This article is more than 5 years old. Information may no longer be current.
Exercise tests less sensitive, more specific after adjustment for referral bias
Click Here to Manage Email Alerts
Exercise tests with echocardiography and myocardial perfusion imaging were less sensitive and more specific after adjustment for referral bias, according to a new meta-analysis.
The findings mean that the tests might do a better job of ruling in CAD in intermediate- and high-risk patients than ruling out CAD in low-risk patients, researchers wrote in the Journal of the American Heart Association.
False-negatives common
Joseph A. Ladapo, MD, PhD, of New York University School of Medicine, told Cardiology Today that cardiologists should think about more sensitive tests to rule patients out. “Cardiac stress testing is just not a very sensitive test for obstructive CAD,” he said. “There are patients who have normal cardiac stress tests, but yet they do have coronary atherosclerosis. The prevalence of … false-negative patients is underestimated in the literature and underestimated in clinical knowledge.”
Joseph A. Ladapo
Referral bias occurs when patients with an abnormal stress test result are referred to cardiac catheterization at a higher rate than those with a normal stress test result, according to the study background. However, the researchers wrote, referral bias is usually not accounted for when analyzing the sensitivity and specificity of stress tests.
American College of Cardiology/American Heart Association guidelines for exercise testing state that physicians should interpret stress test results in the context of the patient’s pretest risk, and that both posttest disease probability and test performance should guide clinical decision-making, according to the study background.
In a meta-analysis of 21 studies covering 49,006 patients (mean age, 60.7 years; 39.6% women; 0.8% prior MI), Ladapo and colleagues found that catheterization referral rates were substantially higher for patients with abnormal exercise tests (42.5%; 95% CI, 36.2-48.9) compared with those with normal results (4%; 95% CI, 2.9-5). Compared with a normal test, the pooled OR for referral after an abnormal test was 14.6 (95% CI, 10.7-19.9).
The researchers adjusted measurements of diagnostic effectiveness for referral bias. For exercise testing with echocardiography, sensitivity decreased from 84% (95% CI, 80-89) to 34% (95% CI, 27-41), but specificity increased from 77% (95% CI, 69-86) to 99% (95% CI, 99-100). For exercise testing with myocardial perfusion imaging, sensitivity fell from 85% (95% CI, 81-88) to 38% (95% CI, 31-44), but specificity rose from 69% (95% CI, 61-78) to 99% (95% CI, 99-100).
Summary receiver operating curve analysis revealed only modest changes in overall discriminatory power, but adjusting for referral increased positive-predictive value and reduced negative-predictive value.
Ruling in vs. ruling out
As a result of the findings, if a physician is looking to rule out CAD in a low-risk patient, alternate noninvasive tests such as coronary CT angiography may be more appropriate to use first; however, if a physician is looking to rule in CAD in an intermediate- or high-risk patient, exercise echocardiography or myocardial perfusion imaging may be appropriate to use first, according to the researchers.
Prior studies and the present meta-analysis “consistently found that the sensitivity of cardiac stress testing is lower and the specificity is higher than whatever the conventional thinking of the time was,” Ladapo told Cardiology Today. “This very important finding for clinical practice and for patient care just has not been able to garner the attention that it needs to change clinical practice. Part of the reason may be the fact that, until recently, there really weren’t any noninvasive options for diagnosing coronary disease. Now there are, and I think [they] need more attention and research. Some examples include coronary CT angiography and gene expression tests.” – by Erik Swain
For more information:
Ladapo JA. J Am Heart Assoc. 2013;doi:10.1161/JAHA.113.000505.
Joseph A. Ladapo, MD, PhD, can be reached at the department of population Health, New York University School of Medicine, 227 E. 30th St., VZ30 Sixth Floor, 614, New York, NY 10016; email: joseph.ladapo@nyumc.org.
Disclosure: The study was funded in part by CardioDx. Some researchers are employees of or consultants for CardioDx.
Click Here to Manage Email Alerts