Am J Physiol Heart Circ Physiol 301: H766–H772, 2011.
First published July 8, 2011; doi:10.1152/ajpheart.00507.2011.
Evidence for greater burden of peripheral arterial disease in lower extremity
arteries of spinal cord-injured individuals
Jeffrey W. Bell,1 David Chen,2 Martin Bahls,1 and Sean C. Newcomer1
1
Department of Health and Kinesiology, Purdue University, West Lafayette, Indiana; and 2Spinal Cord Injury Rehabilitation
Program Rehabilitation Institute of Chicago, Chicago, Illinois
Submitted 18 May 2011; accepted in final form 1 July 2011
atherosclerosis; exercise
(CVD) contributes to more than 40%
of deaths in spinal cord-injured individuals (SCI) (9). Interestingly, the mortality risk from CVD is 5–10% greater in SCI
compared with age- and risk factor-matched able-bodied individuals (22). Specifically, it has been demonstrated that spinal
cord injury leads to greater coronary artery calcification (24)
and T-wave abnormalities (33), but evidence is equivocal
regarding the development of subclinical atherosclerotic markers of carotid artery (CA) intima-media thickening and arterial
stiffness (14, 18). The potential mechanisms for atherosclerosis
in SCI include disorders of carbohydrate metabolism (1),
dyslipidemia (1, 28), lower-extremity physical inactivity, and
the resultant loss of lean body tissue (21), reduced daily energy
expenditure (5), and greater prevalence of type 2 diabetes (2).
Furthermore, especially in the lower extremities, altered hemo-
CARDIOVASCULAR DISEASE
dynamics reported post-spinal cord injury may place additional
risk of arterial wall endothelial dysfunction leading to subclinical atherosclerosis (13, 35). Indeed, for SCI compared with
able-bodied controls, reduced vascular function has been reported in the posterior tibialis of the legs but not in the radial
artery of the arms, and this decline worsens with greater
duration postinjury (31, 32). Others have reported that vascular
function is not altered in the superficial femoral artery (SFA) of
SCI, and, therefore, it remains unclear whether arterial function
declines in other arteries of the lower extremities. (38) The
reported CVD rates in SCI, however, may underestimate the
prevalence of this disease because of the fact that the epidemiological and experimental studies describing the incidence
of atherosclerotic CVD in SCI have focused on the coronary
circulation. It is currently not known whether these risks may
also contribute to a greater prevalence of peripheral arterial
disease (PAD) in SCI. Evidence for increased PAD after spinal
cord injury may also be derived from amputation data, indicating that the cause of tissue death was attributed to atherosclerotic plaques in 40% of amputation cases (11).
It is well known that exercise exerts its antiatherogenic
effects through modifications of traditional risk factors, but
other factors such as hemodynamics during exercise may also
play a role (17, 19). Exercise has been shown to increase blood
flow and shear rates by 20-fold or more in the legs during
lower-body exercise (26), but the necessary threshold of these
forces in creating antiatherosclerotic environments is not
known. Interestingly, upper-body exercise has been shown to
increase blood flow in the legs of both able-bodied and SCI but
to less of an extent as possible for leg exercise (34, 37). In
addition, upper-body exercise improves symptoms of PAD in
able-bodied individuals, but it is unclear whether upper-body
exercise may benefit the lower-extremity vasculatures of SCI.
Therefore, the purpose of this investigation was to investigate
the prevalence of PAD in SCI using the ankle-brachial index
(ABI) and intima-media thickness (IMT), measured via ultrasonography. We hypothesized that the incidence of lower-body
PAD as determined by ABI and IMT would be greater in SCI
compared with able-bodied controls because of higher incidence of traditional CVD risk factors, reduced physical activity, and lower daily peak shear stress in the quiescent lower
extremities of SCI. A second aim for this study was to determine whether chronic upper-body physical activity in SCI
would reduce the prevalence of this disease compared with
their sedentary counterparts. We hypothesized that physical
activity would confer a benefit to the arteries in the legs of SCI.
METHODS
Address for reprint requests and other correspondence: S. C. Newcomer,
Purdue Univ., Dept. of Health and Kinesiology, Lambert Fieldhouse, 800 W.
Stadium Ave., West Lafayette, IN, 47907 (e-mail: [email protected]).
H766
Subjects. Studies were performed on 105 SCI and 156 controls.
Subject characteristics are reported in Table 1. A medical history
0363-6135/11 Copyright © 2011 the American Physiological Society
http://www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.32.247 on June 18, 2017
Bell JW, Chen D, Bahls M, Newcomer SC. Evidence for greater
burden of peripheral arterial disease in lower extremity arteries of spinal
cord-injured individuals. Am J Physiol Heart Circ Physiol 301: H766–H772,
2011. First published July 8, 2011; doi:10.1152/ajpheart.00507.2011.—Spinal cord injury leads to increased risk for cardiovascular disease and
results in greater risk of death. Subclinical markers of atherosclerosis
have been reported in carotid arteries of spinal cord-injured individuals (SCI), but the development of lower extremity peripheral arterial
disease (PAD) has not been investigated in this population. The
purpose of this study was to determine the effect of spinal cord injury
on ankle-brachial index (ABI) and intima-media thickness (IMT) of
upper-body and lower-extremity arteries. We hypothesized that the
aforementioned measures of lower-extremity PAD would be worsened in SCI compared with controls and that regular participation in
endurance exercise would improve these in both groups. To test these
hypotheses, ABI and IMT were determined in 105 SCI and compared
with 156 able-bodied controls with groups further subdivided into
physically active and sedentary. ABIs were significantly lower in SCI
versus controls (0.96 ⫾ 0.12 vs. 1.06 ⫾ 0.07, P ⬍ 0.001), indicating
a greater burden of lower-extremity PAD. Upper-body IMTs were
similar for brachial and carotid arteries in controls versus SCI. Lower
extremity IMTs revealed similar thicknesses for both superficial
femoral and popliteal arteries, but when normalized for artery diameter, individuals with SCI had greater IMT than controls in the
superficial femoral (0.094 ⫾ 0.03 vs. 0.073 ⫾ 0.02 mm/mm lumen
diameter, P ⬍ 0.01) and popliteal (0.117 ⫾ 0.04 vs. 0.091 ⫾ 0.02
mm/mm lumen diameter, P ⬍ 0.01) arteries. The ABI and normalized
IMT of SCI compared with controls indicate that subclinical measures
of lower-extremity PAD are worsened in individuals with SCI. These
findings should prompt physicians to consider using the ABI as a
screening method to detect lower-extremity PAD in SCI.
H767
PERIPHERAL ARTERIAL DISEASE IN SPINAL CORD INJURY
Table 1. Descriptive data by group for active controls, sedentary controls, active SCI, and sedentary SCI
Injury duration, yr
Age, yr
Sex, %men
Height, cm
Weight, kg
BMI, kg/m2
Smoking, packs/day
Relatives with CVD
Diabetes, %
Systolic BP, mmHg
Active Controls
Sedentary Controls
Active SCI
Sedentary SCI
P Value
NA
40.5 ⫾ 12.7
90.1 ⫾ 30.3
179.7 ⫾ 8.4
83.1 ⫾ 17.1
25.6 ⫾ 3.4
0.002 ⫾ 0.02
1.0 ⫾ 1.1
4.2 ⫾ 18.6%
122 ⫾ 11.5
NA
42.6 ⫾ 11.6
84.7 ⫾ 33.6
176.9 ⫾ 10.3
88.5 ⫾ 20.6
28.1 ⫾ 6.0
0.121 ⫾ 0.32
1.1 ⫾ 1.0
8.2 ⫾ 24.6%
125 ⫾ 12.8
20.0 ⫾ 9.0
39.3 ⫾ 11.7
87.3 ⫾ 30.8
176.7 ⫾ 11.4
79.6 ⫾ 14.3
25.5 ⫾ 5.3
0.064 ⫾ 0.19
0.8 ⫾ 0.9
7.2 ⫾ 30.7%
125 ⫾ 15.4
18.4 ⫾ 8.8
42.0 ⫾ 9.9
90.0 ⫾ 39.3
179.0 ⫾ 7.8
80.2 ⫾ 22.0
25.3 ⫾ 3.4
0.167 ⫾ 0.30
0.6 ⫾ 0.7
12.0 ⫾ 34.4
129 ⫾ 21.3
NA
0.37
0.72
0.20
0.03
⬍0.01
⬍0.01
0.05
0.47
0.07
Values are means ⫾ SD. SCI, spinal cord-injured individuals; NA, not applicable; BMI, body mass index; CVD, cardiovascular disease; Systolic BP, average
systolic blood pressure for right and left brachial arteries.
Table 2. Subject self-report of at least one medication used
by general drug class
Medication Category
Controls, %
SCI, %
Anticoagulants
Antihypertensives/antiarrhythmics
Antilipidemics
Insulin/glucophages
CNS-D/MR/anti-convulsants/sedatives
Antidepressants
NSAIDs/analgesics
Bladder
Urinary tract-related antibiotics
Laxatives/antiperistaltic agents
Gastric reflux
Hormone replacement/birth control
Thyroid analogs
Other antibiotics and antiviral
Antipsoriatics
Erectile dysfunction
5.77
9.62
13.46
3.21
5.77
6.41
9.62
5.13
0.00
0.00
8.33
1.92
2.56
1.92
1.28
0.00
5.71
11.43
4.76
3.81
43.81
8.57
21.90
28.57
9.52
16.19
11.43
3.81
0.95
1.90
0.00
13.33
CNS-D, central nervous system depressants; MR, muscle relaxants; NSAIDs,
nonsteroidal anti-inflammatory drugs.
AJP-Heart Circ Physiol • VOL
their physical activity habits after which they transferred to a padded
examination table and rested supine for a minimum of 5 min. An ABI
was performed for both sides of the body, followed by vascular
imaging via ultrasound on the side with the lowest ABI. The same
investigator performed all ABIs and ultrasonography.
Measurements. The ABI is commonly used in clinical and research
settings to diagnose and measure severity of lower extremity PAD.
The ABI as a screening tool for PAD has been validated in a small
spinal cord-injured population without overt CVD risk factors (10).
To obtain an ABI, our subjects were fitted with a limb-appropriate
blood pressure cuff (Hokanson SC10/SC12, Bellevue WA). A bidirectional Doppler probe (Hokanson MD6) was placed at a 60° angle
to the artery of interest and used to detect systolic pressure during cuff
deflation. Systolic pressures were determined in the brachial, anterior
tibialis, and posterior tibialis arteries. The ABI was calculated as the
ratio of systolic pressures in the ankles to the brachial artery (BA),
using the higher ABI from the two tibialis arteries. The higher
pressure method has been proposed to most accurately detect clinically relevant PAD in the general population (16).
The IMT of four arteries was measured via ultrasound (Terason
T3000, Burlington MA), using a variable frequency transducer (5–12
MHz). The BA, CA, and SFA were measured with the subject supine,
and the popliteal artery (PA) was measured with the subject lying
prone. Each artery was measured in the longitudinal plane, using
B-mode with a high-frequency setting (25). Specific anatomical landmarks were used at each site of measurement to ensure that the
measurements were consistent between all subjects. The BA was
imaged above the antecubital fossa, and the CA images were taken
adjacent to the inferior border of the carotid bulb. In the leg, the SFA
images were obtained in the subsartorial canal distal to the bifurcation
of the common femoral artery, whereas the PA was measured in the
popliteal fossa. A 10-s video file was recorded, and the IMT was
continuously analyzed during the duration of the video along the most
clearly visible region of interest that allowed automated edge-detection software (Medical Imaging Applications LLC Carotid Analyzer
5, Coralville, IA) to easily track the media-adventitia border and the
intima-lumen border of the far wall. It is appropriate to compare IMT
of differently sized arteries by normalizing per millimeter of internal
diameter (23). Therefore, arterial diameter was determined using a
low-frequency setting and automated detection software (Medical
Imaging Applications LLC Brachial Analyzer 5), allowing the calculation of IMT normalized per millimeter of lumen diameter. IMT and
artery diameter were determined to the nearest hundredth of a millimeter, and IMT normalized to artery diameter was calculated to the
nearest thousandth of a millimeter per millimeter of artery lumen
diameter. Two separate investigators analyzed all IMT and artery
diameter measurements with an average IMT coefficient of variation
of 6.68% for absolute IMT and 10.06% for normalized IMT across all
four arteries measured.
301 • SEPTEMBER 2011 •
www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.32.247 on June 18, 2017
questionnaire including items regarding traditional cardiovascular risk
factors was completed: this included asking subjects to report regular
use of medications, cholesterol, smoking behavior, and history of
hypertension. Medication usage is reported in Table 2. Furthermore,
participants were asked questions about their physical activity habits
including exercise type, minutes per day, days per week, and weeks
per year: these questions were used to determine physical activity
status and further subdivided the groups into active SCI (SCI-A, n ⫽
55), sedentary SCI (SCI-S, n ⫽ 50), active controls (ABC-A, n ⫽ 71),
and sedentary controls (ABC-S, n ⫽ 85). Active groups met or
exceeded 150 min/wk of moderate-intensity exercise recommended
by the American College of Sports Medicine/American Heart Association. (12) The spinal cord injury group consisted of individuals
with paraplegia (n ⫽ 74) or tetraplegia (n ⫽ 31) who were at least 5
years postinjury. Only American Spinal Injury Association Impairment Scale grades A, B, or C were included in the study and thus
limited to subjects who primarily used wheelchairs for mobility
activities. SCI were recruited from the patient population, served by
the Rehabilitation Institute of Chicago and across the midwest and
southern United States through recreation and sporting groups. Control subjects were tested from similar geographic areas. Before participation, each subject was verbally informed of potential risks and
discomforts associated with the study and signed written, informed
consent. The study was approved by Institutional Review Boards of
Purdue University and Northwestern University.
Experimental protocol. An individual testing session was completed within the same day. On the day of the study, subjects reported
to testing, completed a medical history, and answered questions about
H768
PERIPHERAL ARTERIAL DISEASE IN SPINAL CORD INJURY
Table 3. Percent race distribution by group for active controls, sedentary controls, active SCI, and sedentary SCI
Race
Active Controls
Sedentary Controls
Active SCI
Sedentary SCI
P Value
Caucasian, %
Hispanic, %
African-American, %
Asian, %
80.3 ⫾ 35.9
7.0 ⫾ 26.5
14.1 ⫾ 28.7
4.2 ⫾ 20.8
74.1 ⫾ 41.9
11.8 ⫾ 32.4
15.3 ⫾ 32.4
2.4 ⫾ 15.2
70.9 ⫾ 45.8
14.5 ⫾ 35.0
12.7 ⫾ 33.6
3.6 ⫾ 18.9
46.0 ⫾ 50.3
14.0 ⫾ 35.6
40.0 ⫾ 49.5
0.00
⬍0.01
0.54
⬍0.01
0.53
Values are means ⫾ SD. Multiracial subjects were able to claim more than one racial identity.
RESULTS
Group characteristics. For SCI, the mean duration of injury
was 19.3 ⫾ 8.9 yr with a range of 5.0 – 41.2 yr. There were no
significant differences between groups for age, sex, height,
prevalence of diabetes, number of reported relatives with CVD,
or average systolic brachial blood pressure (Table 1). When
compared with controls, SCI exhibited greater tobacco smoking behavior, lower body weight, and lower body mass index
(Table 1). For race, there were significant differences in percentage of Caucasian and African-American subjects by group
(Table 3).
Ankle-brachial index. ABIs (Table 4) were larger in controls
versus SCI (1.06 ⫾ 0.07 vs. 0.96 ⫾ 0.12, P ⬍ 0.001). Two
control individuals and 29 SCI had ABIs measuring below the
clinically accepted diagnostic threshold of 0.90. Additionally,
ABI in 0 controls and 10 SCI were below 0.80. When the
effects of physical activity in these groups were compared, an
Table 4. Ankle-brachial index in controls versus SCI
Group
Ankle-Brachial Index
All controls
All SCI
Active controls
Sedentary controls
Active SCI
Sedentary SCI
1.06 ⫾ 0.07*
0.96 ⫾ 0.12
1.08 ⫾ 0.07†
1.05 ⫾ 0.07†
0.94 ⫾ 0.11
0.97 ⫾ 0.10
Values are means ⫾ SD. *P ⬍ 0.01, significant difference between control
and SCI group; †P ⬍ 0.01, significant difference between control and active or
sedentary SCI group values.
AJP-Heart Circ Physiol • VOL
analysis revealed a significant interaction between group and
physical activity level (P ⬍ 0.01) and a trend (P ⫽ 0.06) for
physical activity in controls. Post hoc analysis indicated a trend
for larger ABI in ABC-A compared with ABC-S (1.08 ⫾ 0.07
vs. 1.05 ⫾ 0.07, P ⫽ 0.06), whereas no differences were seen
for physical activity in the SCI group ABI (0.94 ⫾ 0.11 vs.
0.97 ⫾ 0.10, P ⫽ 0.28). In SCI, ABI had a moderate negative
correlation with number of years postinjury (r ⫽ 0.40, P ⬍
0.01). This association remained after controlling for age (r ⫽
0.39, P ⬍ 0.01).
Intima-media thickness. Upper-body artery IMT and diameter measurements are summarized in Table 5. BA-IMT was
similar for control and SCI when expressed as absolute thickness (0.35 ⫾ 0.07 vs. 0.36 ⫾ 0.08 mm, P ⫽ 0.76). Interestingly, there was a significant beneficial effect of spinal cord
injury in controls versus SCI for normalized BA-IMT (0.079 ⫾
0.02 vs. 0.077 ⫾ 0.02 mm/mm lumen diameter, P ⫽ 0.02). BA
diameter was smaller in control compared with SCI (4.47 ⫾
0.69 vs. 4.79 ⫾ 0.88 mm, P ⬍ 0.01). In the BA, IMT was
significantly correlated with diameter in control individuals
(r ⫽ 0.33, P ⬍ 0.01). Absolute IMT measurements for CA were
not significantly different in controls compared with SCI (0.61 ⫾
0.14 vs. 0.59 ⫾ 0.15 mm, P ⫽ 0.60) or for normalized IMT
(0.083 ⫾ 0.02 vs. 0.081 ⫾ 0.02 mm/mm lumen diameter, P ⫽
0.37). Similarly, CA diameter measurements were not different
for controls compared with SCI (7.34 ⫾ 0.67 vs. 7.23 ⫾ 0.65
mm, P ⫽ 0.19). There was, however, a significant effect of
activity level with a smaller CA-IMT (0.57 ⫾ 0.14 vs. 0.62 ⫾
0.14 mm, P ⬍ 0.02) and a smaller normalized CA-IMT (0.079 ⫾
0.02 vs. 0.085 ⫾ 0.02, P ⬍ 0.05) in active subjects compared
Table 5. Arterial structure in upper-extremity arteries of
controls and SCI
Artery/Group
Brachial
All controls
Active
Sedentary
All SCI
Active
Sedentary
Carotid
All controls
Active
Sedentary
All SCI
Active
Sedentary
IMT, mm
IMT, per mm lumen
Diameter, mm
0.35 ⫾ 0.07
0.35 ⫾ 0.08
0.35 ⫾ 0.07
0.36 ⫾ 0.08
0.35 ⫾ 0.07
0.36 ⫾ 0.07
0.079 ⫾ 0.02
0.078 ⫾ 0.02
0.080 ⫾ 0.02
0.077 ⫾ 0.02*
0.074 ⫾ 0.02
0.080 ⫾ 0.02
4.47 ⫾ 0.69
4.53 ⫾ 0.69
4.42 ⫾ 0.69§
4.79 ⫾ 0.88†
4.93 ⫾ 0.91‡
4.64 ⫾ 0.83
0.61 ⫾ 0.14
0.58 ⫾ 0.15
0.63 ⫾ 0.15
0.59 ⫾ 0.15
0.56 ⫾ 0.10
0.62 ⫾ 0.11
0.083 ⫾ 0.02
0.080 ⫾ 0.02
0.085 ⫾ 0.02
0.081 ⫾ 0.02
0.078 ⫾ 0.02
0.085 ⫾ 0.02
7.34 ⫾ 0.67
7.30 ⫾ 0.56
7.38 ⫾ 0.75
7.23 ⫾ 0.65
7.18 ⫾ 0.56
7.28 ⫾ 0.75
Values are means ⫾ SD. IMT, intima-media thickness. *P ⬍ 0.05,
significant difference between control and SCI group values; †P ⬍ 0.01, significant
difference between control and SCI group values; ‡P ⬍ 0.05, significant difference
between active SCI and active controls; §P ⬍ 0.01, significant difference
between active SCI and sedentary controls.
301 • SEPTEMBER 2011 •
www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.32.247 on June 18, 2017
Statistical analysis. Statistics were performed using commercially
available programs (SAS versions 9.1 and 9.2, Cary, NC). One-way
ANOVA was used with covariates added to the model when indicated
to determine group differences between SCI and controls. Covariates
were tested for model significance only when they significantly
differed between SCI and controls. To determine differences for
chronic exercise, a two-level factor for activity was included. Least
square means was used post hoc to determine within group differences. To investigate associations between variables, correlations
were performed via Pearson product-moment correlation analysis. All
data are presented as means ⫾ SD. Statistical significance was
established a priori as P ⬍ 0.05.
During analysis of reported traditional cardiovascular risk factors,
covariates emerged that were included within each model. BA-IMT
and BA-IMT normalized per millimeter of diameter included white
racial identity in each model, whereas normalized BA-IMT also
included body mass. CA-IMT included the factor of body mass,
whereas CA-IMT normalized to artery diameter included black racial
identity. Both absolute and normalized SFA-IMT models included the
number of relatives reported to have CVD, and the absolute SFA-IMT
model also included body mass index. The absolute PA-IMT indicated
body mass to be a significant covariate, whereas normalized PA-IMT
had no significant model covariates.
PERIPHERAL ARTERIAL DISEASE IN SPINAL CORD INJURY
Table 6. Arterial structure in lower-extremity arteries of
controls and SCI
Artery/Group
Superficial femoral
All controls
Active
Sedentary
All SCI
Active
Sedentary
Popliteal
All controls
Active
Sedentary
All SCI
Active
Sedentary
IMT, mm
IMT, per mm lumen
Diameter, mm
0.50 ⫾ 0.09
0.49 ⫾ 0.10
0.50 ⫾ 0.08
0.49 ⫾ 0.15
0.46 ⫾ 0.13§
0.53 ⫾ 0.18
0.073 ⫾ 0.02*
0.069 ⫾ 0.02†
0.077 ⫾ 0.02‡
0.094 ⫾ 0.03
0.086 ⫾ 0.02a
0.102 ⫾ 0.05
6.89 ⫾ 1.04*
7.11 ⫾ 0.96†
6.71 ⫾ 1.07†
5.38 ⫾ 0.94
5.37 ⫾ 0.86
5.38 ⫾ 1.02
0.60 ⫾ 0.14
0.58 ⫾ 0.12
0.62 ⫾ 0.15
0.60 ⫾ 0.18
0.57 ⫾ 0.18
0.62 ⫾ 0.23
0.091 ⫾ 0.02*
0.087 ⫾ 0.02†
0.094 ⫾ 0.02†
0.117 ⫾ 0.04
0.116 ⫾ 0.04
0.118 ⫾ 0.04
6.71 ⫾ 1.15*
6.79 ⫾ 1.20†
6.65 ⫾ 1.11†
5.20 ⫾ 0.91
5.11 ⫾ 0.96
5.30 ⫾ 0.86
Values are means ⫾ SD. *P ⬍ 0.01, significant difference between control
and SCI group values; †P ⬍ 0.01, significant difference between control and
both SCI activity group values;. ‡P ⬍ 0.01, significant difference between
sedentary controls and sedentary SCI; §P ⬍ 0.05, significant difference
between active SCI and sedentary SCI; aP ⬍ 0.01, significant difference
between active SCI and sedentary SCI.
AJP-Heart Circ Physiol • VOL
1.15 mm, P ⬍ 0.01). When arterial lumen diameter was
normalized, the SCI group IMT was significantly larger compared with controls (0.117 ⫾ 0.04 vs. 0.091 ⫾ 0.02 mm/mm
lumen diameter, P ⬍ 0.01). In SCI, there was a trend for
greater normalized PA-IMT with longer injury duration (r ⫽
0.16, P ⫽ 0.06). Additionally, there was no effect of physical
activity for the absolute IMT of the PA. Similar to the SFA
normalized IMT, the larger IMT in ABC-S compared with
ABC-A was not significant (0.095 ⫾ 0.02 vs. 0.087 ⫾ 0.02
mm/mm lumen diameter, P ⫽ 0.44). ABC-A-normalized PAIMT and ABC-S-normalized PA-IMT were smaller than in
both SCI groups (P ⬍ 0.01), although both SCI groups were
similar (P ⫽ 0.97). PA-IMT was significantly correlated with
diameter for controls (r ⫽ 0.40, P ⬍ 0.01).
DISCUSSION
The elevated CVD risk factors and reduced physical activity
of the legs in SCI suggest that the burden of PAD may be
higher in these individuals compared with able-bodied controls. Therefore, the objective of this study was to determine
whether persons with spinal cord injury had worsened clinical
measures for PAD when assessed by the ABI and IMT.
Additionally, we sought to determine whether exercise would
improve the occurrence of PAD as assessed by these measures.
The primary new findings of this study are as follows: 1) the
ABI is lower in SCI compared with controls, potentially
indicating a greater prevalence of lower extremity PAD occurring in chronic SCI; 2) the normalized IMT of SFA and PA in
SCI is larger than control populations, indicating the possibility
for greater atherosclerotic disease; 3) able-bodied individuals
who performed chronic physical activity had improvements in
ABI, but these benefits of greater physical activity did not
remain in the ABI of SCI groups; and 4) chronic exercise
benefits the SFA-IMT in SCI but does not appear to protect the
PA: this is in contrast to the improved PA-IMT in ABC-A. To
our knowledge, this is the first study to document a greater
subclinical prevalence of lower extremity PAD for SCI.
Spinal cord injury effects. As previously stated, the ABI is a
clinical measure for diagnosing lower-extremity PAD that has
been validated in SCI. In the current investigation ABI was
lower in SCI compared with ABC by 0.10. These results
provide evidence for a greater burden of lower-extremity PAD
in SCI than ambulatory populations. This is highlighted by the
fact that 29 SCI, or 27.6% of SCI, in this study had ABI
measurements below the clinically accepted threshold of 0.90,
whereas only two control individuals, or 1.3%, met this threshold. However, others have suggested that using the higher
pressure method for determining ABI that we used may underestimate the incidence of PAD compared with using the
lower ankle systolic pressure for ABI calculation (20). Using
the lower pressure from the two different tibialis arteries
(ABI-L) method, we determined that the mean ABI-L was
below (0.89 ⫾ 0.13) the well-accepted clinical threshold of
0.90, and therefore this method may be valuable to clinicians
wishing to screen their patients for PAD. Additionally, in
active versus sedentary controls, we detected a significantly
higher ABI-L (1.04 ⫾ 0.07 vs. 0.99 ⫾ 0.07, P ⫽ 0.02) but no
differences in ABI-L in active or sedentary SCI (0.89 ⫾ 0.13
vs. 0.90 ⫾ 0.13, P ⫽ 0.90). To our knowledge, the only study
to report ABI in SCI was performed in a small spinal cord-
301 • SEPTEMBER 2011 •
www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.32.247 on June 18, 2017
with sedentary subjects. When the activity level by group was
compared, there were no significant differences for absolute or
normalized CA-IMT. In the CA, IMT was significantly correlated with diameter in controls (r ⫽ 0.33, P ⬍ 0.01).
SFA-IMT measurements are summarized in Table 6. SCI
and control groups had similar absolute IMT measurements for
the SFA (0.49 ⫾ 0.12 vs. 0.50 ⫾ 0.11 mm, P ⫽ 0.68), and
when normalized to arterial lumen diameter, the SCI group
IMT was significantly larger compared with controls (0.094 ⫾
0.03 vs. 0.073 ⫾ 0.02 mm/mm lumen diameter, P ⬍ 0.01).
Interestingly, the normalized SFA-IMT was not associated
with number of years postinjury (r ⫽ ⫺0.04, P ⫽ 0.67). The
SFA diameters were smaller in SCI compared with controls
(5.38 ⫾ 0.94 vs. 6.89 ⫾ 1.04 mm, P ⬍ 0.001). Additionally,
SFA-IMT was significantly correlated with diameter of control
subjects (r ⫽ 0.19, P ⫽ 0.02). When absolute IMT based on
activity grouping was compared, the IMT in SFA of active
compared with sedentary subjects was smaller (0.47 ⫾ 0.11 vs.
0.51 ⫾ 0.12 mm, P ⬍ 0.01). Post hoc analysis revealed that
this effect was primarily located within the difference between
SCI-A and SCI-S for SFA (0.46 ⫾ 0.12 vs. 0.53 ⫾ 0.17 mm,
P ⬍ 0.02). Furthermore, these differences are amplified in the
normalized IMT measurements of SFA-IMT with greater IMT
in SCI compared with controls (0.094 ⫾ 0.03 vs. 0.074 ⫾ 0.02
mm/mm lumen diameter, P ⬍ 0.01). SCI-S had the largest
normalized IMT (0.102 ⫾ 0.04 mm/mm lumen diameter),
followed by SCI-A (0.086 ⫾ 0.02 mm/mm lumen diameter),
ABC-S (0.077 ⫾ 0.02 mm/mm lumen diameter), and smallest
normalized IMT in ABC-A (0.069 ⫾ 0.02). Differences were
significant (P ⬍ 0.01) in ABC-A versus SCI-A, SCI-A versus
SCI-S, ABC-S versus SCI-S, and ABC-A versus SCI-S. The
larger normalized IMT in the ABC-S versus ABC-A was not
significantly different (P ⫽ 0.07).
PA-IMT measurements are summarized in Table 6. Consistent with those of the SFA, PA absolute IMT measurements
were similar between SCI and control groups (0.60 ⫾ 0.18 vs.
0.60 ⫾ 0.14 mm, P ⫽ 0.84). The diameters of the PA were
smaller in SCI compared with controls (5.20 ⫾ 0.91 vs. 6.71 ⫾
H769
H770
PERIPHERAL ARTERIAL DISEASE IN SPINAL CORD INJURY
AJP-Heart Circ Physiol • VOL
diameter or whether thickening is occurring because of atherosclerotic processes in the intima. Although there was a trend in
the PA, injury duration in our study was not significantly
correlated to greater normalized IMT, and therefore these
measurements should be investigated during the time of arterial
remodeling immediately after the spinal cord injury has occurred to determine the role of remodeling on these measures.
Regardless of the underlying causes, the normalized IMT in the
lower extremity arteries of our SCI, particularly in the PA,
mirrors the pattern of PAD determined by the ABI.
Exercise effects. In control subjects who were chronic exercisers, both ABI methods yielded significantly higher indexes
compared with those in sedentary controls. This pattern did not
Downloaded from http://ajpheart.physiology.org/ by 10.220.32.247 on June 18, 2017
injured population (n ⫽ 15) that was free of overt CVD risk
factors (10). To get a cross-sectional sampling of the spinal
cord-injured population, our study did not exclude individuals
on the basis of CVD risk profiles. Our higher pressure method
yielded a mean ABI that was lower by 0.19 than that reported
in SCI free of CVD risk factors. In the previously mentioned
study, a negative association has been reported between ABI
and greater number of years postinjury (10). Therefore, the
greater average duration postinjury in the current study compared
with the previous (19.3 ⫾ 8.9 vs. 7.0 yr) may account for a portion
of the lower ABI measurements in our SCI. This is consistent with
our current finding that injury duration was correlated with reduced ABI in our SCI. It is possible with the clustering of risk
factors occurring in SCI that these inclusion criteria could increase
the likelihood of measuring a reduced ABI. This increased CVD
risk would be the natural consequence of SCI in our population
and therefore more accurately depict the prevalence of lowerextremity PAD.
As expected, the upper-body artery IMT measurements were
not significantly different between groups. This is similar to
results demonstrating no difference in CA-IMT between physically active SCI and age-matched controls. (14) Our absolute
CA-IMT measurements for all SCI were slightly larger than
those previously reported in 28 wheelchair athletes, ⬃0.10
mm, but were 0.15 mm smaller than those reported for 34 SCI,
including injuries causing paraplegia and tetraplegia (18). Interestingly, the normalized IMT in the BA was smaller in our
SCI group. This small benefit may have appeared due to a
larger BA diameter in this group, and the outward remodeling
of this artery lumen is most likely a consequence of manual
wheelchair usage and the repetitive activity of the arms in the
majority of our subject population. Unfortunately, we did not
collect data on manual versus power wheelchair use and cannot
fully explore the differences seen in the BA. This activityspecific adaptation, however, is consistent with the report that
the BA of the preferred arm in racquet sports players is
significantly larger than the BA of the nonpreferred arm (27).
The lower extremity arteries were not different between
groups when considering the absolute IMT. These IMTs are
similar to previous reports of common femoral artery IMTs not
differing between paraplegic, sedentary, and athletic individuals (29). As previously suggested, the paradox of these similar
lower-extremity IMT measurement lies in the fact that arteries
below the level of injury have a reduced diameter and therefore
should have a reduced wall thickness in accordance with
Laplace’s Law (29). Herein lies a major tenant of this investigation in that artery diameter is related to the absolute IMT.
In our able-bodied control individuals, correlations between
artery diameter and IMT were significant. This suggests that
diameter is responsible for a portion of the variance in the
measurement and that other factors related to group conditions,
such as greater atherosclerotic disease, explain most of the
remaining variance. Indeed, in lower-extremity arteries, when
normalized IMT are compared, SCI show greater thickening of
the arterial intima-media. The artery diameter in the lower
extremities is reduced to nearly half its original size within
weeks of the injury (7). This suggests that a greater proportion
of lumen is occupied by the IMT in SCI after the reductions in
internal diameter occur. In our study, because of subject
inclusion of at least 5 years postinjury, it is unclear whether the
media fails to remodel to meet the reductions in internal
Fig. 1. Ankle-brachial index (A), normalized superficial femoral artery intimamedia thickness (SFA-IMT, mm/mm lumen diameter), and normalized popliteal artery IMT (PA-IMTC, mm/mm lumen diameter) in active able-bodied
controls, sedentary able-bodied controls, active spinal cord-injured individuals
(SCI), and sedentary SCI in Caucasian and African-American subjects. Data
from least square means post hoc analysis are presented: *P ⬍ 0.01, control
significantly different than active and sedentary SCI; †P ⬍ 0.05, active control
significantly different than active SCI; ‡P ⬍ 0.01, sedentary control significantly different than sedentary SCI; §P ⬍ 0.01, active control significantly
different than active SCI; aP ⬍ 0.05, sedentary control significantly different
than active SCI; and bP ⬍ 0.05, active controls significantly different than
active SCI. Error bars represent means ⫾ SD.
301 • SEPTEMBER 2011 •
www.ajpheart.org
PERIPHERAL ARTERIAL DISEASE IN SPINAL CORD INJURY
AJP-Heart Circ Physiol • VOL
(AD), a reflexively activated large increase in systemic blood
pressure, may have played a role in our results. AD is often
triggered when SCI have pressure sores or experience other
painful stimuli inferior to their injury level. Only 7.6% of our
SCI reported current pressure sores with only one-third of
those reporting pressure sores having injuries where they
would be likely to experience AD. Therefore, we feel that it is
unlikely that AD played a significant role in our results.
Clinical relevance. Greater rates of PAD are associated with
an increased risk of future heart attack (4⫻) and stroke (3⫻) in
able-bodied individuals (6). If these risk associations hold true
for SCI, the elevated mortality rates in SCI (22) may partially
be explained by the normalized IMT in their lower extremity
arteries. Additionally, with many SCI being insensate in the
lower extremities, typical symptoms for PAD are not detectable by the individual. Therefore, these data indicate that
physicians who treat SCI should consider using the ABI for
regular PAD screening in this population. The diagnosis and
subsequent treatment of early stage, lower-extremity PAD
would in turn have a direct positive impact on a spinal
cord-injured individual’s quality of life and health care costs
associated with this disease. For instance, skin blood flow is
reduced in the lower extremities of SCI and may be related to
the development of pressure ulcerations (8). One can then
speculate that PAD in the lower extremities of SCI may be
related to the development of or increased healing time for skin
breakdown lesions and that PAD identification may aid physicians in determining appropriate treatment.
Summary. Results from this investigation suggest that elevated CVD risk for SCI includes a greater incidence of lowerextremity PAD. Measurements of ABI were worsened for
persons with spinal cord injuries and were different for physically active SCI. This corresponds with a greater normalized
IMT for SCI in the PA and SFA with evidence of improved
normalized IMT in the SFA of active compared with sedentary
SCI.
GRANTS
This project was funded, in part, with support from the Indiana Clinical and
Translational Sciences Institute, National Center for Research Resources Grant
RR-025761, and a Clinical and Translational Sciences Award.
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the author(s).
REFERENCES
1. Bauman W, Kahn N, Grimm D, Spungen A. Risk factors for atherogenesis and cardiovascular autonomic function in persons with spinal cord
injury. Spinal Cord 37: 601–616, 1999.
2. Bauman W, Spungen A. Disorders of carbohydrate and lipid metabolism
in veterans with paraplegia or quadriplegia: a model of premature aging.
Metabolism 43: 749 –756, 1994.
3. Bernardi M, Guerra E, Di Giacinto B, Di Cesare A, Castellano V,
Bhambhani Y. Field evaluation of paralympic athletes in selected sports:
implications for training. Med Sci Sports Exerc 42: 1200 –1208, 2010.
4. Brawarsky P, Brooks D, Wilber N, Gertz RJ, Klein Walker D.
Tobacco use among adults with disabilities in Massachusetts. Tob Control
11, Suppl 2: ii29 –ii33, 2002.
5. Buchholz A, McGillivray C, Pencharz P. Physical activity levels are low
in free-living adults with chronic paraplegia. Obes Res 11: 563–570, 2003.
6. Criqui M, Langer R, Fronek A, Feigelson H, Klauber M, McCann T,
Browner D. Mortality over a period of 10 years in patients with peripheral
arterial disease. N Engl J Med 326: 381–386, 1992.
301 • SEPTEMBER 2011 •
www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.32.247 on June 18, 2017
persist with the SCI. It is likely that exercise cannot ameliorate
the causes of worsened ABI in the lower-extremity arteries of
SCI. The effects of activity grouping were interesting when
comparing the normalized IMT of the lower-body arteries.
Similar to the ABI, the normalized PA-IMT had the best
outcome in the ABC-A group, followed by the ABC-S. The
normalized PA-IMT was not different when comparing active
versus sedentary SCI. The current data do suggest that upperbody exercise is a strong enough stimulus for improved arterial
structure in some arteries of SCI. In our SCI-A, both absolute
and normalized SFA-IMT were smaller than sedentary SCI.
These data suggest that the increased blood flow and shear
reported during upper-body exercise (34, 36, 37) may confer
some benefit in the SFA. The shear stimulus in the PA during
upper-body wheelchair exercise is currently not known. It is
possible that the distal location of this artery along with the
position of this limb during physical activity may alter this
hemodynamic pattern so that the benefit seen in SFA is not
apparent in PA.
Limitations. There are several limitations to this study that
may limit our ability to generalize these results to all populations of SCI. Blood analysis for lipids was not performed in
this study. We requested that blood lipids profiles obtained by
the subject’s physician within 12 months of the study date be
reported, but subject participation was low (n ⫽ 41). We
cannot rule out that differences in total, low-density lipoprotein, high-density lipoprotein, and triglyceride components,
which have been reported to be altered in the spinal cord injury
literature (1, 2, 28), may be driving the differences between our
groups. However, we did not find differences between groups
in CA-IMT measurements as would likely be the effect of
elevated blood lipids circulating systemically. Additionally,
subjects were categorized into active and sedentary groups
based on self-reported minutes of endurance training activity.
Although a parity of upper-body to lower-body exercise routines for atheroprotection is unlikely, recent reports for a
wheelchair team sports participation has shown that individuals
in these programs exercised at intensities that are within
appropriate ranges for improving cardiorespiratory fitness (3).
Furthermore, race differences existed between groups in our
study. In particular, SCI-S was represented by a larger percentage of African-American individuals compared with the three
other groups, including the SCI-S. Disability literature reflects
greater rates of physical inactivity in mobility-limited minorities than able-bodied minorities, mobility-limited nonminorities, or able-bodied nonminorities (15). These differences,
however, did not result in large differences in ABI or IMT
between Caucasian and African-American individuals in this
sample (Fig. 1, A–C). The lack of finding statistical differences
in the African-American ABI and IMT measurements appears
to be sample size related since patterns are quite similar.
Additionally, greater smoking behavior was reported by our
SCI with the greatest differences between active control and
SCI groups. Greater smoking behavior has been reported in
individuals with orthopedic disabilities (4) and American male
veterans with spinal cord injuries ages 25 to 44 (30), but to our
knowledge, there are no data that support the incidence of
greater smoking behavior in physically active SCI compared
with the general population. Also, with 50.4% of SCI in this
study having injury levels above the splanchnic innervation,
we cannot rule out the possibility that autonomic dysreflexia
H771
H772
PERIPHERAL ARTERIAL DISEASE IN SPINAL CORD INJURY
AJP-Heart Circ Physiol • VOL
23. Nishiyama S, Wray D, Richardson R. Aging affects vascular structure
and function in a limb-specific manner. J Appl Physiol 105: 1661–1670,
2008.
24. Orakzai S, Orakzai R, Ahmadi N, Agrawal N, Bauman W, Yee F,
Adkins R, Waters R, Budoff M. Measurement of coronary artery
calcification by electron beam computerized tomography in persons with
chronic spinal cord injury: evidence for increased atherosclerotic burden.
Spinal Cord 45: 775–779, 2007.
25. Pignoli P, Tremoli E, Poli A, Oreste P, Paoletti R. Intimal plus medial
thickness of the arterial wall: a direct measurement with ultrasound
imaging. Circulation 74: 1399 –1406, 1986.
26. Proctor D, Koch D, Newcomer S, Le K, Smithmyer S, Leuenberger U.
Leg blood flow and V̇O2 during peak cycle exercise in younger and older
women. Med Sci Sports Exerc 36: 623–631, 2004.
27. Rowley NJ, Dawson EA, Birk GK, Cable NT, George KP, Whyte G,
Thijssen DH, Green DJ. Exercise and arterial adaptation in humans:
uncoupling localized and systemic effects. J Appl Physiol 110: 1190 –
1195, 2011.
28. Schmid A, Halle M, Stutzle C, Konig D, Baumstark MW, Storch MJ,
Schmidt-Trucksass A, Lehmann M, Berg A, Keul J. Lipoproteins and
free plasma catecholamines in spinal cord injured men with different
injury levels. Clin Physiol 20: 304 –310, 2000.
29. Schmidt-Trucksäss A, Schmid A, Brunner C, Scherer N, Zäch G,
Keul J, Huonker M. Arterial properties of the carotid and femoral artery
in endurance-trained and paraplegic subjects. J Appl Physiol 89: 1956 –
1963, 2000.
30. Spungen A, Lesser M, Almenoff P, Bauman W. Prevalence of cigarette
smoking in a group of male veterans with chronic spinal cord injury. Mil
Med 160: 308 –311, 1995.
31. Stoner L, Sabatier M, VanhHiel L, Groves D, Ripley D, Palardy G,
McCully K. Upper vs. lower extremity arterial function after spinal cord
injury. J Spinal Cord Med 29: 138 –146, 2006.
32. Stoner L, Sabatier MJ, Mahoney ET, Dudley GA, McCully KK.
Electrical stimulation-evoked resistance exercise therapy improves arterial
health after chronic spinal cord injury. Spinal Cord 45: 49 –56, 2007.
33. Szlachcic Y, Carrothers L, Adkins R, Waters R. Clinical significance of
abnormal electrocardiographic findings in individuals aging with spinal
injury and abnormal lipid profiles. J Spinal Cord Med 30: 473–476, 2007.
34. Tanaka H, Shimizu S, Ohmori F, Muraoka Y, Kumagai M, Yoshizawa M, Kagaya A. Increases in blood flow and shear stress to
nonworking limbs during incremental exercise. Med Sci Sports Exerc 38:
81–85, 2006.
35. Theisen D, Vanlandewijck Y, Sturbois X, Francaux M. Central and
peripheral haemodynamics in individuals with paraplegia during light and
heavy exercise. J Rehabil Med 33: 16 –20, 2001.
36. Thijssen D, Green D, Steendijk S, Hopman M. Sympathetic vasomotor
control does not explain the change in femoral artery shear rate pattern
during arm-crank exercise. Am J Physiol Heart Circ Physiol 296: H180 –
H185, 2009.
37. Thijssen D, Steendijk S, Hopman M. Blood redistribution during exercise in subjects with spinal cord injury and controls. Med Sci Sports Exerc
41: 1249 –1254, 2009.
38. Thijssen DH, Kooijman M, de Groot PC, Bleeker MW, Smits P,
Green DJ, Hopman MT. Endothelium-dependent and -independent vasodilation of the superficial femoral artery in spinal cord-injured subjects.
J Appl Physiol 104: 1387–1393, 2008.
301 • SEPTEMBER 2011 •
www.ajpheart.org
Downloaded from http://ajpheart.physiology.org/ by 10.220.32.247 on June 18, 2017
7. de Groot P, Bleeker M, van Kuppevelt D, van der Woude L, Hopman
M. Rapid and extensive arterial adaptations after spinal cord injury. Arch
Phys Med Rehabil 87: 688 –696, 2006.
8. Deitrick G, Charalel J, Bauman W, Tuckman J. Reduced arterial
circulation to the legs in spinal cord injury as a cause of skin breakdown
lesions. Angiology 58: 175–184, 2007.
9. Garshick E, Kelley A, Cohen S, Garrison A, Tun C, Gagnon D, Brown
R. A prospective assessment of mortality in chronic spinal cord injury.
Spinal Cord 43: 408 –416, 2005.
10. Grew M, Kirshblum S, Wood K, Millis S, Ma R. The ankle brachial
index in chronic spinal cord injury: a pilot study. J Spinal Cord Med 23:
284 –288, 2000.
11. Grundy D, Silver J. Amputation for peripheral vascular disease in the
paraplegic and tetraplegic. Paraplegia 21: 305–311, 1983.
12. Haskell WL, Lee IM, Pate RR, Powell KE, Blair SN, Franklin BA,
Macera CA, Heath GW, Thompson PD, Bauman A, American College
of Sports Medicine; American Heart Association. Physical activity and
public health: updated recommendation for adults from the American
College of Sports Medicine and the American Heart Association. Circulation 116: 1081–1093, 2007.
13. Hopman M, Nommensen E, van Asten W, Oeseburg B, Binkhorst R.
Properties of the venous vascular system in the lower extremities of
individuals with paraplegia. Paraplegia 32: 810 –816, 1994.
14. Jae SY, Heffernan KS, Lee M, Fernhall B. Arterial structure and
function in physically active persons with spinal cord injury. J Rehabil
Med 40: 535–538, 2008.
15. Jones G, Sinclair L. Multiple health disparities among minority adults
with mobility limitations: san application of the ICF framework and codes.
Disabil Rehabil 30: 901–915, 2008.
16. Lange SF, Trampisch HJ, Pittrow D, Darius H, Mahn M, Allenberg
JR, Tepohl G, Haberl RL, Diehm C; getABI Study Group. Profound
influence of different methods for determination of the ankle brachial
index on the prevalence estimate of peripheral arterial disease. BMC
Public Health 7: 147, 2007.
17. Laughlin M, Roseguini B. Mechanisms for exercise training-induced
increases in skeletal muscle blood flow capacity: differences with interval
sprint training versus aerobic endurance training. J Physiol Pharmacol 59,
Suppl 7: 71–88, 2008.
18. Matos-Souza J, Pithon K, Ozahata T, Oliveira R, Téo F, Blotta M,
Cliquet AJ, Nadruz WJ. Subclinical atherosclerosis is related to injury
level but not to inflammatory parameters in spinal cord injury subjects.
Spinal Cord 48: 740 –744, 2010.
19. McAllister RM, Newcomer SC, Laughlin MH. Vascular nitric oxide:
effects of exercise training in animals. Appl Physiol Nutr Metab 33:
173–178, 2008.
20. McDermott M, Criqui M, Liu K, Guralnik J, Greenland P, Martin G,
Pearce W. Lower ankle/brachial index, as calculated by averaging the
dorsalis pedis and posterior tibial arterial pressures, and association with
leg functioning in peripheral arterial disease. J Vasc Surg 32: 1164 –1171,
2000.
21. McDonald C, Abresch-Meyer A, Nelson M, Widman L. Body mass
index and body composition measures by dual x-ray absorptiometry in
patients aged 10 to 21 years with spinal cord injury. J Spinal Cord Med 30,
Suppl 1: S97–S104, 2007.
22. Myers J, Lee M, Kiratli J. Cardiovascular disease in spinal cord injury:
an overview of prevalence, risk, evaluation, and management. Am J Phys
Med Rehabil 86: 142–152, 2007.