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

The likelihood of finding significant coronary obstruction on elective diagnostic angiography varies considerably across hospitals in the United States. Moreover, rates of imaging-revealed disease correlate with patient selection and preprocedural assessment strategies rather than institutional characteristics, according to findings published in the August 16, 2011, issue of the Journal of the American College of Cardiology.

Investigators led by Pamela S. Douglas, MD, of the Duke University School of Medicine (Durham, NC), reviewed data from the National Cardiovascular Data Registry (NCDR) CathPCI Registry from 565,504 patients without known CAD who underwent elective diagnostic angiography at 691 participating centers between 2005 and 2008. The results were correlated with patient demographics, clinical risk factors, symptom status, and noninvasive test findings.

Little Change from Year to Year

Rates of imaging-identified CAD varied markedly among hospitals, ranging from 23% to 100%, with a median of 45% (interquartile range 39%-52%). Ninety-one hospitals had a positive diagnostic yield of less than 35% while 82 had a yield of at least 75%. The overall rates held steady over the study period (44.4% in 2005 and 45.6% in 2008). In addition, individual hospital rates changed little from year to year.

Compared with institutions with CAD yields in the highest quartile (52%-100%), those in the lowest quartile (23%-39%) more often performed angiography in younger patients, women, African Americans, and outpatients as well as those with lower Framingham risk scores, atypical symptoms, and negative or equivocal stress tests (table 1).

Table 1. Lowest vs. Highest Obstructive CAD Rates by Patient, Clinical Characteristics

 
In addition, patients at hospitals with low rates of finding obstructive CAD were less likely to have been prescribed cardiac medications before angiography, such as aspirin, beta blockers, statins, or platelet inhibitors (all P < 0.0001). On an institutional level, CAD rates tended to be lower at smaller hospitals and those with lower annual cath lab volumes and higher at centers in the Western United States and at clinical trial sites. On the other hand, whether or not a hospital had a teaching program, was private or university-affiliated, or was located in an urban or rural setting had no bearing on CAD rates.

The investigators used a logistic regression model incorporating Framingham risk score, other clinical variables, chest pain characteristics, and noninvasive testing results to predict the likelihood of obstructive CAD and compared those results with the actual rate at each institution. Hospitals in the lowest quartile had only a 39% median predicted likelihood for obstructive CAD compared with 55% for centers in the highest quartile (P < 0.0001).

Altering CAD Cutpoints Did Not Change Findings

A sensitivity analysis that altered the definition of obstructive CAD, ranging from 20% to 70% stenosis, did not reduce the wide variation among hospital rates.

Finally, the researchers investigated the potential impact of applying the preprocedural CAD evaluation criteria and referral strategies of the hospitals in the highest quartile to those with lower rates. They calculated that if the median 70% rate of CAD findings in the highest quartile hospitals were achieved by the rest of the centers, the number of patients undergoing angiography without positive findings in the latter could be reduced by 70%. Furthermore, if the CAD rate at the lowest 2 quartiles of hospitals were to rise to the median US level of 45%, the number of patients without obstructive disease having angiography would be reduced by 23%.

The study data cannot establish the “optimal” CAD rate for elective coronary angiography, Dr. Douglas and colleagues acknowledge, but the correlations highlighted suggest that improved patient selection could increase the rate in the lower-quartile institutions. “Consistent use of clinical risk stratification algorithms and improvement in the use, accuracy, and quality of preprocedural noninvasive testing, and perhaps future coronary angiography appropriate-use criteria, may help raise the rate of finding obstructive CAD,” they write.

Moreover, the investigators point out, the consistency of each hospital’s CAD rate over time suggests that it reflects institutional more than individual practice. Thus, CAD rate is an appropriate metric for assessing use of diagnostic angiography and local clinical practice patterns are the best target for quality-improvement efforts, the authors contend.

Showing Differences, Not Making Judgments

“These are good data to have,” said Christopher J. White, MD, of the Ochsner Clinic Foundation (New Orleans, LA), and president of the Society for Cardiovascular Angiography and Interventions, in a telephone interview with TCTMD. “But it is important to understand that the paper simply says, ‘Look at this broad variation, which is perhaps more than we should have, and here are some [factors] that are correlated with it.’ It doesn’t make judgments about what’s good or bad.

“The trick is to figure out what the right number or range is and then, using that information, steer our practices toward a common goal,” he continued. “But right now there is no point in scolding anybody. Besides, the problem is not individual physician behavior, it’s systematic behavior by institutions.”

Dr. White said the varying rates mainly reflect aggressive vs. conservative strategies for investigating possible CAD. And both may be driven in part by patients’ attitudes and/or fears. Sometimes ordering an angiogram to prove that the coronary arteries are normal is appropriate patient management—as long as it does not become standard practice, he added.

Finding a ‘Sweet Spot’

“We should leave room for clinical judgment,” Dr. White observed. With regard to CAD rates, “we need to find a ‘sweet spot’ so you don’t miss the rare case of [obstruction in] a low-risk patient—a situation where you could really save a life—but still do most angiographies in the higher-risk, higher-yield population,” he said. With more research, Dr. White predicted, there will likely be different acceptable ranges of diagnostic yield for different categories of patients depending on their pretest probability of CAD.

The American College of Cardiology is currently considering appropriate-use criteria for diagnostic angiography, Dr. White reported. Meanwhile, papers like this that raise awareness of discrepancies among hospitals help spur efforts to improve practice, he added, noting that it will require repeated cycles of encouraging careful patient evaluation and then measuring the impact to shrink variation to an optimal level.

Obstructive CAD was defined as being at least 50% stenosis in any major epicardial vessel or branch vessel greater than 2.0 mm in diameter and that is associated with a fractional flow reserve of less than 0.80 in 63% of patients.

Related Stories:

• Coronary CTA May Offer Cost-Effective Triage Test in Asymptomatic Patients