Muscle mass is an important biomarker of survival from a critical illness, because lower muscle mass at intensive care unit (ICU) admission is associated with higher mortality and ICU-acquired weakness [1, 2]. Recently, the notion that a state of low muscle mass is representative of sarcopenia has been gaining increased attention . The identification of sarcopenia is important in planning nutrition and rehabilitation management strategies. Both are needed to maintain muscle mass during ICU hospitalization. However, to date, there is no widely accepted method to assess low muscle mass and sarcopenia at the time of ICU admission.
Several methods have been used to assess muscle mass . Generally, dual-energy X-ray absorptiometry and bioelectrical impedance analyses have been used to assess whole-body muscle mass and sarcopenia . However, these indirect muscle mass assessments are inaccurate in critically ill patients because these are influenced by dynamic fluid changes [6–8]. During a critical illness, computed tomography (CT) is considered the gold standard to assess whole-body muscle mass because it can visually separate muscle mass from other tissues . Although CT is a reliable method to measure muscle mass, prospective CT evaluation is infeasible because of patient transportation risks and radiation exposure . Alternatively, ultrasound is an emerging tool used to measure muscle mass noninvasively at the bedside . Although ultrasound is used to assess limb muscles, it is unclear whether the partial muscle mass assessments reflect whole-body muscle mass in critically ill patients. To validate ultrasound assessments of whole-body muscle mass, it is important to show the measurement correlation between ultrasound and CT.
The rectus femoris muscle is assessed commonly using ultrasound, in which muscle thickness or cross-sectional area measurements are conducted. The measurement of the cross-sectional area is preferable because it correlates with the patient’s physical functions [12, 13]. However, it is unclear if these mass measurements reflect whole-body or partial-body muscle mass in critically ill patients. Given that a previous study reported that the cross-sectional area of the rectus femoris is preferable for muscle mass assessments, we hypothesized that this area (not the thickness) is associated with whole-body muscle mass. We retrospectively evaluated the muscle mass area at the level of the third lumbar vertebra using CT, and compared the outcomes with those obtained from prospectively obtained ultrasound data at ICU admission. This study aimed to investigate whether ultrasound measurements can replace CT regarding whole-body muscle mass measurements at ICU admission.
This two-center retrospective study was conducted in the mixed medical/surgical ICUs of Tokushima University Hospital and Tokushima Prefectural Central Hospital. The study was based on Declaration of Helsinki, and approved by the clinical research ethics committees at Tokushima University Hospital (approval number 2593) and Tokushima Prefectural Central Hospital (approval number 1739). Prospectively obtained data from May 2016 to June 2020 were retrospectively analyzed. This trial was retrospectively registered as a clinical trial (UMIN-Clinical Trials Registry: 000044032). At the time of data acquisition, written informed consent was obtained from patients or their relatives. One part of this study was published previously [7, 14].
We included patients who met the following criteria: (1) adults (≥ 18 years) admitted to the ICU; (2) those expected to stay in the ICU for more than 5 days; (3) those who underwent ultrasound assessments of the rectus femoris muscle at the day of ICU admission; and (4) those who underwent CT assessments of the third lumbar vertebra within 2 days before and after ICU admission. The following patients were excluded from the studies conducted previously: those with (1) primary neuromuscular disease and (2) obstacles at the ultrasound measurement site.
We used a linear transducer and conducted B-mode imaging. The measurements were taken at the dominant limb, with elbows and knees extended in the spine position. The transducer was placed perpendicular to the long axis of the limbs. The thickness and cross-sectional area of the rectus femoris were measured. Measurements were taken midway between the anterior superior iliac spine and the proximal end of the patella. The thickness, including the underlying vastus intermedius muscle, was measured from the superficial fascia of the rectus femoris to the uppermost part of the femur. The cross-sectional area was measured by outlining the area shown in the transverse plane. The biceps brachii muscle was measured at a distance equal to two-thirds of the distance from the acromion to the antecubital crease. The thickness, including the underlying brachialis muscle, was defined as the depth between the superficial fascia of the biceps brachii muscle and the uppermost part of the humerus. The diaphragm was measured at the end expiration on the right chest wall at the zones of apposition 0.5–2 cm below the costophrenic sinus between the antero-axillary and the midaxillary lines. Ultrasound measurements were taken by a physician (N.N.) three times, and the median value was used for evaluation. The reliability of measurements was confirmed by another ICU physician. The intraclass and interclass correlation coefficients were 0.96–0.99 and 0.99 for limbs and 0.92 and 0.96 for the diaphragm, respectively, as reported previously .
CT was used to evaluate whole-body muscle mass. The CT image at the level of the third lumbar vertebra is reported to correlate with whole-body muscle mass. A board-certified radiologist (A.Y.) retrospectively measured the total muscle mass in the CT image at the middle point of the third lumbar vertebra (L3) where transverse processes were visualized. At this slice level, the total muscle area included the psoas, quadratus lumborum, transversus abdominis, external and internal obliques, and rectus abdominis muscles. CT images acquired within 2 days before and after ICU admission were included in the analyses, and examinations conducted close to the day of ICU admission were used for comparisons in patients with multiple CT examinations. The radiologist was blinded to all clinical characteristics. Images were analyzed using ImageJ software (National Institutes of Health, Bethesda, MD, USA) . The reliability of measurements was confirmed in 10 patients by two examiners (Y.A. and N.N.). The intraclass correlation coefficient was ρ = 0.98 (p < 0.01) and the interclass value was ρ = 0.94 (p < 0.01). The Bland–Altman plot yielded a bias of − 1.24 ± 1.58 and − 4.83 to 2.34 at the 95% limits of agreement regarding intraobserver reproducibility, and a bias of − 0.94 ± 2.67 and − 6.98 to 5.10 at the 95% limits of agreement regarding interobserver reproducibility.
We used the skeletal muscle index to discriminate sarcopenia at ICU admission. The sex-specific cutoff point was set to 29.0 cm2/m2 for males and 36.0 cm2/m2 for females, as one of the commonly used cutoff points for sarcopenia in the Asian population . This cutoff point was previously reported to be important in the Japanese population .
The primary outcome was the relationship between ultrasound assessments of the rectus femoris muscle mass and CT assessments. We also assessed whether ultrasound assessments of the rectus femoris muscle can predict sarcopenia in the same manner as that assessed by CT. Secondary outcomes of this study included the relationship between ultrasound assessments at the biceps brachii muscle and diaphragm and CT assessments.
Continuous variables were presented as the mean (standard deviation) or median values [interquartile range (IQR)], whereas categorical data were presented as counts and proportions. Variables were compared using either the t-test or the Mann–Whitney U-test. The Spearman correlation coefficient was used to investigate relationships in primary and secondary outcomes. The area under the receiver operating characteristic curve (AUC) was generated to determine the cutoff values of ultrasound assessments for sarcopenia. For reproducibility, the Spearman correlation coefficient and Bland–Altman plot were determined using JMP statistical software, version 13.1.0 (SAS Institute Inc., Cary, NC, USA).