'Normal' Angiogram Rates for Suspected CAD Differ Widely Among Hospitals

Shelley Wood

October 10, 2013

TORONTO, ON — New data from Ontario show that almost 42% of patients with suspected coronary artery disease (CAD) who underwent a diagnostic angiogram within a 17-hospital network were found to have none[1].

The results imply that whatever criteria physicians are using to send a patient for the invasive test are often failing to identify patients who actually warrant further testing, according to Dr Kevin Levitt (University of Toronto, ON) and colleagues.

Strikingly, rates of angiograms turning up normal were not uniform across the hospitals studied, ranging from as low as 18% at some hospitals to as high as 77%.

"This is a huge amount of variation, so I was very surprised," senior author on the study, Dr Jack Tu (University of Toronto, ON), told heartwire . "It suggests to me that physicians are not selecting patients [for invasive testing] in the same way across centers, and we need a better approach, a more specific approach. And the overall rate of 41.9% is on the high side, and that's a big issue. We don't expect the normal rates to be zero, but most physicians will be surprised to see that it is this many."

Abnormal Rates of Normal?

The study by Tu and colleagues, published in the October 2013 issue of the American Heart Journal, drew on the Ontario Cardiac Care Network Variations in Revascularization Practice database, focusing on 2718 patients who underwent cardiac catheterization for stable angina between April 2006 and March 2007 at one of the 17 hospitals, none of which are mentioned by name in the paper.

"We didn't get into the business of naming names," Tu said in an interview yesterday. "The hospitals have since been notified of the data, and it is likely part of their ongoing quality assurance. This is a type of indicator that we should pay more attention to. We also know that hospitals do change their behavior [in response to data like these], so it will be interesting to revisit this issue over time and see if the situation changes."

Of note, teaching hospitals typically had lower rates of normal angiograms than community hospitals: 35.4% vs 47.1% (p<0.001).

Predictors of Test Outcomes

Other predictors of normal angiograms included atypical symptoms of ischemia or no symptoms; the absence of diabetes, hyperlipidemia, smoking history, and peripheral vascular disease; and testing at a nonteaching hospital. The strongest predictor of a normal angiogram, however, was female gender, Levitt and colleagues report.

Other recent studies have suggested that women are less often referred for angiography, so at least intuitively, one might expect women who actually get angiograms to be more likely to have disease. "It could be that the wrong females are being suggested for this test, because in general they do appear to be less worked up then men, but in this test, they still have a much higher normal rate than men," Tu said. "It might be that women have more microvascular disease that's not being picked up by an angiogram."

Of note, noninvasive testing was performed in 70.6% of patients who went on to have angiography. A positive noninvasive test was seen in 93.8% of patients who eventually were found to have an abnormal angiogram, but also in 87.7% of those whose angiograms turned out to be normal. "The stress testing and nuclear medicine tests didn't do a great job of predicting which angiograms would be normal," Tu commented. "We were surprised by those findings as well."

Canada and the US: Similarities

The Ontario data are remarkably similar to that of a US study of almost 400 000 patients from 663 hospitals by Dr Manesh Patel (Duke Clinical Research Institute, Durham, NC) and colleagues published in 2010, as reported by heartwire . That study drew on the American College of Cardiology National Cardiovascular Data Registry and found that 38% of angiograms turned out to be normal.

It's notable that a Canadian study, with a national healthcare payment scheme, was seen to have similar rates of seemingly unwarranted angiography, Tu said but added: "I'm not sure reimbursement played a role directly here. These are fee-for-service procedures in that the more physicians do the more they get paid. Having said that, everyone in Ontario is paid in a similar manner. I think it is probably other factors like physician selection criteria."

Commenting on Levitt et al's study for heartwire , Patel also highlighted the similarities in the findings between the US and Canadian studies. "This is not surprising to me, in that clinicians in many areas face the same difficulties in diagnosing and managing patients with symptoms concerning for obstructive coronary disease," he said. New studies and tools may soon help with this, he added. "Upcoming randomized studies such as the PROMISE trial with [coronary CT angiography ] CTA vs usual stress testing should be informative. In addition, it maybe that strategies of going earlier to the cath lab with invasive [fractional flow reserve/coronary flow reserve] FFR/CFR might be ways to be more efficient if they can be done safely."

Tu hopes these data spark improvements to this triage process and that the proportion of normal/abnormal angiograms for stable CAD becomes a performance indicator.

"This does add up in terms of costs, but also for patients this is not necessarily a good thing. It may be good for reassurances, but on the other hand, some patients do get serious complications. If you can avoid an invasive test, that's to be preferred."

Levitt and Tu had no financial conflicts; disclosures for the coauthors are listed in the paper.


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.