Page 16 - Cardiac testing for coronary artery disease in potential kidney transplant recipients

Dobutamine stress echocardiography (DSE)
stenosis and avoided partial verification were included, the pooled
DSE was compared with coronary angiography in 13 studies (745 sensitivity was 0.78 (95% CI 0.59 to 0.89), specificity 0.88 (95%
participants) (Bates 1996; Brennan 1997; Cai 2010; De Lima CI 0.76 to 0.94), positive likelihood ratio 6.44 (95% CI 3.03 to
2003; Ferreira 2007; Gang 2007; Garcia-Canton 1998; Herzog 13.70) and negative likelihood ratio 0.26 (95% CI 0.13 to 0.50)
1999; Jassal 2007; Reis 1995; Sharma 2005; Sharma 2009; West with pooled DOR 25.22 (95% CI 7.68 to 82.80) and AUC 0.90.
2000). Using induced wall motion abnormalities during dobu- Overall, there was very strong evidence of heterogeneity among the
tamine stress as a positive result indicating CAD, the sensitivity 13 studies (Figure 6). This remained highly statistically significant
of DSE varied from 44% to 96% and the specificity from 60% even after accounting for differences in reference standard thresh-
to 100%. Overall, DSE had a DOR of 29.98 (95% CI 12.17 to old (Figure 7) and partial verification (Figure 8). The remaining
73.89) and area under the curve (AUC) of 0.91 (95% CI 0.85 studies were similar in the performance of index test and interpre-
to 0.95). The pooled sensitivity was 0.79 (95% CI 0.67 to 0.88), tation of test results, but two studies (Sharma 2005; Sharma 2009)
specificity 0.89 (95% CI 0.81 to 0.94). One study also investigated were responsible for most of the heterogeneity. There was no sta-
the relationship between peak systolic velocity during DSE for tistical evidence of heterogeneity in six studies (De Lima 2003;
CAD (Sharma 2009). This study reported that ≥ 50% elevation Ferreira 2007; Gang 2007; Garcia-Canton 1998; Herzog 1999;
in peak systolic velocity with exercise during DSE was associated West 2000). Sharma 2005 and Sharma 2009 differed from other
with ≥ 70% stenosis on coronary angiography (sensitivity 86%, studies in that they originated from a single research group and
specificity 88%). had the highest proportion of patients who were symptomatic for
Not all patients who underwent index testing proceeded to have chest pain. Despite the hypothesis that prevalence of CAD may
these test results verified by the reference standard. Partial veri- have accounted for heterogeneity, we could not investigate any
fication was made in three studies (Bates 1996; Brennan 1997; relationship between diagnostic accuracy and prevalence of CAD
Cai 2010). Furthermore, four studies (Bates 1996; Brennan 1997; more formally because of the small number of studies, lack of
Jassal 2007; Reis 1995) used a reference test diagnostic threshold subgrouped patient data, and five studies (Bates 1996; Cai 2010;
of ≥ 50% stenosis. In the nine studies that used the higher thresh- Garcia-Canton 1998; Jassal 2007; West 2000) did not report pro-
old of ≥ 70% stenosis, the pooled sensitivity was 0.76 (95% CI portions of symptomatic patients. Two studies (Bates 1996; Gang
0.60 to 0.87) and specificity 0.88 (95% CI 0.78 to 0.94) with 2007) enrolled only patients with DM, and sensitivity was found
pooled DOR 23.01 (95% CI 8.08 to 65.51) and AUC 0.90. When to range from 47% to 90% and specificity from 86% to 95%.
only studies that applied a reference standard threshold of ≥ 70%

Cardiac testing for coronary artery disease in potential kidney transplant recipients (Review) 14
Copyright © 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.