Perioperative Cardiac Risk Assessment

> Research explainer: This briefing examines verified primary research published 61 days before the briefing date. It is not a same-day research update and does not provide medical advice.

This research explainer examines a prospective observational cohort of adults undergoing elective non-cardiac surgery at a university hospital in Bahia, Brazil. It compared two preoperative cardiovascular risk tools—the Revised Cardiac Risk Index (RCRI) and American College of Physicians (ACP) score—with cardiologist clinical judgment. The study’s central question was how well these approaches distinguished patients who did and did not experience postoperative cardiovascular outcomes in that setting. [pmid:42384973]

Evidence

The cohort included 420 adults assessed for cardiovascular risk before elective non-cardiac surgery between April 2019 and December 2023. Participants had a mean age of 64.1 years; 47.1% were female and 71.2% were Black/Hispanic. The reported prevalence of hypertension was 74.3%, diabetes 32.9%, heart failure 13.1%, coronary artery disease 11.2%, and prior stroke 7.9%. [pmid:42384973]

The investigators calculated RCRI and ACP scores according to the cited Brazilian perioperative cardiovascular assessment guideline and also recorded intuitive clinical judgment from cardiologists. They evaluated discrimination with C-statistics, or areas under receiver operating characteristic curves, for death and combined in-hospital cardiovascular outcomes. The supplied report characterizes values below 0.70 as poor discrimination in this analysis. [pmid:42384973]

During hospitalization, the study recorded 11 deaths, three myocardial infarctions, and one episode of acute pulmonary edema. The outcomes of interest were aligned with the score definitions and included fatal and non-fatal acute myocardial infarction, acute pulmonary edema, ventricular fibrillation or primary cardiac arrest, and complete heart block; outcomes were assessed through postoperative day 5, with additional events recorded through 30 days. [pmid:42384973]

Both the RCRI and ACP score had C-statistics below 0.70 for death and for combined in-hospital cardiovascular outcomes. Cardiologist judgment performed slightly better than the two scores. The report also states that findings were similar when surgeries were divided into major and non-major procedures and when analysis was limited to people aged 50 years or older. [pmid:42384973]

Risk categories differed across the approaches. RCRI classified 88.0% of participants as low risk, 10.5% as medium risk, and 1.4% as high risk. ACP classified 71.2%, 27.4%, and 1.4% in those respective categories, while clinical judgment classified 64.5%, 33.1%, and 2.4%. These distributions describe how the approaches sorted this cohort; they do not by themselves establish which categorization is preferable. [pmid:42384973]

Analysis — Interpreting discrimination in context

The key result is comparative and setting-bound: the two assessed scores did not discriminate strongly between patients with and without the measured in-hospital outcomes in this cohort, while cardiologist judgment was only slightly better. A C-statistic concerns discrimination—how well a method ranks or separates outcome groups—not whether an approach improves decisions or outcomes. Accordingly, this evidence supports a narrow conclusion about observed predictive performance at one hospital, rather than a claim that either score is broadly ineffective or that judgment should replace structured assessment. [pmid:42384973]

The subgroup consistency adds context but not broad validation. Similar results across major and non-major procedures and among adults aged 50 years or older indicate that the reported pattern persisted in those analyses within the study population. The cohort’s demographic and clinical profile, referral pathway for preoperative cardiovascular assessment, and elective non-cardiac surgery setting all matter when interpreting transferability. [pmid:42384973]

The comparison also separates two related questions that are often conflated. Risk tools can standardize information and risk categories, whereas this study tested their observed discriminatory accuracy against selected outcomes. A tool may be used within a clinical workflow, yet its performance must still be evaluated in the population and outcome horizon in question. This cohort supplies evidence for that evaluation in its own context only. [pmid:42384973]

Limitations

This was a single-center study at a referral university hospital in Northeastern Brazil, and participants were recruited through convenience sampling. Those features limit how confidently the findings can be generalized to other hospitals, health systems, populations, or surgical case mixes. [pmid:42384973]

The cohort consisted of adults referred for cardiovascular risk assessment who underwent elective non-cardiac surgery. It therefore does not directly establish score performance in emergency surgery, cardiac surgery, or people outside that referral context. [pmid:42384973]

The study was observational and primarily assessed discrimination with C-statistics. It does not show that substituting cardiologist judgment for either score would improve patient outcomes. Formal blinding between the preoperative assessors and outcome evaluators was not applied. In addition, some 30-day outcomes were collected by telephone, and participants without obtainable 30-day data were assessed only for in-hospital outcomes. [pmid:42384973]