Background and Goal of Study: The Revised Cardiac Risk Index (RCRI) is a validated index that uses preoperative data to predict postoperative cardiac complications with moderate accuracy. The Surgical Apgar Score (SAS) is a validated prognostic index based on intraoperative factors (hypotension, bradycardia, estimated blood loss). We conducted a retrospective cohort study to determine if combining the 2 indices allows for more accurate estimation of perioperative cardiac risk. 
Materials and Methods: Following research ethics approval, we conducted a retrospective cohort study of adults (≥18 y) who underwent major elective noncardiac surgery from 2009-2014 at the Toronto General Hospital (Toronto, ON, Canada). The exposure variables were the RCRI (classified as 0, 1, 2 or ≥3) and SAS (classified as 0-2, 3-4, 4-5, 6-7, 7-8, or 9-10). The primary outcome was myocardial injury, defined a postoperative elevation in troponin I above the 99th percentile. We assessed the correlation between the RCRI and SAS using the Spearman rho. The area under the curve (AUC) of the receiver-operating characteristic (ROC) curve was used to measure the association of the RCRI and SAS with myocardial injury. Finally, ROC curve and net risk reclassification analyses were used to assess if combining the RCRI and SAS resulted in more accurate prediction of myocardial injury. Four risk categories (< 1%, 1-5%, 5-10% and >10%) were used for reclassification analyses.
Results and Discussion: The cohort included 16,841 patients, of whom 183 (1.1%) had myocardial injury. The proportions with RCRI scores of 0, 1, 2 and ≥3 were 46.5%, 38.9%, 11.3% and 3.3% respectively. The proportions with SAS of 0-2, 3-4, 4-5, 6-7, 7-8, and 9-10 were 0.5%, 4.4%, 21.3%, 55.3% and 18.6% respectively. There was weak correlation between the RCRI and SAS (rho 0.13; 95% CI 0.12-0.15). As individual indices, the RCRI (AUC 0.73; CI 0.70-0.76) predicted myocardial injury better than the SAS (AUC 0.63; CI 0.59-0.67). The combination of the indices resulted in improved predictive accuracy (AUC 0.77; CI 0.74-0.80) than the RCRI (P< 0.001). The improvement was driven by better risk classification of the 16,658 patients without myocardial injury (net 1325 patients reclassified to lower risk categories).
Conclusions: Supplementation of a preoperative index (RCRI) with data from an intraoperative index (SAS) results in improved estimation of perioperative cardiac risk. Further research is needed to confirm these findings.

Does supplementation by the Surgical Apgar Score improve the prognostic accuracy of the Revised Cardiac Risk Index?