This review assessed computed tomographic angiography for the evaluation of coronary artery bypass grafts. The authors concluded that single-slice computed tomography (CT) can be used to detect graft patency but not occlusion, while 4- and 16-slice CT can be used for the detection of both occlusion and patency. However, the findings should be interpreted with caution given the limitations of the review, in particular the lack of any formal quality assessment.

To review the sensitivity and specificity of computed tomographic angiography for the evaluation of coronary artery bypass grafts (CABGs).

MEDLINE, OLDMEDLINE, Ovid and the Cochrane Library were searched via PubMed (1950 to March 2005); the search terms were listed. Bibliographies of the included studies and recent reviews were screened for additional studies. Studies in any language were eligible for inclusion. Abstracts, case reports, letters, comments, reviews, animal studies and in vitro studies were excluded.

Study designs of evaluations included in the review

The studies had to include more than 10 patients.

Specific interventions included in the review

Studies of multi-slice computed tomography (CT) were eligible for inclusion. The studies had to describe the type of imaging unit used for CT to be included. The readers of the CT scans had to be blinded to the results of the selective angiography. Significant stenosis was defined as greater than 50% stenosis, while high-grade stenosis was defined as greater than 75% stenosis. Investigations limited to methods or techniques and electron-beam CT were excluded. The studies included in the review assessed single-slice, 4-slice and 16-slice CT.

Reference standard test against which the new test was compared

Studies that used selective angiography of the CABG as the reference standard were eligible for inclusion.

Participants included in the review

Studies of patients who had undergone CABGs (arterial and or saphenous vein grafts) were eligible for inclusion.

Outcomes assessed in the review

The studies had to report sufficient data to calculate the sensitivity and specificity of CT. The outcomes reported in the review were the sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios (LRs).

How were decisions on the relevance of primary studies made?

The electronic search and reference lists were reviewed. Full articles were obtained if a determination of the inclusion criteria could not be made from the title or abstract.

The authors did not state that they assessed validity.

The authors did not state how the data were extracted for the review, or how many reviewers performed the data extraction. The sensitivity, specificity, and positive and negative LRs were calculated. Ninety-five per cent confidence intervals (CIs) around these estimates were calculated, based on binomial expansion. LRs were only calculated if studies reported the data used for the calculation of both sensitivity and specificity.

How were the studies combined?

The authors stated that the average sensitivity and specificity were based on pooled data, but did not provide further details on how the pooling was carried out. Based on the data reported in the table, it appears that simple pooling was based on adding the numbers of true positives, false positives, true negatives and false negatives across studies and then using these to calculate pooled values of sensitivity, specificity, positive and negative predictive values, and positive and negative LRs. The chi-squared test was used to compare proportions.

How were differences between studies investigated?

The authors did not report any methods for investigating differences between the studies. However, a number of subgroup analyses were reported in the results but these were not discussed in the 'Methods' section.

Thirty-one studies were included. There were 17 studies of single-slice CT (985 patients, 2,200 CABGs), 12 studies of 4-slice CT (441 patients, 1,246 CABGs) and 2 studies of 16-slice CT (144 patients, 416 CABGs).

Single-slice CT (n=17).

The sensitivity ranged from 0 to 100%; the pooled sensitivity was 81% (95% CI: 77, 84). The specificity ranged from 80 to 100%; the pooled specificity was 89% (95% CI: 87, 90).

Four-slice CT (n=12).

The sensitivity ranged from 83 to 100%; the pooled sensitivity was 93% (95% CI: 85, 95). The specificity ranged from 88 to 100%; the pooled specificity was 96% (95% CI: 94, 97).

Sixteen-slice CT (n=2).

Estimates of sensitivity were 95 and 100%; the pooled sensitivity was 99% (95% CI: 92, 99). Estimates of specificity were also 95 and 100%; the pooled specificity was 98% (95% CI: 96, 99).

The authors stated that computed tomographic angiography is less expensive than conventional, selective coronary angiography.

Single-detector CT had poor sensitivity for graft occlusion but was able to detect graft patency. Four- and 16-slice CT can be used for the detection of complete graft occlusion or graft patency of CABG, but significant stenosis is better assessed with 16-slice CT than with 4-slice CT.

This review answered a clearly stated review question. Inclusion criteria were reported, and these were supported by a flow diagram to illustrate the flow of studies through the review process. The authors stated that studies had to include multi-slice CT, but studies of single-slice CT were also included. A reasonable literature search was conducted but this was limited to MEDLINE and the Cochrane Library, with no attempts to locate unpublished studies. Relevant studies might, therefore, have been missed and the review may be subject to publication bias. Very few details of the review process, such as the number of reviewers involved in assessing the inclusion criteria and the data extraction, were reported; it was therefore not possible to determine whether appropriate steps were taken to minimise bias. No formal quality assessment was undertaken and very limited details of the included studies were reported. Thus, it was not possible to determine whether the results of the original studies might have been biased, and hence whether the conclusions of this review are also biased. Studies that used selective angiography of the CABG as the reference standard were eligible for inclusion.

The authors reported very few details of the statistical methods used, but it appears that a simple pooling was carried out without any investigation of heterogeneity. It would have been preferable to have used more statistically rigorous methods and to have included some investigation of the heterogeneity between studies. This is especially important for studies of single-slice CT where there appears to have been considerable heterogeneity in the estimates of sensitivity. Given the limitations of this review, in particular the lack of a formal quality assessment, the authors' conclusions should be interpreted with caution.

The authors did not state any implications for practice or further research.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.