Background: Patients with chronic kidney disease (CKD) are at increased risk of coronary artery disease (CAD) and adverse cardiac events. Screening for CAD is therefore an important part of preoperative evaluation for kidney transplant candidates. There is significant interest in the role of non-invasive cardiac investigations and their ability to identify patients at high risk of CAD.
Objectives: We investigated the accuracy of non-invasive cardiac screening tests compared with coronary angiography to detect CAD in patients who are potential kidney transplant recipients.
Search methods: MEDLINE and EMBASE searches (inception to November 2010) were performed to identify studies that assessed the diagnostic accuracy of non-invasive screening tests, using coronary angiography as the reference standard. We also conducted citation tracking via Web of Science and handsearched reference lists of identified primary studies and review articles.
Selection criteria: We included in this review all diagnostic cross sectional, cohort and randomised studies of test accuracy that compared the results of any cardiac test with coronary angiography (the reference standard) relating to patients considered as potential candidates for kidney transplantation or kidney-pancreas transplantation at the time diagnostic tests were performed.
Data collection and analysis: We used a hierarchical modelling strategy to produce summary receiver operating characteristic (SROC) curves, and pooled estimates of sensitivity and specificity. Sensitivity analyses to determine test accuracy were performed if only studies that had full verification or applied a threshold of ≥ 70% stenosis on coronary angiography for the diagnosis of significant CAD were included.
Main results: The following screening investigations included in the meta-analysis were: dobutamine stress echocardiography (DSE) (13 studies), myocardial perfusion scintigraphy (MPS) (nine studies), echocardiography (three studies), exercise stress electrocardiography (two studies), resting electrocardiography (three studies), and one study each of electron beam computed tomography (EBCT), exercise ventriculography, carotid intimal media thickness (CIMT) and digital subtraction fluorography (DSF). Sufficient studies were present to allow hierarchical summary receiver operating characteristic (HSROC) analysis for DSE and MPS. When including all available studies, both DSE and MPS had moderate sensitivity and specificity in detecting coronary artery stenosis in patients who are kidney transplant candidates [DSE (13 studies) - pooled sensitivity 0.79 (95% CI 0.67 to 0.88), pooled specificity 0.89 (95% CI 0.81 to 0.94); MPS (nine studies) - pooled sensitivity 0.74 (95% CI 0.54 to 0.87), pooled specificity 0.70 (95% CI 0.51 to 0.84)]. When limiting to studies which defined coronary artery stenosis using a reference threshold of ≥ 70% stenosis on coronary angiography, there was little change in these pooled estimates of accuracy [DSE (9 studies) - pooled sensitivity 0.76 (95% CI 0.60 to 0.87), specificity 0.88 (95% CI 0.78 to 0.94); MPS (7 studies) - pooled sensitivity 0.67 (95% CI 0.48 to 0.82), pooled specificity 0.77 (95% CI 0.61 to 0.88)]. There was evidence that DSE had improved accuracy over MPS (P = 0.02) when all studies were included in the analysis, but this was not significant when we excluded studies which did not avoid partial verification or use a reference standard threshold of ≥70% stenosis (P = 0.09).
Authors' conclusions: DSE may perform better than MPS but additional studies directly comparing these cardiac screening tests are needed. Absence of significant CAD may not necessarily correlate with cardiac-event free survival following transplantation. Further research should focus on assessing the ability of functional tests to predict postoperative outcome.