Atherosclerosis 232 (2014) 305e312
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Review
Serum lipid concentrations among persons with spinal cord injury e A systematic review and meta-analysis of the literature Olivia Gilbert a, *, James R. Croffoot a, Allen J. Taylor b, Mark Nash c, Katherine Schomer d, Suzanne Groah e a
Department of Medicine, MedStar Georgetown University Hospital, 3800 Reservoir Road, Washington DC 2007, USA Cardiology Division, MedStar Georgetown University Hospital, 3800 Reservoir Road, Washington DC 2007, USA University of Miami, 1400 NW 12th Avenue, Miami, FL 33136, USA d University of Washington, 1959 Northeast Pacific Street, Seattle, WA 98195, USA e MedStar National Rehabilitation Hospital, 102 Irving Street NW, Washington DC 20010, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 January 2013 Received in revised form 24 October 2013 Accepted 5 November 2013 Available online 18 November 2013
Background: Lipid optimization comprises a therapeutic cornerstone of primary and secondary cardiovascular disease prevention. This systematic review and meta-analysis sought to clarify patterns of lipid profiles in spinal cord injury (SCI) patients compared to able-bodied individuals as well as among subgroups of SCI patients stratified by sex, activity level, race, and level of injury. Methods: Searches were conducted in PubMed, CINAHL, PsycINFO, and EMBASE. The initial literature search broadly identified peer-reviewed studies that examined cardiovascular risk factors in SCI. A total of 50 studies were ultimately identified that focused on lipid levels in SCI. Demographic data (including subject age, duration of injury, height, weight, and body mass index [BMI]) and lipid values were extracted for able-bodied individuals and subjects with SCI. Statistical analyses included t-testing and analysis of variance (ANOVA). Results: Compared with controls, individuals with SCI had significantly lower total cholesterol (TC) (183.4 mg/dL versus 194.9 mg/dL, p ¼ 0.019) and high-density lipoprotein cholesterol (HDL-C) (41.0 mg/ dL versus 49.6 mg/dL, p < 0.001) and higher TC/HDL-C ratios (4.5 versus 4.0, p ¼ 0.002), though no significant differences were found for triglyceride (TG) and non-HDL-C values. Conclusions: SCI represents an increasingly common chronic condition, now secondarily characterized by heightened CVD risk potentially in part due to unique lipid profiles characterized primarily by low HDL-C and an increased TC/HDL-C ratio. As other at-risk patient populations have received increased acknowledgment with more stringent lipid panel screening at earlier ages and increased frequency, we would propose that the same be implemented for the SCI population until more-specific CVD risk stratification guidelines are established for this population. Ó 2013 Elsevier Ireland Ltd. All rights reserved.
Keywords: Spinal cord injury Lipid metabolism disorders Cardiovascular risk factors
Contents 1. 2.
3. 4. 5.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 2.1. Data sources and searches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 2.2. Study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 2.3. Data extraction and quality assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 2.4. Data synthesis and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
* Corresponding author. 310 West 4th Street, #1502, Winston-Salem, NC 27101, USA. E-mail address: [email protected] (O. Gilbert). 0021-9150/$ e see front matter Ó 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2013.11.028
306
O. Gilbert et al. / Atherosclerosis 232 (2014) 305e312
Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Supplementary data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
1. Introduction Paralysis of varying types affects 6 million people in the United States. Emerging data indicate that patients with paralysis involving spinal cord injury (SCI), who now live longer with the aid of advances in neurological care, have increased risk for cardiovascular disease (CVD) compared to the general population [1e3]. Yekutiel et al. found a 2.5 fold higher incidence of ischemic heart disease in SCI patients (16.9%) compared to the general population (6.9%) [3]. Furthermore, Whiteneck et al. found an annual mortality rate from cardiac causes of 144.8 per 1000 persons with SCI compared to 100.2 per 1000 persons in the general population [2,4]. Moreover, ischemic heart disease in SCI disproportionately affects the young, with a nearly 4-fold higher rate of cardiac death before the age of 45 compared to the general population [2]. Just as concerning as their risk for CVD, SCI patients often have decreased awareness of the presence of their disease. Related to lack of physical activity to precipitate angina and disruption of sympathetic afferent fibers associated with cardiac pain, there is an increased incidence of asymptomatic disease compared to the general population [5]. In a small cross-sectional study, Baumen et al. demonstrated that 13 of 20 (65%) asymptomatic SCI patients with benign electrocardiograms had evidence of ischemia with thallium-201 single photon emission computerized tomography (SPECT) [6]. Similarly, in another study by Fleg et al., 55 of 407 (14%) paralyzed patients with benign electrocardiograms were found to have ischemia revealed by thallium-201 SPECT imaging [7]. Behavioral, metabolic, anthropomorphic, neurocardiogenic, and inflammatory reasons have all been implicated in the increased risk for CVD in SCI. In particular, SCI patients have a propensity to develop metabolic syndrome, a clustering of CVD risk factors including obesity, dyslipidemia, elevated blood pressure, and insulin resistance, contributed to by reduced activity levels and relative obesity [8e10]. Determination of obesity for these individuals using traditional BMI cutpoints presents challenges because of profound body composition changes characterized by loss of lean mass and increases in fat mass contributing to a significant underestimation of fat mass [11]. Where the BMI cutoff for obesity in able-bodied individuals is 30 kg/m2, it has been postulated that the BMI cutoff for obesity in SCI should be closer 22 kg/ m2 [11]. Thus, with increased tendency for CVD, unique risk factor considerations, and challenges to CVD detection from decreased personal awareness of cardiac symptoms, individuals with SCI are at risk for earlier, more frequent, and more severe cardiac presentations than the general population. Accordingly, there is need to improve the characterization of cardiovascular risk factor profiles in such a population. To date, SCI individuals have not received specific consideration within national guidelines regarding screening and treatment of CVD risk. The intent of the current study is to define lipid profiles unique to the SCI population as compared to able-bodied individuals in order to address their specific risk factors associated with the development of cardiovascular disease at a younger age. As concluded by a recent meta-analysis by Wilt et al. that examined 58 articles, “the role of lipid disorders to alter the risk of cardiovascular morbidity and mortality in SCI adults has not been adequately
addressed [12].” Thus, this meta-analysis will attempt to clarify patterns of lipid profiles in SCI patients compared to able-bodied individuals as well as among subgroups of SCI patients in order to inform discussions on the need to adjust lipid screening and treatment guidelines for SCI patients. We also sought to understand the patterns of dyslipidemia within SCI subgroups of sex, activity level, race, and injury level. 2. Methods This was a collaborative study of MedStar Georgetown University Hospital, MedStar National Rehabilitation Hospital, University of Miami Miller School of Medicine, and University of Washington. 2.1. Data sources and searches The criteria used to search for published studies included peerreviewed studies that met the following inclusion criteria: (1) of adults with a spinal cord injury; (2) investigating risk factors or markers of cardiovascular/heart disease; (3) published since 1960; and (4) written in English. For the initial database search, all study types such as review papers and meta-analyses were included. Searches in PubMed, CINAHL, PsycINFO, and EMBASE located 2190 articles (Fig. 1). 2.2. Study selection Once the search for published articles was complete, more specific inclusion criteria as specified in Fig. 1 were created to find the most relevant of the 2190 articles with studies that had cardiovascular disease as either a primary or secondary outcome. This was done by two expert reviewers at the University of Washington Model Systems Knowledge Translation Center (MSKTC). Discrepancies were resolved by consensus of the reviewers. Case reports were excluded. After reviewing the abstracts, 115 of the 2190 articles appeared to meet the inclusion criteria. Two physician reviewers from MedStar Georgetown University Hospital examined these selected articles to identify ones in which dyslipidemia was specifically highlighted as a clinical risk variable. Forty-two articles were identified with an additional eight articles found with review of reference lists for a total of fifty references (See Fig. 1). It should be clarified that the same inclusion/exclusion criteria for studies in general were applied to the acquisition of controlled studies with eighteen of the total fifty studies including control groups. A list of included articles can be found in the Appendix 1. 2.3. Data extraction and quality assessment Demographic data (including age, duration of injury, height, weight, and BMI) and lipid values were extracted from each of the 50 articles for able-bodied individuals, SCI patients, and SCI subgroups (including sex, race, activity level, and level of injury). If data were missing for one or more of the lipid values, it was calculated when possible, using the Friedewald Formula: LDL ¼ TCeHDL-Ce TG/5 (mg/dL) [13]. Additionally, non-HDL-C was calculated for all studies, such that non-HDL-C ¼ TCeHDL-C. For studies that
O. Gilbert et al. / Atherosclerosis 232 (2014) 305e312
307
Review of abstracts to identify studies meeting inclusion criteria. Searches were conducted in PubMed, CINAHL, PsycINFO, and EMBASE. Search terms used for each database: a) “cardiovascular disease, coronary heart disease, vascular inflammation, c-reactive protein, interleukin-6, postprandial dyslipidemia, atherosclerosis, and cardiovascular;” b) “diabetes, diabetes mellitus, metabolic syndrome X, insulin resistance, pre-diabetes, and glucose intolerance;” c) “obesity, overweight, and BMI;” d) “nutrition, fitness, exertion, physical activity, and exercise” and spinal cord injury (including spina bifida, transverse myelitis, and meningomyelocele). (n=2190)
Studies excluded due to not meeting the following more specific inclusion criteria 1) An observational or experimental (controlled or uncontrolled) study in which one of the primary outcomes is related to discovering the risk factors or markers of CVD in SCI 2) An observational or experimental (controlled or uncontrolled) study in which one of the secondary outcomes is reporting some possible risk factors or markers of CVD in SCI 3) A prevalence/incidence study, where the main focus of the study is to learn about overall CVD in SCI (n=2075)
Full review of articles and data extraction. Searches identified experimental and observational research studies where risk factors or markers of cardiovascular disease were either a primary or secondary outcome. (n=115)
Studies excluded due to not meeting secondary inclusion criteria 1) A study which one of the primary or secondary outcomes are any of the four major lipid components – TC, LDL, HDL, or TG – and that the specific numerical values of the relevant lipid components be provided. (n=65)
Final Articles Included (n=50)
Fig. 1. Study selection process.
evaluated treatment of dyslipidemia, only baseline, pretreatment data were used. For studies that were longitudinal, if two data points were given, the second one was used; if more than two data points were given, the data were averaged. Data were compiled in a custom web-based database specifically developed for systematic reviews.
whether data on SCI subjects in controlled versus uncontrolled studies were similar, an initial T-test was performed comparing age, duration, BMI, and lipid values (Table 1). As no significant differences were found in baseline lipid variables between controlled and uncontrolled studies, we combined data from these studies for comparison to non-SCI controls (Table 1).
2.4. Data synthesis and analysis
3. Results
SPSS statistical software (version 16, IBM Corporation, Armon NY) was used. T-testing compared lipid values in continuous variable groups (age, duration, BMI) as well as discrete variable groups (activity, level of injury, sex). ANOVA compared lipid values among different races (African American, Caucasian, Hispanic). Level of significance was determined at the 0.05 level. To determine
There were 37 observational studies (19 uncontrolled crosssectional, 1 uncontrolled case series, and 17 controlled crosssectional) and 13 experimental studies (8 uncontrolled crossover, 2 uncontrolled cohort, 1 controlled cohort, 1 randomized controlled, and 1 time series), which included 1252 control individuals and 4512 SCI patients (Appendix 1). In terms of
308
O. Gilbert et al. / Atherosclerosis 232 (2014) 305e312
Table 1 Controlled and uncontrolled SCI subgroup data comparison. Controlled SCI
N (studies) n (subjects) Age (years) Duration (years) BMI (kg/m2) TC (mg/dL) LDL (mg/dL) HDL (mg/dL) TG (mg/dL) non-HDL (mg/dL)
Uncontrolled SCI
Means (SD)
Ranges
Mean (SD)
Ranges
18
e
32 105.0 38.1 11.2 24.8 183.4 115.5 41.0 134.0 142.1
e
65 35.5 10.8 24.1 180.8 113.9 39.5 134.9 140.9
(79) (9.4) (5.0) (2.5) (17.8) (18.5) (7.6) (71.7) (17.9)
3e320 21e59 2e19 20e29 142e205 77e151 31e60 57e383 98e165
(146.0) (7.0) (5.6) (2.6) (15.1) (14.6) (5.8) (45.6) (14.1)
5e600 18e50 0.3e18 22e34 142e205 77e151 31e60 57e383 98e165
T-test
e e 0.28 0.85 0.24 0.40 0.57 0.18 0.93 0.70
demographics, there were no significant differences between ablebodied individuals and SCI patients with respective average ages of 34.6 and 37.1 years (p ¼ 0.310) and BMI’s of 25.1 and 24.8 kg/m2 (p ¼ 0.740) (Table 2). All abstracted data are available in the Appendix 1. There was a significant difference in mean TC for 1203 controls from 16 studies compared with 4449 SCI patients from 47 studies (183.4 mg/dL versus 194.9 mg/dL, p ¼ 0.019). There was no significant difference in mean LDL for 1211 controls from 16 studies compared with 4132 SCI patients from 46 studies (118.0 mg/dL versus 115.5 mg/dL, p ¼ 0.640). There was a significant difference in mean HDL-C for 1231 controls from 17 studies compared with 4479 SCI patients from 48 studies (41.0 mg/dL versus 49.6 mg/dL, p < 0.001). There was no significant difference in mean TG for 1204 controls from 16 studies compared with 4062 SCI patients from 45 studies (125.0 mg/dL versus 134.0 mg/dL, p ¼ 0.570). There was no significant difference in mean non-HDL-C for 1183 controls from 15 studies compared with 4059 SCI patients from 44 studies (145.9 mg/dL versus 142.1 mg/dL, p ¼ 0.530). The above values for TC and HDL-C were combined to reveal a significant difference in the TC/HDL-C ratios (4.5 versus 4.0, p ¼ 0.002) (Table 2). Comparing female versus male SCI subgroups, data were extracted from 17 studies with lipid data for 1603 men and 7 studies for 271 women. Accordingly, HDL-C was significantly higher (38.9 mg/dL versus 51.0 mg/dL, p < 0.001) and TC/HDL-C was significantly lower (4.8 versus 3.8, p ¼ 0.001) for women compared to men with SCI (Table 3). Four studies examined activity level and lipids among SCI participants. Combining the data from these studies, lipid data from 156 inactive individuals were compared with 64 active individuals. It should be noted that different levels of activity were defined by the individual studies. Comparing these groups, TC/HDL-C ratio was found to be significantly lower in the active subgroup compared to the inactive subgroup (3.6 versus 4.7, p ¼ 0.033) (Table 3). While, LDL and TG were generally lower as well as HDL-C, generally higher in active versus inactive subjects, these findings were not found to be significant. Table 2 Able-bodied and SCI data comparison. Means (SD)
Able-bodied
SCI
N (studies) Age (years) BMI (kg/m2) TC (mg/dL) LDL (mg/dL) HDL (mg/dL) TG (mg/dL) non-HDL (mg/dL) TC/HDL
18 34.6 25.1 194.9 118.0 49.6 125.0 145.9 4.0
50 37.1 24.8 183.4 115.5 41.0 134.0 142.1 4.5
(9.7) (2.3) (19.7) (25.4) (6.8) (74.2) (19.6) (0.6)
Bold values indicate statistical significance.
T-test (8.0) (2.6) (15.1) (14.7) (5.8) (45.6) (14.1) (0.6)
0.31 0.74 0.019 0.64
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Review
Serum lipid concentrations among persons with spinal cord injury e A systematic review and meta-analysis of the literature Olivia Gilbert a, *, James R. Croffoot a, Allen J. Taylor b, Mark Nash c, Katherine Schomer d, Suzanne Groah e a
Department of Medicine, MedStar Georgetown University Hospital, 3800 Reservoir Road, Washington DC 2007, USA Cardiology Division, MedStar Georgetown University Hospital, 3800 Reservoir Road, Washington DC 2007, USA University of Miami, 1400 NW 12th Avenue, Miami, FL 33136, USA d University of Washington, 1959 Northeast Pacific Street, Seattle, WA 98195, USA e MedStar National Rehabilitation Hospital, 102 Irving Street NW, Washington DC 20010, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 January 2013 Received in revised form 24 October 2013 Accepted 5 November 2013 Available online 18 November 2013
Background: Lipid optimization comprises a therapeutic cornerstone of primary and secondary cardiovascular disease prevention. This systematic review and meta-analysis sought to clarify patterns of lipid profiles in spinal cord injury (SCI) patients compared to able-bodied individuals as well as among subgroups of SCI patients stratified by sex, activity level, race, and level of injury. Methods: Searches were conducted in PubMed, CINAHL, PsycINFO, and EMBASE. The initial literature search broadly identified peer-reviewed studies that examined cardiovascular risk factors in SCI. A total of 50 studies were ultimately identified that focused on lipid levels in SCI. Demographic data (including subject age, duration of injury, height, weight, and body mass index [BMI]) and lipid values were extracted for able-bodied individuals and subjects with SCI. Statistical analyses included t-testing and analysis of variance (ANOVA). Results: Compared with controls, individuals with SCI had significantly lower total cholesterol (TC) (183.4 mg/dL versus 194.9 mg/dL, p ¼ 0.019) and high-density lipoprotein cholesterol (HDL-C) (41.0 mg/ dL versus 49.6 mg/dL, p < 0.001) and higher TC/HDL-C ratios (4.5 versus 4.0, p ¼ 0.002), though no significant differences were found for triglyceride (TG) and non-HDL-C values. Conclusions: SCI represents an increasingly common chronic condition, now secondarily characterized by heightened CVD risk potentially in part due to unique lipid profiles characterized primarily by low HDL-C and an increased TC/HDL-C ratio. As other at-risk patient populations have received increased acknowledgment with more stringent lipid panel screening at earlier ages and increased frequency, we would propose that the same be implemented for the SCI population until more-specific CVD risk stratification guidelines are established for this population. Ó 2013 Elsevier Ireland Ltd. All rights reserved.
Keywords: Spinal cord injury Lipid metabolism disorders Cardiovascular risk factors
Contents 1. 2.
3. 4. 5.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 2.1. Data sources and searches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 2.2. Study selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 2.3. Data extraction and quality assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306 2.4. Data synthesis and analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
* Corresponding author. 310 West 4th Street, #1502, Winston-Salem, NC 27101, USA. E-mail address: [email protected] (O. Gilbert). 0021-9150/$ e see front matter Ó 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atherosclerosis.2013.11.028
306
O. Gilbert et al. / Atherosclerosis 232 (2014) 305e312
Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Supplementary data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
1. Introduction Paralysis of varying types affects 6 million people in the United States. Emerging data indicate that patients with paralysis involving spinal cord injury (SCI), who now live longer with the aid of advances in neurological care, have increased risk for cardiovascular disease (CVD) compared to the general population [1e3]. Yekutiel et al. found a 2.5 fold higher incidence of ischemic heart disease in SCI patients (16.9%) compared to the general population (6.9%) [3]. Furthermore, Whiteneck et al. found an annual mortality rate from cardiac causes of 144.8 per 1000 persons with SCI compared to 100.2 per 1000 persons in the general population [2,4]. Moreover, ischemic heart disease in SCI disproportionately affects the young, with a nearly 4-fold higher rate of cardiac death before the age of 45 compared to the general population [2]. Just as concerning as their risk for CVD, SCI patients often have decreased awareness of the presence of their disease. Related to lack of physical activity to precipitate angina and disruption of sympathetic afferent fibers associated with cardiac pain, there is an increased incidence of asymptomatic disease compared to the general population [5]. In a small cross-sectional study, Baumen et al. demonstrated that 13 of 20 (65%) asymptomatic SCI patients with benign electrocardiograms had evidence of ischemia with thallium-201 single photon emission computerized tomography (SPECT) [6]. Similarly, in another study by Fleg et al., 55 of 407 (14%) paralyzed patients with benign electrocardiograms were found to have ischemia revealed by thallium-201 SPECT imaging [7]. Behavioral, metabolic, anthropomorphic, neurocardiogenic, and inflammatory reasons have all been implicated in the increased risk for CVD in SCI. In particular, SCI patients have a propensity to develop metabolic syndrome, a clustering of CVD risk factors including obesity, dyslipidemia, elevated blood pressure, and insulin resistance, contributed to by reduced activity levels and relative obesity [8e10]. Determination of obesity for these individuals using traditional BMI cutpoints presents challenges because of profound body composition changes characterized by loss of lean mass and increases in fat mass contributing to a significant underestimation of fat mass [11]. Where the BMI cutoff for obesity in able-bodied individuals is 30 kg/m2, it has been postulated that the BMI cutoff for obesity in SCI should be closer 22 kg/ m2 [11]. Thus, with increased tendency for CVD, unique risk factor considerations, and challenges to CVD detection from decreased personal awareness of cardiac symptoms, individuals with SCI are at risk for earlier, more frequent, and more severe cardiac presentations than the general population. Accordingly, there is need to improve the characterization of cardiovascular risk factor profiles in such a population. To date, SCI individuals have not received specific consideration within national guidelines regarding screening and treatment of CVD risk. The intent of the current study is to define lipid profiles unique to the SCI population as compared to able-bodied individuals in order to address their specific risk factors associated with the development of cardiovascular disease at a younger age. As concluded by a recent meta-analysis by Wilt et al. that examined 58 articles, “the role of lipid disorders to alter the risk of cardiovascular morbidity and mortality in SCI adults has not been adequately
addressed [12].” Thus, this meta-analysis will attempt to clarify patterns of lipid profiles in SCI patients compared to able-bodied individuals as well as among subgroups of SCI patients in order to inform discussions on the need to adjust lipid screening and treatment guidelines for SCI patients. We also sought to understand the patterns of dyslipidemia within SCI subgroups of sex, activity level, race, and injury level. 2. Methods This was a collaborative study of MedStar Georgetown University Hospital, MedStar National Rehabilitation Hospital, University of Miami Miller School of Medicine, and University of Washington. 2.1. Data sources and searches The criteria used to search for published studies included peerreviewed studies that met the following inclusion criteria: (1) of adults with a spinal cord injury; (2) investigating risk factors or markers of cardiovascular/heart disease; (3) published since 1960; and (4) written in English. For the initial database search, all study types such as review papers and meta-analyses were included. Searches in PubMed, CINAHL, PsycINFO, and EMBASE located 2190 articles (Fig. 1). 2.2. Study selection Once the search for published articles was complete, more specific inclusion criteria as specified in Fig. 1 were created to find the most relevant of the 2190 articles with studies that had cardiovascular disease as either a primary or secondary outcome. This was done by two expert reviewers at the University of Washington Model Systems Knowledge Translation Center (MSKTC). Discrepancies were resolved by consensus of the reviewers. Case reports were excluded. After reviewing the abstracts, 115 of the 2190 articles appeared to meet the inclusion criteria. Two physician reviewers from MedStar Georgetown University Hospital examined these selected articles to identify ones in which dyslipidemia was specifically highlighted as a clinical risk variable. Forty-two articles were identified with an additional eight articles found with review of reference lists for a total of fifty references (See Fig. 1). It should be clarified that the same inclusion/exclusion criteria for studies in general were applied to the acquisition of controlled studies with eighteen of the total fifty studies including control groups. A list of included articles can be found in the Appendix 1. 2.3. Data extraction and quality assessment Demographic data (including age, duration of injury, height, weight, and BMI) and lipid values were extracted from each of the 50 articles for able-bodied individuals, SCI patients, and SCI subgroups (including sex, race, activity level, and level of injury). If data were missing for one or more of the lipid values, it was calculated when possible, using the Friedewald Formula: LDL ¼ TCeHDL-Ce TG/5 (mg/dL) [13]. Additionally, non-HDL-C was calculated for all studies, such that non-HDL-C ¼ TCeHDL-C. For studies that
O. Gilbert et al. / Atherosclerosis 232 (2014) 305e312
307
Review of abstracts to identify studies meeting inclusion criteria. Searches were conducted in PubMed, CINAHL, PsycINFO, and EMBASE. Search terms used for each database: a) “cardiovascular disease, coronary heart disease, vascular inflammation, c-reactive protein, interleukin-6, postprandial dyslipidemia, atherosclerosis, and cardiovascular;” b) “diabetes, diabetes mellitus, metabolic syndrome X, insulin resistance, pre-diabetes, and glucose intolerance;” c) “obesity, overweight, and BMI;” d) “nutrition, fitness, exertion, physical activity, and exercise” and spinal cord injury (including spina bifida, transverse myelitis, and meningomyelocele). (n=2190)
Studies excluded due to not meeting the following more specific inclusion criteria 1) An observational or experimental (controlled or uncontrolled) study in which one of the primary outcomes is related to discovering the risk factors or markers of CVD in SCI 2) An observational or experimental (controlled or uncontrolled) study in which one of the secondary outcomes is reporting some possible risk factors or markers of CVD in SCI 3) A prevalence/incidence study, where the main focus of the study is to learn about overall CVD in SCI (n=2075)
Full review of articles and data extraction. Searches identified experimental and observational research studies where risk factors or markers of cardiovascular disease were either a primary or secondary outcome. (n=115)
Studies excluded due to not meeting secondary inclusion criteria 1) A study which one of the primary or secondary outcomes are any of the four major lipid components – TC, LDL, HDL, or TG – and that the specific numerical values of the relevant lipid components be provided. (n=65)
Final Articles Included (n=50)
Fig. 1. Study selection process.
evaluated treatment of dyslipidemia, only baseline, pretreatment data were used. For studies that were longitudinal, if two data points were given, the second one was used; if more than two data points were given, the data were averaged. Data were compiled in a custom web-based database specifically developed for systematic reviews.
whether data on SCI subjects in controlled versus uncontrolled studies were similar, an initial T-test was performed comparing age, duration, BMI, and lipid values (Table 1). As no significant differences were found in baseline lipid variables between controlled and uncontrolled studies, we combined data from these studies for comparison to non-SCI controls (Table 1).
2.4. Data synthesis and analysis
3. Results
SPSS statistical software (version 16, IBM Corporation, Armon NY) was used. T-testing compared lipid values in continuous variable groups (age, duration, BMI) as well as discrete variable groups (activity, level of injury, sex). ANOVA compared lipid values among different races (African American, Caucasian, Hispanic). Level of significance was determined at the 0.05 level. To determine
There were 37 observational studies (19 uncontrolled crosssectional, 1 uncontrolled case series, and 17 controlled crosssectional) and 13 experimental studies (8 uncontrolled crossover, 2 uncontrolled cohort, 1 controlled cohort, 1 randomized controlled, and 1 time series), which included 1252 control individuals and 4512 SCI patients (Appendix 1). In terms of
308
O. Gilbert et al. / Atherosclerosis 232 (2014) 305e312
Table 1 Controlled and uncontrolled SCI subgroup data comparison. Controlled SCI
N (studies) n (subjects) Age (years) Duration (years) BMI (kg/m2) TC (mg/dL) LDL (mg/dL) HDL (mg/dL) TG (mg/dL) non-HDL (mg/dL)
Uncontrolled SCI
Means (SD)
Ranges
Mean (SD)
Ranges
18
e
32 105.0 38.1 11.2 24.8 183.4 115.5 41.0 134.0 142.1
e
65 35.5 10.8 24.1 180.8 113.9 39.5 134.9 140.9
(79) (9.4) (5.0) (2.5) (17.8) (18.5) (7.6) (71.7) (17.9)
3e320 21e59 2e19 20e29 142e205 77e151 31e60 57e383 98e165
(146.0) (7.0) (5.6) (2.6) (15.1) (14.6) (5.8) (45.6) (14.1)
5e600 18e50 0.3e18 22e34 142e205 77e151 31e60 57e383 98e165
T-test
e e 0.28 0.85 0.24 0.40 0.57 0.18 0.93 0.70
demographics, there were no significant differences between ablebodied individuals and SCI patients with respective average ages of 34.6 and 37.1 years (p ¼ 0.310) and BMI’s of 25.1 and 24.8 kg/m2 (p ¼ 0.740) (Table 2). All abstracted data are available in the Appendix 1. There was a significant difference in mean TC for 1203 controls from 16 studies compared with 4449 SCI patients from 47 studies (183.4 mg/dL versus 194.9 mg/dL, p ¼ 0.019). There was no significant difference in mean LDL for 1211 controls from 16 studies compared with 4132 SCI patients from 46 studies (118.0 mg/dL versus 115.5 mg/dL, p ¼ 0.640). There was a significant difference in mean HDL-C for 1231 controls from 17 studies compared with 4479 SCI patients from 48 studies (41.0 mg/dL versus 49.6 mg/dL, p < 0.001). There was no significant difference in mean TG for 1204 controls from 16 studies compared with 4062 SCI patients from 45 studies (125.0 mg/dL versus 134.0 mg/dL, p ¼ 0.570). There was no significant difference in mean non-HDL-C for 1183 controls from 15 studies compared with 4059 SCI patients from 44 studies (145.9 mg/dL versus 142.1 mg/dL, p ¼ 0.530). The above values for TC and HDL-C were combined to reveal a significant difference in the TC/HDL-C ratios (4.5 versus 4.0, p ¼ 0.002) (Table 2). Comparing female versus male SCI subgroups, data were extracted from 17 studies with lipid data for 1603 men and 7 studies for 271 women. Accordingly, HDL-C was significantly higher (38.9 mg/dL versus 51.0 mg/dL, p < 0.001) and TC/HDL-C was significantly lower (4.8 versus 3.8, p ¼ 0.001) for women compared to men with SCI (Table 3). Four studies examined activity level and lipids among SCI participants. Combining the data from these studies, lipid data from 156 inactive individuals were compared with 64 active individuals. It should be noted that different levels of activity were defined by the individual studies. Comparing these groups, TC/HDL-C ratio was found to be significantly lower in the active subgroup compared to the inactive subgroup (3.6 versus 4.7, p ¼ 0.033) (Table 3). While, LDL and TG were generally lower as well as HDL-C, generally higher in active versus inactive subjects, these findings were not found to be significant. Table 2 Able-bodied and SCI data comparison. Means (SD)
Able-bodied
SCI
N (studies) Age (years) BMI (kg/m2) TC (mg/dL) LDL (mg/dL) HDL (mg/dL) TG (mg/dL) non-HDL (mg/dL) TC/HDL
18 34.6 25.1 194.9 118.0 49.6 125.0 145.9 4.0
50 37.1 24.8 183.4 115.5 41.0 134.0 142.1 4.5
(9.7) (2.3) (19.7) (25.4) (6.8) (74.2) (19.6) (0.6)
Bold values indicate statistical significance.
T-test (8.0) (2.6) (15.1) (14.7) (5.8) (45.6) (14.1) (0.6)
0.31 0.74 0.019 0.64
