For Angiographic Data, Only Moderate Agreement Between NCDR, Core Lab Readings

Download this article's Factoid (PDF & PPT for Gold Subscribers)

While the implications for patient management are uncertain, data from a national registry demonstrate that the interpretation of angiograms in clinical practice often diverges from the conclusions of an independent core laboratory, according to a study published online February 5, 2014, ahead of print in Circulation: Cardiovascular Interventions.

Researchers led by C. Michael Gibson, MS, MD, of Beth Israel Deaconess Medical Center (Boston, MA), evaluated 2,013 patient records from the National Cardiovascular Data Registry (NCDR); ultimately 46.3% of the cohort underwent PCI and 53.7% underwent CABG at 54 hospitals from 2004 to 2007. Concordance was assessed between the clinical sites and an independent angiographic core laboratory as to whether a significant (< 50%) lesion was present by visual assessment anywhere in the vessel.

There was only moderate agreement between the NCDR and core laboratory, with the most concordance seen for RCA lesions.  For left main disease, 11.2% of the cases described as such by the core laboratory were deemed normal by registry data, while 56.7% of cases that were read as diseased by the registry were shown to have no left main lesion by the core laboratory.

Overall, 20.9% of cases read as 3-vessel disease by NCDR sites were read as less extensive by the core laboratory, and 23.0% of cases read as 3-vessel by the core laboratory were read as less extensive by the NCDR sites.

Comparing patients for whom there was concordance and those that the NCDR registry overestimated there were no differences in death (11.3% vs 12.7%; P = 0.21) or MI (3.7% vs 2.5%; P = 0.47) at 3 years. Similarly no differences in outcomes were seen based on left main disease agreement.

Substantial Implications

“These findings demonstrate at best moderate consistency between registry data and core laboratory analysis with regard to patients with 2- and 3-vessel coronary artery disease, a cohort of patients in which selection of revascularization strategy is imperative,” Dr. Gibson and colleagues write, adding that the sample size was large enough to “provide excellent stability for the analysis.”

The authors observe that the results have substantial implications for research and, potentially, clinical practice, although they do not specify how. “Validation of the angiographic data from the NCDR is necessary for a meaningful nationwide study of the optimal revascularization strategy for patients with multivessel coronary artery disease,” they say.

However, the authors caution against generalization of the results to non-NCDR centers “because nonstandardized angiograms may not demonstrate the same level of agreement.”

What About Ischemia?

In a telephone interview with TCTMD, James K. Min, MD, of Weill Cornell Medical College (New York, NY), said inevitably “there is always some variability with these tests,” whether it lies with the definition of the reference vessel, severity of stenosis, or visual versus core lab interpretation. “There’s going to be interobserver variability, as well,” he added.

To improve on diagnostic agreement, Dr. Min said, “we need to figure out what exactly is driving [the variability] and try to improve the concordance. The core lab readings are probably more accurate because they spend a lot of time [reading] and they do it quantitatively rather than visually.”

Therefore, there may be several methods that can help clinical sites improve accuracy, Dr. Min commented, including automated software and more quantitative as opposed to visual analysis. “The caveat should be that at the end of the day, it’s just a stenosis that they are looking at and we should probably also assess concomitant ischemia associated with these,” he added. “It may be that when you add in ischemia, the differences are less robust.”

As for future studies, Dr. Min recommended a quality improvement initiative that educates physicians on how to demonstrate more concordance with the core lab. “The next step, if they could find the data, would be to see whether or not these differences in angiographic stenosis severity correlate to ischemia measures,” he said. “If they do, then there’s an issue because they would be erroneously taking people off the cath lab table who do have symptomatic ischemia. . . . If it doesn’t matter, then that’s probably a good thing, but I think someone needs to look through that.”

Study Details

The only baseline difference between patients with data agreed upon by both the NCDR and the core laboratory and those with at least 1 disagreement was history of chronic lung disease (16.7% vs 22.0%; P = 0.003).

Source:

Chakrabarti AK, Grau-Sepulveda MV, O’Brien S, et al. Angiographic validation of the American College of Cardiology Foundation—The Society of Thoracic Surgeons collaboration on the comparative effectiveness of revascularization strategies study. Circ Cardiovasc Interv. 2014;Epub ahead of print.

Related Stories

• SCAI Paper Outlines ‘Best Practices’ for FFR, IVUS, OCT
• Angiography May Overestimate Disease Severity vs. FFR in Women
• Visual Assessment Overestimates Lesion Severity vs. QCA