(PC030) Frailty Indices for Predicting Outcomes After Peripheral Vascular Interventions: Comparative Validation To Standardized Preoperative Screening

Objectives: Frailty, which is a biologic syndrome of reduced reserve and resistance to stressors, has been associated with adverse outcomes after vascular surgery. Quantitative Frailty indices such as the NSQUIP Modified Frailty Index (mFI), the Veterans Affairs Risk Analysis Index (RAI), the Vascular Quality Initiative Frailty Index (VQI-FI), and Modified Vascular Quality Initiative Frailty Index (mVQI-FI) have independently been associated with adverse outcomes, but the predictive ability of these indices has not been directly compared in the vascular surgery population. This study aims to validate and compare he predictive performance of these indices in our institution.

Methods: A retrospective cohort analysis evaluated the performance of mFI, RAI, VQI-FI, and the mVQI-FI, in predicting postoperative outcomes after vascular surgery. Unadjusted logistic and linear regression modeled the associations of frailty scores with our primary and secondary outcomes, including 30-day readmission, 60-day readmission, and one year mortality. Secondary outcomes included surgical and vascular-specific complications.

Results: The VQI-derived frailty index (VQI-FI) was associated with 8 adverse outcomes. It demonstrated moderate predictive capability and large effect size for new dialysis within 30 days (AUC: 0.83, OR: 1.75, CI: 1.25–2.45, p=.0011), and one-year mortality (AUC: 0.75, OR: 1.66, CI: 1.29–2.15, p=.0001) (Figure.1). There was also significant association with 30-day readmission, 60-day readmission, and survival at discharge, although with a smaller effect size. (Table 1) The NSQUIP modified frailty index was associated with 6 adverse outcomes with a large effect size for 30-day readmission (OR: 1.50, CI: 1.14-1.97, p=.003), dialysis initiation within 30 days (OR: 1.75, CI: 1.16–2.65, p=.0081), and one-year mortality (OR: 1.59, CI: 1.14–2.22, p=.0058). There was a smaller effect size for 60-day readmission, acute kidney injury, and all reinterventions. The Risk Analysis Index (RAI) was associated with 7 outcomes with a modest effect size. (Table 1) However, it was the only frailty index that was predictive of major amputation (OR: 1.05, CI: 1.01-1.08, p=.013) and major adverse limb events (OR: 1.03, CI: 1.0-1.07, p=.03). The modified VQI frailty index was associated with 6 adverse outcomes, however, all with had a small effect size (OR < 1.02).

Conclusions: This study demonstrates the utility of quantitative frailty scoring in the vascular surgery population. However, there is variability in the strength of these associations between indices. These demonstrate that quantitative frailty scores are a useful tool for preoperative risk stratification. Future directions include automating frailty assessments within electronic medical records.