This policy was created with literature review in 2011. The policy has been updated with a literature review using MEDLINE. Following is a summary of the key literature to date.
Articles were retrieved that included primary data on the accuracy, predictive value, or clinical utility of multifunction cardiogram for the diagnosis of coronary artery disease (CAD). The available evidence on the accuracy of multifunction cardiogram consists of several cross-sectional studies that evaluate the performance characteristics of the test in patients with suspected CAD, using coronary angiography as the gold standard.
Review of Evidence
Grube et al. (1) is the largest study on the accuracy of multifunction cardiogram for diagnosing CAD. The study population consisted of 562 patients with no prior history of coronary revascularization who were scheduled to receive coronary angiography over a 1-year period at one institution in Germany. All patients underwent multifunction cardiogram and coronary angiography, with results of each modality interpreted independently and blinded to the results of the other test(s). A total of 139 (24.7% of total) patients were excluded from analysis, 17 because of poor-quality electrocardiogram (ECG) tracing and 122 because full risk-factor data were not available, leaving 423 patients in the final analysis. Obstructive coronary disease, defined as at least 1 stenosis greater than 70%, was diagnosed in 47.5% of patients (201/423). The reported sensitivity and specificity of multifunction cardiogram were 89.1% and 81.1%, respectively. The positive predictive value (PPV) was 79.4%, and the negative predictive value (NPV) was 90.0%. The calculated area under the curve by receiver operating characteristic (ROC) analysis was 84.3% (95% confidence interval [CI]: 80.2-88.4%).
Grube et al. (2) published a companion article on 213 patients scheduled for angiography who had previously undergone revascularization. The protocol and analysis for this study was identical to the first article, except for the presence or absence of prior revascularization. A total of 41 patients were excluded from analysis, leaving a final sample of 172 patients. In this sample, obstructive coronary disease, defined as at least 1 stenosis greater than 70%, was diagnosed in 32% of patients (55/172). The estimated sensitivity and specificity were 90.9% and 88.9%, respectively. The PPV was 62.7% and the NPV was 97.8%.
Weiss et al. (3) included 200 ambulatory patients who were scheduled to undergo coronary angiography at one institution in New York. All patients underwent multifunction cardiogram; however, 64 patients (32% of total) had ECG tracings of insufficient quality and were excluded from analysis, leaving 136 patients in the final sample. The authors did not state that the test results were interpreted in an independent and blinded manner. Obstructive coronary disease, defined as at least 1 stenosis greater than 60%, was diagnosed in 57.4% of patients (78/136). The reported sensitivity and specificity of multifunction cardiogram were 89.1% and 81.1%, respectively. The PPV was 79.4%, and the NPV was 90.0%. The calculated area under the curve by ROC analysis was not reported.
Hosokawa et al. (4) enrolled 222 patients who were scheduled to receive coronary angiography over an approximately 6-month period from 5 medical centers in Asia. All patients underwent multifunction cardiogram and coronary angiography, with results of each modality interpreted independently and blinded to the results of the other test(s). A total of 33 patients (14.9% of total) were excluded from analysis, 3 because of poor-quality ECG tracing and 30 because coronary angiograms were not available for interpretation, leaving 189 patients in the final analysis. Obstructive coronary disease, defined as at least 1 stenosis greater than 60%, was diagnosed in 40.7% of patients (77/189). The reported sensitivity and specificity of multifunction cardiogram were 94.8% and 86.6%, respectively. The PPV was 78.4% and the NPV was 97.1%. The calculated area under the curve by ROC analysis was 91.4% (95% CI: 86.8-96.1%).
A meta-analysis of these studies was published by Strobeck et al.(5) This combined analysis included 1,072 patients from the 4 studies described above. Hemodynamically significant CAD was diagnosed in 43.4% of patients (467/1,072). The calculated sensitivity and specificity were 91.2% and 84.6%, respectively. The PPV was 78% and the NPV 94%. The area under the curve by ROC analysis was 88.1% (95% CI: 86–90.3%). Using a severity score of 4.0 as the cutoff for a positive test, the likelihood ratio positive was 5.9, and the likelihood ratio negative was 0.10. There were only minor differences between centers in the sensitivity and specificity; the statistical significance of these differences was not tested.
In 2012, an additional study of diagnostic accuracy was published by Strobeck et al. that compared the accuracy to SPECT MPI (myocardial perfusion imaging) using angiography as the gold standard. (6). This study enrolled 165 consecutive patients with suspected coronary disease and/or valvular heart disease who agreed to participate. Of the 165 patients, 49 did not undergo angiography due to a normal SPECT exam and no other indications for angiography. Of these 49 patients, 8% (4/49 patients) had an abnormal computerized 2-lead ECG. The 49 patients who did not undergo angiography were excluded from further analysis, leaving 116 patients in the study who received all 3 tests (computerized 2-lead resting ECG, SPECT MPI). The sensitivity and specificity of the computerized 2-lead resting ECG was 91% (95% CI: 0.79-0.97) and 87% (95% CI: 0.76-0.94) respectively, compared to a sensitivity and specificity of 85% (95% CI 0.72-0.93) and 14% (95% CI 0.07-0.25) respectively, for SPECT MPI. Subgroup analyses revealed similar accuracy by gender, severity of coronary obstruction, and age. The specificity for SPECT MPI in this study was markedly lower than that reported previously. In a recent meta-analysis of 13 studies including 1,323 individuals, the pooled estimate for specificity of SPECT was 0.77 (95% CI: 0.64-0.86), and the pooled sensitivity was 0.83 (CI: 0.81-0.91).(7) The reduced specificity reported in the Strobeck study may have been related to the performance of angiography in patients with valvular disease, since these patients may have higher rates of false-positive SPECT exams compared to patients with suspected ischemia.
There were no published articles that directly addressed the clinical utility of a computerized 2-lead resting ECG. The impact of this technology on patient management decisions is uncertain. While it is possible that the results of this test may influence patient management, for example, the decision to perform angiography, the evidence on this question is incomplete. As a result, clinical utility has not been demonstrated and the impact on health outcomes is unknown.
Practice Guidelines and Position Statements
A total of 5 studies from 4 patient cohorts report on the accuracy of multifunction cardiogram for diagnosing CAD. These studies report sensitivities and specificities that are in the high range, with sensitivity ranging from 89.1–94.8% and specificity in the range of 81.1–88.9%. However, these studies have several limitations that limit their internal and external validity. In all of the studies, the population is a convenience sample of patients who underwent angiography. These patient populations are thus subject to a referral or “work-up” bias in that the population of patients that might be considered for the multifunction cardiogram in clinical practice is not the same population that is being referred for angiography. Also, the number of patients enrolled but not included in the analysis was relatively high, ranging from 14.9–32% of the total number of patients enrolled. This high rate of exclusion from analysis leaves the potential for a biased estimate of the sensitivity and specificity of the test. Finally, in one of the cohorts, the angiogram and multifunction cardiogram were not interpreted in an independent, blinded manner, thus potentially leading to additional bias.
There are no studies that attempt to determine the clinical utility of the multifunction cardiogram. Even if this test does have good accuracy for diagnosing CAD, its role in clinical practice would still need to be determined. Use of the multifunction cardiogram to screen for CAD would be a departure from usual practice, as screening for CAD has not been shown to improve outcomes. In the non-acute setting, the traditional resting ECG has a limited role in diagnosing CAD. The most common method for diagnosing CAD for this purpose is stress testing. There is no evidence comparing the accuracy of multifunction cardiogram to stress testing. The comparison to angiography, while useful from a research perspective, has a limited role in determining clinical utility given that multifunction cardiogram would not be used as a replacement for angiography.
Because of these limitations, the evidence is not sufficient to determine the impact of the computerized 2-lead resting electrocardiogram analysis (e.g., multifunction cardiogram) on health outcomes, and therefore the use of this device is considered experimental, investigational and unproven.
Disclaimer for coding information on Medical Policies
Procedure and diagnosis codes on Medical Policy documents are included only as a general reference tool for each policy. They may not be all-inclusive.
The presence or absence of procedure, service, supply, device or diagnosis codes in a Medical Policy document has no relevance for determination of benefit coverage for members or reimbursement for providers. Only the written coverage position in a medical policy should be used for such determinations.
Benefit coverage determinations based on written Medical Policy coverage positions must include review of the member’s benefit contract or Summary Plan Description (SPD) for defined coverage vs. non-coverage, benefit exclusions, and benefit limitations such as dollar or duration caps.