Quantitative Peripheral Muscle Ultrasound in Sepsis: Muscle Area Superior to Thickness

Abstract

Purpose

The objective of this study is to describe the relationship between two quantitative muscle ultrasound measures, the rectus femoris cross-sectional area (RF-CSA) and quadriceps muscle thickness, with volitional measures of strength and function in critically ill patients with sepsis.

Materials and Methods

We performed a prospective study of patients admitted to a medical ICU with sepsis and shock or respiratory failure. We examined the association of two ultrasound measurements – the RF-CSA and quadriceps muscle thickness – with strength and function at day 7. Strength was determined using the Medical Research Council Score and function using Physical Function in the ICU Test, scored.

Results

Twenty-nine patients were enrolled; 19 patients had outcome testing performed. Over 7 days, RF-CSA and thickness decreased by an average of 23.2% and 17.9%, respectively. The rate of change per day of RF-CSA displayed a moderate correlation with strength (ρ 0.51, p-value 0.03) on day 7. Baseline and day 7 RF-CSA did not show a significant correlation with either outcome. Quadriceps muscle thickness did not significantly correlate with either outcome.

Conclusions

Muscle atrophy as detected by the rate of change in RF-CSA moderately correlated with strength one week after sepsis admission.

Introduction

Skeletal muscle dysfunction develops early and rapidly during the course of critical illness.[1] This dysfunction, defined clinically as intensive care unit acquired weakness (ICU-AW), is associated with prolonged mechanical ventilation, prolonged intensive care unit (ICU) stays, and increased hospital and one-year mortality.[2–5] Weakness is also a major driver of morbidity after hospital discharge in survivors of critical illness.[6–8] Critically ill patients with sepsis are at increased risk for reduced muscle strength and impaired exercise tolerance after discharge.[9] The diagnosis of skeletal muscle dysfunction in critically ill septic patients, however, remains a challenge. The most widely accepted method of diagnosing ICU-AW requires volitional participation by the patient.[2, 10, 11] Critical illness itself as well as concomitant delirium or sedation precludes such participation in the majority of patients.[2] As such, the true incidence of skeletal muscle dysfunction in sepsis remains unknown and risk assessment early in the course of critical illness is rarely possible.

Peripheral muscle ultrasound is appealing as a diagnostic and risk stratification tool for skeletal muscle dysfunction in critical illness. Ultrasonography is widely available, performed at the bedside, and does not require patient participation. Quantitative muscle ultrasound measurements have been shown to be reliable in critically ill patients.[1, 12] Challenges with this modality, however, still exist. There is inconsistency in the literature about which quantitative ultrasound measurement to use.[13] In addition, it is unknown if the absolute muscle thickness or area on ultrasound or the rate of muscle loss over time more strongly correlates with muscle strength and function.[14] Patients with sepsis are at increased risk for skeletal muscle dysfunction[9, 15, 16] and, if further developed, peripheral quantitative muscle ultrasound may be an important tool to describe and follow muscle dysfunction specifically in this subset of patients. Prior research has shown that a reduction in cross-sectional area correlates with muscle weakness in a mixed cohort of mechanically ventilated patients.[17] In a cohort including only patients with sepsis, however, static measures of muscle thickness at 10 days post-ICU admission did not correlate with weakness.[18]

The objective of this study is to describe the relationship between two quantitative muscle ultrasound measures, the rectus femoris cross-sectional area (RF-CSA) and quadriceps muscle thickness (Q-MT), with volitional measures of strength and function in a cohort of patients with sepsis complicated by shock or respiratory failure, a very high risk population. In contrast to prior studies, we examined these relationships at specific time points within one week after ICU admission to limit the variability in time points assessed.

Material and Methods

Study Design

We conducted a single-center prospective cohort study of patients with sepsis or septic shock conducted in the medical ICU at an urban academic referral center. The Institutional Review Board at the University of Pennsylvania approved the study, Protocol #820585. This study was approved with a waiver of timely consent. Consent was obtained from patients or surrogates as soon as was feasible.

Study Population

Eligible patients were admitted to the medical ICU at the Hospital of the University of Pennsylvania between June and December 2015 with a diagnosis of sepsis. In order to identify the most severely ill patients with sepsis, we also required patients to have sepsis complicated by respiratory failure or shock requiring vasopressors for a minimum of 6 hours, and an anticipated ICU length of stay > 48 hours. Sepsis was defined by the American College of Chest Physicians (ACCP) and Society of Critical Care Medicine (SCCM) consensus criteria for severe sepsis and septic shock.[19] Respiratory failure was defined as the need for invasive mechanical ventilation, non-invasive mechanical ventilation, or high flow nasal cannula with fraction of inspired oxygen > 50%. Patients were sampled consecutively. To minimize the impact of premorbid muscle dysfunction, patients were excluded if they were transferred to our ICU from a long-term acute care facility or outside hospital more than 48 hours after the onset of their critical illness, if they had a tracheostomy in place at the time of admission, or were admitted to the ICU for at least 7 days in the prior 3 months. Patients who were non-English speaking or cognitively impaired and unable to follow physical therapy commands were also excluded from the study. Finally, we excluded patients with preexisting neuromuscular disease including multiple sclerosis, amyotrophic lateral sclerosis, preexisting paresis of bilateral lower extremities, and acute spinal cord injury.

Ultrasound Protocol

A standardized ultrasonography protocol was adapted from previously published protocols (see supplementary appendix).[1, 20–22] A portable ultrasound machine (SonoSite_TM Edge) was used with a 5-10 MHz linear transducer and 2-5 MHz curvilinear transducer. One experienced critical care physician obtained all of the images (JP). This physician underwent training by a fellowship-trained musculoskeletal radiologist and a subset of the images was reviewed with the radiologist to ensure accurate image capture. Initial ultrasound measurements were performed within 48 hours of ICU admission. Follow-up images were obtained within 48 hours of day 7 after ICU admission. All ultrasound measurements were taken with the transducer placed perpendicular to the long axis of the thigh. The site at two-thirds of the distance between the anterior superior iliac spine and the superior patellar border was marked with a skin marker. Light pressure was used and an image was obtained with the entire rectus femoris muscle within view. At each time point, measures were repeated until three measures were made within 10% of each other. These three measures were then averaged for analyses. Ultrasonography was performed on both lower extremities and the right lower extremity was used in the analysis unless there was technical inferiority of the right-sided images.

Quantitative Muscle Ultrasound Measurements

We examined two quantitative peripheral ultrasound measurements (Figure 1). Quadriceps muscle thickness (Q-MT) was defined as the thickness of the rectus femoris-vastus intermedius complex and measured as the linear distance from below the subcutaneous tissue down to the level of the femur. Rectus femoris cross-sectional area (RF-CSA) was defined as the area of the rectus femoris muscle calculated by planimetric technique. The linear array probe was used to obtain the Q-MT measurements. We chose to use the curvilinear probe for the RF-CSA measurement to allow for greater window width and depth penetration because during pilot testing it was noted that the entire muscle was often not visible in a single view with the linear array probe. Utilization of the curvilinear probe for RF-CSA measurement has been previously validated.[23]

Figure 1. Representative Ultrasound Images.

A. Ultrasound Image of Rectus Femoris Cross-Sectional Area (RF-CSA) using curvilinear probe. The area is outlined by dotted line. B. Ultrasound Image of Quadriceps Muscle Thickness (Q-MT) using linear array probe. The thickness measure is the linear distance from below the subcutaneous tissue to the level of the femur as labeled with the arrow.

Outcomes

Strength was determined with manual muscle testing using the Medical Research Council Score (supplementary Table E1).[24] Function was determined using the Physical Function in the ICU Test, scored (PFIT-s) (supplementary table E2).[25, 26] MRC and PFIT-s were obtained on day 7 by critical care trained physical therapists using a standardized approach. All scores were recorded on a standardized score report form. Physical therapists were blinded to ultrasound assessments. Patients received a score of 0 on MRC and PFIT testing for the inability to follow commands or hemodynamic instability that persisted at the time of MRC assessment. PFIT was recorded on the ordinal scoring system (0-12) and converted to the equivalent interval score using the previously published algorithm for all analyses.[26]

Statistical Analysis

The Wilcoxon signed rank test was used to determine if there was a significant change in Q-MT or RF-CSA measurements between admission and day 7. Spearman correlation coefficients were used to determine correlation between change in Q-MT or RF-CSA and MRC or PFIT score.[27] As exploratory analyses, we also tested for correlation between change in RF-CSA and either fluid balance or sequential organ failure assessment score. All statistical analyses were performed with Stata/IC 13.0 (StataCorp LP). All tests were two-sided with a significance level of 0.05. With a sample size of 29 subjects, we had 80% power to detect correlation coefficients of 0.50 or stronger (assuming a 2-tailed alpha of 0.05). With a sample size of 19 subjects, we had 80% power to detect a correlation coefficient of 0.60 or stronger.

Results

A total of 92 patients were eligible for enrollment and 29 patients were enrolled (Figure 2). Forty patients were missed and 15 patients or surrogates declined consent. Initial ultrasonography was attempted in 31 patients. However, image quality was poor in 2 patients with morbid obesity; thus measurements were not possible in these patients. A total of 29 patients had RF-CSA and Q-MT measures obtained at admission. All patients had the initial ultrasound performed within 40 hours of ICU admission (median 22.1 hours, IQR 17.4 to 28.7 hours). A total of 19 patients had follow-up ultrasound images obtained at day 7. The mean duration between the initial ultrasound and follow-up ultrasound was 6.0 days (IQR 5.9 to 7.0 days). Follow-up images were not obtained in 10 patients for the following reasons: death prior to day 7 (n=6), surrogate refusal (n=1), discharged home prior to day 7 (n=1), and goals of care transitioned to comfort (n=1). Follow-up images were not analyzed in one patient after radiologist review given profound cachexia and inability to visualize borders of RF-CSA. Volitional strength measures were completed in 18 patients. One patient was enrolled in hospice on day 7 immediately prior to physical therapy assessment thus MRC and PFIT-s were missing in this patient.

Figure 2. Flowchart of Patients.

Abbreviations: LTACH, long-term acute care hospital; OSH, outside hospital; DNR/DNI, do-not-resuscitate/do-not-intubate

Patient characteristics of the 19 patients with both admission and follow-up ultrasound measurements are included in Table 1. This is a severely ill patient population with a median APACHE III score of 91 (IQR 76, 99) and median ICU length of stay of 8.0 days (IQR 2.7, 11.2). The majority of patients received steroids at some point during their ICU stay (n=13). Thirteen patients (45%) of the 29 enrolled patients died during the hospital stay and 6 of the 19 patients with follow-up images died (n=4) or were transferred to hospice (n=2) within 30 days.

Table 1.

Baseline Characteristics of Patients

Patient Characteristics (n=19)
Age (median, IQR)	66 (54, 75)
Male sex (n, %)	12 (63)
Race (n, %)
White	12 (63)
Black	4 (21)
Unknown/Other	3 (16)
Comorbidities (n, %)
Immunocompromised	13 (68)
History of leukemia or lymphoma	7 (37)
History of metastatic cancer	3 (16)
History of liver disease	1 (5)
History of COPD	4 (21)
History of heart disease	5 (26)
History of diabetes	5 (26)
APACHE III (median, IQR)	91 (76, 99)
Day 0 SOFA score (median, IQR)	13 (12,16)
Medication Use (n, %)
Steroids prior to admission	3 (16)
Steroids during ICU stay	13 (68)
Neuromuscular blockade	1 (5)
Shock requiring vasopressors (n, %)	14 (74)
Respiratory Support (n, %)
Invasive mechanical ventilation	17 (89)
High flow nasal cannula	2 (11)
Duration of ICU stay, days (median, IQR)	8.0 (2.7, 11.2)
Discharge location
Home	3 (16)
Home with services	6 (32)
Rehabilitation center	2 (11)
Skilled nursing facility	2 (11)
Hospice	2 (11)
Death	4 (21)

Median MRC score was 40 (IQR 0, 54) in the 19 patients with follow-up images. Four (21%) patients received a score of 0 and 10 patients (53%) had a score < 48 consistent with the diagnosis of ICU-AW. Of the four patients that scored 0 on MRC testing, three were unable to follow commands and one patient had unstable hemodynamics. Median PFIT-s was 4 (IQR 0, 6.4). Five (26%) of patients received a PFIT-s of 0. One patient was enrolled in hospice on day 7 immediately prior to physical therapy assessment thus MRC and PFIT-s were missing in this patient.

Rectus Femoris Cross-Sectional Area

Median RF-CSA at admission was 4.33 cm² (IQR 2.35 cm², 5.27 cm²) in the 29 enrolled patients and 4.52 cm² (IQR 2.53 cm², 6.77 cm²) in the 19 patients with follow-up images. In the 10 patients with only admission images, the median RF-CSA was smaller at 2.78 cm² (IQR 1.53, 4.60) though the difference between the median RF-CSA in those with only admission images and those with images at two time points did not reach statistical significance (p-value for Wilcoxon rank sum test was 0.06). Baseline RF-CSA was not correlated with APACHE III (ρ 0.05, p=0.80). The median reduction in RF-CSA per day was −0.13 cm² (IQR −0.30cm², −0.08cm²) and the median percentage of RF-CSA lost over the first week of critical illness was 23.2% (IQR − 32.6%, −11.9%) (Figure 3). There was a statistically significant difference between RF-CSA at admission and day 7 (p ≤ 0.001). There was no correlation between day 0 Sequential Organ Failure Assessment (SOFA) score and reduction in RF-CSA over the first week in this cohort (ρ 0.07, p=0.78).[28] Day 0 fluid balance was not correlated with RF-CSA at admission (0.12, p=0.63). There was a negative correlation between reduction in RF-CSA and total fluid balance over the first 3 days that did not reach statistical significance (ρ=−0.41, p=0.08).

Figure 3. Change in Quantitative Ultrasound Measurements Over 7 Days.

Each line represents one patient, connecting the admission ultrasound value for that patient to the value on day 7.

In the 18 patients with outcome measures, the reduction per day of RF-CSA over the first week of critical illness displayed a moderate correlation with MRC score (ρ 0.51, p=0.03). The correlation with PFIT-s was also moderate but did not achieve statistical significance (ρ 0.40, p=0.10) (Table 2). There was no significant correlation between static RF-CSA at admission or day 7 and MRC or PFIT-s (Table 2).

Table 2.

Correlation of Muscle Measurement with MRC and PFIT Scored at Day 7, n=18

	MRC Score		PFIT Scored
	Spearman	P-value	Spearman	P-value
RF-CSA
RF-CSA at ICU admission (cm2)	−0.24	0.34	−0.12	0.63
RF-CSA at Day 7 (cm2)	−0.07	0.79	0.03	0.90
Rate of loss of RF-CSA per day (cm2/day)	0.51*	0.03*	0.40	0.10
Q-MT
Q-MT at ICU admission (cm)	−0.16	0.52	−0.06	0.82
Q-MT at Day 7 (cm)	−0.18	0.48	−0.07	0.80
Rate of loss of QT per day (cm/day)	−0.07	0.77	−0.11	0.68

Abbreviations: MRC, Medical Research Council; PFIT, Physical Function in the ICU Test; Q-MT, Quadriceps Muscle Thickness; RF-CSA, Rectus Femoris Cross-Sectional Area

^*Significant at alpha 0.05 level

Quadriceps Muscle Thickness

Median baseline quadriceps muscle thickness was 2.23 cm (IQR 1.39 cm, 2.82 cm) in the 29 patients enrolled and 2.46 (IQR 1.71, 3.00) in the 19 patients with follow-up images. In the 10 patients with images only at admission, the median Q-MT was smaller at 1.43 cm (IQR 0.99 cm, 2.72 cm) though the difference between the median Q-MT in those with only admission images and those with images at two time points did not reach statistical significance (p-value for Wilcoxon rank sum test was 0.07). Baseline Q-MT was not correlated with APACHE III (ρ 0.002, p=0.99). The median reduction in Q-MT per day was −0.06 cm (IQR −0.09 cm, −0.04 cm) and the median percentage of muscle thickness lost over the first week of critical illness was 17.9% (IQR, −27.7%, −8.2%) (Figure 3). There was a statistically significant difference between Q-MT at admission and Q-MT at day 7 (p=0.02). There was no significant correlation between day 0 Sequential Organ Failure Assessment (SOFA) score and reduction in Q-MT over the first week in this cohort (ρ 0.11, p=0.64). Neither day 0 fluid balance nor fluid balance over first 3 days of ICU admission were correlated with muscle thickness (ρ=0.24, p=0.33 and ρ =0.10, p=0.68, respectively).

In the 18 patients with outcome measures, the reduction per day in quadriceps muscle thickness over the first week of critical illness was not correlated with MRC score (ρ −0.07, p= 0.77) or PFIT-s (ρ −0.11, 0.68) (Table 2). Q-MT at admission or day 7 did not significantly correlate with MRC or PFIT-s (Table 2).

We performed two sensitivity analyses. In a sensitivity analysis excluding subjects who received steroids during their ICU stay, leaving only 6 subjects, the correlation between RF-CSA reduction per day and MRC and P-FIT appeared stronger than in the overall population (ρ 0.67 for both), though these results were not statistically significant (p=0.15). Furthermore, excluding patients with MRC score of 0, correlation coefficients attenuated only slightly for RF-CSA (ρ 0.38) and Q-MT (ρ 0.04).

Discussion

Our study adds to the literature describing the relationship of quantitative ultrasound measurements with volitional measures of strength and function in a cohort of patients with sepsis. Our study confirms that patients with sepsis suffer substantial muscle atrophy and this occurs early in the course of critical illness. Patients lost an average of 23.2% of their RF-CSA in just the first week after ICU admission. While this is greater than the loss in RF-CSA detected by ultrasound in other studies of critically ill patients, all patients in our cohort had sepsis. [1, 14] Sepsis is a risk factor for the development of neuromuscular dysfunction[2, 10, 30–33] and patients with sepsis have been shown to develop neuromuscular dysfunction disproportionate to what is expected from immobilization alone.[9, 15] Our cohort also included patients with high severity of illness, an independent risk factor for muscle loss in several studies.[29, 30] While prior studies have shown a greater degree of muscle loss in patients with multi-organ dysfunction, the rate of muscle loss determined by RF-CSA or Q-MT did not correlate with admission SOFA score in our cohort. Our cohort also included a significant proportion of immunocompromised patients and patients with underlying malignancy. There is a growing interest in understanding the intersection of premorbid illness and function with ICU-AW.[33] Pre-illness function and frailty have been shown to associate with loss of function after critical illness.[18, 34, 35] The role of pre-ICU frailty and disease in the development of neuromuscular dysfunction during sepsis remains unknown. We had limited information regarding clinical frailty in our cohort and suggest future studies examining muscle wasting in sepsis prospectively collect frailty scores in addition to comorbidities.

In our cohort of patients with sepsis, the rate of change in RF-CSA over 7 days moderately correlated with strength one week after admission. However, static measurements of RF-CSA at admission or day 7 did not correlate with strength or function at one week. Baseline measurements were numerically smaller in patients with only admission images compared to patients alive at day 7 with follow-up images and we cannot exclude the possibility that with increased power, we may have detected smaller baseline muscle size in those who die before day 7 compared to those who did not. In our cohort, the rate of change of Q-MT, baseline Q-MT, and day 7 Q-MT did not correlate with strength or function at one week. In a cohort of 22 mixed medical/surgical patients, Parry and colleagues reported a moderate correlation between the static measure of RF-CSA and PFIT-s at the time of ICU discharge (r=0.71) and a fair correlation between RF-CSA and MRC and PFIT-s at the time of ICU awakening (r=0.47 and 0.44 respectively) though this did not reach statistical significance in their cohort.[14] They did not specifically examine the association between the admission muscle measurement nor the rate of change in muscle measurement with strength and function. We chose to evaluate the correlation of quantitative measurements with volitional measures at day 7 in all patients as the time of ICU awakening can vary significantly and is impacted by multiple factors including but not limited to severity of illness, sedation, and delirium. ICU awakening occurred at a median of 10 days after admission (IQR 7 to 16 days) in the Parry et al. cohort. Our goal was to determine the relationship between ultrasound measures and volitional measures at a fixed time point with the ultimate goal of using quantitative peripheral muscle ultrasound to risk stratify patients early in the course of critical illness for subsequent skeletal muscle dysfunction.

Our results suggest that the rate of change in RF-CSA has the stronger relationship with volitional outcome measures compared to static measures of area at admission or day 7. While baseline muscle mass itself did not have a strong relationship with strength or function on day 7, it was important to obtain this measure to determine the rate of change. Our findings have practical implications with regard to the use of quantitative peripheral muscle ultrasound. The admission image can be challenging to obtain in the first 48 hours when the patient is in the acute phase of critical illness and there are many competing demands, such as travel for radiology scans or urgent bedside procedures. Follow-up assessment may also be challenging given the significant mortality in this patient population and transitions to hospice or early discharge. Obtaining this measure routinely may have implications for study efficiency. This will be important to consider if this modality is used as an outcome measure to assess the effect of an early intervention designed to prevent skeletal muscle dysfunction. Larger cohort studies are needed to understand the relationship between quantitative ultrasound measurements and muscle dysfunction in distinct patient groups (e.g. those with baseline frailty or comorbid disease) and to determine in which patient groups this measurement is most useful. Future studies should also examine if the change in RF-CSA over shorter time intervals can predict subsequent weakness or limitation. For example, does the change in RF-CSA over the first 3 or 5 days of illness predict muscle weakness at hospital discharge or at 3-month follow-up?

We did not find a significant relationship between quadriceps muscle thickness, either as a static measure or its change, with strength or function at any time point in our cohort. While this measurement has been more commonly used in the literature [15, 21, 22, 36–38], it includes multiple individual muscles and may be more sensitive to edema in the muscle and subcutaneous tissue. It is also challenging to ensure that this measurement is taken from a consistent point each time. Our results support findings reported in prior studies. Baldwin et al. examined the relationship of static quadriceps muscle thickness to knee extension force, a volitional measure of strength, in a small cohort of septic patients at a single time point a median of 16 days (IQR of 11 to 29 days) after ICU admission.[15] They concluded that patients had muscle weakness even when normalized to thickness thus thickness may not be an appropriate surrogate of strength in critically ill patients. Concordant with our findings, Puthucheary and colleagues described that reduction in RF-CSA was greater in those with weakness than those without at 10 days following ICU admission in their cohort of mechanically ventilated patients.[17] However, reduction in thickness over the same time period did not differ between the group with lower extremity weakness and the group without weakness in their cohort. The authors concluded that change in RF-CSA may be a more reliable measure for volitional and non-volitional muscle strength than muscle thickness or change in thickness. We also advocate using RF-CSA in future studies of quantitative peripheral muscle ultrasound.

Finally, although we determined that the rate of change of RF-CSA correlated with muscle strength at one week, it is important to note that this was only a moderate correlation. The reasons for this finding are multifactorial. Tests of muscle strength and function are complicated outcomes that require intact cognition and patient motivation as well as functioning muscle. In addition, the relationship between muscle size and function is complex and it is unlikely that any measure of muscle mass, such as ultrasound, will be a strong independent assay for muscle function in critically ill patients. While muscle mass and function are related, muscle function depends on many additional factors including neural input, fiber type composition, muscle membrane excitability, excitation-contraction coupling and others.[39] Also, the pathophysiology underlying ICU-AW includes a variety of sub-phenotypes including nerve injury and loss of muscle specific force, which may be independent of atrophy.[40] Our study does show that atrophy was a near ubiquitous feature of the muscle wasting phenotype in our cohort, and therefore remains an attractive target for therapy in this population.[41]

There are several limitations to this study. First, it is a single center study and one clinician acquired all the images analyzed in this study. This may limit its generalizability. The sample size also prevents adjustment for potential confounders. While we were able to perform ultrasonography in all enrolled patients, quantitative measurements could not be obtained in two morbidly obese patients at admission and one patient with profound cachexia at follow-up. We used a curvilinear probe to obtain our RF-CSA measurements. While this has been previously validated, analyzing spliced images obtained from the linear array probe would have been an alternative option. We did not examine all individual muscles that comprise the quadriceps complex; however, our priority was to identify easily obtained ultrasound measures. Future studies should consider examining additional muscles (i.e. tibialis anterior muscle depth) and additional ultrasound measures (i.e. echogenicity or angle of pennation of muscle). While we obtained all baseline images within 40 hours of ICU admission (mean of 22.4 hours), muscle wasting occurs early in critical illness and we may not have captured a true baseline measurement for many patients. MRC and PFIT-s were missing in one patient due to hospice enrollment. Finally, survival bias may have affected our study. The relationship between the change in quantitative peripheral muscle ultrasound measurements and strength and function was only determined in those patients that survived until day 7 of their critical illness. It is possible the muscle loss was greater or the relationship different in those patients that succumbed to sepsis soon after admission.

The pathophysiology underlying the development of muscle weakness during sepsis is heterogeneous and complex. Quantitative peripheral muscle ultrasound detects one aspect of the neuromuscular dysfunction seen in sepsis – muscle atrophy. While muscle atrophy does not explain all of the weakness seen during or after critical illness, the pathways that mediate muscle atrophy are understood [42–44] and remain an important target for therapeutic interventions aimed at reducing weakness. Attenuation of this atrophy with a medication or tailored rehabilitation holds the potential to significantly improve the long-term outcomes of patients with sepsis. Serial quantitative peripheral muscle ultrasound may provide a useful tool for selecting high risk patients and monitoring response to therapy.

Conclusions

Quantitative peripheral muscle ultrasound is feasible but challenging in patients with sepsis. In this cohort of critically ill patients with sepsis, the rate of change in RF-CSA over 7 days moderately correlated with strength one week after sepsis admission and may complement volitional measures of strength or function in describing skeletal muscle dysfunction in sepsis. Future research is needed to determine if quantitative peripheral muscle ultrasound can be used to predict subsequent muscle weakness and functional limitation.

Highlights.

• Muscle atrophy occurs early in sepsis and can be detected by peripheral muscle ultrasound.

• Change in rectus femoris cross-sectional area as detected by ultrasound over the first week of sepsis admission moderately correlates with strength at one week.

• Quantitative peripheral muscle ultrasound may complement volitional measures of strength and function in describing skeletal muscle dysfunction in patients with sepsis.

Abbreviations

ICU-AW

intensive care unit acquired weakness

RF-CSA

rectus femoris cross-sectional area

Q-MT

quadriceps muscle thickness

MRC

Medical Research Council

PFIT

Physical Function in the ICU Test

APACHE

Acute Physiology and Chronic Health Evaluation