Page 20 - Cardiac testing for coronary artery disease in potential kidney transplant recipients

suboptimal stress capacity. Sensitivity for this study was 1.0 (95%
Myocardial perfusion scintigraphy (MPS)
CI 0.29 to 1.0) and specificity 0 (95% CI 0 to 0.97). In Sharma
MPS was compared with coronary angiography in nine stud- 2005, which enrolled 125 participants, sensitivity was 0.36 (95%
ies (582 participants) (Boudreau 1990; De Lima 2003; Garcia- CI 0.21 to 0.54) and specificity 0.91 (95% CI 0.83 to 0.96).
Canton 1998; Garg 2000; Gowdak 2010; Krawczynska 1988; • One study (97 participants) (Rosario 2010) compared
Marwick 1990; Vandenberg 1996; Worthley 2003). Sensitivity of EBCT with coronary angiography. This study reported that
MPS varied from 29% to 100% and specificity from 31% to 88%. when a calcium score threshold of 1330.72 Agatston units was
The pooled summary estimates showed that MPS had a DOR used as a cut-off point, sensitivity was 0.64 (95% CI 0.43 to
6.69 (95% CI 2.35 to 19.03) and AUC 0.78 (95% CI 0.64 to 0.82) and specificity 0.65 (95% CI 0.53 to 0.76), using a
0.88). The pooled sensitivity was 0.74 (95% CI 0.54 to 0.87), reference standard threshold of ≥ 70% stenosis to diagnose
and specificity 0.70 (95% CI 0.51 to 0.84). CAD.
All but one study (Krawczynska 1988) avoided partial verification • One study (35 participants) (Vandenberg 1996) compared
bias. Twostudies(Garg 2000; Krawczynska 1988)usedathreshold exercise radionuclide ventriculography with coronary
of ≥ 50% stenosis and not the reference threshold of ≥ 70% angiography showing a sensitivity of 0.50 (95% CI 0.23 to 0.77)
stenosis. Whenthese studieswere removedfromthe analysis, DOR and a specificity of 0.67 (95% CI 0.43 to 0.85)
remained unchanged at 6.70 (95% CI 1.84 to 24.41) and AUC • One study (86 participants) (Marwick 1989) compared
0.78. The pooled sensitivity was 0.67 (95% CI 0.48 to 0.82), DSF with coronary angiography, showing a sensitivity of 0.78
specificity 0.77 (95% CI 0.61 to 0.88), with positive and negative (95% CI 0.61 to 0.90) and a specificity of 0.68 (95% CI 0.51 to
likelihood ratios of 2.89 (95% CI 1.39 to 5.99) and 0.43 (95% 0.79).
CI 0.23 to 0.80) respectively. • One study (105 participants) (Modi 2006) compared
There was very strong evidence of heterogeneity among the nine CIMT with coronary angiography, showing a sensitivity of 0.90
studies (Figure 6). Heterogeneity remained even after account- (95% CI 0.77 to 0.97) and a specificity of 0.78 (95% CI 0.66 to
ing for differences in reference standard threshold (Figure 7) and 0.87).
partial verification (Figure 8). Of the studies that had reference • Three studies (Garg 2000; Sharma 2005; Sharma 2009)
standards of ≥ 70% stenosis and avoided verification bias, four correlated echocardiography findings with CAD. Two studies
(Boudreau 1990; Garg 2000; Gowdak 2010; Vandenberg 1996) (Sharma 2005; Sharma 2009) used resting wall motion
enrolled only patients with DM. Heterogeneity among these four abnormality to define an abnormal index test. These studies,
studies of patients with diabetes remained strongly significant, al- which were performed by the same authors on similar
though heterogeneity of the other four studies (De Lima 2003; populations, had very similar sensitivity and specificity (Sharma
Garcia-Canton 1998; Marwick 1990; Worthley 2003) decreased 2005 reported sensitivity of 0.31 (95% CI 0.16 to 0.48) and
when they were excluded. One study (Worthley 2003) that em- specificity of 0.96 (95% CI 0.89 to 0.99); Sharma 2009 found
ployed tachycardia pacing in some patients to ensure diagnostic sensitivity of 0.33 (95% CI 0.19 to 0.49) and specificity of 0.95
MPS had a much higher sensitivity and specificity compared with (95% CI 0.89 to 0.98)). Sharma 2005 also compared mitral
the other studies and accounted for much of the remaining het- annular calcification and CAD and reported that this
erogeneity. echocardiographic finding had a sensitivity of 0.61 (95% CI
Meaningful investigation into whether prevalence of angina and/ 0.43 to 0.77) and specificity of 0.72 (95% CI 0.61 to 0.81).
or ischaemic heart disease symptoms on diagnostic test perfor- Garg 2000 used echocardiographic criteria of left ventricular
mance was not possible as four studies (Garcia-Canton 1998; Garg dysfunction or cardiomegaly to define test positivity, and
2000; Gowdak 2010; Krawczynska 1988) did not provide any in- reported sensitivity of 0.30 (95% CI 0.14 to 0.50) and specificity
formation regarding prevalence of angina or ischaemic heart dis- of 0.80 (95% CI 0.59 to 0.93).
ease symptoms in their study populations. • Three studies (Gang 2007; Garg 2000; Sharma 2005)
investigated resting ECG for CAD diagnosis. In these studies,
abnormal resting ECG was defined as the presence of
Other tests pathological Q waves, left ventricular hypertrophy, ST depression
≥ 1 mm, ST elevation ≥ 1 mm, T wave inversion or bundle
• Two studies (129 participants) (Bennett 1978; Sharma
branch block. However, results differed. Gang 2007 reported
2005) compared EST with coronary angiography. In Bennett
sensitivity of 0.47 (95% CI 0.24 to 0.71) and specificity of 0.43
1978, only 4/7 participants were able to achieve an adequate
(95% CI 0.22 to 0.66); Garg 2000 identified sensitivity of 0.70
heart rate and had a diagnostic exercise stress test; the three
(95% CI 0.58 to 0.80) and specificity of 0.96 (95% CI 0.80 to
remaining participants underwent non-diagnostic tests due to
Cardiac testing for coronary artery disease in potential kidney transplant recipients (Review) 18
Copyright © 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.