P66 Inter-observer reliability of ultrasound to measure rectus femoris cross-sectional area in critically ill patients

Abstract

Introduction: Ultrasound is a relatively simple, non-invasive, nonirradiating effort-independent tool to measure quadriceps rectus femoris cross-sectional area (RFCSA) in critically ill patients. We investigated the inter-observer reliability of the technique to validate its clinical utility in this group of patients. Methods: Critically ill patients either in, or within 48 h discharge from, the Intensive Care Unit (ICU) underwent measurement of RFCSA using real-time B-mode ultrasonography using an 8MHz 5.6 cm linear transducer (PLM805, Toshiba Medical Systems Ltd, Crawley, UK) at a distance three-fifths from the anterior superior iliac spine to the superior patellar border. Where complete visualisation of RFCSA was not possible at this point, a more distal point of 2/3 of this distance was used. Ultrasound measurements were performed in turn by two critical care clinicians trained in ultrasound in a random order. The average of three consecutive measurements within 10% was taken as RFCSA for each patient. Both clinicians were blinded to the results of the other. Results: 24 patients had RFCSA measurements performed using ultrasound (M:F 14:10; mean age 55.3620.1 years). Inter-observer reliability was assessed by considering the level of agreement between RFCSA measurements for each patient between the two clinicians using intra-class correlation coefficients (ICC) adopting a two-way, random effects model for absolute agreement. An ICC of 0.99 (95% CI 0.97 to 0.99) was observed. Abstract P66 Figure 1 shows RFCSA images from both clinicians for one patient. (Figure presented) Conclusion: These data demonstrate high levels of inter-observer reliability between two trained critical care clinicians using ultrasound as a measurement technique for RFCSA in critically ill patients. RFCSA can be used as a novel, reproducible technique to track the trajectory of muscle loss in critically ill patients.