Simple frailty score predicts postoperative complications across surgical specialties

Abstract

Background: Our purpose was to determine the relationship between preoperative frailty and the occurrence of postoperative complications after colorectal and cardiac operations.

Methods: Patients 65 years or older undergoing elective colorectal or cardiac surgery were enrolled. Seven baseline frailty traits were measured preoperatively: Katz score less than or equal to 5, Timed Up and Go test greater than or equal to 15 seconds, Charlson Index greater than or equal to 3, anemia less than 35%, Mini-Cog score less than or equal to 3, albumin less than 3.4 g/dL, and 1 or more falls within 6 months. Patients were categorized by the number of positive traits as follows: nonfrail: 0 to 1 traits, prefrail: 2 to 3 traits, and frail: 4 or more traits.

Results: Two hundred one subjects (age 74 ± 6 years) were studied. Preoperative frailty was associated with increased postoperative complications after colorectal (nonfrail: 21%, prefrail: 40%, frail: 58%; P = .016) and cardiac operations (nonfrail: 17%, prefrail: 28%, frail: 56%; P < .001). This finding in both groups was independent of advancing age. Frail individuals in both groups had longer hospital stays and higher 30-day readmission rates. Receiver operating characteristic curves examining frailty's ability to forecast complications were colorectal (.702, P = .004) and cardiac (.711, P < .001).

Conclusions: A simple preoperative frailty score defines older adults at higher risk for postoperative complications across surgical specialties.

Keywords: Geriatric surgery; Preoperative risk assessment; Surgical morbidity.

Figure 1. Summing Geriatric Assessment Deficits to Determine Post-Operative Risk: The Balance Scale Analogy

The analogy of a bar scale that weighs an individual’s surgical risk is useful to understand the concept of summing abnormal geriatric assessment domains (or deficits) to predict postoperative risk. Blocks represent geriatric assessment domains (e.g., function, cognition, nutrition, co-morbidity burden and geriatric syndromes). Blocks are place opposite one another on the bar scale depending on whether the domain was measured as a normal characteristic or an abnormal deficit. Two older adults’ assessments are depicted in this graphic. On the left is a non-frail individual; the majority of the measured geriatric domains are normal resulting in tilting the scale toward “better” post-operative outcomes. In contrast, on the right is a frail individual; the majority of geriatric domains are abnormal (or deficits) resulting in tilting the scale toward “poor” post-operative outcomes.

Figure 2. Receiver Operating Characteristic (ROC) Curves

The area under the curve is compared to the null hypothesis (diagonal line) where the true area is 0.5. For colorectal operations group, area under the curve equals 0.702 (95% CI: 0.576, 0.828; p=0.004). For the cardiac operations group, area under the curve equals 0.711 (95% CI: 0.606, 0.817; p<0.001).