Cardiac Biomarker Measurement After Elective PCI in Older Patients
Cardiac Biomarker Measurement After Elective PCI in Older Patients
Our study shows that most hospitals do not routinely measure postprocedure biomarkers after elective PCI. Despite guideline recommendation, a substantial proportion of patients with intraprocedural complications or complex interventions did not have postprocedure biomarker measurements made. Hospitals that did routinely perform postprocedure biomarker measurement were found to have lower associated long-term mortality risk compared with hospitals that did not perform these measurements. Patients who underwent biomarker testing had greater use of evidence-based therapies.
Current ACC/AHA/SCAI guidelines provide a Class IIb recommendation for routine measurement of biomarkers after elective PCI. The rate of biomarker surveillance in the older patients examined here (26%) is similar to that of an all-age elective PCI cohort (25%) examined in a previous study. Patients with biomarker measurement are only modestly more complex in terms of clinical and angiographic characteristics compared with nonmeasured patients. In fact, a multivariable model incorporating a comprehensive list of clinical, angiographic, and intraprocedural characteristics was poorly predictive of postprocedure biomarker measurement (c index 0.56), but model performance improved substantially with the addition of "hospital" as a random effect (c index 0.92). These results suggest that the measurement of biomarkers after PCI is a decision that is largely determined by hospital PCI protocols or individual operator preference.
In contrast, current guidelines provide a class I recommendation for biomarker measurement among patients who have signs or symptoms suggestive of MI during elective PCI, and in asymptomatic patients with angiographic complications such as embolization, dissection, no-reflow phenomenon, side-branch occlusion, or coronary thrombosis. Although this data set does not capture intraprocedural symptoms of myocardial ischemia, notably, only 37% of patients with reported intraprocedural coronary complications in our study had postprocedure biomarker testing. We also examined a broad spectrum of complex lesions attempted during PCI in this older population, such as ACC/AHA type C (high-risk) lesions, total occlusions, bifurcation lesions, or multivessel/multistent PCI, all of which were associated with low rates (<33%) of postprocedure biomarker surveillance. These data call attention to a potential gap in the translation of guidelines into clinical practice.
Patients who did not have postprocedure biomarker measurement had modestly worse long-term outcomes compared with those who did receive postprocedure biomarker testing, despite more complex characteristics in the latter group. Although there is no physiologic mechanism linking biomarker measurement (in and of itself) to better outcomes, a likely explanation is that providers who conduct post-PCI biomarker testing are more attentive to care quality. Although most patients in our study received guideline-recommended therapy, we observed greater use of these therapies among patients who had post-PCI biomarker testing. These results echo earlier findings by Eagle et al, who found that hospitals with standardized care algorithms and order sets are more likely to prescribe recommended treatments where indicated. Previous studies have also shown better adherence with evidence-based therapies to be associated with lower long-term mortality.
The decision to measure cardiac biomarkers after PCI is complex, largely because of the evolving definition of periprocedural MI and uncertainty about its prognostic significance, particularly when the event is asymptomatic. The 2007 European Society of Cardiology/ACC/AHA/World Heart Foundation Universal Definition of MI defines a PCI-related MI as an increase in biomarker levels ≥3× ULN among patients with normal baseline levels. This definition was supported by delayed enhancement magnetic resonance imaging studies confirming myonecrosis associated with this degree of biomarker elevation; however, recently published studies show a discrepancy in prognosis associated with periprocedural MI. Notably, the recently published 2012 Universal Definition of MI raised the threshold of a periprocedural MI diagnosis to >5x ULN, stating that this threshold was "arbitrarily chosen based on clinical judgment and societal implications of the label of periprocedural MI." These definitions also express a preference for troponin over CK-MB for detecting myocardial necrosis. A meta-analysis of 7,578 patients undergoing elective PCI found postprocedure troponin levels ≥3× ULN to be correlated with increased risks of adverse events out to 18 months. Other studies found no correlation between troponin elevation and cardiac magnetic resonance imaging-diagnosed MI, infarct size, or mortality, but did find a correlation with CK-MB elevation. In our study, we observed that older patients undergoing PCI in the United States were less likely to get postprocedure troponin than CK-MB measurement. Among patients with CK-MB values ≥3× ULN, 75% also had troponin levels ≥3× ULN; however, only 39% of patients with troponin values ≥3× ULN had CK-MB levels ≥3× ULN.
Our results should be interpreted in light of several considerations. First, although this observational study allows us to examine biomarker surveillance patterns across a large number of institutions and PCI patients in the United States, decisions for biomarker measurement were individualized without a universal set of testing criteria; therefore, we cannot exclude the possibility of confounding from unmeasured covariates even after rigorous multivariable adjustment, and any association with outcomes cannot be causally attributed. Second, although the CathPCI Registry provides an enriched data set of clinical variables, ischemic symptoms during PCI are not captured; therefore, we cannot fully reconstruct the definition of symptomatic periprocedural MI. Third, definitions for angiographic characteristics and complications are standardized across NCDR sites but are self-reported without core laboratory adjudication. Nevertheless, this results in a bias toward the null because intraprocedural complications are more likely underreported than overreported. Fourth, our study population only allows insight into older patients who are Medicare beneficiaries, so results may not be generalizable to younger patients or those covered by other insurance types. Finally, we do not have information on longitudinal adherence to evidence-based therapies after hospital discharge. As a result, we cannot ascertain the influence of this factor on long-term outcomes.
Discussion
Our study shows that most hospitals do not routinely measure postprocedure biomarkers after elective PCI. Despite guideline recommendation, a substantial proportion of patients with intraprocedural complications or complex interventions did not have postprocedure biomarker measurements made. Hospitals that did routinely perform postprocedure biomarker measurement were found to have lower associated long-term mortality risk compared with hospitals that did not perform these measurements. Patients who underwent biomarker testing had greater use of evidence-based therapies.
Current ACC/AHA/SCAI guidelines provide a Class IIb recommendation for routine measurement of biomarkers after elective PCI. The rate of biomarker surveillance in the older patients examined here (26%) is similar to that of an all-age elective PCI cohort (25%) examined in a previous study. Patients with biomarker measurement are only modestly more complex in terms of clinical and angiographic characteristics compared with nonmeasured patients. In fact, a multivariable model incorporating a comprehensive list of clinical, angiographic, and intraprocedural characteristics was poorly predictive of postprocedure biomarker measurement (c index 0.56), but model performance improved substantially with the addition of "hospital" as a random effect (c index 0.92). These results suggest that the measurement of biomarkers after PCI is a decision that is largely determined by hospital PCI protocols or individual operator preference.
In contrast, current guidelines provide a class I recommendation for biomarker measurement among patients who have signs or symptoms suggestive of MI during elective PCI, and in asymptomatic patients with angiographic complications such as embolization, dissection, no-reflow phenomenon, side-branch occlusion, or coronary thrombosis. Although this data set does not capture intraprocedural symptoms of myocardial ischemia, notably, only 37% of patients with reported intraprocedural coronary complications in our study had postprocedure biomarker testing. We also examined a broad spectrum of complex lesions attempted during PCI in this older population, such as ACC/AHA type C (high-risk) lesions, total occlusions, bifurcation lesions, or multivessel/multistent PCI, all of which were associated with low rates (<33%) of postprocedure biomarker surveillance. These data call attention to a potential gap in the translation of guidelines into clinical practice.
Patients who did not have postprocedure biomarker measurement had modestly worse long-term outcomes compared with those who did receive postprocedure biomarker testing, despite more complex characteristics in the latter group. Although there is no physiologic mechanism linking biomarker measurement (in and of itself) to better outcomes, a likely explanation is that providers who conduct post-PCI biomarker testing are more attentive to care quality. Although most patients in our study received guideline-recommended therapy, we observed greater use of these therapies among patients who had post-PCI biomarker testing. These results echo earlier findings by Eagle et al, who found that hospitals with standardized care algorithms and order sets are more likely to prescribe recommended treatments where indicated. Previous studies have also shown better adherence with evidence-based therapies to be associated with lower long-term mortality.
The decision to measure cardiac biomarkers after PCI is complex, largely because of the evolving definition of periprocedural MI and uncertainty about its prognostic significance, particularly when the event is asymptomatic. The 2007 European Society of Cardiology/ACC/AHA/World Heart Foundation Universal Definition of MI defines a PCI-related MI as an increase in biomarker levels ≥3× ULN among patients with normal baseline levels. This definition was supported by delayed enhancement magnetic resonance imaging studies confirming myonecrosis associated with this degree of biomarker elevation; however, recently published studies show a discrepancy in prognosis associated with periprocedural MI. Notably, the recently published 2012 Universal Definition of MI raised the threshold of a periprocedural MI diagnosis to >5x ULN, stating that this threshold was "arbitrarily chosen based on clinical judgment and societal implications of the label of periprocedural MI." These definitions also express a preference for troponin over CK-MB for detecting myocardial necrosis. A meta-analysis of 7,578 patients undergoing elective PCI found postprocedure troponin levels ≥3× ULN to be correlated with increased risks of adverse events out to 18 months. Other studies found no correlation between troponin elevation and cardiac magnetic resonance imaging-diagnosed MI, infarct size, or mortality, but did find a correlation with CK-MB elevation. In our study, we observed that older patients undergoing PCI in the United States were less likely to get postprocedure troponin than CK-MB measurement. Among patients with CK-MB values ≥3× ULN, 75% also had troponin levels ≥3× ULN; however, only 39% of patients with troponin values ≥3× ULN had CK-MB levels ≥3× ULN.
Study Limitations
Our results should be interpreted in light of several considerations. First, although this observational study allows us to examine biomarker surveillance patterns across a large number of institutions and PCI patients in the United States, decisions for biomarker measurement were individualized without a universal set of testing criteria; therefore, we cannot exclude the possibility of confounding from unmeasured covariates even after rigorous multivariable adjustment, and any association with outcomes cannot be causally attributed. Second, although the CathPCI Registry provides an enriched data set of clinical variables, ischemic symptoms during PCI are not captured; therefore, we cannot fully reconstruct the definition of symptomatic periprocedural MI. Third, definitions for angiographic characteristics and complications are standardized across NCDR sites but are self-reported without core laboratory adjudication. Nevertheless, this results in a bias toward the null because intraprocedural complications are more likely underreported than overreported. Fourth, our study population only allows insight into older patients who are Medicare beneficiaries, so results may not be generalizable to younger patients or those covered by other insurance types. Finally, we do not have information on longitudinal adherence to evidence-based therapies after hospital discharge. As a result, we cannot ascertain the influence of this factor on long-term outcomes.
Source...