PATIENTS undergoing elective major vascular surgery still have substantial perioperative risk of myocardial infarction and cardiac death, despite decades of research focused on risk stratification and implementation of risk-reduction strategies.1–3This significant myocardial risk extends beyond the perioperative period. At 18 months after vascular surgery, 18.7% of patients still experience death or myocardial infarction, i.e. , almost 1 in every 5 subjects studied.1
Recent research has suggested that perioperative intervention may improve outcomes in this high-risk subset of vascular surgical patients.4,5Therefore, enhanced identification of patients at high risk represents a means of targeting perioperative intervention. Serum biomarkers such as C-reactive protein, serum creatinine, brain natriuretic peptides, and/or troponin may collectively represent such a method.1,6–10
In this issue of the Journal, Dr. Mahla et al. 11further examine the utility of cardiac biomarkers to stratify perioperative risk and those patients in whom more targeted long-term interventions should be directed. In this prospective observational study, the investigators tested whether certain cardiac biomarkers (N-terminal pro-brain natriuretic peptide [NT-proBNP], C-reactive protein, fibrinogen, and serum creatinine) were predictors of adverse cardiac outcome. The study cohort consisted of 218 elective vascular surgical patients who all had a preoperative ejection fraction greater than 40% and who had a median follow-up of 2.25 yr. Twenty percent of the study cohort experienced adverse cardiac events as follows: 7% cardiac death, 15% nonfatal myocardial infarction, and 1% emergent coronary artery revascularization.
After univariate analysis, the following serum markers were predictors of adverse cardiovascular outcome: preoperative NT-proBNP, postoperative NT-proBNP, preoperative creatinine, preoperative fibrinogen, and C-reactive protein. After multivariate analysis, there were three independent predictors of adverse cardiovascular events: postoperative NT-proBNP levels of 860 pg/ml or greater (odds ratio, 4.88; 95% confidence interval, 2.43–9.81), occurrence of surgical complications (odds ratio, 2.56; 95% confidence interval, 1.11–5.90), and preoperative creatinine greater than 1.2 mg/dl (odds ratio, 1.92; 95% confidence interval, 1.02–3.62).
The significant ability of preoperative NT-proBNP to predict adverse cardiovascular outcome after noncardiac surgery has been established.8–10The novel observation in this outcome study is that postoperative NT-proBNP is more predictive of short-term and long-term cardiovascular morbidity and mortality than preoperative NT-proBNP.
This important observation must be interpreted in light of the study design. The peptide, NT-proBNP, is released from cardiomyocytes in response to ischemia and myocardial stretch.6–12The investigators chose NT-proBNP, as compared with BNP, because of its longer half-life. Furthermore, they excluded clinical entities associated with higher levels of natriuretic peptides such as atrial fibrillation, aortic stenosis, unstable coronary syndromes, decompensated heart failure, and impaired renal function. From an anesthetic standpoint, all patients in this cohort underwent general anesthesia compared with regional anesthesia, which decreases perioperative levels of natriuretic peptides.13
Where do we go from here? The cumulative evidence strongly suggests that we should include serum markers to better predict the subset of high-risk patients who will experience postoperative adverse cardiovascular outcome. We have no doubt that further perioperative research will identify more candidate markers besides NT-proBNP, including not only peptides but also genetic polymorphisms.
The results from this study beg future investigation in at least three areas: (1) trials to confirm this observation across perioperative populations, (2) trials that target perioperative interventions based on this marker, and (3) trials that target further long-term consultation and/or assessment based on this marker. The remaining discussion will be stratified with respect to these stated groups.
Despite the extremely positive findings from this study, further clinical trials are required to confirm and validate these findings (group 1 of further studies). The predictive value of the generated multivariate model is robust for this data set, but whether this will generalize across different data sets is still unknown. Further studies should examine various noncardiac surgical cohorts to test the reproducibility and validity of postoperative NT-proBNP as a predictive marker of postoperative myocardial risk. Will the odds ratio still be greater than 1? That is, will a postoperative NT-proBNP of 860 pg/ml or greater still correlate with increased cardiovascular risk? Will the odds ratio be 4.88 or greater? That is, will a postoperative NT-proBNP level of 860 pg/ml or greater correlate with a ≥ 5 times higher risk of developing an adverse cardiovascular event? Will the confidence interval of the odds ratio be greater than 1? That is, will it be certain to a 95% confidence limit that a postoperative NT-proBNP level of 860 pg/ml or greater correlates with increased cardiovascular risk? These further studies should also examine the predictive performance of postoperative NT-proBNP with respect to patient age, anesthetic technique, and compromised organ reserve (e.g. , degrees of renal dysfunction, degrees of acute and chronic ventricular dysfunction).
If the elevation of NT-proBNP occurs before irreversible cardiac morbidity, it is conceivable that multimodal perioperative intervention could be implemented to prevent or limit this morbidity (group 2 of further studies). These interventions could be pharmacologic (e.g. , β-blockade, statins, anticoagulation), diagnostic (e.g. , admission to a high-care setting such as an intensive care unit, further noninvasive or invasive testing), and/or therapeutic (e.g. , coronary angioplasty). The study population could be stratified post hoc , i.e. , after postoperative NT-proBNP measurement. The high-risk group would be the subgroup with a postoperative NT-proBNP level of 860 pg/ml or greater. This high-risk group could then receive targeted intervention as detailed above. High-priority interventions to test would be intensive β-blockade, platelet blockade, and/or statin therapy. We anticipate that this group of studies will follow in the near future and are highly likely to document further reduction in cardiovascular risk after vascular surgery.
With regard to long-term strategies, the perioperative period could be viewed as a “stress test.” The presence of markedly elevated NT-proBNP could therefore be viewed as analogous to a positive stress test result and identify high-risk patients for referral to a cardiologist or internist. This referral could trigger specific management to reduce future myocardial risk (group 3 of further studies). This kind of biomarker-driven management already exists in published guidelines for cardiologists (e.g. , serum natriuretic peptide and heart failure,14C-reactive protein and coronary disease).15The referral could direct long-term medical attention to aggressive risk-reduction for atherosclerotic events not only in the heart but also elsewhere in the arterial tree, such as the brain and kidney, given that atherosclerosis is a systemic disease. We anticipate that this group of studies will occur most likely as long-term follow-up of the cohorts from groups 1 and 2 of further studies.
In summary, Dr. Mahla et al. are to be congratulated for further refining identification of patients at high risk for both short-term and long-term postoperative myocardial mortality and morbidity. Their study not only has shaped future outcome research in this important area, but also is another step toward making the immediate perioperative period and beyond safer for our patients.
* Cardiovascular and Thoracic Section, Department of Anesthesiology and Critical Care, † Department of Anesthesiology and Critical Care, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania. firstname.lastname@example.org