Prognostic Utility of Calcium Scoring as an Adjunct to

Stress Myocardial Perfusion Scintigraphy in End-Stage

Renal Disease
William E. Moody, PhD, MBChBa,*, Erica L.S. Lin, MBChBa, Matthew Stoodleya,
David McNulty, PhDa, Louise E. Thomson, MBChBb, Daniel S. Berman, MDb,
Nicola C. Edwards, PhD, MBChBa, Benjamin Holloway, MBChBa, Charles J. Ferro, MDa,
Jonathan N. Townend, MDa, and Richard P. Steeds, MD, MAa, On behalf of the
Birmingham Cardio-Renal Group

Coronary artery calcium score (CACS) is a strong predictor of adverse cardiovascular events
in the general population. Recent data confirm the prognostic utility of single-photon
emission computed tomographic (SPECT) imaging in end-stage renal disease, but whether
performing CACS as part of hybrid imaging improves risk prediction in this population is
unclear. Consecutive patients (n [ 284) were identified after referral to a university hospital
for cardiovascular risk stratification in assessment for renal transplantation. Participants
underwent technetium-99m SPECT imaging after exercise or standard adenosine stress in
those unable to achieve 85% maximal heart rate; multislice CACS was also performed
(Siemens Symbia T16, Siemens, Erlangen, Germany). Subjects with known coronary artery
disease (n [ 88) and those who underwent early revascularization (n [ 2) were excluded.
The primary outcome was a composite of death or first myocardial infarction. An abnormal
SPECT perfusion result was seen in 22% (43 of 194) of subjects, whereas 45% (87 of 194) had
at least moderate CACS (>100 U). The frequency of abnormal perfusion (summed stress
score ‡4) increased with increasing CACS severity (p [ 0.049). There were a total of 15
events (8 deaths, and 7 myocardial infarctions) after a median duration of 18 months
(maximum follow-up 3.4 years). Univariate analysis showed diabetes mellitus (Hazard ratio
[HR] 3.30, 95% CI 1.14 to 9.54; p [ 0.028), abnormal perfusion on SPECT (HR 5.32, 95% CI
1.84 to 15.35; p [ 0.002), and moderate-to-severe CACS (HR 3.55, 95% CI 1.11 to 11.35; p [
0.032) were all associated with the primary outcome. In a multivariate model, abnormal
perfusion on SPECT (HR 4.18, 95% CI 1.43 to 12.27; p [ 0.009), but not moderate-to-severe
CACS (HR 2.50, 95% CI 0.76 to 8.20; p [ 0.130), independently predicted all-cause death or
myocardial infarction. The prognostic value of CACS was not incremental to clinical and
SPECT perfusion data (global chi-square change [ 2.52, p [ 0.112). In conclusion, a
perfusion defect on SPECT is an independent predictor of adverse outcome in potential renal
transplant candidates regardless of the CACS. The use of CACS as an adjunct to SPECT
perfusion data does not provide incremental prognostic utility for the prediction of mortality
and nonfatal myocardial infarction in end-stage renal disease. Ó 2016 The Authors.
Published by Elsevier Inc. This is an open access article under the CC BY license (http://
creativecommons.org/licenses/by/4.0/). (Am J Cardiol 2016;117:1387e1396)

Renal transplantation remains the most successful and (CV) outcomes compared with maintenance dialysis.1 Even
cost-effective treatment for patients with end-stage renal after transplantation, however, patients remain at high risk of
disease (ESRD), significantly improving cardiovascular long-term CV complications. To ensure that graft survival is
not limited by premature CV death, both US and UK regu-
latory bodies recommend noninvasive CV assessment of
a
Birmingham Cardio-Renal Group, Department of Cardiology, Institute those transplant candidates with multiple risk factors or
of Cardiovascular Sciences, Nuffield House, Queen Elizabeth Hospital diabetes, although there is no clear guidance on which im-
Birmingham, Edgbaston; and bDepartment of Cardiac Imaging and Nuclear aging method to use.2,3 The current suggestion is to adopt an
Cardiology, S. Mark Taper Foundation Imaging Center Los Angeles, imaging protocol for CV risk stratification according to “best
California. Manuscript received November 2, 2015; revised manuscript local expertise.” Accordingly, many transplant centers
received and accepted February 8, 2016.
continue to use stress myocardial perfusion scintigraphy
Funding: Dr. Moody is supported by a British Heart Foundation Clin-
ical Research Fellowship (FS/11/17/28700).
because of longstanding data supporting its prognostic utility
See page 1395 for disclosure information. in ESRD.4e7 Despite this, myocardial perfusion scintigraphy
*Corresponding author: Tel: (þ44) (0)-121-371-6130; fax: (þ44) (0)- has poor positive predictive value for identifying coronary
121-371-4629. artery stenosis on invasive angiography.5,8 Moreover, it is
E-mail address: william.moody@nhs.net (W.E. Moody). not able to detect subclinical atherosclerosis, potentially

0002-9149/16/$ - see front matter Ó 2016 The Authors. Published by www.ajconline.org

Elsevier Inc. This is an open access article under the CC BY license (http://
creativecommons.org/licenses/by/4.0/).
http://dx.doi.org/10.1016/j.amjcard.2016.02.003
1388 The American Journal of Cardiology (www.ajconline.org)

Figure 1. Study consort diagram. *Six of 88 subjects (7%) excluded from the analysis because of previous coronary atheroma, PCI, or CABG underwent early
revascularization. Two further subjects without a baseline diagnosis of coronary atheroma underwent early revascularization (1 percutaneous coronary
intervention and 1 coronary artery bypass graft surgery) driven by the SPECT/CT result.

predisposing the patient in the longer term to subsequent was guided by the Strengthening the Reporting of Obser-
obstructive CV events. Hybrid single-photon emission vational Studies in Epidemiology statement.14
computed tomographic (SPECT)/CT imaging offers an Demographic and anthropometric data were collected on
attractive opportunity to combine anatomic measures of all patients through review of patient electronic records. In
coronary artery calcification alongside a functional assess- addition, a standard prescan assessment involving a detailed
ment of myocardial ischemia. Coronary artery calcium score patient interview was performed to obtain information on
(CACS) is a surrogate marker of atherosclerotic burden and a symptoms, CV risk factors, previous CV events, and
strong predictor of adverse CV events in subjects at inter- medication. A Duke pretest probability of coronary artery
mediate risk from the general population.9 The predictive disease (CAD) was calculated at the time of the imaging
role of CACS in subjects with ESRD, however, is less study.15 Routine hematology and biochemistry at the time of
certain.10,11 Despite the very high burden of coronary the test were also recorded. Diabetes mellitus (DM) was
calcification in this population, there is only a modest asso- defined as a fasting glucose >126 mg/dl, history of DM,
ciation between CACS and perfusion defects.12,13 In the diabetic nephropathy, or currently receiving hypoglycemic
present study, we hypothesize that CACS will provide an treatment. Hypertension was defined as an office blood
incremental benefit for the prediction of death and first pressure >140/90 mm Hg or currently taking antihyper-
myocardial infarction (MI) in patients with ESRD beyond tensive medication. Hypercholesterolemia was defined as a
that provided by perfusion defect scores on myocardial serum cholesterol of >193 mg/dl or currently taking lipid
perfusion imaging. reduction therapy. A history of CV disease was defined as
having any of the following: CAD (MI, previous percuta-
neous, or surgical revascularization), heart failure, stroke,
Methods
and peripheral vascular disease. Significant family history of
Consecutive patients with chronic kidney disease (CKD) CV disease was defined as a first degree relative with a
stage 4 to 5D were identified after referral to Queen Eliz- history of MI or ischemic stroke in men younger than
abeth Hospital Birmingham for CV risk stratification as part 55 years and in women younger than 65 years.
of a pretransplant screening work-up from January 2011 to Patients were asked to discontinue b blockers, rate-
December 2013. In accordance with current guidelines limiting calcium channel blockers, and caffeine products
compiled by a Joint Working Party of The British Trans- 24 hours before testing, and nitrate compounds were dis-
plantation Society and The Renal Association,3 subjects continued >6 hours before testing. All participants under-
were referred for noninvasive CV risk assessment if they went 2-day stress-rest technetium-99m SPECT imaging
fulfilled any of the following criteria: age 50years, dia- with exercise treadmill or standard adenosine stress (140 mg/
betes, suspected angina, or known ischemic heart disease. kg/min for 6 minutes) in those unable to achieve 85%
Those subjects with a history of MI, coronary atheroma or maximal heart rate; and multislice CACS was performed as
stenosis on angiography, or previous percutaneous or sur- routine. CT-based attenuation correction was performed in
gical revascularization were excluded from the present study all patients during reconstruction of the SPECT data
(Figure 1). Formal ethical approval was not required (Symbia T16, Siemens, Erlangen, Germany).
because this study was a retrospective assessment of solely SPECT myocardial perfusion images were visually
clinical data and was therefore regarded as a health out- analyzed by 2 experienced observers (RPS and BH) blinded to
comes evaluation. The conduct and reporting of this study outcome variables (Quantitative Perfusion SPECT; Hermes
 Coronary Artery Disease/Prognostic Value of CACS in ESRD 1389

Medical Solutions, Stockholm, Sweden). In addition to ex- Disease, Tenth revision).23,24 With regard to outcome anal-
amination of raw images in cine mode, both nonattenuated ysis, HES data alone have the limitation of only capturing
and attenuated images were reviewed, and a report produced deaths occurring in a hospital setting. To obtain the complete
consistent with recommendations outlined in the American mortality list, the study cohort was also cross-referenced with
Society of Nuclear Cardiology Imaging Guidelines for Nu- mortality data from the Office for National Statistics, which
clear Cardiology Procedures.16 Short-axis and vertical long- collects information on all registered deaths in the UK. All
axis tomograms were divided into 17 segments for each outcomes were further verified by cross-referencing with in-
study,16 and segmental tracer uptake was evaluated using a dividual hospital case notes held electronically.
validated semiquantitative 5-point scoring system (0, normal; Statistical analyses were performed with Stata, version
1, equivocal; 2, moderate; 3, severe reduction of radioisotope 12 (StataCorp LP, College Station, Texas) and SAS (Sta-
uptake; and 4, absence of detectable tracer uptake).17 The tistical Analysis System, SAS Institute Inc., Cary, North
summed stress and rest scores were obtained by adding the Carolina). Data are expressed as mean  SD, median
scores of the 17 segments of the respective images. The sum of (interquartile range), or frequency (%), unless otherwise
the differences between each of the 17 segments from these stated. The normality of distribution for continuous vari-
images was defined as the summed difference score, repre- ables was determined using normality plots and the Kol-
senting the amount of ischemia. These indexes were con- mogoroveSmirnov test. Baseline characteristics of the
verted to the percentage of total myocardium involved with population were examined by CACS category and SPECT
stress, ischemic, or fixed defects by dividing the summed results. The KruskaleWallis analysis of variance was used
scores by 68 (the maximum potential score ¼ 4  17) and to identify significant differences in central tendencies of
multiplying by 100. The presence of abnormal perfusion was continuously scaled variables between groups. Contingency
defined as a summed stress score of 4 or greater.18 A stress- table analysis was performed using the chi-square or
induced total perfusion defect size (PDS) >15% or an Fisher’s exact tests where appropriate.
ischemic PDS >10% defined high risk for cardiac events.19 Annualized event rates are expressed as the number of
Cardiac volumes and left ventricular (LV) ejection fraction patients having first MI or all-cause death as a proportion of
were also calculated from the gated SPECT images. the number of patients at risk divided by the number of
The CACS was calculated according to Agatston et al20 patient-years follow-up. KaplaneMeier analysis of out-
by the same 2 independent observers blinded again to comes were based on discrete CACS categories (0 to 10, 11
outcome data. Lesions were manually traced on CT images to 100, 101 to 400, and >400 U) and SPECT categories
before semiautomatic quantification-derived vessel-specific (normal, total LV PDS dichotomized at 15%, ischemic PDS
scores were summated to yield the total CACS (syngo.via; dichotomized at 10%). The date of the imaging test was
Leonardo; Siemens Medical Solutions, Forchheim, Ger- used as time zero. In view of the beneficial CV effects of
many). Minimal, mild, moderate, and severe coronary renal transplantation, those patients undergoing renal
calcification were defined as Agatston scores of 0 to 10 U, transplantation were censored at the time of the procedure.25
11 to 100 U, 101 to 400 U, and >400 U, respectively.19 Two-sided log-rank tests were used to determine signifi-
The primary outcome was a composite of all-cause death cance. Univariate and multivariate Cox proportional hazards
or MI. Myocardial infarction was defined as a clinical (or models were used to identify the association between time-
pathologic) event caused by myocardial ischemia where to-event and baseline clinical characteristics, SPECT and
there is evidence of myocardial injury or necrosis as defined CACS results. Multivariate Cox regression analyses were
by an increase and/or decrease of cardiac biomarkers in the also repeated using follow-up data not censored for trans-
presence of typical symptoms or electrocardiographic plantation. The change in the global chi-square statistic was
changes, or imaging evidence of new loss of viable calculated to determine the incremental prognostic value of
myocardium or new regional wall motion abnormality.21 clinical, SPECT, and CACS data. A p value <0.05 was
Patients who had revascularization within 90 days of the considered statistically significant for all analyses.
imaging study were identified and excluded from the anal-
ysis to avoid inclusion of outcomes that may have been
Results
driven temporally by the SPECT/CT result.6 The event of
all-cause death was examined separately as a secondary In total, 284 consecutive patients (CKD stage 4 to 5D)
outcome. Patients who underwent renal transplant surgery with imaging performed from March 2011 to December
during the study period were also identified. 2013 were identified; of those, 88 had CAD at baseline. A
Patient follow-up data were retrieved by an observer further 2 subjects without a previous diagnosis of coronary
blinded to the clinical and imaging data (WEM). Every patient atheroma underwent early revascularization (1 coronary
in the National Health Service has a unique identifier which artery bypass graft surgery and 1 percutaneous coronary
enables outcomes to be tracked using the Hospital Episodes intervention) after SPECT demonstrated a reversible PDS
Statistics (HES) database, an administrative data warehouse 10%, leaving 194 subjects available for inclusion in the
containing admissions to all National Health Service hospitals present analysis (Figure 1).
in England.22 It contains detailed records relating to individ- The baseline characteristics of the study cohort are
ual patient treatments, with data extraction facilitated using summarized in Table 1. Mean age was 56 years, 60% were
codes on procedural classifications (Office of Population men, 33% were diabetic, and 82% were hypertensive. Most
Censuses and Surveys Classification of Interventions and patients were asymptomatic (75%). Two-thirds of patients
Procedures, Fourth revision) and medical classifications had at least mild CACS (65%), and over a quarter had se-
(World Health Organization International Classification of vere CACS (27%). In those with an abnormal SPECT result
1390 The American Journal of Cardiology (www.ajconline.org)

Table 1 (28%), almost half (42%) had a total PDS 15% and a third
Baseline demographics and clinical characteristics for study cohort (30%) had an ischemic PDS 10%.
Variable n ¼ 194 Patients with a large total or ischemic PDS were older,
less likely to be able to perform exercise treadmill stress and
Age (years) 56.3  10.2 more likely to have accompanying LV dysfunction
Male 117 (60%)
(Table 2). There was no difference in the mean number of
White 128 (66%)
Asian 49 (25%)
cardiac risk factors between subjects with a normal SPECT
Afro-Caribbean 12 (6%) result and those with a large perfusion defect.
Other ethnicity 4 (2%) As depicted in Table 3, subjects with a higher CACS were
Body mass index (kg / m2) 27.5  5.0 older and more frequently men and diabetic. There was a
Diabetes mellitus 64 (33%) graded association between increasing CACS and worsening
Hypertension* 159 (82%) LV function. There was no significant association between the
Hypercholesterolemia† 133 (69%) frequency of symptomatic chest pain and CACS severity.
Current smoker 36 (19%) Subjects with a normal SPECT result had a lower median
Family history of coronary artery 38 (20%) Agatston score compared to those with abnormal perfusion
disease
(35 U [IQR 0 to 349 U] vs 306 U [IQR 14 to 912 U]; p
Number of cardiac risk factors 2.3  1.0
Duke pre-test probability (%) 5 (3 e 8)
<0.01). There was a weak-graded association between the
Symptomatic chest pain 48 (25%) increasing proportion of patients with abnormal perfusion
Typical angina / atypical / 10 (5%) / 20 (10%) / 18 (9%) and increasing CACS severity (p ¼ 0.049; Figure 2). There
non-cardiac was, however, no significant association between CACS
Hemoglobin (g/ L) 111  16 severity and the frequency of a large stress-induced total
Total cholesterol (mg / dL) 185  46 (10%) or ischemic (10%) PDS. An abnormal SPECT
Calcium (mg / dL) 9.00  0.64 result was observed in 12% of subjects (8 of 68) with a
Phosphate (mg / dL) 4.30  1.24 CACS 0 to 10 and in 23% of subjects (9 of 39) with a CACS
Parathyroid hormone, (median pg / mL 21.8 (13.1 e 39.9) 11 to 100 U. In 4% of patients with only minimal CACS (3
[IQR])
of 68), a high-risk SPECT profile was demonstrated based
Uric acid (mg / dL) 7.13  1.98
CACS (median Agatston units [IQR]) 52 (0 e 509)
on the stress-induced total PDS.
CACS severity There were a total of 15 primary events (8 deaths and 7
0 e 10 68 (35%) MIs) after a median duration of 18 months (maximal follow-
11 e 100 39 (20%) up 3.4 years). Forty-one patients (21%) underwent renal
101 e 400 35 (18%) transplantation during the study period, one of whom died.
>400 52 (27%) This posttransplant death occurred 3 months after surgery in
Ability to perform exercise stress 112 (58%) a subject with hypertension and type 2 DM; SPECT/CT
METS achievedz 6.7  3.4 imaging had demonstrated severe CACS (2,376 U) but no
Stress electrocardiogram result 130 (67%) / 39 (20%) / 25 (13%) evidence of a perfusion defect. Two further patients who
Normal / Equivocal / Abnormal
underwent transplant suffered a nonfatal MI, 1 subject with
Left ventricular ejection fraction 56 (50 e 62)
(median % [IQR])
severe CACS >400 U and 1 subject with a detectable
Abnormal SPECTx 43 (22%) perfusion defect.
Total perfusion deficit score (% LV) 3.9  8.9 Univariate predictors of the primary outcome were DM,
Ischemic perfusion deficit score (% LV) 1.6  3.8 abnormal perfusion on SPECT, and an Agatston score of
Total perfusion deficit score  15% 18 (9%) >100 U (all p <0.05; Table 4). In a multivariate model,
Ischemic perfusion deficit score  10% 13 (7%) abnormal perfusion on SPECT and diabetes, but not CACS
Medications independently predicted all-cause death/nonfatal MI. The
Aspirin 71 (37%) results from multivariate Cox regression analyses performed
Thienopyridine 9 (5%) using data not censored for transplantation showed no sig-
Beta-blocker 79 (41%)
nificant difference in the models shown (data not shown).
ACE inhibitor / angiotensin 86 (44%)
receptor blocker
The risk for all-cause death/nonfatal MI increased
Calcium channel blocker 97 (50%) significantly with the presence and extent of SPECT
Loop diuretic 66 (33%) abnormality (Figure 3) and with the presence of moderate-
Statin 123 (63%) to-severe CACS (Figure 4). In subjects with abnormal
Insulin 42 (22%) perfusion by SPECT (summed stress score 4), the annu-
alized event rate for the primary outcome of all-cause death/
Data are number (%) or mean  SD unless otherwise stated.
nonfatal MI was 13.8% versus 2.8% in those with normal
ACE ¼ angiotensin-converting enzyme; CACS ¼ coronary artery calcium
score; IQR ¼ interquartile range; LV ¼ left ventricular; METS ¼ metabolic
perfusion. Similarly, the incident rate of all-cause death/
equivalents of task; SPECT ¼ single-photon emission computed tomography. nonfatal MI was 12.8% for those subjects with moderate-to-
* Defined as an office blood pressure of >140/90 mm Hg or currently severe CACS compared with 7.6% in those with a CACS
taking antihypertensive medications. <100 U. The value of integrating SPECT and CACS results
†
Defined as a fasting serum cholesterol of >193 mg/dl or currently for risk prediction is depicted in Figure 5.
taking lipid reduction therapy. The incremental value of CACS and stress SPECT re-
z
In the 112 subjects capable of treadmill exercise. sults to predict the primary event over clinical data by global
x
Defined as a summed stress score of 4. chi-square analysis is depicted in Figure 6. There was a
 Coronary Artery Disease/Prognostic Value of CACS in ESRD 1391

Table 2
Baseline demographics, clinical characteristics, and stress test differences by single-photon emission computed tomography results (n ¼ 194)
Variable Normal PDS <15% PDS 15% p Value* IPDS <10% IPDS 10% p Value†
(n ¼ 151) (n ¼ 25) (n ¼ 18) (n ¼ 30) (n ¼ 13)

Age 56.0  10.2 53.9  10.7 62.8  7.3 0.01 54.8  10.0 65.0  8.1 <0.01
Male 88 (58%) 20 (80%) 8 (44%) 0.046 22 (73%) 7 (54%) 0.27
Diabetes mellitus 46 (31%) 15 (60%) 7 (39%) 0.02 10 (33%) 7 (54%) 0.22
Hypertension 124 (82%) 21 (84%) 13 (72%) 0.56 26 (87%) 8 (64%) 0.14
Hypercholesterolemia 104 (69%) 17 (68%) 12 (67%) 0.98 20 (67%) 9 (73%) 0.97
Smoker 71 (47%) 11 (44%) 8 (44%) 0.95 15 (50%) 5 (36%) 0.78
Number of risk factors 2.3  0.1 2.4  1.0 2.4  1.4 0.55 2.4  1.0 2.3  1.5 0.66
Duke pre-test probability (%) 6 (3 e 8) 5 (3 e 8) 7 (3 e 18) 0.01 4 (3 e 7) 10 (4 e 20) 0.02
Symptomatic chest pain 35 (23%) 8 (32%) 5 (28%) 0.61 8 (27%) 5 (36%) 0.46
Ability to perform exercise stress 96 (64%) 11 (44%) 5 (28%) <0.01 13 (43%) 3 (23%) <0.01
LV ejection fraction (%) 57 (51 e 63) 55 (50 e 60) 46 (29 e 51) <0.001 51 (45 e 57) 50 (34 e 61) <0.001

Data are number (%), mean  SD or median (interquartile range).

IPDS ¼ ischemic perfusion defect size; PDS ¼ perfusion defect size.
* Normal SPECT versus total PDS <15%, total PDS 15%.
†
Normal SPECT versus ischemic PDS <10%, ischemic PDS 10%.

Table 3
Baseline demographics, clinical characteristics, and stress test differences by coronary artery calcium score severity
Variable CACS Severity Groups (n ¼ 194)

0 e 10 11 e 100 101 e 400 >400 P Value

(n ¼ 68) (n ¼ 39) (n ¼ 35) (n ¼ 52)

Age (years) 51.8  11.5 58.5  7.6 58.1  7.9 59.1  9.6 <0.001
Male 29 (43%) 26 (67%) 21 (60%) 42 (79%) <0.001
Diabetes mellitus 15 (22%) 12 (31%) 18 (51%) 19 (36%) 0.02
Hypertension 59 (87%) 33 (85%) 26 (74%) 42 (79%) 0.44
Hypercholesterolemia 46 (68%) 25 (64%) 25 (71%) 37 (70%) 0.88
Smoker 29 (43%) 22 (56%) 15 (43%) 24 (45%) 0.55
Number of risk factors 2.2  1.0 2.4  1.0 2.5  1.0 2.3  1.0 0.50
Duke pre-test probability (%) 4 (2 e 5) 5 (4 e 8) 5 (3 e 8) 7 (4 e 9) 0.06
Symptomatic chest pain 24 (35%) 11 (28%) 7 (20%) 8 (15%) 0.07
Ability to perform exercise stress 44 (65%) 23 (59%) 20 (57%) 26 (49%) 0.45
LV ejection fraction (%) 57 (54 e 64) 58 (49 e 62) 57 (48 e 64) 53 (44 e 59) 0.047

Data are number (%), mean  SD or median (interquartile range).

significant improvement in risk prediction with the addition adjusting for clinical data and the SPECT perfusion result.
of abnormal perfusion on SPECT to clinical information Most patients with ESRD had at least mild coronary calci-
(chi-square change ¼ 8.06, p ¼ 0.005). The prognostic fication (CACS >10 U), but there was a significant pro-
value of CACS was not incremental to clinical and SPECT portion (12%) with only minimal CAC who had an
perfusion data (global chi-square change ¼ 2.52, p ¼ abnormal SPECT perfusion result, which continued to
0.112). confer a higher event rate. This finding demonstrates that the
absence of CAC does not eliminate the potential for
obstructive CAD in ESRD.
Discussion
Our study is the first to identify that abnormal perfusion
This study suggests that quantification of CACS along- is the more important factor in identifying adverse CV event
side SPECT imaging does not provide incremental prog- rates in ESRD relative to the impact of CACS. One previous
nostic utility for prediction of mortality and nonfatal MI in study in 411 patients with ESRD (86% dialysis dependent)
potential renal transplant candidates. SPECT imaging identified a modest association between increasing CACS
continued, however, to be a useful method in identifying and abnormal perfusion, as found in our study, but did not
those subjects with ESRD at high CV risk. In those with examine the association with clinical outcomes.13 In general
abnormal perfusion, the risk for all-cause death/nonfatal MI population subjects without advanced CKD, there are con-
increased significantly with the presence and extent of flicting reports regarding the ability of hybrid imaging to
SPECT abnormality. Although a CACS >100 U was predict CV outcomes. Our data are consistent with those of
associated with a worse outcome, the presence of moderate- Rozanski et al26 which suggest that when perfusion is
to-severe CAC did not independently predict outcome after normal, elevated CACS does not confer an increased risk of
1392 The American Journal of Cardiology (www.ajconline.org)

Figure 2. Relation between CACS and SPECT results. Relation between CACS severity and stress SPECT results (n ¼ 194). The percentage of subjects with an
abnormal SPECT result significantly increased with increasing CACS severity (p ¼ 0.049). There was no significant association between the frequency of a
large stress-induced total (>15%) or ischemic (>10%) LV perfusion defect and CACS severity. Twelve percent of subjects with minimal CACS (8 of 68) had
abnormal perfusion on SPECT.

Table 4
Univariate and multivariate predictors of events
Variable Death or Non-fatal Myocardial Infarction All-cause Mortality

Univariate Analysis Multivariate Analysis Univariate Analysis Multivariate Analysis

HR (95% CI) P Value HR (95% CI) P Value HR (95% CI) P Value HR (95% CI) P Value

Age 0.99 (0.95 - 1.05) 0.829 0.98 (0.92 - 1.04) 0.463

Gender (female) 0.85 (0.29 - 2.44) 0.758 0.37 (0.09 - 1.55) 0.173
Diabetes 3.30 (1.14 - 9.54) 0.028 2.57 (0.87 - 7.59) 0.088 2.46 (0.61 - 9.87) 0.203 1.99 (0.48 - 8.20) 0.339
Current smoker 2.21 (0.74 - 6.62) 0.155 4.34 (1.08 - 17.41) 0.038
Hypercholesterolemia 0.63 (0.22 - 1.81) 0.390 0.79 (0.19 - 3.31) 0.746
LV ejection fraction < 55%* 2.44 (0.84 - 7.05) 0.099 3.20 (0.76 - 13.42) 0.112
Ability to exercise 0.31 (0.10 - 0.98) 0.046 0.45 (0.11 - 1.89) 0.275
Abnormal perfusion† 5.32 (1.84 - 15.35) 0.002 4.18 (1.43 - 12.27) 0.009 5.32 (1.84 - 15.35) 0.002 3.00 (0.72 - 12.46) 0.131
At least moderate CACSz 3.55 (1.11 - 11.35) 0.032 2.50 (0.76 - 8.20) 0.130 2.23 (0.53 - 9.4) 0.273 1.62 (0.37 - 7.13) 0.524

Multivariate regression models were adjusted for age, gender, and diabetes.
* Defined by gated single-photon emission computed tomography imaging.
†
Defined as summed stress score 4.
z
Defined as coronary artery calcium score >100 U.

CV events. In a further study of 695 consecutive subjects in subjects with a normal perfusion result.19 The relative
with intermediate risk, abnormal perfusion was associated increase in all-cause death/MI was limited to those with
with adverse CV events even in those subjects with no CACS >400 U and survival curves only began separating
calcification, albeit with a lower event rate than in those after 3 years, raising the possibility that the impact of CACS
subjects with higher CACS.27 However, in a study of 1,126 on outcome may only be seen after longer follow-up than in
largely asymptomatic patients, after a much longer duration our study.
of follow-up (median 6.9 years), Chang et al were able to A second possible explanation for the failure of CACS to
demonstrate that CACS offered incremental risk prediction provide incremental risk predictive value over SPECT in the
 Coronary Artery Disease/Prognostic Value of CACS in ESRD 1393

Figure 3. KaplaneMeier curves comparing time to death or first MI according to stress SPECT results: (A) Perfusion abnormality; (B) total PDS; and (C)
ischemic PDS. Two-sided log-rank tests were used to determine significance.
1394 The American Journal of Cardiology (www.ajconline.org)

Figure 4. KaplaneMeier curves comparing time to death or first MI according to the presence or absence of severe CACS. Two-sided log-rank tests were used
to determine significance.

Figure 5. KaplaneMeier curves comparing time to death or first MI according to integrated results of SPECT/CT. p Value shown corresponds to a significance
difference between all 4 survival curves. There is also a significant difference in the survival curves for “abnormal perfusion/CACS 100” and “abnormal
perfusion/CACS >100” (p <0.01). Two-sided log-rank tests were used to determine significance.

present study relates to the pathophysiology of arterial risk by confirming the absence of calcification,28 but pa-
calcification in ESRD.11 One of the major uses of CACS in tients with ESRD represent a different challenge. Our study
the general population has been to identify those at very low is consistent with others in identifying a remarkably high
 Coronary Artery Disease/Prognostic Value of CACS in ESRD 1395

Figure 6. Incremental predictive value of CACS and stress SPECT results over clinical information. The clinical data entered into the global chi-square analysis
model included age, gender, and the presence or absence of diabetes. Abnormality on SPECT (defined as SSS >4) and at least moderate calcification (CACS
>100 U) were entered as binary variables.

prevalence of moderate and severe CACS, which may be a hospital abroad, although it would be unusual for patients
consequence of other factors including abnormal cal- on renal replacement therapy to leave the country,
ciumephosphate handling in ESRD rather than reflecting particularly around the time of work-up for potential
atherosclerosis alone. Indeed, a strong correlation between transplant. Age did not appear to have a significant in-
decreasing glomerular filtration rate and increasing CACS fluence on adverse outcomes in this cohort. This finding
has been demonstrated, such that 3 of 4 subjects with ESRD may in part, reflect the relatively narrow age range of our
have a CACS above the 75th centile for gender-and age- population. There are data that demonstrate traditional CV
matched subjects without ESRD.29 Moreover, arterial risk factors are very poor predictors of cardiac events in
calcification in ESRD is not limited to the intima (athero- ESRD.30 Annual CV mortality for those receiving main-
sclerosis) causing obstructive coronary disease but also af- tenance hemodialysis is from 10 to 20 times that of the
fects the media (arteriosclerosis), which is associated with general population, and younger adults have the greatest
pressure overload and heart failure.11 CACS using 16-slice increase in CV risk.1 Thus, time on maintenance dialysis
CT without noninvasive angiography is unable to discrim- rather than age may be a more important factor in pre-
inate between intimal and medial calcification, which may dicting adverse outcomes, and the lack of data on this
be a further factor contributing to the lack of data associ- variable is an important limitation of our analysis.
ating increasing CACS with an increased CV event rate in
ESRD.11
There are a number of limitations to our study. These Disclosure
data are from consecutive patients but recruited from a
The authors have no conflicts of interest to disclose.
single center with retrospective analysis. The relatively
low number of events during follow-up that was limited to
1. Jardine AG, Gaston RS, Fellstrom BC, Holdaas H. Prevention of car-
a median of 18 months (maximum 3.4 years) may have diovascular disease in adult recipients of kidney transplants. Lancet
impacted on our ability to demonstrate an independent 2011;378:1419e1427.
association of CACS with hard clinical outcomes. By 2. Lentine KL, Costa SP, Weir MR, Robb JF, Fleisher LA, Kasiske BL,
combining HES with Office for National Statistics data Carithers RL, Ragosta M, Bolton K, Auerbach AD, Eagle KA. Cardiac
disease evaluation and management among kidney and liver trans-
sources, our data linkage process created a complete plantation candidates: a scientific statement from the American Heart
dataset with regard to mortality. It is possible that events Association and the American College of Cardiology Foundation. J Am
may have been missed for those subjects admitted to Coll Cardiol 2012;60:434e480.
1396 The American Journal of Cardiology (www.ajconline.org)

3. Dudley C, Harden P. Joint UK Renal Association and British Trans- 18. Hachamovitch R, Berman DS, Shaw LJ, Kiat H, Cohen I, Cabico JA,
plant Society Clinical Practice Guidelines: Assessment of the Potential Friedman J, Diamond GA. Incremental prognostic value of myocardial
Kidney Transplant Recipient. Hampshire, UK: UK Renal Association, perfusion single photon emission computed tomography for the pre-
2010. diction of cardiac death: differential stratification for risk of cardiac
4. Venkataraman R, Hage FG, Dorfman T, Heo J, Aqel RA, de Mattos death and myocardial infarction. Circulation 1998;97:535e543.
AM, Iskandrian AE. Role of myocardial perfusion imaging in patients 19. Chang SM, Nabi F, Xu J, Peterson LE, Achari A, Pratt CM, Mah-
with end-stage renal disease undergoing coronary angiography. Am J marian JJ. The coronary artery calcium score and stress myocardial
Cardiol 2008;102:1451e1456. perfusion imaging provide independent and complementary prediction
5. Rabbat CG, Treleaven DJ, Russell JD, Ludwin D, Cook DJ. Prognostic of cardiac risk. J Am Coll Cardiol 2009;54:1872e1882.
value of myocardial perfusion studies in patients with end-stage renal 20. Agatston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M Jr,
disease assessed for kidney or kidney-pancreas transplantation: a meta- Detrano R. Quantification of coronary artery calcium using ultrafast
analysis. J Am Soc Nephrol 2003;14:431e439. computed tomography. J Am Coll Cardiol 1990;15:827e832.
6. Bhatti S, Hakeem A, Dhanalakota S, Palani G, Husain Z, Jacobsen G, 21. Alpert JS, Thygesen K, Antman E, Bassand JP. Myocardial infarction
Ananthasubramaniam K. Prognostic value of regadenoson myocardial redefinedea consensus document of the Joint European Society of
single-photon emission computed tomography in patients with different Cardiology/American College of Cardiology Committee for the
degrees of renal dysfunction. Eur Heart J Cardiovasc Imaging redefinition of myocardial infarction. J Am Coll Cardiol 2000;36:
2014;15:933e940. 959e969.
7. Hakeem A, Bhatti S, Dillie KS, Cook JR, Samad Z, Roth-Cline MD, 22. National Health Service HSCIC. Hospital Episodes Statistics Online
Chang SM. Predictive value of myocardial perfusion single-photon Database 2014. Available at: www.hesonline.nhs.uk. Accessed on
emission computed tomography and the impact of renal function on November 1, 2015.
cardiac death. Circulation 2008;118:2540e2549. 23. National Health Service. OPCS-4 Classification e NHS Connecting for
8. Wang LW, Fahim MA, Hayen A, Mitchell RL, Lord SW, Baines LA, Health 2014. Available at: www.connectingforhealth.nhs.uk. Accessed
Craig JC, Webster AC. Cardiac testing for coronary artery disease in on November 1, 2015.
potential kidney transplant recipients: a systematic review of test ac- 24. World Health Organisation. International Classification of Diseases
curacy studies. Am J Kidney Dis 2011;57:476e487. (ICD) 2014. Available at: www.who.int/classifications/icd/en.
9. Yeboah J, McClelland RL, Polonsky TS, Burke GL, Sibley CT, O’Leary Accessed on November 1, 2015.
D, Carr JJ, Goff DC, Greenland P, Herrington DM. Comparison of novel 25. Patel RK, Mark PB, Johnston N, McGeoch R, Lindsay M, Kingsmore
risk markers for improvement in cardiovascular risk assessment in DB, Dargie HJ, Jardine AG. Prognostic value of cardiovascular
intermediate-risk individuals. JAMA 2012;308:788e795. screening in potential renal transplant recipients: a single-center pro-
10. Fensterseifer DM, Karohl C, Schvartzman P, Costa CA, Veronese FJ. spective observational study. Am J Transplant 2008;8:1673e1683.
Coronary calcification and its association with mortality in haemo- 26. Rozanski A, Gransar H, Wong ND, Shaw LJ, Miranda-Peats R, Polk
dialysis patients. Nephrology (Carlton) 2009;14:164e170. D, Hayes SW, Friedman JD, Berman DS. Clinical outcomes after both
11. Bashir A, Moody WE, Edwards NC, Ferro CJ, Townend JN, Steeds coronary calcium scanning and exercise myocardial perfusion scintig-
RP. Coronary artery calcium assessment in chronic kidney disease: raphy. J Am Coll Cardiol 2007;49:1352e1361.
utility in cardiovascular disease risk assessment and treatment? Am J 27. Schenker MP, Dorbala S, Hong EC, Rybicki FJ, Hachamovitch R,
Kidney Dis 2015;65:937e948. Kwong RY, Di Carli MF. Interrelation of coronary calcification,
12. Jug B, Kadakia J, Gupta M, Papazian J, Derakhshani A, Koplik S, myocardial ischemia, and outcomes in patients with intermediate
Karlsberg RP, Budoff MJ. Coronary calcifications and plaque charac- likelihood of coronary artery disease: a combined positron emission
teristics in patients with end-stage renal disease: a computed tomo- tomography/computed tomography study. Circulation 2008;117:
graphic study. Coron Artery Dis 2013;24:501e508. 1693e1700.
13. Karohl C, D’Marco L, Bellasi A, Raggi P. Hybrid myocardial imaging 28. Greenland P, Bonow RO, Brundage BH, Budoff MJ, Eisenberg MJ,
for risk stratification prior to kidney transplantation: added value of Grundy SM, Lauer MS, Post WS, Raggi P, Redberg RF, Rodgers GP,
coronary calcium and epicardial adipose tissue. J Nucl Cardiol Shaw LJ, Taylor AJ, Weintraub WS. ACCF/AHA 2007 clinical expert
2013;20:1013e1020. consensus document on coronary artery calcium scoring by computed
14. von Elm E, Altman DG, Egger M, Pocock SJ, Gotzsche PC, Van- tomography in global cardiovascular risk assessment and in evaluation
denbroucke JP. The Strengthening the Reporting of Observational of patients with chest pain: a report of the American College of Car-
Studies in Epidemiology (STROBE) statement: guidelines for reporting diology Foundation Clinical Expert Consensus Task Force (ACCF/
observational studies. Epidemiology 2007;18:800e804. AHA writing committee to update the 2000 expert consensus document
15. Pryor DB, Harrell FE Jr, Lee KL, Califf RM, Rosati RA. Estimating the on electron beam computed tomography) developed in collaboration
likelihood of significant coronary artery disease. Am J Med 1983;75: with the Society of Atherosclerosis Imaging and Prevention and the
771e780. Society of Cardiovascular Computed Tomography. J Am Coll Cardiol
16. Holly TA, Abbott BG, Al-Mallah M, Calnon DA, Cohen MC, DiFi- 2007;49:378e402.
lippo FP, Ficaro EP, Freeman MR, Hendel RC, Jain D, Leonard SM, 29. Raggi P, Boulay A, Chasan-Taber S, Amin N, Dillon M, Burke SK,
Nichols KJ, Polk DM, Soman P. Single photon-emission computed Chertow GM. Cardiac calcification in adult hemodialysis patients. A
tomography. J Nucl Cardiol 2010;17:941e973. link between end-stage renal disease and cardiovascular disease? J Am
17. Berman DS, Kiat H, Friedman JD, Wang FP, van Train K, Matzer L, Coll Cardiol 2002;39:695e701.
Maddahi J, Germano G. Separate acquisition rest thallium-201/stress 30. Israni AK, Snyder JJ, Skeans MA, Peng Y, Maclean JR, Weinhandl
technetium-99m sestamibi dual-isotope myocardial perfusion single- ED, Kasiske BL. Predicting coronary heart disease after kidney
photon emission computed tomography: a clinical validation study. transplantation: Patient Outcomes in Renal Transplantation (PORT)
J Am Coll Cardiol 1993;22:1455e1464. study. Am J Transplant 2010;10:338e353.