Cardiac Imaging Tests Beget More Tests, but No Benefit

February 10, 2014

NEW HAVEN, CT – An analysis of more than 220 US hospitals reveals wide variation in the use of noninvasive imaging in patients presenting to the emergency department with chest pain[1]. Despite the variation, hospitals with high rates of imaging did not have lower MI readmission rates than those with lower rates of noninvasive cardiac imaging.

Patients treated at hospitals that were larger users of noninvasive imaging were more likely to be admitted to the hospital and to undergo coronary angiography, however.

"Our experience is that there are a lot of different approaches being employed by hospitals around the country, some with protocol-based chest-pain centers and others with more ad hoc approaches," senior investigator Dr Harlan Krumholz (Yale University School of Medicine, New Haven, CT) told heartwire . "And it seems like there is much less discretion in terms of patients who would merit imaging and how best to apply it. We wanted to get some perspective on how much variation there was and whether there was any relationship between the variation and patient outcomes."

Krumholz said the study, published online February 10, 2014 in JAMA: Internal Medicine, was intended to provide some data on the nationwide evaluation of patients with suspected ischemic heart disease, specifically those with acute coronary syndromes. These patients are often difficult to track because they can be lost to follow-up once they enter into the hospital.

Lots of Variation, No Difference in Readmissions

The present analysis is based on data from the PREMIER database of 2700 acute-care hospitals in the US. In total, 549 078 patients from 224 hospitals were included in the study, and the use of noninvasive imaging ranged from 0.2% to 55.7%. When the hospitals were stratified by quartiles (Q), the use of noninvasive testing was performed 6.0%, 15.9%, 23.5%, and 34.8% in Q1, Q2, Q3, and Q4, respectively. Myocardial perfusion imaging and stress echocardiograms were the most commonly used imaging tests. In total, 80.4% of the 113 602 imaging tests performed were myocardial perfusion tests, 16.6% were echocardiograms, and 1.2% were computed tomography coronary angiograms (CTCAs).

"We found a remarkable variation in the use of imaging, which is an expensive intervention, and its use was strongly linked to what happened to the patient subsequently," said Krumholz. "The testing cascade has been discussed in other articles, but this is more proof that the more expensive tests you do the more likely you are to pursue additional tests."

Use and Outcomes Compared by Hospital Imaging Quartiles

Use and outcomes Q1 (n=129 021) Q2 (n=123 648) Q2 (n=149 668) Q4 (n=146 741) p
Imaging (%) 6.0 15.9 23.5 34.8
Inpatient admission (%) 32.1 37.0 37.8 40.0 <0.001
Coronary angiogram (%) 1.2 2.2 3.3 4.9 <0.001
Revascularization (%) 0.5 0.9 1.2 1.9 <0.001
Revascularization per imaging study (n) 7.6 5.6 4.9 5.4 <0.001
Revascularization per angiogram (n) 41.2 40.9 36.4 38.8 <0.001
Readmission with AMI (%) 0.3 0.3 0.3 0.3 NS

Compared with those in Q1, hospitals with the highest rates of imaging (Q4) were more likely to admit patients and perform coronary angiography. Hospitals with lower rates of imaging performed angiography in 1.2% of patients vs 4.9% in patients treated at hospitals with higher use of noninvasive imaging. The rate of coronary revascularization was also higher among patients treated at hospitals with higher rates of noninvasive imaging. However, in terms of revascularizations per imaging study and revascularizations per angiogram, hospitals in Q1 had significantly better yield than those in Q4.

Despite the differences in care among hospitals more likely to use imaging, there was no difference in the patients readmitted to hospital within the month or the subsequent month. "We couldn't find any evidence that patients are being benefited by the approach," said Krumholz, referring to the higher rates of imaging at some hospitals.

Variation Attributable to Hospital Factors

To heartwire , Krumholz said that a lot of hospitals don't have any feedback in terms of where they stand in relationship to other hospitals with their use of noninvasive imaging tests. In analyzing the patient characteristics and hospital factors, the group also found that nearly 25% of the between-hospital variation is attributable to institutional factors and not the types of patients treated. Regarding best clinical practices, the study was not designed to determine whether imaging was appropriate, but the researchers contend that patient-case mix would unlikely account for the variation in cardiac imaging rates.

In an editorial[2], Drs Ezra Amsterdam and Edris Aman (University of California, Davis) point out that accelerated diagnostic protocols (ADPs) involve identifying low-risk patients based on clinical stability, a normal ECG, and a negative biomarker test. Predischarge testing is then used to identify patients for early discharge, and this could entail anything from an exercise treadmill test to CCTA.

"At the University of California, Davis, Medical Center in Sacramento, we practice physician discretion in selecting patients for predischarge testing," they write. "In more than 500 patients discharged directly from the unit after evaluation consisting of normal results of electrocardiograms and cardiac troponin tests, there has been only one adverse cardiac event (0.2%) at the 30-day follow-up."

Such a strategy, however, depends on obtaining a detailed family history, an accurate assessment of the ECG, and a reliable assay for cardiac troponin, as well as the willingness of clinicians to adopt such an algorithm, especially given the hazards of a missed acute coronary syndrome.

Krumholz has received a research grant from Medtronic, through his institution, to study methods of clinical-trial data sharing. The editorialists report no conflicts of interest.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.