You are on page 1of 10

Reviews/Commentaries/ADA Statements

R E V I E W A R T I C L E

Cardiac Imaging for Risk Stratification in


Diabetes
JEROEN J. BAX, MD1 EUGENE J. BARRETT, MD5 HOW IS CAD DIAGNOSED?
SILVIO E. INZUCCHI, MD2 ON BEHALF OF THE GLOBAL DIALOGUE The “gold standard” for detection of CAD
ROBERT O. BONOW, MD3 GROUP FOR THE EVALUATION OF remains invasive angiography with ves-
JOANNE D. SCHUIJF, MSC1 CARDIOVASCULAR RISK IN PATIENTS WITH sel-selective contrast injection of the cor-
MICHAEL R. FREEMAN, MD4 DIABETES onary arteries. Both spatial (0.2 mm) and
temporal (5 ms) resolution of the tech-
nique are extremely high, and the degree

W
orldwide, 200 million individu- often experience silent ischemia. In addi- of luminal narrowing can be quantified
als currently have diabetes, and tion, diabetic patients generally have a precisely. This is an invasive and expen-
projections by the World Health less favorable response to revasculariza- sive procedure with a small but definite
Organization and others suggest that its tion (with frequent need for repeat percu- risk for complications. Noninvasive test-
prevalence will exceed 300 million by taneous coronary intervention or coronary ing is increasingly used to assess CAD,
2025 and 360 million by 2030 (1,2). artery bypass grafting) and a reduced and multiple methods are now unavail-
More than 90% of these individuals will long-term survival. able. These can be divided into functional
have type 2 diabetes. Management guide- Accordingly, early accurate diagnosis imaging, which detects the hemodynamic
lines in Europe (3) and the U.S. (4) of CAD in patients with diabetes is consequences of CAD (i.e., ischemia),
consider type 2 diabetes to be a cardiovas- needed, and reliable prognostication is and anatomical imaging, which detects
cular disease equivalent. These patients mandatory. The American Diabetes Asso- atherosclerosis and permits direct visual-
have a two- to fourfold higher risk of a ciation has recommended an algorithm ization of the coronary arteries.
cardiovascular event than nondiabetic pa- whereby symptomatic diabetic patients
tients. Importantly, cardiovascular death would be referred for either stress perfu- Functional imaging
is the most common cause of mortality in
sion imaging or stress echo or evaluation The basis of functional imaging is the de-
the type 2 diabetic population (5). It has
by a cardiologist. The exception would be tection of CAD by assessing the hemody-
been estimated that after a myocardial in-
individuals with atypical chest pain and a namic consequences (i.e., ischemia) of
farction, 79% of diabetic patients die of
normal electrocardiogram who might un- CAD rather than direct visualization of
cardiac complications (6). Accordingly,
dergo a simple exercise stress test unless the coronary arteries. A sequence of
accurate cardiovascular risk stratification
they have multiple other cardiovascular events occurs during induction of isch-
of patients with type 2 diabetes is needed.
risk factors, in which case imaging studies emia, referred to as “the ischemic cascade”
This can be problematic in that the clini-
would be preferred (8). (9). Early (within seconds) in the isch-
cal presentation and progression of coro-
The purpose of the present review is emic cascade, perfusion abnormalities
nary artery disease (CAD) differs between
to discuss the available imaging tech- occur, and systolic wall motion abnor-
diabetic and nondiabetic patients. In ad-
niques in assessing CAD in symptomatic malities follow within 10 –20 s. Electro-
dition to a higher prevalence of CAD (7),
patients with diabetes (and compare ob- cardiogram (ECG) changes and angina
patients with diabetes experience more
servations to the accuracy of the techniques occur only at the end of the cascade. Ac-
diffuse, calcified, and extensive CAD,
in the general population). In addition, cordingly, exercise ECG is predictably not
more often have left ventricular dysfunc-
the most sensitive technique, and its di-
tion, often have more advanced coronary the issue of screening CAD in asymptom-
agnostic accuracy has been demonstrated
disease at the time of diagnosis, and more atic diabetic patients is discussed.
to be low in patients with diabetes (10).
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
Conversely, abnormalities in perfusion
From the 1Department of Cardiology, Leiden University Medical Center, Leiden, The Netherlands; 2Section and systolic wall motion are early markers
of Endocrinology, Department of Medicine, Yale University School of Medicine, New Haven, Connecticut;
3
Division of Cardiology, Northwestern University Feinberg School of Medicine, Chicago, Illinois; the 4Di- of ischemia. While perfusion abnormali-
vision of Cardiology, University of Toronto, Toronto, Ontario, Canada; and the 5Department of Internal ties should be the more sensitive of the
Medicine, University of Virginia, Charlottesville, Virginia. two for assessment of ischemia, in daily
Address correspondence and reprint requests to Jeroen J. Bax, MD, PhD, Department of Cardiology, practice both phenomena are similarly
Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, Netherlands. E-mail: j.j.bax@lumc.nl.
Received for publication 11 October 2006 and accepted in revised form 18 January 2007.
sensitive.
Published ahead of print at http://care.diabetesjournals.org on 26 January 2007. DOI: 10.2337/dc06- A number of imaging techniques can
2094. assess myocardial perfusion, including
J.J.B. has received research grants from GE Healthcare and BMS Medical, and R.O.B. is a consultant for nuclear techniques (i.e., positron emis-
Bristol-Meyers Squibb Medical Imaging. sion tomography [PET] or single photon
Additional information for this article can be found in an online appendix available at http://dx.doi.org/
10.2337/dc06-2094. emission computed tomography
Abbreviations: CAD, coronary artery disease; EBCT, electron beam computed tomography; DIAD, De- [SPECT]), first-pass perfusion imaging
tection of Silent Myocardial Ischemia in Asymptomatic Diabetics; ECG, electrocardiogram; MRI, magnetic with magnetic resonance imaging (MRI),
resonance imaging; MSCT, multislice computed tomography; PET, positron emission tomography; SPECT, and myocardial contrast echocardiogra-
single photon emission computed tomography.
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion
phy. For assessment of systolic wall mo-
factors for many substances. tion, the following techniques are used:
© 2007 by the American Diabetes Association. two-dimensional stress echocardiogra-

DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007 1295


Cardiac imaging for risk stratification in diabetes

Table 1—Diagnostic accuracy imaging tests specificity would be the normalcy rate.
This is the percentage of normal SPECT
General population Diabetic patients studies in a population with a low likeli-
hood of CAD. SPECT has a normalcy rate
Sensitivity Specificity Sensitivity Specificity of 89% (based on pooled analysis of 10
Functional imaging (ref.) studies, 543 patients) (11). With the abil-
Nuclear imaging (11–14) 86 74 80–97 56–88 ity to acquire ECG-gated images, simulta-
Stress echocardiography (15–18) 71–84 82–93 81–82 54–88 neous assessment of regional and global
Contrast echocardiography (21,22) 89 63 89 52 function is obtainable, which increases
First-pass perfusion MRI (21) 84 85 NA NA diagnostic accuracy (12,13).
Stress cine MRI (21) 89 84 NA NA Considerably less information on di-
Anatomical imaging (ref.) agnostic accuracy is available in diabetic
CAC score NA NA NA NA patients, and studies specifically dedi-
MRI angiography (27) 72 86 NA NA cated to the diagnostic accuracy of nu-
MSCT angiography (28,29) 91 96 95 95 clear perfusion imaging in patients with
EBCT angiography (30) 87 91 NA NA diabetes are scarce. Kang et al. (14) eval-
Data are percentages. CAC, coronary artery calcium. NA, not available. uated 138 patients with diabetes who also
underwent invasive angiography and re-
ported a sensitivity of 86% with a lower
phy, cine stress MRI, and stress-gated also allow assessment of plaque composi- specificity of 56%. The normalcy rate,
SPECT or PET imaging. tion in the near future. however, was 89% (online appendix Fig.
Most importantly, for ischemia as- 2). Most important, the accuracy of
sessment, imaging needs to be performed SPECT was not different between patients
during stress and at rest. Comparison of with and without diabetes.
DIAGNOSTIC ACCURACY OF Stress echocardiography. Stress echo-
the stress and rest images reveals whether CAD
stress-inducible perfusion or systolic wall cardiography is the most frequently used
motion abnormalities are present, indi- technique to assess systolic wall motion.
Functional imaging
cating ischemia. The stress can be per- Both physical exercise and pharmacolog-
Nuclear imaging. In the clinical setting,
formed using bicycle or treadmill exercise ical stress can be used. Resting wall mo-
nuclear imaging (mainly with SPECT) is
or (in patients unable to exercise) phar- tion abnormalities mainly represent
the most frequently used technique to as-
macological agents. Pharmacological infarcted myocardium, while those in-
sess perfusion as a marker of CAD (Table
stressors include dobutamine (a ␤-1– 1). Three radiopharmaceuticals are used:
duced by stress reflect ischemia.
specific agonist), which increases heart In the general population, as com-
thallium-201, technetium-99m sesta-
rate, contractility, and arterial blood pres- mibi, and technetium-99m tetrofosmin. pared with invasive angiography, the sen-
sure, resulting in increased myocardial Two sets of images are obtained, after sitivity and specificity of exercise
oxygen demand, and adenosine (a direct stress and at rest. Perfusion defects can be echocardiography for the detection of
vasodilator) or dipyridamole, which act divided into reversible (stress-induced) CAD are 84 and 82%, respectively
indirectly by inhibiting cellular uptake defects (reflecting ischemia) and irrevers- (pooled analysis of 15 studies, 1,849 pa-
and breakdown of adenosine. ible (fixed) defects (indicating infarcted tients) (15). The sensitivity and specificity
Functional imaging performed using myocardium). An example is provided in of dobutamine stress echocardiography
gated SPECT (contrast) stress echocardi- online appendix Fig. 1 (available at http:// are 80 and 84%, respectively (pooled
ography and MRI allow integrated assess- dx.doi.org/10.2337/dc06-2094). analysis of 28 studies, 2,246 patients)
ment of perfusion and function at rest and In the general population, the sensi- (15). Though less extensively studied, the
after stress. tivity and specificity of SPECT for detec- sensitivity and specificity for dipyridam-
tion of CAD (defined typically as ⬎50% ole stress echocardiography (71 and 93%,
Anatomical imaging stenosis on coronary angiography) are 86 respectively, in 12 studies of a total of 818
Anatomical imaging assesses atheroscle- and 74%, respectively (based on pooled patients) appear comparable (16).
rosis by direct visualization of the coronary analysis of 79 studies, 8,964 patients), as Studies that specifically addressed the
arteries. The several imaging modalities compared with invasive angiography topic of detection of CAD with stress
available include MRI, multislice com- (11). These data reflect potential patient echocardiography in patients with diabe-
puted tomography (MSCT), and electron selection biases, as patients are referred tes are limited to a few with small num-
beam computed tomography (EBCT). for coronary angiography after abnormal bers of patients. Hennessy et al. (17)
Since the coronary arteries are small, tor- SPECT findings. In contrast, coronary an- evaluated 52 patients with diabetes with
tuous, and move substantially during the giography is usually not performed in pa- dobutamine stress echocardiography and
cardiac cycle, imaging remains technically tients with normal SPECT findings. This reported a sensitivity of 82% with a spec-
challenging. As a result, all techniques have post-test referral bias will artificially lower ificity of 54%. Elhendy et al. (18) evalu-
shortcomings and limitations, but with the specificity, as a higher percentage of ated 50 patients with diabetes and 240
recent and ongoing technical advances, patients with normal coronary angio- nondiabetic patients with stress echocar-
image quality and diagnostic accuracy are grams will have abnormal SPECT findings diography and invasive angiography. The
continuously improving. Besides nonin- in these studies than in the general popu- sensitivity and specificity in the patients
vasive angiography, these techniques may lation with no CAD. A better indicator for with diabetes were 81 and 85%, respec-

1296 DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007


Bax and Associates

tively, as compared with 74 and 87% in are currently available with MRI. Disad- further improve diagnostic accuracy.
the nondiabetic patients. vantages of the technique include the rel- Dedicated studies in patients with diabe-
Myocardial contrast echocardiography. atively high costs as well as the time- tes have not been published.
With recent developments in echocardio- consuming nature of the examination. Noninvasive angiography with MSCT.
graphic equipment and microbubble con- At present, MSCT is the technique of
trast agents, real-time perfusion imaging Anatomical imaging choice for noninvasive angiography (on-
is now feasible (19). The infused micro- Coronary artery calcium scoring. The line appendix Fig. 5, lower panels). The
bubbles remain in the vascular space until two computed tomography techniques, technique is simple, fast, and reproduc-
they dissolve, reflecting the microvascular EBCT and MSCT, both permit noninva- ible. The technique is rapidly developing,
circulation. As with nuclear perfusion im- sive detection and quantification of coro- and 64-slice MSCT is currently the clini-
aging, resting perfusion defects suggest nary artery calcium (online appendix Fig. cal standard. In the general population,
infarcted myocardium, whereas stress- 5, upper panels). The vast majority of stud- the sensitivity and specificity to detect
induced perfusion defects indicate isch- ies published have been performed with CAD are 91 and 96%, respectively (nine
emia (online appendix Fig. 3). The EBCT, which has a lower radiation dose studies, 542 patients) (28). The percent-
agreement between SPECT and myocar- and possibly superior reproducibility age of noninterpretable segments on 64-
dial contrast echocardiography for detec- (Table 1). The Agatston score is the pre- slice MSCT has varied from 0 to 12%,
tion of perfusion abnormalities is good ferred score to quantify coronary artery with a mean value of 4%.
(20). In the general population, the sen- calcium (23). Scores ⬍10 represent non- At present, one study has specifically
sitivity and specificity of contrast echocar- significant coronary artery calcium, 11– addressed diagnostic accuracy in patients
diography for the detection of CAD are 89 100 mild calcium, 101– 400 moderate with diabetes. Schuijf et al. (29) evaluated
and 63%, respectively (based on pooled calcium, 401–1,000 severe calcium, and 30 patients with type 2 diabetes. Signifi-
analysis of seven studies, 245 patients), as ⬎1,000 extensive calcium. Although the cant stenoses (ⱖ50% luminal narrowing)
compared with invasive angiography presence of coronary artery calcium is on MSCT were compared with invasive
(21). One study has specifically addressed closely correlated with the total athero- angiography. A total of 220 of 256 coro-
the value of contrast echocardiography in sclerotic burden, it is not predictive of sig- nary artery segments (86%) were inter-
the detection of CAD in patients with di- nificant coronary stenoses and is not site pretable on MSCT. In these segments,
abetes. Elhendy et al. (22) evaluated 128 specific (24). This approach is generally sensitivity and specificity for detection of
patients with contrast echocardiography; not used for diagnosing CAD, but rather coronary artery stenoses were both 95%.
in 101 (79%) patients, invasive angiogra- to provide an estimate of the total athero- When the uninterpretable segments were
phy detected CAD. The sensitivity and sclerotic burden for prognostic and risk included, sensitivity and specificity
specificity were 89 and 52%, respectively. stratification purposes (see below). Ob- dropped to 81 and 82%, respectively. Pa-
MRI. Myocardial perfusion is evaluated servational studies revealed that diabetic tients with diabetes frequently have ex-
by injecting a bolus of contrast agent fol- patients have significantly higher coro- tensive calcifications in the coronary
lowed by continuous data acquisition as nary artery calcium scores than nondia- arteries, and this hampers the interpreta-
the contrast passes through the cardiac betic patients (25). However, coronary tion of stenosis severity.
chambers and into the myocardium. Per- calcium scoring may be most valuable in Noninvasive angiography with EBCT.
fusion defects are characterized as regions risk stratification, in order to determine Due to high spatial and temporal resolu-
of low signal intensity within the myocar- the intensity of primary prevention treat- tion, this technique appears particularly
dium (online appendix Fig. 4). The high ments. In patients with diabetes, who are useful for the imaging of coronary arter-
spatial resolution of MRI permits differen- already considered a coronary risk equiv- ies. Instead of a mechanically rotating X-
tiation between subendocardial and alent and treated with secondary preven- ray tube (as with MSCT), X-rays are
transmural perfusion defects. Resting de- tion guidelines, assessment of advanced generated through an electron beam that
fects indicate infarction, and stress- obstructive CAD may be more relevant. is guided along a 210° tungsten target ring
induced defects indicate ischemia. In the Noninvasive angiography with MRI. in the gantry. As a result, a high-
general population, the sensitivity and For more than a decade, MRI has at- resolution image is acquired in 50⫺100
specificity for detection of CAD are 84 tempted to provide noninvasive images of milliseconds. In the general population,
and 85%, respectively (based on pooled the coronary arteries. While an initial re- the sensitivity and specificity to detect
analysis of 17 studies, 502 patients), as port in 39 patients suggested a sensitivity CAD are 87 and 91%, respectively (10
compared with invasive angiography and specificity of 90 and 92%, respec- studies, 583 patients) (30). No specific
(21). tively (26), additional reports were less studies in patients with diabetes are avail-
In addition to myocardial perfusion, optimistic. Recent developments, includ- able. Data are summarized in Table 1.
global and regional systolic left ventricu- ing free breathing, navigator techniques,
lar function can also be assessed with high and three-dimensional acquisition tech- DETECTION OF CAD
accuracy using MRI. As with stress echo- niques, permit superior visualization of
cardiography, resting systolic wall motion the coronary arteries. In the general pop- Functional versus anatomical
abnormalities indicate infarcted myocar- ulation, the sensitivity and specificity for imaging
dium and stress-induced abnormalities the detection of CAD are 72 and 86%, When interpreting the data above, it is
indicate ischemia. In the general popula- respectively (28 studies, 903 patients) important to realize that the original gold
tion, the sensitivity and specificity of (27). However, up to 30% of all segments standard (invasive angiography) defines
stress cine MRI are 89 and 84%, respec- had to be excluded due to uninterpret- CAD when stenoses ⱖ50% luminal nar-
tively (10 studies, 654 patients) (21). No ability. The introduction of three Tesla rowing are present. In contrast, the func-
specific studies in patients with diabetes imaging and newer contrast agents may tional imaging techniques define CAD as

DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007 1297


Cardiac imaging for risk stratification in diabetes

Table 2—Nuclear imaging studies on prognosis in symptomatic patients with diabetes (based on ref. 39)

Mean Hard events Hard events


Patients Abnormal follow-up in abnormal in normal
Year Author (ref.) (n) Tracer Stressor MPI (%) (months) MPI (%/year) MPI (%/year)
201
1987 Felsher et al. (66) 123 TL Exercise 56 36 4.8 1.3
1999 Kang et al. (14) 1,271 201
TL, MIBI Exercise, adenosine 41 24 ⫾ 8 3.9–7.9 1.2
2002 Schinkel et al. (67) 207 MIBI Dobutamine 64 49 ⫾ 29 6.6* 0.7*
2002 Giri et al. (40) 929 201
TL, MIBI Exercise, adenosine 48 36 ⫾ 18 5.0–6.4 3.6–3.9
2003 Berman et al. (68) 5,333 201
TL, MIBI Adenosine 37–62 27 ⫾ 9 4.7–9.0* 1.8–2.5
201
2004 Zellweger et al. (59) 911 TL, MIBI Exercise, adenosine 44–51 24 5.6–13.2 2.0–3.3
2004 Miller et al. (69) 2,998 201
TL, MIBI Exercise, adenosine, 60 70 ⫾ 42 3.6–5.9 NA
dipyridamole, dobutamine
Data are means ⫾ SD unless otherwise indicated. Hard events include cardiac death or nonfatal myocardial infarction. 201
TL, thallium-201 chloride; MIBI,
technetium-99m sestamibi; MPI, myocardial perfusion imaging; NA, not available. *Only cardiac death.

the induction of ischemia (reflected in studies are not performed for diagnostic size, increasing defect size, defect revers-
stress-induced perfusion or systolic func- but rather for prognostic purposes. The ibility, defects in multiple vascular terri-
tion abnormalities). It has been demon- prognostic value of these imaging modal- tories, increased tracer lung uptake, and
strated in various studies that stenoses ities is addressed below. transient ischemic dilatation of the left
ⱖ50% luminal narrowing are not always ventricle. Additionally, in patients who
associated with stress-inducible ischemia, PROGNOSIS OF CAD were unable to perform exercise and un-
while in some cases ⬍50% luminal nar- For prognostication, patients are gener- derwent pharmacological stress, the event
rowing may be. This has been highlighted ally classified into three categories. The rates of both normal and abnormal scans
recently by Salm et al. (31), demonstrat- low-risk patients are those with an annual were higher than in patients able to exer-
ing that almost 50% of the intermediate cardiac mortality ⬎1%; the high-risk pa- cise (online appendix Fig. 7, upper panel).
stenoses (40 –70% luminal narrowing) in tients are those with an annual cardiac The prognostic value of a normal scan
bypass grafts were not associated with mortality ⬎3% per year. Intermediate- is maintained over a long period. Schinkel
ischemia on SPECT. risk patients are considered those with an et al. (37) evaluated 531 patients with
With the introduction of noninvasive annual mortality between 1 and 3%. SPECT over a follow-up period of 8.0 ⫾
angiography, this problem has been re- A wealth of prognostic data has been 1.5 years. The authors reported an annual
emphasized. In addition to significant ste- gathered with nuclear imaging and stress cardiac death rate of 0.9%, with an annual
noses (ⱖ50% luminal narrowing), the echocardiography, whereas little prog- cardiac death/infarction rate of 1.2% in
computed tomography techniques also nostic data with the other functional im- the presence of a normal scan. This an-
identify stenoses ⬍50%. In general, these aging techniques are available. Also, nual rate of coronary events in patients
techniques detect any level of atheroscle- extensive prognostic data on coronary ar- with normal scans is much higher in those
rosis. Many of these lesions will not be tery calcium scoring are available, but vir- with diabetes as discussed below.
associated with stress-inducible ischemia. tually no prognostic data on noninvasive Further risk stratification became
Indeed, Schuijf et al. (32) recently evalu- angiography have been published. possible when gated SPECT was intro-
ated 114 patients with MSCT and SPECT duced. The work from Sharir et al. (38)
and demonstrated that 55% of the pa- Nuclear imaging demonstrated that integration of perfu-
tients with atherosclerosis on MSCT do The vast majority of studies on noninva- sion data with left ventricular ejection
not have ischemia on SPECT (online ap- sive imaging for prognosis have used fraction and end-systolic volume resulted
pendix Fig. 6). Similar percentages have SPECT; a meta-analysis of 31 studies in- in superior discrimination of low- and
been reported in other studies (33,34). cluding 69,655 patients was reported re- high-risk patients.
Thus, as a result of the recent availability cently (36). These data indicate that a Seven studies with ⬎100 patients
of noninvasive anatomical imaging, a par- normal SPECT study is associated with an each specifically addressed the prognostic
adigm shift in the definition of CAD is excellent prognosis. The average annual value of SPECT imaging in symptomatic
occurring, shifting away from stenosis se- hard event rate (cardiac death or myocar- patients with diabetes using either thalli-
verity and stress-inducible ischemia to dial infarction) was 0.85%; this number is um-201 and/or technetium-99m sesta-
atherosclerosis in general. In addition, pa- comparable with the annual event rate in mibi (Table 2) (39). Two studies used
tients with diabetes frequently have an- the general population without CAD. In pharmacological stress only, and the
other form of vascular malfunctioning, contrast, the annual hard event rate was other studies used either exercise or phar-
referred to as microvascular disease (35). 5.9% in patients with a moderate-severe macological stress. The prevalence of ab-
This is not assessed by anatomic imaging abnormal SPECT study. The likelihood of normal perfusion studies was high,
and may or may not be assessed with an event increases in parallel to the extent ranging from 37 to 64%. The results
functional imaging. of abnormalities on a SPECT study. Vari- clearly confirm the higher event rate in
Apart from the discussion on the op- ous predictive parameters on SPECT have the presence of an abnormal scan com-
timal definition of CAD, one needs to re- been identified; these include (with in- pared with a normal scan, similar to non-
alize that most noninvasive imaging creasing risk for events) small fixed defect diabetic patients. The event rate in the

1298 DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007


Bax and Associates

Table 3—Stress echocardiographic studies on prognosis in symptomatic patients with diabetes

Hard event in Hard event in


Abnormal stress Mean abnormal stress normal stress
Patients echocardiography follow-up echocardiography echocardiography
Year Author (ref.) (n) Stressor (%) (months) (%/year) (%/year)
2001 Elhendy et al. (48) 563 Exercise 60 36 4.7 1.5
2001 Bigi et al. (70) 259 Dobutamine, dipyridamole 42 24 ⫾ 22 7.9 3
2001 Marwick et al. (46) 937 Exercise, dobutamine 40 3.9 ⫾ 2.3 10 4
years
2001 Sozzi et al. (71) 396 Dobutamine 82 36 6.2 4.8
2003 D’Andrea et al. (72) 325 Dobutamine, dipyridamole 46 34 13.8 4.8
Data are means ⫾ SD unless otherwise indicated. Hard events include cardiac death or nonfatal myocardial infarction.

presence of a normal scan also appears Stress echocardiography was specifically studied by Kamalesh et al.
higher compared with the general popu- A large number of studies have used stress (47), who performed a follow-up study
lation. Giri et al. (40) evaluated 4,755 pa- echocardiography to assess prognosis in (mean 25 months) in 233 patients (144
tients (including 929 diabetic patients) the general population. Similar to nuclear nondiabetic and 89 diabetic) with a neg-
with SPECT; the patients were prospec- data, stress echocardiography can differ- ative stress echocardiogram. The diabetic
tively followed for 2.5 ⫾ 1.5 years. Eighty entiate between low- and high-risk pa- patients had a significantly higher inci-
hard events occurred in the diabetic pa- tients. A negative stress echocardiogram is dence of nonfatal infarctions (6.7 vs.
tients (8.6%, 39 deaths and 41 infarc- associated with an excellent prognosis. A 1.4%), with a higher annual hard event
tions), as compared with 172 (4.5%, 69 recent meta-analysis of 13 studies and rate (6.0 vs. 2.7%).
deaths and 103 infarctions) in the nondi- 32,739 patients reported an annual hard The issue of the warranty period of a
abetic patients. The event rate was high- event rate (death or myocardial infarc- normal study was addressed by Elhendy
est, both for diabetic and nondiabetic tion) of 1.2% for subjects with a normal et al. (48). The authors evaluated 563 pa-
patients, in the presence of reversible de- stress echocardiogram (43). In contrast, tients with diabetes with exercise echo-
fects in two or more vascular territories, the hard event rate for those with an ab- cardiography with follow-up of up to 5
with an infarction rate of 17.1% in the normal study was 7.0% (online appendix years. Although the 1-year event rate was
diabetic patients. Women with diabetes Fig. 7, lower panel). Importantly, a recent 0%, there was a gradual increase up to
and ischemia on SPECT in two or more study demonstrated a comparable prog- 7.6% at the 5-year follow-up. Consider-
vascular territories were at the highest nostic accuracy of nuclear imaging and ing an event rate ⬍1% indicative for a
risk, with a 3-year survival rate of 60% in stress echocardiography (44). Similar to low-risk group, the warranty period of a
diabetic women. The authors subse- the nuclear studies, the severity of abnor- normal stress echo is 2 years. In addition,
quently demonstrated that the SPECT re- malities determines the prognosis (44). the authors confirmed the high event rate
sults provided significant incremental Five studies with ⬎100 patients have in patients with multivessel abnormalities
prognostic value over the clinical vari- studied the prognostic value of stress on stress echocardiography. In the same
ables. They also observed that for subjects echocardiography in diabetic patients study, Elhendy et al. (48) confirmed the
with normal SPECT studies, the event with CVD symptoms using either exercise incremental prognostic value of stress
rates were significantly higher in diabetic or pharmacological stress (Table 3) (45). echocardiography over clinical variables.
than in nondiabetic patients. The cardiac The prevalence of abnormal studies
death and infarction rates were 3.9 and ranged from 40 to 60%, in line with the Coronary artery calcium scoring
3.6%, respectively, in diabetic patients nuclear data. These results confirm the In the general population, extensive data
compared with 1.4 and 2.1%, respec- higher event rate in the presence of an have been gathered regarding the prog-
tively, in nondiabetic patients. When the abnormal study compared with a normal nostic value of coronary artery calcium
survival curves for patients with a normal study, similar to nondiabetic patients (on- but mainly in asymptomatic individuals.
SPECT were compared, survival was line appendix Fig. 8, middle panel). The In one of the largest studies thus far, more
comparable for the first 2 years after the largest cohort of diabetic patients under- than 10,000 asymptomatic patients were
SPECT study (online appendix Fig. 8, up- going stress echocardiography has been evaluated with EBCT and followed for the
per panels). Thereafter, however, diabetic published by Marwick et al. (46). These occurrence of all-cause death for 5 years
patients exhibited a sharp increase in authors evaluated the prognostic value of (49). In patients without or with minimal
events. This could possibly be explained stress echocardiography in 937 diabetic coronary artery calcification, excellent
by the more rapid progression in athero- patients. As observed with nuclear perfu- survival (99%) was demonstrated. In con-
sclerosis in patients with diabetes (41). sion studies, survival was related to trast, a 5-year all-cause mortality of
Based on this observation, Hachamovitch whether the patients were able to exer- 12.3% was witnessed in patients with ex-
et al. (42) proposed that the “warranty cise, with those not able having a worse tensive (⬎1,000) coronary artery calcifi-
period” of a normal scan may be limited in survival (online appendix Fig. 8, lower cation. Importantly, risk-adjusted
high-risk subsets (e.g., diabetic patients); panel). analysis revealed that coronary artery cal-
these patients may need repeat testing af- This issue of a higher event rate with a cium provided information incremental
ter 2 years. normal study in patients with diabetes to traditional risk assessment. In individ-

DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007 1299


Cardiac imaging for risk stratification in diabetes

Table 4—Evidence for (silent) ischemia or atherosclerosis in studies with asymptomatic diabetic patients (Only studies with >500 patients are
included.)

Patients Patient Abnormal


Author (ref.) (n) characteristics Technique study Details
Anand et al. (56) 510 Type 2 diabetes EBCT calcium scoring 46.3% 19.6% mild calcium (score 11–100 AU);
5.5% extensive calcium (score ⬎1,000
AU)
Sconamiglio et al. (58) 1,899 Type 2 diabetes MCE; dipyridamole 60% 59.4% of 1,121 patients with more than
two risk factors; 60% of 778 patients
with at least one risk factor
Wackers et al. (57) 522 Type 2 diabetes Nuclear imaging, SPECT; 21% 16% of perfusion abnormalities involved
adenosine, low-level ⬎5% of the left ventricle
exercise
Miller et al. (69) 1,738 Diabetic patients Nuclear imaging, SPECT; 59% 20% considered to represent high risk
exercise, pharmacologic
Zellweger et al. (59) 1,737 Diabetic patients Nuclear imaging, SPECT; 39–51% 39% of 826 asymptomatic patients; 51%
exercise, pharmacologic of 151 patients short of breath; 44% of
760 patients with angina
Rajagopalan et al. (60) 1,427 Diabetic patients Nuclear imaging, SPECT; 58% 20% considered to represent high risk
exercise, pharmacologic
MCE, myocardial contrast echocardiography.

uals with an intermediate risk (according tantly, in patients without coronary artery Detection of Silent Myocardial Ischemia
to the Framingham score), the 5-year calcium, survival was similar for individ- in Asymptomatic Diabetics (DIAD) study,
mortality was 1.1% for individuals with uals with and without diabetes (98.8 and showing a prevalence of 21% abnormal
minimal or no calcium, as compared with 99.4%, respectively). Qu et al. (53) per- SPECT studies. The perfusion defect in-
9.0% in individuals with a similar risk formed coronary artery calcium scoring volved ⬎5% of the left ventricle in 40% of
profile but extensive calcifications. Even in 1,312 high-risk individuals (with 269 patients with an abnormal SPECT study.
in patients with low risk (according to the diabetic patients) with an average fol- Of note, conventional risk factors did not
Framingham score), the coronary artery low-up of 6.3 years but failed to demon- predict perfusion abnormalities on
calcium score allowed further risk modi- strate the incremental value of coronary SPECT. A possible exception was the
fication, with a 3.9% mortality rate in in- artery calcium score over diabetes for pre- higher prevalence of cardiac neuropathy
dividuals with extensive calcifications as diction of events. Raggi et al. (54) pointed in patients with an abnormal SPECT
compared with 0.9% with minimal or no out that the discrepancy may be related to study.
calcifications. Accordingly, the coronary differences in sample size and risk profile Three additional studies used nuclear
artery calcium score provides incremental of the different studies. Accordingly, imaging to assess ischemia in asymptom-
prognostic information over traditional more studies are needed to determine atic diabetic patients and reported perfu-
risk stratification (50,51). Still, contro- whether calcium scoring allows more ro- sion abnormalities in 39 to 59% of
versy persists regarding the threshold for bust identification of high-risk patients patients (Table 4). One study used echo-
a calcium score that should be used to with diabetes compared with current risk cardiography with myocardial contrast to
designate increased risk. In contrast, assessment strategies. assess perfusion in 1,899 asymptomatic
absence of calcification is consistently as- diabetic patients (58). The population
sociated with excellent survival, empha- ASYMPTOMATIC DIABETIC was divided into patients with two or
sizing the power of this technique to PATIENTS more risk factors for CAD (n ⫽ 1,121) or
identify low-risk patients. Many diabetic patients with CAD are one or no risk factors (n ⫽ 778). Interest-
Thus far, limited data are available on asymptomatic or present with atypical ingly, the prevalence of perfusion abnor-
coronary artery calcium scoring in dia- symptoms (55). The prevalence of athero- malities was almost 60% and comparable
betic patients. In a large observational sclerosis was evaluated using EBCT in between both groups. In the patients with
study of 10,377 individuals, including 510 asymptomatic diabetic patients, and an abnormal contrast echocardiogram,
⬎900 asymptomatic diabetic patients, significant atherosclerosis (score ⬎10 Ag- invasive angiography was performed.
coronary artery calcium was the best pre- atston units) was noted in 46.3% (Table These results demonstrated that the se-
dictor of all-cause mortality in both dia- 4) (56). Various studies have evaluated verity of CAD was less in patients with
betic and nondiabetic individuals (52). the prevalence of silent ischemia (using one or no risk factor, with a lower preva-
Furthermore, a highly significant interac- either nuclear imaging or echocardiogra- lence of three-vessel disease (7.6 vs.
tion between coronary artery calcium phy) in both retrospective and prospec- 33.3%), diffuse CAD (18.0 vs. 54.9%),
score and diabetes was observed, with a tive settings (39). Wackers et al. (57) and vessel occlusion (3.8 vs. 31.2%).
greater increase in mortality rate for every evaluated 522 asymptomatic patients Overall, the widely differing estimates of
increase in calcium score in diabetic com- with at least two risk factors using gated CAD in asymptomatic patients most
pared with nondiabetic patients. Impor- technetium-99m sestamibi SPECT in the likely reflect differences in study design

1300 DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007


Bax and Associates

(retrospective vs. prospective) and inclu- presence of atherosclerosis with coronary present no data are available, but it is
sion criteria. artery calcium scoring using computed likely that a stepwise protocol as outlined
The prognostic value of nuclear imag- tomography techniques (either EBCT or above (EBCT first, followed by SPECT if
ing in asymptomatic diabetic patients has MSCT). In patients with extensive coro- needed) may be more cost-effective than
been addressed in few studies. Zellweger nary artery calcium, nuclear imaging with referring all patients to SPECT immedi-
et al. (59) studied three subsets of patients SPECT could be used to detect the pres- ately; data to support this hypothesis are
(without symptoms, with angina, and ence or absence of ischemia. A potential needed.
with dyspnea) and reported that the an- algorithm illustrating a stepwise screen-
nual hard event rates (cardiac death or ing approach is demonstrated in online
infarction) were approximately threefold appendix Fig. 9 (39). Based on the step- SUMMARY AND
higher in patients with abnormal SPECT wise approach, patients with severe ath- CONCLUSIONS
studies (5.4 vs. 1.9%). The event rates erosclerosis on EBCT (calcium score With the alarming worldwide increase in
were not different between asymptomatic ⬎400 AU) could be referred for SPECT. diabetes, and the associated high cardio-
patients and patients with angina. Simi- In patients with moderate calcium (be- vascular morbidity/mortality, adequate
larly, Rajagopalan et al. (60) studied tween 100 and 400 AU), referral may de- diagnostic tools are needed to detect CAD
1,427 asymptomatic diabetic patients and pend on the presence of certain patient and risk stratify patients. On the one
reported that the prevalence of abnormal characteristics or comorbidities, includ- hand, functional imaging tools (nuclear
SPECT scans was 58% with an annual ing the presence of metabolic syndrome, techniques, echocardiography, and MRI)
hard event rate of 5.9% for those with an duration of diabetes ⬎10 years, or reti- are available, which allow assessment of
abnormal scan versus 1.6% for those with nopathy, as patients with these character- ischemia. In general, which particular
a normal scan. In a smaller study, De istics may represent elevated risk, similar technique is preferred depends on local
Lorenzo et al. (61) reported an abnormal to those with extensive calcium scores. expertise and accordingly varies among
SPECT in 26% of 180 asymptomatic dia- Subsequently, in the presence of institutions. The choice for each tech-
betic patients, with annual hard event moderate-severe ischemia on SPECT, an- nique may vary among institutions, and
rates of 9 versus 2% for abnormal and giography could be considered, whereas local expertise may be the best guide. On
normal scans, respectively. those with small perfusion defects should the other hand, anatomical imaging tools
be clinically evaluated by a cardiologist (computed tomography techniques) are
whether invasive coronary angiography is now available, which allow assessment of
Should asymptomatic diabetic indicated or not. Patients without isch- atherosclerosis. Although there are less
patients undergo screening for CAD? emia should have aggressive medical ther- data concerning the diagnostic accuracy
Based on the high prevalence of athero- apy, risk factor modification, and careful of functional and anatomical testing in pa-
sclerosis and silent ischemia (Table 4), monitoring. This stepwise approach tients with diabetes, available information
and the high risk for cardiovascular needs further evaluation in future studies. suggests similar accuracies in diabetic pa-
events, the issue of screening for CAD in Moreover, before screening can be tients compared with the general popula-
asymptomatic diabetic patients has been advised, the following criteria need to be tion. The advantage of anatomical testing
raised and debated intensively met (63). 1) The prevalence in the popu- is that both obstructive and nonobstruc-
(39,55,62,63). lation should be high enough. The exact tive (subclinical) CAD can be visualized,
At present, the American Diabetes As- percentage of asymptomatic diabetic pa- allowing detection of atherosclerosis at an
sociation consensus guidelines for screen- tients with CAD is unknown; large retro- early stage. However, information on the
ing of asymptomatic patients recommend spective studies (59,60) reported homodynamic consequences of the de-
stress imaging in patients with abnormal abnormal SPECT studies in 39 and 58% tected lesions (needed to determine fur-
resting ECG (ischemia, infarction) but not of asymptomatic patients; the only pro- ther management) is not obtained.
in patients with, for example, cerebral/ spective study (DIAD) (57) reported Integration of these imaging techniques
peripheral vascular disease or two or 21%. 2) The screening test needs to accu- therefore may provide optimal informa-
more risk factors (8). In these latter cir- rately differentiate low- and high-risk pa- tion to guide patient management. In
cumstances, only an exercise test (ECG) is tients. In the diabetic population, SPECT asymptomatic patients with diabetes,
recommended, which is known to have a can identify the high-risk patients, but the studies have observed a considerably ele-
low diagnostic accuracy. Moreover, the low-risk patients cannot be identified ac- vated prevalence of silent ischemia and
available evidence has shown that many curately; patients with a normal SPECT atherosclerosis, suggesting the need for
diabetic patients with less than two con- study still had a fairly high event rate (i.e., screening in this population. However, no
ventional risk factors have perfusion ab- ⬎1% in the available studies) (59 – 61). 3) prospective data are currently available,
normalities on either nuclear imaging or Identification of asymptomatic diabetic and improved outcome based on screen-
contrast echocardiography (Table 4). Un- patients should lead to treatment with ing has not yet been demonstrated. Large,
fortunately, clinical variables (including better outcomes. At present, no prospec- randomized, prospective trials are there-
risk factors) do not predict which patients tive data on this topic are available, but fore required to determine the potential
will have an abnormal stress imaging re- the results from the DIAD study should role of screening asymptomatic patients
sult (57). However, nuclear imaging and provide some clues. In addition, data with diabetes for CAD.
stress echocardiography may not be the from the Mayo Clinic showed that pa-
ideal screening tools in terms of cost ef- tients with a high-risk SPECT study had
fectiveness. Anand et al. (56) have pro- better outcomes after CABG as compared Acknowledgments — The contributors to the
posed a stepwise screening approach— with medical therapy (64). 4) The screen- Global Dialogue on the Evaluation of Cardio-
first, patients are screened for the ing strategy should be cost-effective. At vascular Risk in Patients With Diabetes were

DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007 1301


Cardiac imaging for risk stratification in diabetes

Cliff Bailey, PhD, FRCP (Aston University, Bir- (Suppl. 2):S14 –S21, 2001 phy for the diagnosis of coronary artery
mingham, U.K.); Eugene Barrett, MD (Univer- 6. Jacoby RM, Nesto RW: Acute myocardial disease in diabetic patients unable to per-
sity of Virginia, Charlottesville, VA); Jeroen J. infarction in the diabetic patient: patho- form an exercise stress test. Diabetes Care
Bax, MD (Leiden University Medical Center, physiology, clinical course and prognosis. 21:1797–1802, 1998
Leiden, The Netherlands); Robert O. Bonow, J Am Coll Cardiol 20:736 –744, 1992 19. Elhendy A, O’Leary EL, Xie F, McGrain
MD (Northwestern University Feinberg 7. Hammoud T, Tanguay JF, Bourassa MG: AC, Anderson JR, Porter TR: Comparative
School of Medicine, Chicago, IL); Carlos A. Management of coronary artery disease: accuracy of real-time myocardial contrast
Buchpiguel (University of São Paulo Medical therapeutic options in patients with dia- perfusion imaging and wall motion anal-
School, São Paulo, Brazil); Terrance Chua, MD betes. J Am Coll Cardiol 36:355–365, 2000 ysis during dobutamine stress echocardi-
(National Heart Centre, Singapore); Alberto 8. Consensus development conference on ography for the diagnosis of coronary
Cuocolo, MD (University of Naples Federico the diagnosis of coronary heart disease in artery disease. J Am Coll Cardiol 44:2185–
II, Nuclear Medicine Center of the National people with diabetes: 10 –11 February 2191, 2004
Research Council, Naples, Italy); Michael R. 1998, Miami, Florida: American Diabetes 20. Jucquois I, Nihoyannopoulos P, D’Hondt
Freeman, MD (University of Toronto, To- Association. Diabetes Care 21:1551– AM, Roelants V, Robert A, Melin JA, Glass
ronto, Ontario, Canada); Silvio E. Inzucchi, 1559, 1998 D, Vanoverschelde JL: Comparison of
MD (Yale University School of Medicine, New 9. Nesto RW, Kowalchuk GJ: The ischemic myocardial contrast echocardiography
Haven, CT); Avijit Lahiri, MB, BS (Northwick cascade: temporal sequence of hemody- with NC100100 and (99m)Tc sestamibi
Park Hospital, Harrow, U.K.); Mario Ornelas namic, electrocardiographic and symp- SPECT for detection of resting myocardial
Arrieta, MD (Centro Médico Nacional Siglo tomatic expressions of ischemia. Am J perfusion abnormalities in patients with
XXI, Mexico City, Mexico); Paul Poirier, MD Cardiol 59:23C–30C, 1987 previous myocardial infarction. Heart 83:
(Institut Universitaire de Cardiologie et de 10. Wackers FJ: Diabetes and coronary artery 518 –524, 2000
Pneumologie, Québec, Canada); Gerard disease: the role of stress myocardial per- 21. Schuijf JD, Shaw LJ, Wijns W, Lamb HJ,
Slama, MD (University Pierre et Marie Curie, fusion imaging. Cleve Clin J Med 72:21– Poldermans D, de Roos A, van der Wall
Paris, France); Diethelm Tschöpe, MD (Uni- 33, 2005 EE, Bax JJ: Cardiac imaging in coronary
versität Bochum, Germany). 11. Underwood SR, Anagnostopoulos C, Cer- artery disease: differing modalities. Heart
The Global Dialogue on the Evaluation of queira M, Ell PJ, Flint EJ, Harbinson M, 91:1110 –1117, 2005
Cardiovascular Risk in Diabetes was sup- Kelion AD, Al Mohammad A, Prvulovich 22. Elhendy A, Tsutsui JM, O’Leary EL, Xie F,
ported by an educational grant by Bristol- EM, Shaw LJ, Tweddel AC: Myocardial McGrain AC, Porter TR: Noninvasive di-
Myers Squibb. perfusion scintigraphy: the evidence. Eur agnosis of coronary artery disease in pa-
J Nucl Med Mol Imaging 31:261–291, 2004 tients with diabetes by dobutamine stress
12. DePuey EG, Rozanski A: Using gated real-time myocardial contrast perfusion
References technetium-99m-sestamibi SPECT to imaging. Diabetes Care 28:1662–1667,
1. Wild S, Roglic G, Green A, Sicree R, King characterize fixed myocardial defects as 2005
H: Global prevalence of diabetes: esti- infarct or artifact. J Nucl Med 36:952–955, 23. Agatston AS, Janowitz WR, Hildner FJ,
mates for the year 2000 and projections 1995 Zusmer NR, Viamonte M Jr, Detrano R:
for 2030. Diabetes Care 27:1047–1053, 13. Smanio PE, Watson DD, Segalla DL, Vin- Quantification of coronary artery calcium
2004 son EL, Smith WH, Beller GA: Value of using ultrafast computed tomography. J
2. International Diabetes Federation [Web gating of technetium-99m sestamibi sin- Am Col lCardiol 15:827– 832, 1990
site]. Available at www.idf.org. Accessed gle-photon emission computed tomo- 24. O’Rourke RA, Brundage BH, Froelicher
28 August 2006 graphic imaging. J Am Coll Cardiol VF, Greenland P, Grundy SM, Hachamo-
3. De Backer G, Ambrosioni E, Borch- 30:1687–1692, 1997 vitch R, Pohost GM, Shaw LJ, Weintraub
Johnsen K, Brotons C, Cifkova R, Dallon- 14. Kang X, Berman DS, Lewin H, Miranda R, WS, Winters WL Jr: American College of
geville J, Ebrahim S, Faergeman O, Erel J, Friedman JD, Amanullah AM: Cardiology/American Heart Association
Graham I, Mancia G, Cats VM, Orth- Comparative ability of myocardial perfu- Expert Consensus Document on electron-
Gomer K, Perk J, Pyorala K, Rodicio JL, sion single-photon emission computed beam computed tomography for the diag-
Sans S, Sansoy V, Sechtem U, Silber S, tomography to detect coronary artery dis- nosis and prognosis of coronary artery
Thomsen T, Wood D: European guide- ease in patients with and without diabetes disease. J Am Coll Cardiol 36:326 –340,
lines on cardiovascular disease prevention mellitus. Am Heart J 137:949 –957, 1999 2000
in clinical practice: third joint task force of 15. Bax JJ, Van der Wall EE, De Roos A, 25. Schurgin S, Rich S, Mazzone T: Increased
European and other societies on cardio- Poldermans D: State of the art and future prevalence of significant coronary artery
vascular disease prevention in clinical directions. In Clinical Nuclear Cardiology. calcification in patients with diabetes. Di-
practice (constituted by representatives of 3rd ed. Zaret Bl, Beller GA, Eds. Philadel- abetes Care 24:335–338, 2001
eight societies and by invited experts). Eur phia, Mosby, 2005, p. 535–555 26. Manning WJ, Li W, Edelman RR: A pre-
J Cardiovasc Prev Rehabil 10:S1–S10, 2003 16. Picano E, Bedetti G, Varga A, Cseh E: The liminary report comparing magnetic res-
4. National Cholesterol Education Program/ comparable diagnostic accuracies of do- onance coronary angiography with
Adult Treatment Panel III: Third report of butamine-stress and dipyridamole-stress conventional angiography. N Engl J Med
the Expert Panel on Detection, Evalua- echocardiographies: a meta-analysis. Coron 328:828 – 832, 1993
tion, and Treatment of High Blood Cho- Artery Dis 11:151–159, 2000 27. Schuijf JD, Bax JJ, Shaw LJ, de Roos A,
lesterol in Adults (Adult Treatment Panel 17. Hennessy TG, Codd MB, Kane G, Mc- Lamb HJ, van der Wall EE, Wijns W:
III) [article online], 2002. Available from Carthy C, McCann HA, Sugrue DD: Eval- Meta-analysis of comparative diagnostic
http://www.nhlbi.nih.gov/guidelines/ uation of patients with diabetes mellitus performance of magnetic resonance imag-
cholesterol/atp3_rpt.htm. Accessed 30 for coronary artery disease using dobut- ing and multislice computed tomography
June 2006 (NIH Publ. no. 02-5215, Sep- amine stress echocardiography. Coron Ar- for noninvasive coronary angiography.
tember 2002) tery Dis 8:171–174, 1997 Am Heart J 151:404 – 411, 2006
5. Morrish NJ, Wang SL, Stevens LK, Fuller 18. Elhendy A, Van Domburg RT, Polder- 28. Bax JJ, Schuijf JD: Which patients should
JH, Keen H: Mortality and causes of death mans D, Bax JJ, Nierop PR, Geleijnse ML, be referred for non-invasive angiography
in the WHO Multinational Study of Vas- Roelandt JR: Safety and feasibility of do- with multi-slice CT? Int J Cardiol 114:1–3,
cular Disease in Diabetes. Diabetologia 44 butamine-atropine stress echocardiogra- 2007

1302 DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007


Bax and Associates

29. Schuijf JD, Bax JJ, Jukema JW, Lamb HJ, sion computed tomography. Circulation 50. Arad Y, Goodman KJ, Roth M, Newstein
Vliegen HW, Salm LP, de Roos A, van der 100:1035–1042, 1999 D, Guerci AD: Coronary calcification, cor-
Wall EE: Noninvasive angiography and 39. Scholte AJ, Bax JJ, Wackers FJ: Screening onary disease risk factors, C-reactive pro-
assessment of left ventricular function us- of asymptomatic patients with type 2 di- tein, and atherosclerotic cardiovascular
ing multislice computed tomography in abetes mellitus for silent coronary artery disease events: the St. Francis Heart
patients with type 2 diabetes. Diabetes disease: combined use of stress myocar- Study. J Am Coll Cardiol 46:158 –165,
Care 27:2905–2910, 2004 dial perfusion imaging and coronary cal- 2005
30. Budoff MJ, Achenbach S, Duerinckx A: cium scoring. J Nucl Cardiol 13:11–18, 51. Taylor AJ, Bindeman J, Feuerstein I, Cao
Clinical utility of computed tomography 2006 F, Brazaitis M, O’Malley PG: Coronary
and magnetic resonance techniques for 40. Giri S, Shaw LJ, Murthy DR, Travin MI, calcium independently predicts incident
noninvasive coronary angiography. J Am Miller DD, Hachamovitch R, Borges-Neto premature coronary heart disease over
Coll Cardiol 42:1867–1878, 2003 S, Berman DS, Waters DD, Heller GV: Im- measured cardiovascular risk factors:
31. Salm LP, Bax JJ, Jukema JW, Langerak SE, pact of diabetes on the risk stratification mean three-year outcomes in the Prospec-
Vliegen HW, Steendijk P, Lamb HJ, de using stress single-photon emission com- tive Army Coronary Calcium (PACC)
Roos A, van der Wall EE: Hemodynamic puted tomography myocardial perfusion project. J Am CollCardiol 46:807– 814,
evaluation of saphenous vein coronary ar- imaging in patients with symptoms sug- 2005
tery bypass grafts: relative merits of Dopp- gestive of coronary artery disease. Circu- 52. Raggi P, Shaw LJ, Berman DS, Callister
ler flow velocity and SPECT perfusion lation 105:32– 40, 2002 TQ: Prognostic value of coronary artery
imaging. J Nucl Cardiol 12:545–552, 2005 41. Wackers FJ, Zaret BL: Detection of myo- calcium screening in subjects with and
32. Schuijf JD, Wijns W, Jukema, JW, Atsma cardial ischemia in patients with diabetes without diabetes. J Am Coll Cardiol
DE, Lamb HJ, Stokkel MP, Dibbets- mellitus. Circulation 105:5–7, 2002 43:1663–1669, 2004
Schneider P, Decramer I, de Bondt P, van 42. Hachamovitch R, Hayes S, Friedman JD, 53. Qu W, Le TT, Azen SP, Xiang M, Wong
der Wall EE, Vanhoenacker PK, Bax JJ: Cohen I, Shaw LJ, Germano G, Berman ND, Doherty TM, Detrano RC: Value of
Relationship between non-invasive coro- DS: Determinants of risk and its temporal coronary artery calcium scanning by com-
nary angiography with multi-slice com- variation in patients with normal stress puted tomography for predicting coro-
puted tomography and myocardial myocardial perfusion scans: what is the nary heart disease in diabetic subjects.
perfusion imaging. J Am Coll Cardiol 48: warranty period of a normal scan? J Am Diabetes Care 26:905–910, 2003
2508 –2514, 2006 Coll Cardiol 41:1329 –1340, 2003 54. Raggi P, Bellasi A, Ratti C: Ischemia imag-
33. Hacker M, Jakobs T, Matthiesen F, Voll- 43. Schuijf JD, Poldermans D, Shaw LJ, ing and plaque imaging in diabetes: com-
mar C, Nikolaou K, Becker C, Knez A, Jukema JW, Lamb HJ, de Roos A, Wijns plementary tools to improve cardiovascular
Pfluger T, Reiser M, Hahn K, Tiling R: W, van der Wall EE, Bax JJ: Diagnostic risk management. Diabetes Care 28:
Comparison of spiral multidetector CT and prognostic value of non-invasive im- 2787–2794, 2005
angiography and myocardial perfusion aging in known or suspected coronary ar- 55. Nesto RW, Phillips RT, Kett KG, Hill T,
imaging in the noninvasive detection of tery disease. EurJ Nucl Med Mol Imaging Perper E, Young E, Leland OS Jr: Angina
functionally relevant coronary artery le- 33:93–104, 2006 and exertional myocardial ischemia in di-
sions: first clinical experiences. J Nucl Med 44. Schinkel AF, Bax JJ, Elhendy A, van Dom- abetic and nondiabetic patients: assess-
46:1294 –1300, 2005 burg RT, Valkema R, Vourvouri E, Boun- ment by exercise thallium scintigraphy.
34. Hacker M, Jakobs T, Hack N, Nikolaou K, tioukos M, Rizzello V, Biagini E, Agricola Ann Intern Med 108:170 –175, 1988
Becker C, von Ziegler F, Knez A, Konig A, E, Krenning EP, Simoons ML, Polder- 56. Anand DV, Lim E, Hopkins D, Corder R,
Klauss V, Reiser M, Hahn K, Tiling R: Six- mans D: Long-term prognostic value of Shaw LJ, Sharp P, Lipkin D, Lahiri A: Risk
ty-four slice spiral CT angiography does dobutamine stress echocardiography stratification in uncomplicated type 2 di-
not predict the functional relevance of compared with myocardial perfusion abetes: prospective evaluation of the com-
coronary artery stenoses in patients with scanning in patients unable to perform bined use of coronary artery calcium
stable angina. Eur J Nucl Med Mol Imaging exercise tests. Am J Med 117:1–9, 2004 imaging and selective myocardial perfu-
34:4 –10, 2006 45. Anand DV, Lim E, Lahiri A, Bax JJ: The sion scintigraphy. Eur Heart J 27:
35. Nitenberg A, Valensi P, Sachs P: Impair- role of non-invasive imaging in the risk 713–721, 2006
ment of coronary vascular reserve and stratification of asymptomatic diabetic 57. Wackers FJ, Young LH, Inzucchi SE,
ACh-induced coronary vasodilation in di- subjects. Eur Heart J 27:905–912, 2006 Chyun DA, Davey JA, Barrett EJ, Taillefer
abetic patients with angiographically nor- 46. Marwick TH, Case C, Sawada S, Vasey C, R, Wittlin SD, Heller GV, Filipchuk N,
mal coronary arteries and normal left Short L, Lauer M: Use of stress echocardi- Engel S, Ratner RE, Iskandrian AE: Detec-
ventricular systolic function. Diabetes 42: ography to predict mortality in patients tion of silent myocardial ischemia in
1017–1025, 2003 with diabetes and known or suspected asymptomatic diabetic subjects: the DIAD
36. Shaw LJ, Iskandrian AE: Prognostic value coronary artery disease. Diabetes Care 25: study. Diabetes Care 27:1954 –1961,
of gated myocardial perfusion SPECT. J 1042–1048, 2002 2004
Nucl Cardiol 11:171–185, 2004 47. Kamalesh M, Matorin R, Sawada S: Prog- 58. Scognamiglio R, Negut C, Ramondo A,
37. Schinkel AF, Elhendy A, Van Domburg nostic value of a negative stress echocar- Tiengo A, Avogaro A: Detection of coro-
RT, Bax JJ, Valkema R, Roelandt JR, diographic study in diabetic patients. Am nary artery disease in asymptomatic pa-
Poldermans D: Long-term prognostic Heart J 143:163–168, 2002 tients with type 2 diabetes mellitus. J Am
value of dobutamine stress 99mTc-sesta- 48. Elhendy A, Arruda AM, Mahoney DW, Coll Cardiol 47:65–71, 2006
mibi SPECT: single-center experience Pellikka PA: Prognostic stratification of 59. Zellweger MJ, Hachamovitch R, Kang X,
with 8-year follow-up. Radiology 225: diabetic patients by exercise echocardiog- Hayes SW, Friedman JD, Germano G,
701–706, 2002 raphy. J Am Coll Cardiol 37:1551–1557, Pfisterer ME, Berman DS: Prognostic rel-
38. Sharir T, Germano G, Kavanagh PB, Lai S, 2001 evance of symptoms versus objective evi-
Cohen I, Lewin HC, Friedman JD, Zell- 49. Shaw LJ, Raggi P, Schisterman E, Berman dence of coronary artery disease in
weger MJ, Berman DS: Incremental prog- DS, Callister TQ: Prognostic value of car- diabetic patients. Eur Heart J 25:543–550,
nostic value of post-stress left ventricular diac risk factors and coronary artery cal- 2004
ejection fraction and volume by gated cium screening for all-cause mortality. 60. Rajagopalan N, Miller TD, Hodge DO,
myocardial perfusion single photon emis- Radiology 228:826 – 833, 2003 Frye RL, Gibbons RJ: Identifying high-

DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007 1303


Cardiac imaging for risk stratification in diabetes

risk asymptomatic diabetic patients who coronary artery bypass grafting. Circula- nostic value and effect on patient manage-
are candidates for screening stress single- tion 112:I311–I316, 2006 ment. J Am Coll Cardiol 41:1125–1133,
photon emission computed tomography 65. Schuijf JD, Pundziute G, Jukema JW, 2003
imaging. J Am Coll Cardiol 45:43– 49, Lamb HJ, van der Hoeven BL, de Roos A, 69. Miller TD, Rajagopalan N, Hodge DO,
2005 van der Wall EE, Bax JJ: Diagnostic accu- Frye RL, Gibbons RJ: Yield of stress sin-
61. De Lorenzo A, Lima RS, Siqueira-Filho racy of 64-slice multislice computed to- gle-photon emission computed tomogra-
AG, Pantoja MR: Prevalence and prognos- mography in the noninvasive evaluation phy in asymptomatic patients with
tic value of perfusion defects detected by of significant coronary artery disease. Am J diabetes. Am Heart J 147:890 – 896, 2004
stress technetium-99m sestamibi myocar- Cardiol 98:145–148, 2006 70. Bigi R, Desideri A, Cortigiani L, Bax JJ,
dial perfusion single-photon emission 66. Felsher J, Meissner MD, Hakki AH, Heo J, Celegon L, Fiorentini C: Stress echocardi-
computed tomography in asymptomatic Kane-Marsch S, Iskandrian AS: Exercise ography for risk stratification of diabetic
patients with diabetes mellitus and no thallium imaging in patients with diabetes patients with known or suspected coro-
known coronary artery disease. Am J Car- mellitus: prognostic implications. Arch In- nary artery disease. Diabetes Care 24:
diol 90:827– 832, 2002 tern Med 147:313–317, 1987 1596 –1601, 2001
62. Bax JJ, Bonow RO, Tschope D, Inzucchi 67. Schinkel AF, Elhendy A, Van Domburg 71. Sozzi FB, Elhendy A, Roelandt JR, Van
SE, Barrett E: The potential of myocardial RT, Bax JJ, Vourvouri EC, Sozzi FB, Domburg RT, Schinkel AF, Vourvouri
perfusion scintigraphy for risk stratifica- Valkema R, Roelandt JR, Poldermans D: EC, Bax JJ, De Sutter J, Borghetti A,
tion of asymptomatic patients with type 2 Prognostic value of dobutamine-atropine Poldermans D: Prognostic value of dobut-
diabetes. J Am Coll Cardiol 48:754 –760, stress myocardial perfusion imaging in amine stress echocardiography in patients
2006 patients with diabetes. Diabetes Care 25: with diabetes. Diabetes Care 26:1074 –
63. Miller TD, Redberg R, Wackers FJ: 1637–1643, 2002 1078, 2003
Screening asymptomatic diabetic patients 68. Berman DS, Kang X, Hayes SW, Friedman 72. D’Andrea A, Severino S, Caso P, De Si-
for coronary artery disease: why not? J Am JD, Cohen I, Abidov A, Shaw LJ, Amanul- mone L, Liccardo B, Forni A, Pascotto M,
Coll Cardiol 48:761–764, 2006 lah AM, Germano G, Hachamovitch R: Di Salvo G, Scherillo M, Mininni N, Cala-
64. Sorajja P, Chareonthaitawee P, Rajago- Adenosine myocardial perfusion single- bro R: Prognostic value of pharmacologi-
palan N, Miller TD: Improved survival in photon emission computed tomography cal stress echocardiography in diabetic
asymptomatic diabetic patients with in women compared with men: impact of patients. Eur J Echocardiogr 4:202–208,
high-risk SPECT imaging treated with diabetes mellitus on incremental prog- 2003

1304 DIABETES CARE, VOLUME 30, NUMBER 5, MAY 2007