Upper Arm Muscular Echogenicity Predicts Intensive Care Unit-acquired Weakness in Critically Ill Patients.

Objectives: This retrospective observational study investigated whether the degree of muscular echogenicity in patients admitted to the intensive care unit (ICU) could help with the early detection of ICU-acquired weakness (ICU-AW) and predict physical function at hospital discharge.
Methods: Twenty-five patients who were mechanically ventilated for more than 48 h in the ICU were enrolled. We also enrolled 23 outpatients with nonmuscular diseases as the control group. The target sites for measuring muscular echogenicity were the upper arm and lower leg. First, the muscular echogenicity was compared between surviving nonsurgical patients admitted to the ICU and stable outpatients with nonmuscular diseases. Second, we investigated the relationship between muscular echogenicity and clinical features, e.g., the manual muscle test (MMT), Medical Research Council (MRC) sum score, and Functional Independence Measure (FIM).
Results: Muscular echogenicity in the upper arm in the ICU group was significantly higher than that in the control group. In the ICU group, the degree of muscular echogenicity of the upper arm was inversely correlated with the MMT of elbow flexion (P=0.006; r=-0.532) and the MRC sum score (P=0.002; r=-0.591). However, muscular echogenicity of the upper arm did not correlate with functional FIM (P=0.100; r=-0.344) at hospital discharge. Conclusions: Critically ill patients can experience pathological muscle weakness associated with increased muscular echogenicity in the upper arm. Additionally, the degree of muscular echogenicity in the upper arm correlated with the MRC sum score and can facilitate early detection of ICU-AW. The relationship between echogenicity and functional outcome at discharge requires elucidation. 2022 The Japanese Association of Rehabilitation Medicine.

INTRODUCTION

Intensive care unit-acquired weakness (ICU-AW) refers to new diffuse weakness in critically ill patients.[1]) According to a systematic review, the risk of ICU-AW is nearly 50% and the related mortality is 17%–66%.[2]) After surviving critical illness, patients with ICU-AW often suffer from severe systemic muscle weakness. The concept of ICU-AW involves critical illness myopathy or polyneuropathy or both.[3]) However, the detailed pathogenesis of ICU-AW remains unknown. Ultrasonography in the ICU may be useful for the early detection of ICU-AW. Muscular thickness and the cross-sectional areas (CSAs) of the muscles could be good indicators of ICU-AW.[4]) However, muscular echogenicity has not been fully investigated in patients admitted to the ICU; furthermore, it is still unknown whether the findings of muscular echogenicity can predict functional outcomes. This study aimed to investigate whether the degree of muscular echogenicity in ICU patients can facilitate early detection of ICU-AW and the prediction of physical function at hospital discharge.

MATERIAL AND METHODS

Patients

We extracted clinical and laboratory data from medical record systems. In this retrospective observational study, we enrolled 25 surviving nonsurgical ICU patients who were admitted to Jichi Medical University Hospital between March 2018 and January 2019 and who were mechanically ventilated for more than 48 h in the ICU. As a control group, 23 outpatients with nonmuscular diseases who were independent in activities of daily living and without weakness were also enrolled. We routinely perform muscle ultrasonography for outpatients who present at the rehabilitation clinic. In the current study, we first compared the degree of muscular echogenicity between surviving nonsurgical ICU patients and patients with nonmuscular diseases in the control group. Second, we investigated the relationships between muscular echogenicity and clinical manifestations, such as the Medical Research Council (MRC) sum score, modified Rankin Scale (mRS) score, Functional Independence Measure (FIM) score at discharge, and other laboratory findings. ICU-AW was defined as a mean MRC sum score of <4, in accordance with a previous study.[1]) The MRC sum score was assessed by two physiotherapists. Disseminated intravascular coagulation (DIC) was diagnosed on the basis of the Japanese Association for Acute Medicine’s DIC diagnostic criteria.[5]) The criteria for pancytopenia referenced the 2004 diagnostic criteria for hemophagocytic lymphohistiocytosis.[6]) Inclusion criteria were as follows: (1) patients who underwent mechanical ventilation for more than 48 h in the ICU, (2) patients who underwent ultrasonography of the upper arm and lower leg, and (3) patients aged ≥20 years. The exclusion criteria were as follows: (1) patients who were not independent in activities of daily living before admission, (2) patients who could not fully obey motor commands because of unconsciousness or uncooperativeness, and (3) patients who died during the clinical course.

The study protocol was approved by the institutional review boards of Jichi Medical University Hospitals (Approval No.: 20–190) and was conducted in accordance with the Declaration of Helsinki and good clinical practice. The ethical review board waived informed consent because this was a retrospective observational study based on medical records.

Ultrasound Protocol

A single ultrasound machine with a 5–11 MHz real-time linear array scanner (LOGIQ P7/P9, GE Healthcare, Amersham, England) was employed for all patients. One examiner performed ultrasonography weekly for all patients in the ICU, i.e., on days 3 ± 1, 10 ± 1, 17 ± 1, and so on from admission to discharge from the ICU. Patients were examined in the supine position with the upper and lower limbs extended. The center of the upper arm (the biceps brachii and brachialis muscles) and the thickest part of the lower leg (the tibialis anterior [TA] and extensor digitorum longus muscles) were targeted. Bilateral upper arms were examined, and the one exhibiting a higher echogenicity was selected for further analysis. If patients underwent several ultrasound sessions, the session with the highest muscular echogenicity was used for the analysis. The lower limbs were examined in the same manner. The representative regions of interest (ROIs) are shown in Fig. 1. The degree of gain/depth were fixed as 54/4.0 on the ultrasound settings. Follow-up ultrasonography was discontinued in patients who died. We were careful to avoid oblique scanning or excessive compression of the tissues because of the risk of artifacts. Muscular echogenicity was calculated using grayscale histogram analysis with Adobe Photoshop Elements 2018 software (https://www.adobe.com/jp/products/photoshop.html) using the following procedure: (1) Adobe Photoshop Elements 2018 was launched and an ultrasound scan was opened, (2) the drawing tool listed at the left side of the Adobe Photoshop main screen was selected, (3) the “Window” tag at the top of the screen was selected and “histogram” was checked to show the histogram on another screen, (4) the ROI was drawn manually (Fig. 1), and (5) the mean grayscale level was automatically calculated and was displayed on another screen. The histogram analysis steps 1–4 were performed again by examiner 1 (ultrasound operator) after an interval of 1 month; to assess the intra- and inter-rater reliability of histogram analysis, the same analysis was further performed by examiner 2, who was blinded to the results of examiner 1. Rehabilitation was provided for 2–5 days per week for 20–60 min per session.

Figure 1.

 Measurement ROIs for muscle echogenicity. Echogram ROIs measured for muscle echogenicity in (A, B) the upper arm and (C, D) the lower leg. (A, B) The center of the upper arm includes the biceps brachii and brachialis. The bottom of the region of interest (ROI) was set at the level of the humerus (A; arrowhead) or at 3.0 cm depth in patients whose humerus was deeper than 3.0 cm (a depth of 3.0 cm was determined based on the limitations of the resolution). (B) The entire bilateral sides and bottom were included in the ROI unless there was poor visualization. (C, D) The thickest part of the lower leg in which the tibialis anterior and extensor digitorum longus muscles were included. The entire bilateral sides and bottom were included as the ROI, unless there was poor visualization. Notable vessels were avoided (arrow). The mean muscular echogenicities of the ROIs were 59.87 (A), 74.43 (B), 65.58 (C), and 75.16 (D).

Statistical Analysis

All statistical analyses were performed using IBM SPSS Statistics for Windows, Version 26 (IBM, Armonk, NY, USA). Statistical significance was defined as α <0.05. Normal distribution of the data was assessed using the Shapiro–Wilk test, and all the items were non-normally distributed, except for the echogenicity of the upper arm and the lower leg. Normally distributed values are presented as means with SDs. Non-normally distributed values are presented as medians with interquartile ranges (IQRs). Differences in categorical variables were compared using Fisher’s exact test. The t-test was used to compare differences between normally distributed continuous variables, and the Mann–Whitney U test was used for non-normally distributed continuous variables. Correlation analyses were performed using the Pearson correlation coefficient. The area under the receiver operating characteristic (ROC) curve was calculated to determine whether muscle echogenicity can predict ICU-AW. To determine the cut-off value, we measured the minimum distance from the upper-left corner of the unit square. The setting of ROIs in the upper arm and lower leg were originally developed; therefore, the intra- and inter-rater reliabilities of the histogram analysis were assessed using the intraclass correlation coefficients (ICCs). Additionally, absolute reliability was assessed using the standard error of measurement()[7]) and the Bland–Altman method.[8]) SD is the standard deviation for all observations from analysis and reanalysis. The SEM was then used to estimate the minimal detectable change (.[7]) A two-way random-effects model, absolute agreements, and average measures were used for the analysis. ICC values of ≥0.75, 0.74–0.60, 0.59–0.40, and <0.40 indicated excellent, good, fair, and poor reliability, respectively.[9]) The Bland–Altman method with 95% confidence intervals was assessed for absolute reliability,[8]) and the 95% upper and lower differences were set as mean difference ± 1.96 × SD of the mean difference.

RESULTS

Ultrasonography was performed in 34 ICU patients: 9 of these patients died and 25 (12 women; median age, 67 years [IQR, 60–75 years]) survived. The control group consisted of 23 outpatients (6 women; median age, 68 years [IQR, 56–71.5 years], median serum creatinine kinase 102.5 U/L [IQR, 73–174.5 U/L]). No statistically significant difference was found in the percentage of women (P=0.237) or age (P=0.457) between the ICU group and the control group. In the ICU group, the diagnoses were as follows: pneumonia (n=8), sepsis (n=4), peritonitis (n=3), hemorrhagic shock (n=2), hemophagocytic lymphohistiocytosis (n=1), systemic lupus erythematosus (n=1), myocardial infarction (n=1), dermatomyositis (n=1), symptomatic ventricular tachycardia (n=1), iliopsoas abscess (n=1), pulmonary edema (n=1), and bronchial asthma (n=1). The ICU group showed significantly higher echogenicity in the upper arm than that of the control group (70.07 ± 11.05 vs. 58.45 ± 9.41, P=0.008), but no significant difference was found in the echogenicity in the leg (75.29 ± 11.65 vs. 63.66 ± 8.89, P=0.068). In our histogram analysis, both intra-rater and inter-rater reliability were excellent (Table 1). In Fig. 2, the Bland–Altman plots revealed that all analyses–reanalyses by examiner 1 were within the limits of agreement (LOA). Additionally, the analyses by examiners 1 and 2, except for one patient, were within the LOA.

Table 1.

 Intra- and inter-rater reliability values of histogram analysis

	ICC	SEM	MDC95
Upper arm
Intra-rater reliability	0.995	0.78	2.16
Inter-rater reliability	0.994	0.86	2.37
Lower leg
Intra-rater reliability	0.998	0.02	0.06
Inter-rater reliability	0.992	1.04	2.89

ICC, intraclass correlation coefficients; SEM, standard error of measurement; MDC95, minimal detectable change 95%.

Figure 2.

 Scatter plots and Bland–Altman plots of the analysis and reanalysis of ROIs. (A) Simple scatter plot of the echogenicity analysis and reanalysis by examiner 1. Statistical correlation was observed (P <0.001; r=0.992). (B) Simple scatter plot of analysis by examiners 1 and 2. Statistical correlation was observed (P <0.001; r=0.988). The solid diagonal lines are the reference lines. (C) Bland–Altman plots of the echogenicity analysis and reanalysis by examiner 1. All plots are within the limits of agreement. (D) Bland–Altman plots of the analysis by examiners 1 and 2. All plots except for one are within the limits of agreement. Solid horizontal lines correspond to the mean differences between the measurement of analysis and reanalysis (mean). Dotted lines indicate the limits of agreement; 95% upper and lower of differences (i.e., mean ± 1.96 SD).

Table 2 shows the clinical manifestations of the ICU patients. In the ICU group (n=25), muscular echogenicity in the upper arm was significantly correlated with the Manual Muscle Test (MMT) of elbow flexion (P=0.006; r=−0.532) (Fig. 3A) and with the MRC sum score (Fig. 3B, P=0.002; r=−0.591). The ROC curve for determining ICU-AW indicated a cut-off level for muscular echogenicity in the upper arm of 74.00; this yielded a specificity of 75.0% and a sensitivity of 69.2% (Fig. 3C). Muscular echogenicity in the lower leg did not significantly correlate with the MMT of ankle dorsiflexion (P=0.090; r=−0.346) or the MRC sum score (P=0.107; r=−0.337). Moreover, muscular echogenicity in the upper arm did not significantly correlate with mRS (P=0.095; r=0.415), total FIM (P=0.089; r=−0.356), or functional FIM (P=0.100; r=−0.344) scores at hospital discharge. The degree of muscular echogenicity had no obvious correlation with the reduction in body weight, steroid use, continuous muscle relaxant therapy, or any other laboratory findings, including creatinine kinase levels.

Table 2.

 Baseline clinical manifestations of ICU patients

	n=25
Age (median, IQR)	67 (60–75)
Female (%)	12 (48.0)
ICU-AW (%)	13 (52.0)
Best MRC score (median, IQR)	40 (34–52)
CHDF (%)	4 (16.0)
Dysphagia (%)	10 (40.0)
Tracheotomy (%)	5 (20.0)
Continuous muscle relaxant treatment (%)	9 (36.0)
Steroid treatment (%)	19 (76.0)
Laboratory findings in ICU (median, IQR)
Pancytopenia (%)	10 (40.0)
DIC (%)	13 (52.0)
Maximum D Bil (mg/dL)	0.50 (0.33–1.78)
Maximum T Bil (mg/dL)	1.20 (1.04–2.6)
Maximum AST (U/L)	86 (63–235)
Maximum CK (U/L)	407 (111–1918)
Maximum Cre (mg/dL)	1.71 (1.04–3.59)
Maximum sIL2R (U/mL)	1855 (1680–2378)
Maximum ferritin (ng/mL)	999.5 (363.7–2645.6)
Maximum procalcitonin (ng/mL)	2.76 (0.47–21.25)
Maximum CRP (mg/dL)	18.92 (7.20–31.33)
Minimum albumin (g/dL)	2.0 (1.6–2.2)
Muscular echogenicity (mean grayscale level ± SD)
Upper arm	70.07 ±11.05
Lower leg	75.28 ±11.65
BW admission ICU (median, IQR)	59.7 (55.4–68.6)
Minimum BW (kg) after ICU (median, IQR)	49.9 (40.97–57.7)
Reduction of BW (kg) (median, IQR)	8.8 (4.6–12.0)
Hospital duration (day) (median, IQR)	45 (29–67)
ICU duration (day) (median, IQR)	9 (7–12)
Functional outcome at discharge (median, IQR)
Barthel index	60 (30–90)
Modified Rankin Scale	3 (2–4)
Total FIM	89 (61–112)
Functional FIM	54 (28–77)

AST, aspartate aminotransferase; BW, body weight; CHDF, continuous hemodiafiltration; CK, creatine kinase; Cre, creatinine; CRP, C-reactive protein; D Bil, direct bilirubin; DIC, disseminated intravascular coagulation; sIL2R, soluble interleukin-2 receptor; T Bil, total bilirubin; TG, triglyceride.

Figure 3.

 (A, B) Scatter plot of muscular echogenicity and function of the upper arm. Muscle echogenicity of the upper arm was significantly correlated with MMT (P=0.006; r=−0.532) and the MRC sum score (P=0.002; r=−0.591). (C) The ROC curve for determining ICU-AW. The cut-off level for muscle echogenicity of the upper arm was 74.00 (specificity, 75.0%; sensitivity, 69.2%). The AUC was 0.756.

DISCUSSION

This study demonstrated that critically ill patients can develop significant pathological muscle weakness that is associated with increased muscular echogenicity in the upper arm. The degree of muscular echogenicity in the upper arm was correlated with the MRC sum score and can help with early detection of ICU-AW. However, the relationship between muscular echogenicity and functional outcome at discharge remains to be elucidated.

We suggest that muscular echogenicity in the upper arm is a reliable indicator for ICU-AW screening. To the best of our knowledge, only seven studies have investigated muscular echogenicity in ICU patients (Table 3).[10],[11],[12],[13],[14],[15],[16]) Of the seven studies (Table 3), three suggested that the rectus femoris muscle is the preferable site for investigation.[11],[12],[16]) Additionally, in one study, the vastus intermedius muscle was the preferred site for ICU-AW screening.[13]) However, none of these seven studies have successfully shown a relationship between muscular echogenicity in the upper arm and ICU-AW. Critical illness can rapidly exacerbate weakness of the upper arm muscles: the thickness and CSA of the biceps brachii muscles decreased by 13.2% and 16.9%, respectively over 7 days after ICU admission[4]); mid-arm muscle thickness decreased by 7.6% over 7 days in patients with sepsis[17]); and decreased CSA of the biceps brachii was associated with the MRC sum score (r=0.47; P=0.01), handgrip strength (r=0.50; P=0.01), and functional status score (r=0.56; P <0.01) at ICU discharge.[18]) The degree of muscular echogenicity may differ in ROIs with different sizes, shapes, and locations, even in the same study participant.[19]) Therefore, the lack of evidence regarding muscular echogenicity may result from different ROIs of ultrasound images in each study. In the present study, the increase of muscular echogenicity was heterogeneously or locally developed in some patients early in the ICU stay. Therefore, we consider that ROIs should involve both areas of pathologically increased echogenicity and less affected areas; this approach should correctly reflect the entire muscular function. The muscle boundary was occasionally difficult to demarcate because of attenuation or unclear accumulation of thick fascia (asterisk in Fig. 4A). However, our analysis could be performed simply and showed excellent intra- and inter-rater reliabilities. Our analysis needs ROIs to be drawn manually and possibly includes analysis errors. However, the intra- and inter-rater SEM levels were very small (below 1 degree). Therefore, our method is adequately feasible for the analysis. Additionally, the MDC95 indicated that a change of more than 3 degrees can be interpreted as a real change with 95% confidence. This threshold is possibly critical for clinical estimation in follow-up of echogenicity. The cut-off level of upper arm echogenicity by histogram analysis for predicting ICU-AW was 74.00 (specificity, 75.0%; sensitivity, 69.2%), which may be useful for the early diagnosis of ICU-AW. We suggest that the finding of increased muscular echogenicity in the upper arm predicts difficulty in arm-related activities, such as dressing, grooming, and eating. Therefore, some arm-related activities should be incorporated early into the rehabilitation program.

Table 3.

 Summary of ultrasound studies on muscular echogenicity in ICU patients

Author, year	Number of patients	Targeted disease/condition of recruited ICU patients	Ultrasound equipment	Target muscle	Region of interest	Timing of ultrasound	Main ultrasound findings for echogenicity
The present study	n=34	Adults, ventilated >48 h	5–11 MHz linear	Upper arm (biceps brachii and brachialis) and lower leg (TA and EDL)	See Fig. 1	Weekly in ICU (day 3 ± 1, day 10 ± 1, day 17 ± 1, and so on)	The degree of echogenicity of the upper arm is correlated with MRC sum score, which is useful for ICU-AW screening.
Grimm et al., 2013[10])	n=28	ICU-AW associated with sepsis	9–13 MHz linear	Upper arm, forearm, thigh, and lower leg	GME scale using Heckmatt scale	Days 2–5, Day 14	Ultrasound could detect morphological changes early in the course of sepsis and is useful for ICU-AW screening prior to electromyography or biopsy.
Cartwright et al., 2013[11])	n=16	Adults with acute respiratory failure	18 MHz linear	Biceps, ADM, RF, TA	2 × 2 cm or 1 × 1 cm square	Within 80 h, and on Days 3, 7, and 14	TA and RF had significant decreases in grayscale standard deviation when analyzed over 14 days. No significant echogenic change was observed in the biceps or ADM.
Puthucheary et al., 2015[12])	n=30	Surviving ICU, ventilated >48 h, staying >7 days in ICU	ND	RF, fascial tissue	The region of predemarcated RF-CSA	Days 1, 10	Myofiber necrosis and fasciitis can be detected by ultrasound. Changes in RF muscle echogenicity were greater in patients who developed muscle necrosis.
Parry et al., 2015[13])	n=22	Adults, ventilated >48 h, staying ≥4 days in ICU	8.5 MHz linear	RF, VI	2 × 2 cm square	Days 1, 3, 5, 7, 10, and ICU discharge date	Strong association between muscle function and VI echogenicity (r=−0.77). The VI may be an important muscle for monitoring. No significant echogenic change was observed in RF.
Witteveen et al., 2017[14])	n=71	Ventilated ≥48 h	4–13 MHz linear	Biceps, flexor, carpi radialis, RF, and TA	Contours of each muscle, just below the fascia. Lateral borders were excluded	As soon as the patients were awake and cooperative	Muscular ultrasound was not able to reliably diagnose ICU-AW relatively early in the disease course.
Patejdl et al., 2019[15])	n=15	SOFA score ≥8 for three consecutive days within the first 5 days in the ICU, age ≥18 years	4–14 MHz linear	Biceps, brachioradialis quadriceps, TA	GME scale using Heckmatt scale	Day 3, day 10 after study inclusion	GME scale significantly correlated with mRS at day 100 (r=0.67, P= 0.013)
Mayer et al., 2020[16])	n=41	Sepsis or acute respiratory failure; surviving patients aged ≥18 years, staying >3 days in ICU	8.5 MHz linear	RF, TA	Not described in detail	Days 1, 3, 5, and 7	Change in RF echogenicity in first 7 days in the ICU was a predictor of diagnosis of ICU-AW.

ADM, abductor digiti minimi; GME, global muscle echogenicity; MMT, manual muscle test; TA, tibialis anterior; EDL, extensor digitorum longus; RF, rectus femoris; VI, vastus intermedius; SOFA, sequential organ failure assessment; ND, no data.

Fig. 4.

 Representative ultrasound images of the upper arm and lower leg in patients with ICU-AW. (A) The boundary between the biceps brachii and brachialis was unclear due to accumulated dense endomysium accompanied by high echogenicity (*). (B) A thick perimysium with unclear boundary was noted (arrows). (C, D) The endomysium in the tibialis anterior showed increased echogenicity, which is apparent in the long-axis plane.

In contrast, the utility of muscular echogenicity in the lower leg was controversial. A previous study demonstrated that the CSA of the TA decreased by 21% in the acute phase, and it was considered that the muscle wasting was due to the critical illness.[20]) However, in the current study, the muscular echogenicity in the lower leg had no relationship with muscle power. The muscle power on ankle dorsiflexion was preserved in some ICU patients with high leg muscle echogenicity. Moreover, no significant difference was found between the echogenicity of the lower leg in ICU patients and in the control group. This result may supplement those reported by previous studies; only one of three studies[11]) could show the echogenicity of TA reflected the pathological weakness in Table 3.

We could not demonstrate the pathogenesis of the increased muscular echogenicity because no histopathological investigations were carried out. The boundary layers of the muscle fibers, fascia, fat, fibrous tissue, and interstitial fluid all reflect the ultrasound beam, and the level of the returning echoes per unit area determines the echogenicity of the image.[21]) In this study, no correlation was found between the serum CK level and the degree of muscular echogenicity. However, thick perimysium (Fig. 4B) or accumulated dense endomysium (Fig. 4A) was frequently observed, which contributed to the increase in muscular echogenicity. Outstanding endomysium was also observed in the lower leg (Fig. 4C, D). One possible pathogenesis for the increased muscular echogenicity may include fasciitis; fasciitis can be detected as a thickened fascia that makes the surrounding muscle and fascia less demarcated on the ultrasound image.[21]) In the ICU patients, fasciitis was observed in 60% of biopsies; additionally, ultrasound investigations found that fascial edema was linked to the presence of histologically confirmed fascitis.[12])

This study has several limitations. We were unable to perform multivariate analysis because of the small sample size. Some patients required long-term treatment for primary diseases, such as autoimmune disorders, even after recovering from the critical condition, and this may have affected the functional outcome. Some patients underwent the assessment of muscular echogenicity only once because of their short ICU stay; consequently, intrapatient changes in muscular echogenicity could not be investigated statistically. Furthermore, because of difficulty in the analysis, we excluded patients who died; in fatal cases, the muscular echogenicity decreased inversely with longer stays in the ICU. We speculate that this reduction in ultrasound echogenicity resulted from extreme interstitial edema accompanied by deep attenuation of ultrasound. Echogenicity values are specific to a single ultrasound machine; values cannot be compared to data from another ultrasound machine (unless calibrated).[22]) In addition, the echogenicity may be easily changed by the ultrasound protocol; therefore, each laboratory should have their reference values.

CONCLUSION

Upper arm muscular echogenicity may be useful for the early diagnosis of ICU-AW. ICU-AW is often noticeable after discharge from the ICU and can delay appropriate physical and mental care. Therefore, early diagnosis of ICU-AW is fundamental. Ultrasonography of the muscles is useful for ICU-AW screening, even in unconscious or uncooperative patients, before more invasive techniques such as electromyography and biopsy are performed. Further sonographic studies are needed to establish the ultrasound protocol for detecting ICU-AW and to elucidate its pathogenesis.