charts screened, 167 met inclusion criteria. Three additional patients were excluded due to insufficient data available in the chart. Due to the small number of females (n = 9), the authors decided to only include males in the final analyses. Therefore, a total of 155 patients were included for final analyses. Demographic characteristics of the included patients are displayed in Table 2. Anthropometric and biochemical data were obtained for all patients (n = 155), with the exception of WC (n = 74) and glycated hemoglobin (n = 59). Additional anthropometric and biochemical data are shown in Table 3. A total of 144 out of the 155 veterans (93%) were considered “at risk” or obese by at least one of the measurements or screening tools studied.

Table 1. Strengths and limitations of existing obesity, cardiometabolic risk, and metabolic syndrome classification tools.

Screening tool or classification measurement	Strengths or rationale	Limitations specific to SCI population
Obesity Classification
BMI ≥ 30 (WHO)21	Easily accessible in the clinic Commonly used	Not sensitive to changes in body composition after SCI
Adjusted BMI ≥ 25.3 (Ayas)10,26	Easily accessible in the clinic Adjusted for change in body composition after SCI	Not validated in large SCI population Not specific to level of injury (tetraplegia vs. paraplegia) Not specific to completeness of injury
Adjusted BMI > 22 (Laughton)24	Easily accessible in the clinic Adjusted for change in body composition after SCI Validated in small SCI population	Not validated in large SCI population Validated with CRP for CVD marker which may be affected by other factors
IBW Equation (Hamwi)57,58	Simple calculation	Not adjusted based on level of injury No existing classification scheme May not accurately predict BMI or risk
Adjusted IBW29,30	Simple calculation Can be individualized based on level of injury Allows for clinical judgment No significant difference between adjusting by pounds or percentage	Not validated Limited recent supporting data available No existing classification scheme
Waist Circumference (WC) 27 > 40ʺ or 102 cm (men) WC > 35ʺ or 88 cm (women)	May be more specific to risk in the SCI population compared to BMI Does not require height and weight which may be difficult to obtain	May not be practical to routinely and accurately obtain in the clinic Differences between sitting vs. standing vs. supine are unknown in the SCI population SCI-specific classification are not validated in a large population The effect of abdominal muscle atrophy is unknown Not specific to level of injury (tetraplegia vs. paraplegia) Not specific to completeness of injury
Adjusted WC ≥ 94 cm (Ravensbergen)28	May be more specific to risk in the SCI population compared to BMI Does not require height and weight which may be difficult to obtain Validated in small SCI population	May not be practical to routinely and accurately obtain in the clinic Differences between sitting vs. standing vs. supine are unknown in the SCI population SCI-specific classification are not validated in a large population The effect of abdominal muscle atrophy is unknown Not specific to level of injury (tetraplegia vs. paraplegia) Not specific to completeness of injury
Metabolic Syndrome (MetS) Classification
International Diabetes Federation (IDF)51	Comprehensive approach Most components are easily available and can be a quick screening tool Addresses HTN medications as a potential risk factor Addresses diagnosis of diabetes as a potential risk factor	WC cut-off not validated in large SCI population BMI is not sensitive in SCI BMI >30 does not capture WC ≥94cm POC blood pressure may not accurately reflect risk due to BP variations after SCI Only fair to moderate inter-rater agreement with other MetS definitions Does not incorporate diabetes medications as a risk factor
National Cholesterol Education Program (NCEP)50	Comprehensive approach Most components are easily available and can be a quick screening tool	WC Cut-off not validated in large SCI population POC blood pressure may not accurately reflect risk due to BP variations after SCI Only fair to moderate inter-rater agreement with other MetS definitions Does not incorporate medications as risk factors Does not address diagnosis of diabetes or hypertension as a potential risk factor
National Heart, Lung, and Blood Institute and American Heart Association (NHLBI/AHA)49	Comprehensive approach Most components are easily available and can be a quick screening tool Addresses medications as a potential risk factor	WC Cut-off not validated in large SCI population POC blood pressure may not accurately reflect risk due to BP variations after SCI Only fair to moderate inter-rater agreement with other MetS definitions Does not address diagnosis of diabetes or hypertension as a potential risk factor
World Health Organization (WHO)48	Comprehensive approach Most components are easily available and can be a quick screening tool Addresses medications as a potential risk factor	BMI is not sensitive in SCI POC blood pressure may not accurately reflect risk due to BP variations after SCI Only fair to moderate inter-rater agreement with other MetS definitions
Cardiometabolic Risk Screening Tools
Framingham Risk Score (FRS)35	Comprehensive approach Most components are easily available and can be a quick screening tool Addresses blood pressure medications as a potential risk factor Indicates range of risk instead of a dichotomous classification	Developed using non-diverse sample Risk may be difficult to translate Time consuming to calculate POC blood pressure may not accurately reflect risk due to BP variations after SCI
Cardiometabolic Disease Staging System (CMDS)38	Comprehensive approach Most components are easily available and can be a quick screening tool Addresses medications as a potential risk factor Indicates range of risk instead of a dichotomous classification	Some components may not be easily available in the clinic (e.g. glucose tolerance test) Relatively time consuming Waist Circumference cut-off not validated in large SCI population POC blood pressure may not accurately reflect risk due to BP variations after SCI Only fair to moderate inter-rater agreement with other MetS definitions
Edmonton Obesity Staging System (EOSS)40	Comprehensive approach Indicates range of risk instead of a dichotomous classification Addresses psychological well-being	Requires initial classification of obesity which is poorly defined in SCI Some components may not be easily available in the clinic (e.g. glucose tolerance test) Relatively time consuming Contains subjective judgements (e.g. well-being) that may vary between providers Physical limitations are used as a risk factor; not applicable in SCI population POC blood pressure may not accurately reflect risk due to BP variations after SCI
SCI-Specific Screening Tool Needed
Proposed SCI-Specific Screening Tool	Combining the strengths of the aforementioned screening tools, an SCI-specific tool should be/include: Sensitive to body composition changes after SCI Evidence-based and validated Validated classification scheme Easy to translate and communicate with patients Diagnoses and medications as a risk factor, even if measures are WNL Quick and easy screening tool with easily accessible measures Individualization and allow for clinical judgment Comprehensive (more than just a single value) Different levels of risk (not dichotomous outcome)	After evaluating the limitations of the aforementioned screening tools, an SCI-specific tool should take into account: Level of injury Completeness of injury WC: cut-offs adjusted for SCI BMI: cut-offs adjusted for SCI BP: use caution when interpreting in a screening due to neurological changes after SCI Blood Glucose: use caution when interpreting in a screening due to glucose tolerance correlations after SCI Psychological well-being Physical limitations/physical activity: use caution if including this in a screening tool due to changes after SCI

BMI, body mass index; BP, blood pressure; CRP, C-Reactive Protein; CVD, Cardiovascular Disease; DBP, diastolic blood pressure; FBG, fasting blood glucose; HDL, high-density lipoprotein; IBW, ideal body weight; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; SBP, systolic blood pressure; SCI, spinal cord injury; T2DM, type 2 diabetes mellitus; TG, triglycerides; WC, waist circumference; WHO, World Health Organization; WNL, within normal limits.

Table 2. Demographic and injury characteristics.

	Total (n = 155)
Age (years), mean ± SD (min–max)	61.1 ± 13.7 (24–85)
Race/ethnicity, n (%)
White	88 (56.8%)
Black/African American	53 (34.2%)
Asian	2 (1.3%)
Pacific Islander or Native Hawaiian	3 (1.9%)
Other or unknown	9 (5.7%)
Type of injury and AIS score, n (%)
Paraplegia	60 (38.7%)
ASIA A, n	29
ASIA B, n	5
ASIA C, n	8
ASIA D, n	16
ASIA E, n	2
Tetraplegia	92 (59.4%)
ASIA A, n	16
ASIA B, n	7
ASIA C, n	9
ASIA D, n	59
ASIA E, n	1
Unknown	3 (1.9%)
Age at Injury (years), mean ± SD (min–max)	42.5 ± 18.3 (17–79)
Years since Injury, mean ± SD (min–max)	18.7 ± 14.3 (0–59)

AIS, American Spinal Injury Association (ASIA) Impairment Scale.

Table 3. Anthropometric and biochemical characteristics.

	Total (n = 155)	Paraplegia		Tetraplegia
		Incomplete (n = 30)	Complete (n = 30)	Incomplete (n = 75)	Complete (n = 17)
Height (cm), Mean (SD)	178.0 (10.4)	178.5 (7.0)	180.3 (5.7)	176.8 (13.2)	179.4 (6.3)
Weight (kg), Mean (SD)	90.1 (24.0)	92.5 (27.1)	89.4 (18.0)	91.3 (25.6)	81.7 (22.4)
BMI (kg/m2), Mean (SD)	28.3 (7.1)	29.0 (8.1)	27.9 (5.8)	28.8 (7.5)	25.3 (5.9)
WC (cm), Mean (SD)	111.7 (18.8) (n = 74)	110.0 (22.6) (n = 17)	109.7 (17.0) (n = 14)	112.9 (18.1) (n = 39)	115.3 (18.8) (n = 4)
WC method used (n, %)
Sitting	28 (18.1%)	7 (23.3%)	7 (23.3%)	11 (14.7%)	3 (17.6%)
Standing	15 (9.7%)	4 (13.3%)	0	11 (14.7%)	0
Supine	12 (7.7%)	3 (10.0%)	3 (10.0%)	5 (6.7%)	1 (5.9%)
Unknown	100 (64.5%)	16 (53.3%)	20 (66.7%)	48 (64%)	13 (76.5%)
SBP (mmHg), Mean (SD)	121.3 (15.7)	123.8 (15.8)	120.0 (13.7)	122.6 (16.4)	113.2 (14.9)
DBP (mmHg), Mean (SD)	68.3 (9.4)	70.1 (68.0)	67.3 (8.4)	66.1(9.2)	67.3 (11.4)
TG (mg/dL), Mean (SD)	132.5 (82.8)	140.4 (80.5)	149.1 (94.4)	114.7 (62.2)	154.9 (118.6)
HDL (mg/dL), Mean (SD)	47.9 (15.1)	50.2 (16.2)	43.5 (11.9)	50.0 (16.3)	43.1 (11.8)
LDL (mg/dL), Mean (SD)	79.1 (32.0)	86.4 (41.2)	86.3 (29.3)	70.0 (27.0)	90.2 (30.3)
TChol (mg/dL), Mean (SD)	152.0 (35.2)	163.9 (41.1)	158.5 (31.1)	141.7 (30.9)	162.1 (40.0)
FBG (mg/dL), Mean (SD)	102.9 (31.2)	105.7 (29.4)	95.6 (24.1)	104.4 (31.6)	100.1 (38.5)
A1C (%), Mean (SD)	6.4 (1.0) (n = 59)	6.6 (1.0) (n = 10)	6.1 (1.4) (n = 10)	6.4 (1.2) (n = 32)	6.2 (1.8) (n = 5)
eGFR (mL/min/1.73 m2), Mean (SD)	111.3 (51.0)	98.3 (34.1) (n = 29)	125.2 (36.7)	100.0 (42.1)	164.6 (87.9)

BMI, body mass index; WC, waist circumference; SBP, systolic blood pressure; DBP, diastolic blood pressure; TG, triglycerides; HDL, high-density lipoprotein; LDL, low-density lipoprotein; TChol, total cholesterol; FBG, fasting blood glucose; A1c, glycated hemoglobin; eGFR, estimated glomerular filtration rate.

Obesity classification

The mean (SD) BMI was 28.3 (7.1) kg/m². Figure 1 demonstrates that the proportion of veterans who were classified as obese using BMI cut-off points of 30, 25.3, and 22 kg/m² was 30, 68, and 84%, respectively. Obesity classification per BMI organized by level of injury can be found in Fig. 2.

Figure 1.

Proportion of Veterans Screened as “At Risk,” Proportion of veterans screened as “at risk” according to cardiometabolic risk screening tools, presence of metabolic syndrome per WHO, NHLBI/AHA, NCEP, and IDF criteria, and classified as obese using various waist circumference and BMI cut-off points.

Figure 2.

Obesity Classification per BMI Cut-off by Level of Injury. Percentage of veterans classified as obese using BMI cut-off points per World Health Organization and SCI-Specific per Laughton et al. and Ayas et al. categorized by classification of injury (paraplegia versus tetraplegia) and American Spinal Injury Association (ASIA) Impairment Scale (AIS).

The mean (SD) WC was 111.7 cm (18.8). Sixty-nine percent and 77% of the 74 veterans with available WC measurements were classified as having a high WC per the cut-offs of 102 and 94 cm, respectively (Fig. 1).

Regression of %IBW adjusted by pounds and that adjusted by percent showed that both adjustments are very similar. The R² values for regressions at both ranges (low and high) were higher than 0.99. Regression coefficient for low range was 0.993 (P < 0.000) and that for high range was 1.06 (P < 0.000).

Cardiometabolic risk scores

The cardiometabolic risk screening tools produced varying results; 30, 88, and 68% of veterans were considered “at risk” according to the EOSS, CMDS, and FRS, respectively. The proportion of patients who screened “at risk” by each tool is displayed in Fig. 1.

Metabolic syndrome classification

The prevalence of MetS was 31% (IDF), 17% (NCEP), 53% (NHLBI/AHA), and 19% (WHO). The κ-agreement between the various definitions of MetS ranged from fair to moderate, as displayed in Table 4. The highest prevalence of MetS was per the NHLBI/AHA definition throughout all categories of level of injury and American Spinal Injury Association (ASIA) classification. The distribution of the presence of MetS by ASIA classification is displayed in Fig. 3.

Table 4. Comparison of metabolic syndrome definitions using kappa values.

Combination	κ-value	Agreement-strength	P value
MetS (IDF) and MetS (NCEP)	0.516	Moderate	<0.001
MetS (IDF) and MetS (NHLBI/AHA)	0.47	Moderate	<0.001
MetS (IDF) and MetS (WHO)	0.461	Moderate	<0.001
MetS (NCEP) and MetS (NHLBI/AHA)	0.304	Fair	<0.001
MetS (NCEP) and MetS (WHO)	0.521	Moderate	<0.001
MetS (NHLBI/AHA) and MetS (WHO)	0.327	Fair	<0.001

MetS, metabolic syndrome; IDF, International Diabetes Federation; NCEP, National Cholesterol Education Program; NHLBI/AHA, National Heart, Lung, and Blood Institute and American Heart Association; WHO, World Health Organization.

Figure 3.

Metabolic Syndrome by Level of Injury. Percentage of total sample in each classification with presence of MetS per IDF, NCEP, NHLBI/AHA, and WHO criteria. For example, of the veterans with ASIA A paraplegia (n = 29), the criteria for MetS as defined by the IDF identified presence of MetS in n = 9 (31%).

Additional descriptive frequencies

Out of the veterans who rated their overall health as “good” or “excellent” (n = 106), 62% were taking at least one medication for hypertension, diabetes mellitus, and/or hyperlipidemia. Of those same 106 veterans, 28%, 65%, and 82% were classified as obese per BMI cut-off of ≥30, ≥25.3, or >22 kg/m², respectively.

Discussion

The primary aim of this study was to describe and compare the classification of obesity, cardiometabolic risk, and MetS using existing screening tools that are typically available in a clinical setting among veterans with SCI. Frequencies of those screened as “at risk” or obese differed considerably among the various tools. Physiologic changes unique to SCI may confound commonly-used risk assessment tools, therefore potentially limiting the validity and/or reliability of these tools in the SCI population.

Various parameters included within the different screening tools may be affected by physiological changes specific to SCI. For example, fasting blood glucose levels may be affected by SCI-related complications and may not correlate with glucose tolerance in the SCI population.^7,60,61 Individuals with SCI also experience blood pressure fluctuations due to neurological changes.^62–64 WC may be affected by atrophy of abdominal muscles and increases in visceral adiposity in the SCI population, leading current WC cut-off points to underestimate central adiposity and consequent cardiometabolic risk.^17,27

Similarly, BMI of 30 kg/m² or greater is considered “obese” for the non-injured population.²¹ Although two different SCI-specific BMI cut-off points have been proposed, these have only been validated in small, select populations.^10,24,26 Two different methods of adjusting the traditional Hamwi IBW calculation²⁹ have been suggested for patients with SCI, but the proposed adjustments have not been validated. Nevertheless, results of this study indicate that the two different equations for adjusting IBW produce similar results. Since the intercepts were low and insignificant for both equations, it can be concluded that both adjustments are similar. Therefore, it can be inferred that no one equation is superior to the other and clinicians should continue to use clinical judgment in determining which equation to use until more research is available. Considering both adjusted IBW together with SCI-specific BMI cut-offs may be a prudent approach to the clinical assessment of weight status.

Many of the aforementioned criteria are used as parameters to assess cardiometabolic and CVD risk in screening tools and classification systems developed for the non-SCI population such as the FRS, CMDS, EOSS, and MetS classification. This study found only fair to moderate agreement between the different definitions of MetS in this sample of veterans with SCI, so it is not known which of the four definitions is most accurate.

Inconsistent results among previous studies examining the prevalence of MetS in SCI populations illustrate the complexity of identifying MetS in this population, and this warrants caution in applying standard definitions of MetS to patients with SCI. Results from Finnie et al. (n = 56) revealed a prevalence of 3.6% (WHO), 12.3% (NCEP), 15.8% (NHLBI/AHA), and 19.3% (IDF) compared to our results (n = 155) of 19% (WHO), 17% (NCEP), 53% (NHLBI/AHA), and 31% (IDF).⁹ Gater et al. (n = 473) found a higher prevalence of 57% (IDF) when using a surrogate obesity marker to account for the SCI differences, compared to 22.9% (IDF) when using the standard BMI measurement.⁸ Maruyama et al. (n = 44) found a higher prevalence of 43% (NCEP) when adjusting the WC for the Japanese population.⁵⁶ These differences may be due to different populations, sample sizes, and criteria used within a standard definition. For example, as WC was not available, Gater et al. used BMI to determine central obesity while we incorporated WC if the measure was available.⁸

Of note, while the criteria for the presence of MetS per the IDF requires central obesity as defined by ethnicity-specific WC cut-offs, the 2006 IDF Consensus Statement indicates that if BMI is >30 kg/m² central obesity can be assumed and waist circumference does not need to be measured.⁵¹ Since WC is not always measured in clinical practice, BMI > 30 kg/m² is typically used alone to meet this criterion. Of the 74 patients who had a WC measured in our cohort, 34 (46%) had a WC ≥ 94 cm, but also BMI <30. This means that, if WC was not available, these patients would have automatically screened negative for MetS per IDF using BMI alone even though their WC indicates they actually were at risk. Taken together, the evidence suggests that current definitions of MetS that were developed in the non-injured population cannot be extrapolated to individuals with SCI because they may likely underestimate true risk.⁹ The differences in prevalence of MetS within our own sample according to different definitions and also compared to other studies shows the inconsistencies that may be found in research and clinical practice and may lead to different interpretations of true risk.

The results of this study also demonstrated that patients may not be aware of their health risks. Although >90% of this sample screened as “at risk” by the screening tools investigated, 68% rated their overall health as “good” or “excellent.” Of those 106 patients who rated their health as “good” or “excellent,” 62% were on at least one medication for hypertension, diabetes, or hyperlipidemia. Previous research has pointed out that healthcare providers report challenges caring for people with SCI, and patients have expressed perceptions that their providers do not acknowledge or appreciate the complexity of their conditions.^65–68 Lack of agreement between health risk screening tools can add to further confusion, especially if the same patient visits multiple providers who use different tools that come to discrepant outcomes.

There are limitations of this study that should be noted. The first is the potential lack of generalizability to other patient groups. Since this cohort only included male veterans from the Chicago, Illinois area, findings may not be generalizable to women with SCI or those from other regions. Since each veteran is encouraged to schedule an annual evaluation once a year, only one year of appointments was reviewed. However, this could have limited the sample size as it would not capture patients who did not consistently attend one annual evaluation per calendar year. Inherent limitations of retrospective medical record reviews must be acknowledged, including a lack of data that could have added valuable insight such as physical activity levels, additional medications (e.g. drugs that may impact inflammation), C-reactive protein levels (which have been suggested to be associated with CVD risk),^9,53 or presence of undernutrition.

Similarly, since active prescriptions of medications were recorded as a dichotomous value of “yes” or “no” during chart review, the interactions of specific medications were not analyzed. It also does not consider if medications were updated or if patients were taking the medications as prescribed. Due to the retrospective nature of this study involving variables that were collected by clinic staff as part of routine care, it is unknown whether the anthropometric characteristics, clinical measures, and blood draws were obtained consistently or in the same manner as stated in the criteria for MetS. However, the intention of this study was to describe and compare the realistic use of these measures in the clinical setting, which adds to the strengths of this study.

Although robust measurements of body composition such as MRI, DXA, CT, or air displacement plethysmography are more accurate at measuring visceral adipose tissue and potentially better predictors of cardiometabolic health in the SCI population, they are not routinely available in non-research settings. A strength of this study was the focus on measures that can be easily and routinely obtained in a fast-paced outpatient clinic for the SCI patient. Furthermore, this study compares the variation in outcomes from a wide variety of existing screening tools and compiles the strengths and limitations specific to individuals with SCI for each tool or measure investigated (Table 1). This table can be used to guide the development of an SCI-specific screening tool by adjusting the limitations to better fit the SCI-specific conditions while building upon the strengths of the existing tools. An evidence-based SCI-specific screening tool can address the unique physiologic changes after SCI, allow for consistent measurement in cardiometabolic and/or CVD risk, help guide communication and informed decision-making, assist providers with caring for this vulnerable population, and assure the individual that his/her SCI-specific conditions are being addressed.

Conclusion

The current tools that were developed for the non-injured population to classify obesity, cardiometabolic risk, and MetS produce varying results in the SCI population and may not represent true risk given the unique physiologic changes after injury. Future research on this topic should focus on developing an SCI-specific screening tool in order to more accurately identify risk and provide timely education and intervention. This SCI-specific tool can incorporate strengths of the existing tools, such as including measures that are easily attainable in the clinic, while building upon the limitations in the SCI population such as addressing physiologic changes after SCI (Table 1). It would also be beneficial to investigate current practices and barriers for assessing risk by surveying professionals who work with the SCI population before developing an SCI-specific screening tool. In the interim, the interdisciplinary team should consider treating all patients with SCI as at elevated risk, regardless of screening tool assessments.

Disclaimer statements

Contributors None.

Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflicts of interest Authors have no conflict of interests to declare.

Ethics approval

This study was approved by the Institutional Review Board (IRB) at the Edward Hines, Jr. VA Hospital and University of Alabama.