Use of stress testing within 1 year of stenting has declined modestly in recent years, according to a large registry study of Medicare patients published in the October 2012 issue of JACC: Cardiovascular Imaging. Differences in choice of test modality and variation in the likelihood of resulting procedures suggest that clinical decisions are less than optimal about appropriate testing, the authors say.

Investigators led by Pamela S. Douglas, MD, of the Duke Clinical Research Institute (Durham, NC), looked at 284,971 patients enrolled in the National Cardiovascular Data Registry’s CathPCI registry between January 2005 and December 2008. Patients had all received a stent, were at least 65 years old, and could be linked to procedural and claims data from the Centers for Medicare and Medicaid Services. The analysis ignored stress testing performed within 60 days of PCI in order to exclude tests performed for cardiac rehabilitation, procedure staging, or assessment of functional capacity.

Testing Modalities Differ Widely

Overall, 32.5% of patients underwent a stress test between 61 days and 1 year after PCI. Testing modalities used for a first test were:

The chances of undergoing any stress test after PCI declined from 35.2% in the first quarter of 2006 (59.3 per 100 person-years) to 29.4% (47.1 per 100 person-years) in the fourth quarter of 2008 (unadjusted incidence rate ratio per quarter 0.984; P < 0.001).

Among patients undergoing exercise testing, the likelihood of receiving ECG-only testing compared with a test that included imaging increased slightly, becoming significant after adjustment (OR 1.020; P < 0.001). For those who were tested with either nuclear or echocardiographic imaging, the probability of undergoing pharmacological vs. exercise stress tests increased, although the difference was not significant after adjustment. Also, among patients referred for imaging tests, the probability of receiving echocardiography vs. nuclear imaging decreased after adjustment (OR 0.990; P = 0.04).

Different Predictors for Different Tests

In general, patients receiving exercise ECG were clinically similar to those undergoing an exercise stress test with imaging. However, after adjustment, being older and having diabetes or a history of heart failure predicted a greater likelihood of receiving an exercise ECG compared with an exercise test with imaging, while having prior revascularization or longer time from PCI predicted a lower likelihood (table 1).

Table 1. Adjusted Predictors of Stress Test Modality: Exercise ECG vs. Exercise Imaging

The adjusted probability of receiving an exercise ECG compared with exercise imaging also varied widely by geographic region, ranging from 6.8% in the West North Central census division to 22.8% in the East South Central division.

Among patients receiving imaging, there were clear differences between those who underwent pharmacological vs. exercise stress. After adjustment, increasing age (P < 0.001) and most comorbidities were strong predictors of use of the pharmacological approach. On the other hand, geography played little role, with the exception of New England, where pharmacological testing was less common.

Compared with patients who underwent exercise nuclear testing, those who received echocardiography had a lower burden of cardiovascular risk factors. After adjustment, only a history of heart failure, atypical chest pain, and multivessel disease predicted reduced use of echocardiography. Geographic variation in the pattern of echocardiographic vs. nuclear imaging ranged from 9.4% in the South Atlantic division to 34.1% in the Pacific division.

Repeat Procedures Vary with Type of Original Test

The incidence of downstream procedures performed within 90 days of PCI depended on the type of initial stress test. For example, repeat stress testing was more common after an original exercise ECG vs. an exercise test that included imaging, after exercise echo compared with exercise nuclear imaging, and after exercise with imaging vs. pharmacological stress with imaging (all P < 0.001). The incidence of catheterization also varied considerably, being lower in nonimaging tests than imaging exercise tests, exercise tests than pharmacological tests, and exercise echo than exercise nuclear imaging (all P < 0.001). Despite varying use of additional diagnostic procedures, the rate of repeat revascularization within 90 days of catheterization varied little across testing modalities, with a slightly lower rate among pharmacological imaging tests than exercise imaging (47.8% vs. 51.0%; P = 0.002).

The investigators did a “good job” with the data they had, Mark J. Eisenberg, MD, MPH, of McGill University (Montreal, Canada), told TCTMD in a telephone interview. But the 3-year period covered by the study “is not a lot of time to look at trends, and it’s already a little outdated,” he observed, adding that the number of tests may have decreased of late.

Practice Out of Sync with Appropriateness Criteria?

Dr. Eisenberg said he was “struck” by the relatively high incidence of testing since the appropriate use criteria suggest “you’re not supposed to do any stress testing routinely within the first couple of years after PCI.” Exceptions, he noted, are patients who are having symptoms, are incompletely revascularized, or are diabetic and thus less likely to exhibit chest pain with ischemia.

Nonetheless, since about one-third of patents underwent some kind of stress test, “I suspect a lot of the tests [in the study] are uncalled for, and that’s what this paper is trying to get at,” he said, adding that much of the clinical context is “unfortunately” missing. “In order to get at [appropriateness], you really have to home in on what the patient is being referred for,” he said.

Dr. Eisenberg also suggested that “more granular” data are needed to evaluate the appropriateness of the particular test ordered. For example, for a treadmill test to provide reasonably accurate results, patients must be able to achieve 85% of maximal heart rate, he observed. And ECGs are not interpretable in patients with a left bundle branch block, so a nuclear exercise test may be more appropriate for them.

In addition to the appropriateness issue, it is likely that “[some physicians] are picking the wrong test,” Dr. Eisenberg said. “You can’t get that sort of information in a database study.”

But the biggest problem is that “we’re doing a lot of stress testing and it doesn’t seem to be affecting [patient] management much in terms of downstream procedures,” he said. “[Physicians] did [a test] to see if they needed an angiogram or repeat revascularization, but the number of those procedures was really quite low.”

Cost Containment Likely to Have Greatest Impact

Dr. Eisenberg noted that he led 2 randomized trials of exercise testing with perfusion imaging vs. a conservative strategy, ADORE I and ADORE II, both of which yielded negative results. Further randomized studies are unlikely, he added, especially since most experts now agree on the inappropriateness of stress testing in asymptomatic post-PCI patients.

In the current era of cost containment, payers’ refusal to reimburse for testing—at least without preauthorization—is likely to have a greater impact than publication of more data, Dr. Eisenberg commented.

“Tons of these tests are being ordered, and they’re not providing useful information to change the management of patients,” he concluded. “It’s an enormous use of resources, including patient, physician, and technician time, for very little return.”

Source:

Federspiel JJ, Mudrick DW, Shah BR, et al. Patterns and predictors of stress testing modality after percutaneous coronary stenting: Data from the NCDR. J Am Coll Cardiol Img. 2012;5:969-980.

 Related Stories:

• Post-Revascularization Stress Testing Appears Overused
• Stress Echo Appears Overused in Asymptomatic Patients after Revascularization
• Stress Testing More Likely Among Physicians Who Bill for Such Services