Measuring sarcopenia on pre-operative CT in older adults undergoing emergency laparotomy: a comparison of three different calculations.

ABSTRACT

INTRODUCTION: Sarcopenia is associated with outcomes in older-adults undergoing emergency surgery. Psoas major measurement is a surrogate marker of sarcopenia with multiple calculations existing normalising to body size and no consensus as to which is optimal. We compared three different psoas-major calculations to predict outcomes in older adults undergoing emergency laparotomy. METHODS: Consecutive over 65s were identified from the National Emergency Laparotomy Audit(NELA) database at a single centre between 2014 and 2018. Psoas major was measured at the L3 level and normalised to height (psoas muscle index, PMI), L3 vertebral body (psoas muscleL3 ratio, PML3) or body surface area (psoasbody surface area, PBSA) and each correlated to outcomes. Outcome measures included inpatient, 30-day and 90-day mortality. A comparison of the three calculations was performed using the Mann-Whitney U, chi-squared, receiver operating characteristic curves (ROC) and binary logistic regression. RESULTS: Two hundred and sixty-four older adults underwent emergency laparotomy (median age, 75 years ((IQR, 70-81 years), 50% female)). Inpatient mortality was 19.6%, 30-day mortality was 15.1% and 90-day mortality was 18.5%. A total of 31.1% of males and 30% of females were sarcopenic (30.6% overall). A multivariate analysis confirmed each method of psoas major calculation (p < 0.0001) to be associated with mortality, as was ASA-grade (p < 0.0001). Area under the curve (AUC) was greatest for PML3 in predicting mortality (inpatient PML3, 0.76; PMI, 0.71; PBSA, 0.70; 30-day PML3, 0.74; PMI, 0.68; PBSA, 0.68; and 90-day PML3, 0.78; PMI, 0.71; PBSA, 0.70). ASA-grade, P-POSSUM and PML3 were independently associated with mortality on multivariate analysis. ROC analysis of predictions from logistic regression models demonstrated PML3 to be more closely aligned to mortality than ASA or P-POSSUM (inpatient AUCPML3, 0.807; ASA, 0.783; P-POSSUM, 0.762; 30-dayAUC PML3, 0.799; ASA, 0.784; P-POSSUM, 0.787; and 90-day AUCPML3, 0.805; ASA, 0.781; P-POSSUM, 0.756). CONCLUSIONS: Sarcopenia was present in 30.6% of older adults undergoing emergency surgery and is associated with a significantly increased mortality. PML3 is superior to PMI or PBSA and should be considered the method of calculation of choice. Additionally, PML3 compares favourably to ASA and P-POSSUM.