New Index for Preoperative Cardiovascular Evaluation
Consensus guidelines recommend use of a validated preoperative risk stratification tool to facilitate triage and shared decision-making discussions, but existing tools have limitations; some lack accuracy or precision in specific populations, whereas others require complex data input, or may be time-consuming for the clinician. Can a simple but highly discriminative risk-assessment tool be developed for patients undergoing preoperative evaluation before noncardiac surgery?
A cohort of 3,284 consecutive patients aged ≥40 years, undergoing noncardiac surgery at American University Hospital in Beirut between July 1, 2016–December 30, 2017, were enrolled prospectively and subsequently evaluated for incidence of 30-day postoperative all-cause mortality, myocardial infarction (MI), or stroke, comprising the derivation data set. MI was diagnosed when troponin elevation accompanied ischemic electrocardiographic findings and/or symptoms; in absence of suspected ischemic events, neither troponin nor other myocardial injury biomarker was routinely measured or considered as an isolated endpoint. Stroke diagnosis was based on neurologic specialty assessment plus confirmatory imaging following an acute change in neurologic status. From these 3,284 patients, a multivariate logistic regression-derived model (the Cardiovascular Risk Index [CVRI]) was constructed. For validation, the CVRI was then applied retrospectively to 1,167,414 patients from the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) database who had undergone surgery in the United States between 2008-2012.
Analysis from the derivation cohort, undergoing a variety of surgical procedures, identified six predictors of primary outcome: age ≥75 years, any history of heart disease, symptoms of angina or dyspnea with regular activities, hemoglobin <12 mg/dl, planned vascular surgery, and emergency surgery. Stratification by number of predictors (0, 1, 2, 3, and >3) differentiated patients at low (0-1), intermediate (2-3), or high (>3) risk for suffering the primary outcome, with excellent discrimination (area under the curve [AUC] = 0.90). When the risk index was applied to the 1,167,414 patient population from the ACS NSQIP database, undergoing general, vascular, orthopedic, and other surgical procedures, the CVRI again appeared to show excellent discrimination in predicting risk strata for 30-day postoperative MI, stroke, and all-cause mortality, with similar accuracy to that of the ACS NSQIP tool (AUC = 0.90 vs. 0.89), and greater accuracy than the Revised Cardiac Risk index (AUC = 0.77). Primary outcome occurred in 0.3% of patients with 0/6 versus 17.5% of patients with >3/6 risk factors. The risk factors were determined from simple, commonly shared clinical information, and scoring did not require extensive calculations or data entry into hand-held or desktop computer interfaces.
The CVRI appears to have advantages of simplicity and excellent discrimination of 30-day clinically relevant outcomes. In low-risk patients, its use may reinforce efforts to avoid low-yield, time-consuming cardiovascular testing.
The value of the CVRI, with respect to its accuracy in predicting adverse outcomes in surgical populations as procedures, patient demographics, and medical interventions evolve, will be clarified over time. The concept that the CVRI identifies risk factors that may be amenable to preoperative intervention, such as treatment of anemia or optimization of cardiovascular symptoms, suggests constructive potential in guiding risk reduction efforts. Whether the CVRI, or tools like it, can foster important discussions and clarify issues and options for patients, families, and caregivers facing difficult decisions prior to surgery, will be a more complicated issue to determine.
Keywords: Anemia, Angina Pectoris, Biological Markers, Dyspnea, Electrocardiography, Hemoglobins, Myocardial Infarction, Preoperative Care, Primary Prevention, Quality Improvement, Risk Assessment, Risk Factors, Risk Reduction Behavior, Stroke, General Surgery, Triage, Troponin
< Back to Listings