Accuracy of preoperative ultrasound-derived anterior thigh muscle thickness to identify frailty in older adults undergoing surgery

Frailty is increasingly recognized as a determinant of perioperative risk in older surgical patients. The Clinical Frailty Scale (CFS) is one of the most widely used frailty assessment tools but is based mainly on clinical judgment. Ultrasound-derived anterior thigh muscle thickness (ATMT), defined as the combined rectus femoris and vastus intermedius depth measured bilaterally, may support CFS as an objective adjunct in preoperative assessment, but ATMT cut-offs for identifying CFS-defined frailty have not been established in older Asian surgical patients. This was a single-center prospective cross-sectional diagnostic accuracy study at an academic tertiary hospital in Bangkok, Thailand, conducted from 1 July 2022 to 31 July 2023. Consecutive patients aged 60 years or older scheduled for non-emergency surgery were eligible. Ultrasound-trained investigators measured ATMT within 24 h before surgery using a high-frequency linear transducer, while a separate group of trained and blinded anesthesia nurses scored CFS in the same window based on the patient’s usual baseline condition before hospital presentation. Frailty was defined as CFS of 5 or higher. The protocol prespecified a pilot reliability check, receiver operating characteristic analysis, threshold estimation, and an exploratory multivariable model. We analyzed 216 patients (median age 69 years; 88 men, 128 women), of whom 35 (16.2%) had CFS-defined frailty. ATMT was lower in frail than non-frail patients (mean 16.7 ± 5.3 mm versus 21.8 ± 5.3 mm; P < 0.001). The AUROC was 0.754 (95% CI 0.662 to 0.840). The overall cut-off was ATMT ≤ 19.65 mm, with sensitivity 80.0%, specificity 63.0%, PPV 29.5%, and NPV 94.2%. The exploratory ATMT, ASA physical status, and BMI model had an optimism-corrected AUROC of 0.766 (95% CI 0.692 to 0.867) after internal validation with 1,000 bootstrap resamples. Preoperative ATMT showed fair discrimination and a high negative predictive value for CFS-defined frailty in this cohort. It may be useful as an objective screening adjunct where structured CFS assessment is less practical, but it should not be used as a standalone diagnostic test. External validation is required.