Validation of a modified scoring system for cardiovascular risk associated with major lung resection

The well-known revised cardiac risk index (RCRI) has recently been modified based on factors and outcomes specific to thoracic surgery patients (ThRCRI). We explored the accuracy of this modified scoring system in predicting cardiovascular morbidity after major lung resection. We analyzed outcomes from a prospective database of patients undergoing major lung resection 1980-2009. ThRCRI score was based on weighted factors for serum creatinine, coronary artery disease, cerebrovascular disease and extent of lung resection. Target adverse outcomes included pulmonary embolism, myocardial infarction, cardiac arrest, pulmonary edema and cardiac death. A total of 1255 patients (mean age 61.8 years; 649 men) underwent lobectomy or bilobectomy (1070; 85%) or pneumonectomy (185; 15%) for cancer (1037; 83%) or other problems. Severe cardiovascular complications occurred in 30 patients (2.4%), an incidence similar to that in the published derivation group (3.3%). ThRCRI median scores in patients without and with severe CV complications were 0 and 1.5 (P < 0.001). Score categories yielded incremental risks of cardiovascular complications (0: 0.9%; 1-1.5: 4.5%; >= 2: 12.8%; P < 0.001). The Hosmer-Lemeshow test demonstrated no significant difference between expected and observed outcomes (P = 0.11). The incidences of severe postoperative cardiovascular complications were similar in the published derivation group and the current validation group. The ThRCRI score successfully stratified risk for postoperative cardiovascular events after major lung resection in the validation group. The expected risk in the validation group was similar to the observed risk, indicating that ThRCRI accurately predicted specific risk rather than just relative risk. Further evaluation of the utility of this scoring system is warranted.