Inter- and intra-observer variability of ultrasonographic arm muscle thickness measurement by critical care physicians.

PURPOSE: The aim of this study was to assess inter- and intra-observer variability of arm muscle thickness measured by critical care physicians using bedside ultrasonography (USG).
METHODOLOGY: This prospective study included twenty patients admitted with sepsis. Three measurements of thickness of right arm muscles of each patient using B-mode USG were taken by two critical care fellows, independently. Intra- and inter-observer reliability was tested using intraclass correlation coefficient (ICC).
RESULTS: The mean 1st, 2nd, and 3rd measurements of muscle thickness recorded by observer 1 and 2 were 23.620 (±4.171) versus 23.840 (±3.849) mm, 23.235 (±3.620) versus 23.625 (±4.062) mm, and 24.125 (±4.098) versus 23.965 (±3.651) mm, respectively. The average muscle thickness measured by first and second observer was 23.660 (±3.834) mm and 23.810 (±3.674) mm, respectively. ICC for intra-observer variability for observer 1 and 2 was 0.964 (95% confidence interval [CI] 0.924-0.985) and 0.949 (95% CI 0.892-0.978), respectively. ICC for inter-observer variability was 0.995 (95% CI 0.988-0.998).
CONCLUSIONS: USG is a reliable tool for assessment of arm muscle thickness by critical care physicians.

Introduction

Muscle dysfunction characterized by wasting and decreased strength is common among critically ill patients admitted to Intensive Care Unit (ICU).[12] It has been reported that some form of muscle weakness may be seen in 20%–100% of patients admitted to the ICU.[13] This variation in prevalence rate may be due to variation in the length of hospital stay and sensitivity and specificity of different tools used for assessment of muscle weakness. Delayed extubation as well as reintubation are undesired outcome of ICU-acquired weakness (ICU-AW).[12] Both prolonged intubation and reintubation are associated with increased morbidity, mortality, and consumption of health resources.[124]

Early detection of ICU-AW is cornerstone for prevention and treatment of this condition. The assessment of muscle may include either functions (strength) or mass (muscle thickness or cross-sectional area). Routinely, diagnosis of ICU-AW is based on functional assessment of the muscle by manual muscle testing using Medical Research Council (MRC) score.[5] Manual muscle testing requires patient's active participation, reasonable level of consciousness, understanding, and effort. Alteration of sensorium limits the utility of this tool among critically ill patients. Other tools which may be used for assessment of muscle functions include electrophysiological studies (electromyography) and muscle biopsy. However, both of these have their own limitations when used in ICU.[12]

Considering the impact of ICU-AW on in-hospital outcome as well as consumption of hospital resources and postdischarge morbidity, there were attempts to find an objective, noninvasive, sensitive, and reliable tool to measure lean muscle mass among these patients. Anthropometry can provide information about muscle mass but is dependent on hydration status of the patient. Studies also showed a poor correlation between muscle mass measurements by anthropometry with computerized tomography (CT) scan or dual-energy X-ray absorptiometry (DEXA) scan.[67] Currently, DEXA, CT scan, and magnetic resonance imaging (MRI) are used for assessment of muscle mass, both muscle thickness and cross-sectional area.[1258] However, all of these require shifting of patient out of ICU which many a times may not be possible. There are studies which have shown that assessment of quadriceps muscle by ultrasonography (USG) is comparable to CT scan.[67] Another study showed that USG measurements of rectus femoris muscle were comparable to measurements done by MRI.[9] Data also suggest that the muscle thickness is a good surrogate of overall muscle mass.[10] Measurement of muscle thickness is not dependent on patients’ level of consciousness, understanding, and effort. Therefore, muscle thickness measurement by USG overcomes many limitations of the other tests (DEXA, CT scan, or MRI) for this purpose. Further, USG can be used for serial assessment of muscle thickness without significant additional cost or risk to the patient.

Data suggest that upper limb strength is independently associated with increased hospital mortality among critically ill patients in ICU.[1112] USG can easily be used for assessment of upper limb muscle thickness. One study demonstrated the utility of USG for evaluation of upper limb muscle thickness among critically ill patients.[13] Before we plan further studies to validate these results, intra- and inter-observer reliability of USG should be established for assessment of upper limb muscle thickness. Therefore, we conducted this study with the aim to assess the intra- and inter-observer reliability of USG measurement of muscle thickness of the anterior compartment of arm by critical care physicians.

Methodology

The study was conducted following the Declaration of Helsinki and National Guidelines for study involving human subjects.[1415] The study protocol was approved by the Institutional Review Board.

Patients and setting

We conducted this study between March and May 2016 at a tertiary care center. The study included twenty adult patients (age more than 18 years) admitted with sepsis. Diagnosis of sepsis was based on the criteria suggested by SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference held in 2001.[16] Patients who had a primary neuromuscular disease (such as myopathy, neuropathy, and stroke), an amputated limb, and transferred from other hospital (after 48 h of stay) were not included in the study. Patients currently requiring home ventilation, required mechanical ventilation or hospitalization during previous 3 months or who refused consent were also excluded from the study.

Equipment

The muscle thickness was measured by two critical care fellows (Rohit Kumar, Ashesh Dhungana) who were trained and were using ultrasound in the ICU regularly for various indications. The measurements were taken in B-mode USG using 5.0–13.0 MHz (megahertz) linear array probe (VF 13-5) on the Siemens ACUSON X300™ machine.

Site, posture, and measurements

All measurements of thickness of anterior compartment muscles were done on the right arm of the patients. The measurements were taken while the patient lying supine, the elbow extended, and forearm supinated. The site and position of the ultrasound transducer were marked by the first investigator (Rohit Kumar), and the same point was used for the subsequent recordings by both investigators. A circumferential mark was applied at the midway between the tip of the greater tuberosity and tip of the olecranon process of humerus. The linear probe was placed on this circumferential line, perpendicular to the skin, and the probe was moved along the line drawn till a suitable image was obtained. Then, the point on the circumferential mark which was corresponding to the center of the probe was marked with a vertical line. This point was used as the reference point. For the measurement of muscle thickness, with arm in fixed position, the linear probe was placed on the reference point in a way that the center of the probe always remains the reference point. This ensured that all measurements were made at the same level. For the measurement of muscle thickness, the probe was held perpendicular to the skin to negate any variation in the recording of muscle thickness due to angulation of the probe. During measurement of the muscle thickness, the force applied to the probe was just sufficient to make adequate contact between skin and probe without compressing the underlying muscle. The images were centered on the screen by little manipulation of the probe. Both investigators recorded three measurements each, independently, a few minutes apart in the absence of the other investigator. Both observers were blinded to the each other's result.

The thickness of the flexor compartment was measured between the superficial fat–muscle interface and the humerus [Figure 1]. This includes both the biceps brachialis and the coracobrachialis muscles.

Figure 1

Ultrasonography image demonstrating various structures and measurement of muscle thickness

Statistical analysis

Data were analyzed using SPSS version 20.0 (IBM Corp., Armonk, NY) statistical software for Windows. Data were reported as mean ± standard deviation or frequency (%) wherever applicable. For assessment of intra- and inter-observer reliability of USG measurement of muscle thickness, intraclass correlation coefficient (ICC) was computed. Altman–Bland method was used to plot the difference between average muscle thicknesses measured by both observers.

Results

This study included twenty patients admitted in our ward and ICU with the diagnosis of sepsis. The baseline characteristics are presented in Table 1.

Table 1

Baseline characteristics of the study cohort

Characteristics	Observation
Age (mean±SD), years	52±16.23
Gender (%)
Male	11 (55)
APACHE - II score (mean±SD)	18.65±4.44
Underlying lung disease (%)
COPD	9 (45)
Bronchiectasis	4 (20)
Asthma	2 (10)
Pneumonia	2 (10)
Others*	3 (15)

*Interstitial lung disease (1), allergic bronchopulmonary aspergillosis (1) and urinary tract infection (1). SD: Standard deviation, COPD: Chronic obstructive pulmonary disease, APACHE: Acute Physiology and Chronic Health Evaluation

Muscle thickness measurements

Sixty measurements (three for each patient) of muscle thickness of the flexor compartment of the right arm were recorded by each observer. The mean 1st, 2nd, and 3rd measurements of muscle thickness recorded by first observer (Rohit Kumar) were 23.620 (±4.171) mm, 23.235 (±3.620) mm, and 24.125 (±4.098) mm, respectively. The mean 1st, 2nd, and 3rd measurements of muscle thickness recorded by second observer (Ashesh Dhungana) were 23.840 (±3.849) mm, 23.625 (±4.062) mm, and 23.965 (±3.651) mm, respectively. The average muscle thickness by first and second observer was 23.660 (±3.834) mm and 23.810 (±3.674) mm, respectively.

Intra- and inter-observer reliability

Intra- and inter-observer reliability was checked by computing ICC with 95% confidence interval (CI).

Intra-observer reliability

The ICC for reading 1st and 2nd, 1st and 3rd, and 2nd and 3rd measured by observer 1 (Rohit Kumar) was 0.941 (95% CI: 0.858–0.976; P < 0.0001), 0.821 (95% CI: 0.602–0.925; P < 0.0001), and 0.929 (95% CI: 0.829–0.971; P < 0.0001), respectively. Similarly, ICC for reading 1st and 2nd, 1st and 3rd, and 2nd and 3rd measured by observer 2 (Ashesh Dhungana) was 0.910 (95% CI: 0.787–0.963; P < 0.0001), 0.857 (95% CI: 0.675–0.94; P < 0.0001; P < 0.0001), and 0.809 (95% CI: 0.579–0.920; P < 0.0001), respectively.

Inter-observer reliability

The results of inter-observer reliability of the measurement of muscle thickness using interclass variability coefficient computation are summarized in Table 2. Altman–Bland method was used to depict inter-observer difference [Figure 2]. The mean difference of measurement between two observers was −0.150 (95% CI: −1.160–0.860) mm.

Table 2

Intraclass correlation coefficient of the measurements of muscle thickness

Observer	ICC (95% CI)	Significance
Observer 1	0.964 (0.924-0.985)	<0.0001
Observer 2	0.949 (0.892-0.978)	<0.0001
Inter-observer	0.995 (0.988-0.998)	<0.0001

CI: Confidence interval, ICC: Interclass correlation coefficient

Figure 2

Altman–Bland plot of average of muscle thickness measured by both observers. The mean difference was − 0.150 (95% confidence interval: −1.160–0.860) mm

Discussion

Our study showed that USG is an excellent tool in the hands of critical care physicians, demonstrating high intra- and inter-observer reliability for measurement of muscle thickness of the flexor compartment of the arm. These findings reflect the feasibility of USG for the measurement of thickness of flexor muscles of arm in the hands of minimally trained critical care physicians. This proves that USG is a tool which may provide an objective and reliable measure of upper limb muscle functionality.

Skeletal muscle performs plethora of functions other than mobility including metabolism of glucose and protein and immune functions.[171819] Therefore, assessment of muscle functions is critical in ICU, especially in the presence of sepsis. MRC score and handheld dynamometer have been used for assessment of upper limb muscle functions.[211] Ali et al.[11] in their study used MRC and handheld dynamometer and had shown that upper limb weakness was associated with longer hospital stay and mechanical ventilation requirement. This study also showed that upper limb weakness was independently associated with increased hospital mortality. These observations highlight the importance of upper limb muscle function among critically ill. However, both tools used in this study assess volitional strength and require awareness, cooperation, and motivation from patient, which many a times lacking during severe illness. In addition, MRC score is known to have a ceiling effect and differentiation between subnormal (MRC-4) and normal strength (MRC-5) remains questionable.[20] Furthermore, these tests are insensitive to detect the progression of the muscle weakness. All these suggest that neither MRC score nor handheld dynamometer can be accepted as sensitive, reliable, and objective tool for assessment of upper limb muscle dysfunction among critically ill. Our study results are of clinical significance as USG is available among all ICU. It provides objective parameter (muscle thickness) which can be measured. Baldwin and Bersten used USG for measurement of limb muscle thickness among critically ill patients and demonstrated that limb muscle weakness was an important factor affecting the outcome of critically ill patients.[12] These results further demonstrated the utility of USG. However, there is always a concern that USG measurements are observer dependent and can show intra- as well as inter-observer variation. Our study has established that there are no significant intra- or inter-observer variations.

The strength of our study includes that first, the measurement of muscle thickness was done by critical care physicians and not by radiologists. This observation is important as ICU is manned by critical care physicians and services of dedicated radiologists are usually lacking. Second, in our study, intra-observer variations were assessed for both observers independently which further adds to the reliability of the measurements. There are certain limitations pertaining to our study. First, there is a lack of “gold standard” such as CT scan or MRI for verification of the measurements taken by USG. CT scan and MRI although can verify the measurements as already has been discussed; both have limited utility in day-to-day practice in ICU setting. The second limitation is the absence of a radiologist among observers. Radiologists are more experienced and better while using USG. Their inclusion in the study and demonstration of ICC with them could have strengthened our findings. However, because of the absence of a dedicated radiologist in our ICU we were unable to do so. Another limitation of our study is noninclusion of paramedical staff (nurses or physiotherapists) with minimum or no exposure to use of USG for measurement of muscle thickness. USG measurements are known to improve with practice.[21] Therefore, inclusion of paramedics (inexperienced observer) might have assessed the utility of USG in relatively inexperienced hands. We could not include paramedics in the study because of practical reasons.

Our study has shown that USG measurements of arm muscles in ICU patients by critical care physicians are feasible as they can be easily and promptly undertaken, and available 24/7, when fully trained radiologists are not readily available, especially off-hours. The measurements are highly reproducible as there is high inter- and intra-observer agreement with intraobserver ICC of 0.964 (95% CI: 0.924–0.985) and 0.949 (95% CI: 0.892–0.978).

Conclusion

Our study has shown high ICC for both intra- and inter-observer measurement of flexor muscle thickness of the arm. The results imply that USG may be used as an objective and reliable tool for the assessment of thickness of flexor muscle of the arm by critical care physicians.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.