Abstract
Recently published reports have shown that the American Society of Anesthesiology (ASA) classification system has limited applicability in vascular surgery patients. Most patients undergoing vascular procedures are designated as ASA class III, limiting discrimination in preoperative risk assessment. The 2006 National Surgical Quality Improvement Project (NSQIP), containing over 170,000 surgical cases, demonstrated that functional status is an important predictor of mortality. We propose that dividing ASA class III into 2 subgroups, based on NSQIP-defined functional status, improves the predictive value of the ASA scheme.
The 2006 NSQIP database was queried for ASA class III patients undergoing vascular surgery procedures. Patients were divided into groups IIIA and IIIB based on independent or dependent (partial or complete) functional status, respectively. Difference in 30-day survival between subgroups was evaluated using Kaplan-Meier and logistic regression analyses. Differences in postoperative morbidity and length of stay were compared using the unpaired t-test.
ASA class III patients having undergone vascular surgery procedures numbered 11,555 (68%). Of those 9,913 (85.7%) patients were independent (IIIA), and 1,642 (14.3%) were dependent (IIIB). Mean 30-day mortality was 1.3% in subgroup IIIA, and 6.5% in IIIB (logrank P < 0.001, χ
, 137.8). Mean lengths of stay between subgroups IIIA and IIIB were 5.4 and 13.2 days (P < 0.001). The risk of NSQIP-defined postoperative complications was 0.16 in IIIA and 0.32 in IIIB (P < 0.001).
A 5-fold difference in mortality was observed between patients who were functionally independent and dependent. A significant increase in length of stay and incidence of postoperative complications was also observed in subgroup IIIB. Subdividing ASA class III vascular surgery patients markedly improves the value of the ASA classification system. Given the "high-risk" nature of patients with vascular disease, the addition of functional status to the preoperative assessment will assist in predicting outcomes in this patient population.