Ethical approval

This single-center, prospective, observational study was conducted in the ICU of the Third Affiliated Hospital of Southern Medical University. The local medical ethics review committee approved this study (ID: 2021.013) and waived the requirement for informed consent.

Inclusion and exclusion criteria

All patients were continuously evaluated after being admitted to ICU. Adult patients diagnosed with ICUAW (mean MRC score < 48) were considered for inclusion in the study from December 2020 to October 2021. To determine muscle strength, patients were instructed to perform specific movements such as opening or closing the eyes, sticking out the tongue, nodding the head, and raising the eyebrows, depending on their level of consciousness.

The exclusion criteria included patients aged < 18 and > 80 years, pregnancy, ICU stay < 3 days, spine injury, and severe lower limb trauma that prevented ultrasound exploration.

Study design

Patient characteristics (age, sex, admission diagnoses, comorbidities, and prescription medications) were collected upon admission to the ICU. Laboratory data (creatinine, hemoglobin, sodium, myoglobin, creatine kinase, and lactate dehydrogenase levels), vital signs (the mean systolic blood pressure, heart rate, and body temperature), illness severity (Acute Physiology Chronic Health Evaluation [APACHE] II score), and severity of organ dysfunction (Sequential Organ Failure Assessment [SOFA] score) were recorded.

At the time of hospital discharge, variables, including average hospitalization time in the ICU, average hospitalization time, and MRC score, were collected. The MRC muscle strength score was used to assess the muscle function of the upper (wrist flexion, forearm flexion, and shoulder abduction) and lower extremities (ankle dorsiflexion, knee extension, and hip flexion). The score for each movement ranges from 0 (no muscle contraction) to 5 (active movement against full resistance) [ ]. According to the clinical practice guidelines (2014 ATS), the reference standard of the ICUAW is that the average MRC muscle strength score of all tested muscles should be less than 48. Therefore, patients with an MRC score ≤ 48 were classified into the positive group (ICUAW) and those with an MRC score > 48 were included in the negative group (non-ICUAW). Muscle tests were conducted by an ICU doctor with more than 5 years of clinical experience.

Ultrasound examination

Ultrasound assessments were performed using a LOGIQ E9 (General Electric, USA) equipped with an 8–12 MHz high-frequency linear transducer. One sonographer (with 2 years of expertise in musculoskeletal ultrasound and having completed a month-long training course in ultrasound shear wave elasticity) conducted all ultrasound examinations.

For the procedure, patients were positioned supine with their lower limbs in a naturally relaxed position, exhibiting slight flexion of the hip and knee joints. The lower leg was externally rotated between 30–60° ( Fig. 1 ). The ultrasound settings were adjusted to a depth of 4 cm, gain of 40 dB, and focus targeted at the RF level. The ultrasound probe was positioned midway on the thigh, defined as the midpoint between the greater trochanter and base of the patella. This precise location was marked to ensure consistency in the imaging. For enhanced precision, bilateral measurements were averaged, and each muscle was evaluated at least three times.

Data were measured at the marks located at the middle of the thigh with supine position.

Patient ultrasound data, encompassing both B-mode and SWE details, were recorded at 2-day intervals (e.g., D1, D3, and D5) until the patient’s death or ICU transfer (noted as Dout-of-ICU). To avoid baseline differences, we used the formula (Dout of ICU-D1)/D1 (%) to represent shifts in each muscle parameter. Two postgraduate medical students verified the entered values before exporting them to the statistician for analysis.

B-mode images

Muscle metrics were determined using real-time B-mode ultrasonography. Measurements included the muscle thickness of the rectus femoris (RF-TH), cross-sectional area of the rectus femoris (RF-CSA), pennation angle of the rectus femoris (RF-PA), muscle thickness of the vastus intermedius (VI-TH), and pennation angle of the vastus intermedius (VI-RF) ( Fig. 2 ). The PA is defined as the angle between the muscle fibers and tendon membrane and serves as an indicator of muscle strength. A larger PA implies a greater volume of muscle fibers available for contraction, thereby enabling the muscle to produce greater force.

Schematic diagram of ultrasound measurement of muscle thickness, cross-sectional area, and pennation angle. (a) The black arrow shows the RF-TH and the white arrow shows the VI-TH. (b) The black dotted line shows the RF-CSA. (c) The angle between the two white lines is RF-PA. (d) The angle between the two white lines is VI-PA.

Shear wave ultrasound elastography (SWE)

SWE offers a quantitative measurement of muscle elasticity and provides greater insight into the structural and functional characteristics of the muscles and tendons. To determine the SWE-values for the RF and VI, both longitudinal and cross-sectional images of the muscles were acquired, followed by an initial grayscale ultrasound scan. The shear wave mode was manually activated once the best image parallel to the selected structure was obtained. This process yielded a dual-screen display with an elastography image to the right and a B-mode image to the left. In shear wave mode, the region of interest (ROI) for elasticity measurements was defined as a rectangular area measuring 10 mm × 10 mm. The RF and VI were scanned along their longitudinal axes, and the ROI was positioned at the middle depth of the muscles. Subsequently, a 5 mm diameter Q-Box was designed to capture real-time SWE-values (KPa) ( Fig. 4 ).

SWE measurements were obtained by ensuring alignment perpendicular to the longitudinal axes of the RF and VI. It is essential to avoid exerting pressure on the muscles during the evaluation. The probe, which was generously coated with a water-soluble transmission gel, ensured sound acoustic contact without pressing the skin. Successful examination was marked by a color image occupying more than half of the designated ROI.

Statistical analysis

Data processing and statistical analysis were conducted using SPSS25.0 statistical analysis package, and the measurement data are expressed as mean and standard deviation. The first step was to test whether each index satisfied the normality and homogeneity of variance assumptions. Appropriate statistical methods were selected to examine potential statistical differences between the positive and negative groups. If both the normality and homogeneity of variance assumptions were satisfied, an independent sample t-test was utilized. In cases where the normality assumption was satisfied but not the assumption of homogeneity of variance, a Satterthwaite test was applied. The Mann–Whitney U-test was used in other cases. The second step involved plotting the receiver operating characteristic (ROC) curve and calculating the areas under the ROC curve for the indicators that exhibited differences between the positive and negative groups. Simultaneously, the optimal threshold was determined based on the Youden index to obtain diagnostic results for each indicator. Finally, kappa coefficients were used to assess the consistency of the diagnostic outcomes with the MRC scores. Statistical significance was set at p < 0.05.