Preoperative Cardiac Testing (Controversies)

Controversy: whether or not preoperative cardiac testing can ever be justified. It is clear that preoperative testing can stratify patients with regards to risk [Landesberg G et al. Circulation 108: 177, 2003], however in order for cardiac testing to be defensible, anticipated management changes (i.e. revascularization) must improve outcomes.

Data Supporting Preoperative Cardiac Testing:

There are no prospective studies in support of preoperative cardiac testing. The most widely cited retrospective study comes from Landesberg et al., who retrospectively reviewed the perioperative course of 502 patients undergoing major vascular surgery, and found that mortality was lower in those who had moderate-severe reversible ischemia on thallium scanning and underwent revascularization (n = 74) than those who did not (n = 74, OR 0.52, p = 0.018) [Landesberg G et al. Circulation 108: 177, 2003]

Data Not Supporting Preoperative Cardiac Testing:

The largest prospective study assessing the utility of preoperative revascularization was the CARP study, which randomized 510 elective vascular surgery patients to revascularization (41% CABG, 59% PCI) versus no revascularization. There was no mortality difference between groups. Critically, CARP excluded patients with an EF < 20% or left main stenosis > 50% [McFalls EO et al. NEJM 351: 2795, 2004]. Following CARP, Poldermans et al. questioned whether testing could possibly be of use, since revascularization seemed to make no difference (after all, if intervention isn’t helpful, why would you test?). DECREASE-II therefore randomized 770 vascular surgery patients at risk for cardiac morbidity (based on history) to testing vs. no-testing, and found no mortality difference between groups. Furthermore, in the 34 patients with extensive ischemia on stress testing, revascularization did not improve outcome (underpowered). No patients were excluded from this study [Poldermans D et al. J Am Coll Cardiol 48: 964, 2006]. To further test the effect (or lack thereof) of revascularization, Poldermans et al. followed up with DECREASE-V, in which patients with extensive ischemia were randomized to revascularization versus none. Again, no mortality difference was noted [Poldermans D et al. Am J Cardiol 103: 897, 2009].

Consistent with the above, a recent examination of non-surgical patients with stable coronary artery disease, which randomized 2287 patients with stable CAD to PCI versus pharmacologic and lifestyle interventions showed no difference in event rates or a composite death/MI/stroke outcome. The authors noted that “Thus, unstable coronary lesions that lead to myocardial infarction are not necessarily severely stenotic, and severely stenotic lesions are not necessarily unstable. Focal management of even severely stenotic coronary lesions with PCI in our study did not reduce the rate of death and myocardial infarction, presumably because the treated stenoses were not likely to trigger an acute coronary event” [Boden WE et al. NEJM 356: 1503, 2007]

Summary: while patients can be clearly stratified by risk-factors, the only data in support of perioperative cardiovascular testing is retrospective in nature. Furthermore, this data is refuted strongly by CARP, DECREASE-II, and DECREASE-V, three prospective, randomized, controlled trials which included a total of 1410 patients. Importantly, because the incidence of perioperative myocardial infarction is low, it is possible that these studies were underpowered. Also, CARP excluded patients with low EF and/or significant LAD stenosis (and retrospective subgroup analysis suggested that patients with LAD stenosis benefited from intervention), thus one could potentially argue that if significant LAD stenosis is suspected (< 5% incidence), perioperative testing is indicated, however this would be conjecture at best. The onus is now on advocates of perioperative stress testing to prove its value. In lieu of that, routine perioperative cardiovascular testing may satisfy the anesthesiologist’s curiosity, but appears not to affect outcome.