Page 10 - Cardiac testing for coronary artery disease in potential kidney transplant recipients

We identified a further 11 studies (Caglar 2006; Dahan 1995;
Dahan 1998; Dahan 2002; De Vriese 2009; Fujimoto 2006; Fukui which occurs when the index test is incorporated in a compos-
2005; Nishimura 2004; Ohtake 2005; Robinson 2007; Schmidt ite reference standard, often leading to overestimation of diagnos-
2001) that reported diagnostic test accuracy in patients with CKD. tic test accuracy, was not present in any study. No patients were
However, populations in these studies did not consist entirely of verified with a second or third reference standard because disease
patients who were being considered for kidney transplantation - status (CAD) was diagnosed only by coronary angiography. Dif-
patients on dialysis or with CKD who were not being considered ferential verification was therefore also avoided in all studies. The
for transplantation were also represented. These studies were ex- reference standard was not blinded to investigators in three studies
cluded from the review because we were unable to obtain separate that reported coronary angiography being undertaken although
data for potential kidney transplant recipients only from the au- results of non-invasive index test were known to the investigators
thors of these 11 studies. We excluded a total of 55 studies from (Brennan 1997; De Lima 2003; Gang 2007). It was unclear if
our review (see Characteristics of excluded studies). index test results were known at the time of coronary angiography
in seven studies (Bennett 1978; Cai 2010; Gowdak 2010; Jassal
Methodological quality of included studies
2007; Krawczynska 1988; Reis 1995; West 2000). In one study
Results of the validity assessment are depicted (Figure 2; Figure 3) (De Lima 2003, author communication), coronary angiography
for the 25 included studies, including Sharples 2004, which could results were known to investigators who interpreted the index test.
not contribute data. Only 10 included studies provided sufficient It was also unclear if coronary angiography results were known
information to enable scoring for the 11 nominated QUADAS at the time of the index test in eight studies (Bennett 1978; Cai
methodological items. Seven studies satisfied the QUADAS crite- 2010; Gang 2007; Gowdak 2010; Jassal 2007; Krawczynska 1988;
ria. All included studies satisfied the QUADAS criteria of includ- Rosario 2010; West 2000). Of the 25 included studies, 20 aimed
ing study populations that represented the intended target popu- to provide coronary angiography to all patients who underwent
lation (potential kidney transplant recipients) and an acceptable index testing. However, only some participants who underwent
reference standard (coronary angiography). Incorporation bias; index testing proceeded to the reference test in five studies (Bates
1996; Brennan 1997; Cai 2010; Krawczynska 1988; Reis 1995).

Cardiac testing for coronary artery disease in potential kidney transplant recipients (Review) 8
Copyright © 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.