A Retrospective Cohort Study Examining the Validation of the Modified Duke Activity Status Index in the Non-cardiac Surgical Population

Abstract

Purpose

We aimed to ascertain if the Duke Activity Status Index (DASI) and its simplified variants predict 30-day mortality and myocardial injury after major non-cardiac surgery in at-risk patients.

Design

Retrospective cohort study.

Methods

We included 4,199 patients to validate the DASI and its variants in predicting the same composite outcome in patients with risk factors for coronary artery disease. Additional outcomes included 30-day severe complications, 1-year survival, and the effect of the Area Deprivation Index (ADI) on the DASI score and subsequent outcomes.

Findings

Patients were a median of 66 years old (interquartile range 57.0, 73.0), 47.9% were male, predominantly Caucasian (71.9%), with an American Society of Anesthesiologists score of 3 or greater (80.7%) and a median National ADI of 54.0 (interquartile range 33.0 to 74.0). The 30-day composite outcome was predicted by the original DASI (area under the curve [AUC] 0.82 [CI 0.73, 0.91], P < .001); modified 4-question DASI (AUC 0.82 [CI 0.73, 0.91], P < .048). The original DASI also predicted the 1-year composite outcome (hazard ratio 0.88 [CI 0.84, 0.93], P < .001), as did the modified 4-question DASI (hazard ratio 0.78 [CI 0.69, 0.89], P < .001), but not severe complications (P = .400 and P = .332 respectively). The ADI showed an inverse relationship with all versions of the DASI; there was a 0.8-point DASI decrease ([95% confidence interval −0.96 to −0.59], P < .001) for every 10-point increase in the National ADI.

Conclusions

The DASI is a reliable predictor of long-term postoperative outcomes.