Point-of-Care Ultrasound-Based Assessment of Sarcopenia to Predict Revisits or Mortality in Older Adults at the Emergency Department: A Prospective Observational Study
Article information
Abstract
Background
Older patients visiting the emergency department (ED) are at increased risk of adverse outcomes, including ED revisits and mortality. Sarcopenia quantification by point-of-care ultrasound (POCUS) may be a useful bedside screening tool, especially when traditional frailty screening instruments, reliant on history taking, cannot be used. This study evaluated whether POCUS measurement of rectus femoris cross-sectional area (RFcsa) can predict adverse outcomes in older patients visiting the ED.
Methods
In this single-centre prospective study, patients aged ≥70 years presenting to the ED of a Dutch university hospital and enrolled in the Acutelines data and biobank were included. RFcsa was measured using POCUS. ROC-analysis assessed the overall accuracy of RFcsa for prediction of the primary outcome, which was defined as the composite of ED revisit or death within 3 months. Logistic regression determined the added value of RFcsa to Karnofsky Performance Score (KPS).
Results
During the study period, a total of 68 patients were included. Twenty-six patients (38%) met the primary endpoint. RFcsa showed excellent intra-rater reliability (interclass correlation coefficient=0.98). However, the accuracy to predict the composite endpoint was low, with an area under the curves of 0.53 (0.39–0.66) for unadjusted RFcsa and 0.51 (0.36–0.66) for sex-adjusted RFcsa. The addition of RFcsa in a multivariate logistic regression model with KPS did not increase the overall explained variance in the primary endpoint.
Conclusion
In older patients presenting in the ED, POCUS-measured RFcsa does not predict ED revisits or death within 3 months. (Trial Registration No. NCT05369962 at ClinicalTrials.gov)
INTRODUCTION
Older patients visiting the emergency department (ED) more frequently experience adverse health outcomes (including a higher ED re-admission rate and a higher mortality) in the months after their ED visit than younger patients.1-5) Several screening tools have been developed to identify patients in the ED who are most at risk of developing these adverse outcomes.4,6) A key element in all these screening tools is the history as obtained from the patient or their caregiver(s).7,8) However, sometimes a reliable history cannot be obtained, as communication may be impaired due to trauma, confusion, illness or pre-existing cognitive disorders. In such instances, other screening tools are needed. Quantification of sarcopenia, a syndrome characterized by progressive and generalized loss of skeletal muscle mass and strength,9,10) has been proposed as a reliable alternative.11,12) The presence of sarcopenia is associated with a higher incidence of physical disability, a poorer quality of life, higher financial costs, hospitalization, mortality, and it is a key component of frailty.3,10,11,13) Therefore, assessing sarcopenia in the ED may offer a valuable, objective means of identifying older patients at increased risk of adverse outcomes when other screening tools cannot be used.
Sarcopenia is most often quantified in the quadriceps muscles, as it serves as a representative indicator of muscle mass and exhibits the earliest and most pronounced signs of muscular decline.13,14) Magnetic resonance imaging (MRI) and computed tomography (CT) are regarded as the gold standard for quantifying muscle mass, but they are inconvenient, expensive and impossible to perform in everyday clinical practice, especially in the ED.10,13,14) Point-of-care ultrasound (POCUS) has been shown to potentially be a reliable non-invasive alternative to other diagnostic methods (MRI, CT, bioelectrical impedance analysis, and dual energy X-ray absorptiometry) to quantify skeletal muscle mass.13-18) However, the utility of POCUS sarcopenia measurements as a screening tool for frailty and its associated adverse outcomes in a general population of older patients following ED visits is presently unclear. Therefore, in this proof-of-principle study, we performed a single-centre prospective trial to evaluate whether POCUS of the rectus femoris cross-sectional area (as a marker of sarcopenia) may be utilized to identify older patients who are at risk of an adverse outcome in the 3 months after their ED visit.
MATERIALS AND METHODS
Study Setting and Design
This single-centre prospective study was performed in the ED of the University Medical Centre Groningen (UMCG), a tertiary care hospital in the Netherlands. Data for this study were prospectively collected as part of the Acutelines data and biobank of the Department of Acute Care of the UMCG.19) The study protocol was approved by the central ethical review committee of the UMCG (CTc Protocol No. 202100945). The trial was registered in the Clinical Trials Register (NCT05369962). This study was conducted and reported according to the STROBE (STrengthening the Reporting of OBservational Studies in Epidemiology) guidelines.20) Acutelines has been approved by the UMCG Medical Ethics Committee and is registered on ClinicalTrials.gov under trial registration number NCT04615065. Acutelines’ complete protocol and an overview of the current, full data dictionary is available via www.acutelines.nl.
Selection of Participants
The study population consisted of a convenience sample of patients aged 70 years or older, presenting to the ED with medical (e.g., non-trauma related) conditions for one of the following specialties (based on the inclusion criteria of the Acutelines cohort): emergency medicine, internal medicine, nephrology, geriatric medicine, oncology, haematology, pulmonology, rheumatology, gastrointestinal/liver medicine, or urology. Excluded were patients in whom it was not possible to obtain an ultrasound image of the rectus femoris (e.g., due to an amputation or profound oedema), or patients with paresis or paralysis of one or both legs.
Study Outline and Data Collection
Patients visiting the ED were screened every day, between 9 am and 11 pm for eligibility upon arrival by an ED nurse and a trained research team member. Eligible patients were informed about the study and asked to provide written consent (by proxy if necessary). Patients were approached for consent when a physician was present in the ED who had received dedicated training to perform rectus femoris cross-sectional area (RFcsa) measurements (see below).
For participating patients, bedside monitoring data were automatically captured and stored, and information from the hospital's electronic health records was securely imported into the database. Imported variables included: demographic features (age, sex, living situation), medication use (polypharmacy) and disease severity (as rated with the Emergency Severity Index), and measures of functional performance (Karnofsky Performance Scale [KPS]). All data were collected and managed using REDCap (Vanderbilt University, Nashville, TN, USA) electronic data capture tools hosted at the UMCG.21,22)
POCUS measurements of the RFcsa of the right leg were performed in the ED by a trained (emergency or internal medicine) physician. All physicians performing the POCUS measurements had considerable POCUS experience, having passed the national ED POCUS certification program (attended POCUS courses, built a portfolio of >250 ultrasound exams and passed a written and practical exam).23) In addition, they received dedicated training to perform POCUS of the RFcsa in the month before study inclusions. This training consisted of a presentation by a member of the study team about the study methods (including positioning of the patient and an example of the RFcsa measurement, approximately 10 minutes) and an instruction video (1 minute) on how to perform the RFcsa tracing on the ultrasound machine. The instruction video and study protocol were available for reference during the study period. Thereafter, the physician scanned the RFcsa in a healthy volunteer supervised by a member of the study team.
For all included patients, follow-up data on the composite endpoint (death or ED revisit within 3 months) were collected 3 months after their ED visit.
POCUS Measurements
To facilitate the POCUS exam, patients were positioned supine on the bed with the upper body elevated 30° and their legs extended and their muscles relaxed. The length between the anterior superior iliac crest and the superior border of the patella of the right leg was measured and the curvilinear probe (1.5–5 MHz) was placed at the halfway point perpendicular to the skin, applying no pressure, to visualize the transverse plane of the rectus femoris muscle. The echogenic line of the rectus femoris muscle was traced manually on a frozen image, and the RFcsa was calculated using the planimetric technique (Fig. 1) provided on the Venue ultrasound machine (GE Healthcare, Chicago, IL, USA). Three consecutive RFcsa measurements were performed and the average value was reported as the measured RFcsa. The ultrasound images with tracing were saved, and review after the examination was performed by an expert sonographer of the study team (DB) for image quality (good enough to identify the rectus femoris muscle), and for accuracy of tracing. Patients were excluded from the analysis if images were not saved (precluding quality assessment) or if, during the quality assessment, the tracing was found to be grossly inaccurate (defined as measurement of an incorrect muscle or failure to follow the echogenic line of the rectus femoris).
Example of an ultrasound measurement of the musculus rectus femoris cross-sectional area (blue tracing).
As there is a difference in muscle mass between male and female, the average RFcsa was adjusted for sex by multiplying female RFcsa by the coefficient 1.484 prior to analysis.8)
Outcome Measures
The primary outcome was the ability of the POCUS RFcsa to predict ED revisit or death within 3 months (composite endpoint). The secondary outcome was the added value of RFcsa ultrasound measurement to the KPS in predicting ED revisit or death.
Sample Size Calculation
Prior research in the Netherlands concluded that 9.5% of patients aged ≥70 years died within 3 months after ED visit and 30.6% experienced functional decline or mortality at 90-day follow-up.4) Given a presumed association between functional decline and ED revisit and taking into account that measures of functional decline aren't available for many patients (therefore using revisit as a proxy), we based our sample size on a prevalence of the composed endpoint of 30%. Therefore, this trial was powered to confirm a diagnostic test accuracy with a sensitivity and specificity of >0.8 each for a predefined RFcsa cut-off value with an alpha of 5% and power of 90%. To account for a 25% attrition rate during follow-up, we aimed for a sample size of at least 68 patients with correct RFcsa tracings at the time of their ED visit.
Data Analysis
Continuous variables are expressed as means with 95% confidence intervals (CIs) when normally distributed and medians with two times the interquartile range (2 × IQR) when skewed. Categorical data were presented as absolute numbers and percentages. Intra-rater reliability for ultrasound RFcsa measurement was measured by the intraclass correlation coefficient (ICC).24) The Mann-Whitney U test was used to compare means in RFcsa between patients with and without the composite endpoint. ROC-analysis was performed to calculate the area under the curve (AUC) of (sex-adjusted) RFcsa as an overall measure of diagnostic accuracy to predict the primary endpoint. A logistic regression model was used to evaluate the added value of RFcsa to KPS (when obtained) in the prediction of the composite endpoint (ED revisit or death). A survival analysis (Kaplan–Meier survival analysis and Cox regression analysis for hazard ratios) was performed to evaluate differences in timing of ED revisit or death.
A two-tailed p-value <0.05 was considered statistically significant. Missing data are reported according to the STROBE 2015 guideline.20) All analyses were performed with IBM SPSS Statistics version 23 for Windows (IBM, Armonk, NY, USA).
RESULTS
Characteristics of Study Subjects
During the study period (from May 2022 to February 2023) a total of 82 patients provided informed consent for participation at a moment that one of the trained physicians was present. In seven patients, RFcsa measurements were not saved. Post-hoc expert review of RFcsa tracings after inclusion learned that in another seven patients RFcsa tracing was incorrect (in two patients the wrong muscle was traced, and in five patients tracing was grossly incorrect), precluding representative quadriceps surface area calculations, leaving 68 patients for further analysis (Fig. 2). Baseline characteristics of the study population are presented in Table 1. The median age at the day of presentation was 77 years, and the majority of the enrolled patients were male (63%). Most patients were living independently (91%), with a mean KPS of 75. Based on KPS, 10 patients required at least considerable assistance and frequent medical care (score 50 or lower).
Patient flowchart.
Baseline participant characteristics (n=68)
Primary Outcome
Within 3 months after their initial ED visit, 19 patients (28%) revisited the ED and 12 patients (18%) died. In total, 26 patients (38%) met the composite endpoint. The mean sex-adjusted RFcsa was 8.25 cm2 (95% CI 7.41–9.09). The ICC for RFcsa measures was 0.98 (95% CI 0.97–0.99), indicating excellent intra-rater reliability. The ICC did not change when the five patients that were excluded after post-hoc expert review for poor tracing, were included in this analysis, 0.98 (95% CI 0.97–0.98).
The average unadjusted RFcsa was not different for patients who met the composite endpoint compared to those who did not meet the endpoint, nor was it related to either ED revisit or death within 3 months (Table 2, Supplementary Table S1). Based on prior literature8) using a 5.2 cm2 cut-off for sarcopenia, 19 patients (28%) were sarcopenic using average unadjusted RFcsa, and 15 patients (22%) using sex-adjusted RFcsa; of whom 7 (37%) and 4 (27%) experienced an adverse outcome, respectively.
Rectus femoris cross-sectional area (RFcsa) stratified by outcome
ROC-curves of unadjusted RFcsa and KPS are presented in Fig. 3. The AUC of unadjusted RFCsa was 0.53 (95% CI 0.39–0.66), this did not change significantly after sex-adjustment (AUC=0.51, 95% CI 0.36–0.66). Sex-adjusted AUC was lower than the AUC of KPS for the prediction of the composite endpoint ED revisit or death within 3 months (AUC=0.68, 95% CI 0.54–0.83). Due to the very low AUC, no optimal cut-off could be determined for RFcsa to discriminate patients with and without the primary endpoint.
Receiver operating characteristic curves (n=61) for accuracy of unadjusted RFcsa and KPS score in the prediction of the composite endpoint (emergency department revisit and death within 3 months). RFcsa, rectus femoris cross-sectional area; KPS, Karnofsky Performance Scale.
In a stepwise logistic regression model (including 61 patients) containing age, KPS (continuous) and RFcsa as clinically relevant covariates, only KPS was identified as an independent predictor of the composite endpoint (13.6% explained variance). Adding unadjusted RFcsa or sex-adjusted RFcsa to the regression model did not significantly improve the overall explained variance (15.3% and 15.5% explained variance, respectively).
A survival analysis was executed, which did not demonstrate a significant difference for the composite endpoint (p=0.68) (Supplementary Fig. S1). Compared to the reference group (RFcsa >8.04 cm²), the hazard ratio for participants with RFcsa <5.4 cm² was 1.15 (95% CI 0.43–3.07; p=0.78), and for those with RFcsa between 5.4 and 8.04 cm², the hazard ratio was 1.50 (95% CI 0.59–3.81; p=0.39).
DISCUSSION
In this proof-of-principle study, we explored whether ED POCUS measurements of the RFcsa, as a marker of muscle mass, can predict ED revisits or mortality within 3 months in a general geriatric population. We found that RFcsa had a low accuracy in predicting ED revisit or death, and that POCUS-measured RFcsa is not a suitable frailty screening tool when traditional instruments, reliant on history taking, cannot be used.
Previous studies have explored the use of POCUS for sarcopenia quantification in the hospital setting. Whilst some of these studies have demonstrated a positive association of muscle mass with adverse outcomes (such as discharge disposition, intensive care unit and hospital length-of-stay, falls, readmission or mortality),7,8,12,25) others found no such relation.26,27) These discrepancies may be explained by the different prevalence of sarcopenia in the populations under study—in an intensive care setting, where patients are bed-bound, sarcopenia develops rapidly, not only in the older adults,8) whereas in our cohort, the prevalence of sarcopenia was low, or many patients were living independently and had a relatively high KPS. The quadriceps muscle mass in our cohort was relatively high (average RFcsa 8.25 cm2) and only 15 patients met RFcsa criteria for sarcopenia used in previous studies.8,28,29) This may indicate that the degree of muscle wasting in our population may simply not have been severe enough to reach a certain threshold below, which adverse outcomes are deemed more likely to occur.8) An alternative explanation for our negative result could be inherent to the method of sarcopenia quantification, as in our study we only assessed muscle quantity, not muscle quality (e.g., muscle echogenicity). As muscle strength and physical performance are influenced by both muscle size and quality,30) focussing on muscle quantity alone may not yield a reliable surrogate for sarcopenia.14,30-32) A more comprehensive approach combining multiple ultrasound parameters including cross-sectional area and echo intensity may be more effective in predicting adverse outcomes.
It is unlikely that the absence of a relationship between sarcopenia and outcome in our study results from measurement errors. To ensure accuracy, we provided standardized training for the POCUS-certified ED physicians who performed the measurements. Sarcopenia can be quantified using several metrics, most commonly muscle thickness and cross-sectional area.32) We used RFcsa to minimize variability due to excess compression by the ultrasound probe, and because it demonstrated a higher diagnostic value than muscle thickness.13,33) Notably, previous research has established a strong correlation between muscle thickness and cross-sectional area.34-36) Additionally, RFcsa values were adjusted for sex. We also conducted a retrospective quality control of all RFcsa measurements by reviewing the saved images to include only representative measurements. However, it should be noted that significant variation exists between various measurement protocols (measurement site, patient posture, probe positioning, transducer type), which may be a factor contributing to inconsistencies with some other studies.13,14,17,30)
Our study reached the pre-specified sample size and the assumed prevalence of the composite endpoint, making a type II error unlikely. However, the study was powered to detect a relatively high overall accuracy, but with the AUC close to 0.5 it is very unlikely that RFcsa sarcopenia measurements can be used in any purposeful way to predict outcome in ED populations.
Our study had several limitations. Although our study was performed in a general and heterogeneous cohort of ED patients, our results cannot be generalized to other populations, since our study was conducted in a specific sample of patients attending an academic ED. Furthermore, the use of convenience sampling may have introduced selection bias, potentially limiting the representativeness of our sample. Secondly, the definition of the (composite) endpoint may have been too broad and may have compromised our results. Further, comparisons with existing frailty measures were limited to the KPS, which was collected for most patients. Alternative frailty and functional scores (such as Katz Activities of Daily Living, Six-Item Cognitive Impairment Test, Clinical Frailty Scale, and Acutely Presenting Older Patient screener) were only available for a limited number of patients and therefore could not be included in the analysis. Further, the correction factor used to adjust RFcsa for sex was derived from an American/Canadian cohort, aged 18–88 years, which may not have been the appropriate factor for our (exclusively older adults) population. Finally, in 9% of cases, RFcsa measurement were either performed on the wrong muscle or the tracing was not performed correctly. Although not a limitation as such (as these patients were excluded from analysis), this finding highlights that obtaining reliable RFcsa measurements in the ED setting can be challenging.
In conclusion, in older adults patients presenting in the ED, POCUS-measured RFcsa is not a predictor of ED revisits or mortality within 3 months of ED visit.
Notes
The data and biological samples used in this manuscript are provided by Acutelines. The authors would like to thank Acutelines and all its participants. We thank all physicians for participating in this study.
CONFLICT OF INTEREST
The researchers claim no conflicts of interest.
FUNDING
The establishment of Acutelines has been made possible by funds from the University Medical Centre Groningen. No funding was received for this specific study.
AUTHOR CONTRIBUTIONS
Conceptualization, SH, LG, DB, HB, HL, JM, EA; Data curation, SH, LG, DB, HB, JM, EA; Formal analysis, SH, LG, DB, JM, EA; Writing-original draft, SH, LG, DB, HL, JM, EA; Writing-review & editing, SH, LG, DB, HL, JM, EA, HB.
SUPPLEMENTARY MATERIALS
Supplementary materials can be found via https://doi.org/10.4235/agmr.25.0106.
Rectus femoris cross-sectional area (RFcsa) stratified by the individual components of the endpoint
Survival analysis for the composite endpoint (emergency department revisit and death within 3 months). RFcsa, rectus femoris cross-sectional area.