Study design, setting, and participants
This single-center prospective observational cohort study is registered at ClinicalTrials.gov (NCT03865095).
This study was approved by the Ethics Committee of University Hospital Brno, Czech Republic (reference number 05-130219/EK). Informed consent was obtained from all enrolled patients or an independent physician.
All patients admitted to one of the four ICUs of the Department of Anaesthesiology and Intensive Care Medicine of University Hospital Brno from March 2019 to September 2020 were screened for eligibility for enrollment in this study. Inclusion criteria at ICU admission included ≥18 years of age and the physician’s subjective evaluation that the patient would require mechanical ventilation for at least 48 hours. Exclusion criteria included age <18 years, a Clinical Frailty Score >7 prior to admission, a past medical history of neuromuscular disease, amputated lower extremities, prior trauma to the lower extremities involving thighs and inability to cooperate with ultrasound examinations.
Variables, data sources, and measurements
The primary outcome of the study was the quantitative assessment of decreased RFcsa, defined as the percentage reduction in RFcsa based on the results of ultrasound evaluation on day 1 and day 7 of the patient’s ICU stay. We considered a decrease in RFcsa >10% to be clinically significant muscle wasting. The secondary outcome of the study was 28-day mortality. We also examined risk factors associated with a clinically significant decrease in RFcsa.
Ultrasound measurements (Vivid S6, GE Healthcare) of RFcsa were performed by the same investigator within the first 24 hours after ICU admission (day 1) and again on day 7 of the patient’s ICU stay. The physical point at which the ultrasound measurement was performed was drawn on the patient’s skin as a line perpendicular to the long axis of the thigh, three-quarters of the distance between the anterior superior iliac spine and the middle of the upper part of the patella. A linear ultrasound probe with a frequency of 3–9 MHz was placed on the marked line perpendicular to the skin using an excess of gel and minimal pressure. Six images of the rectus femoris muscle cross were collected and stored. An ultrasound machine tool for area measurement was used for RFcsa measurements. The inner side of the fascia of the rectus femoris muscle was manually traced in each image. Following the method described by Turton et al. [15], we omitted the smallest and largest values of RFcsa and calculated the average of the remaining four values. The member of the study team who performed the examination entered the ultrasound measurements into the case report form.
All study-related patient data (age, sex, APACHE II score, SOFA score, Frailty scale value, and ICU Mobility score on admission and day 7) were obtained from medical records and entered into the case report form by a study team member.
Bias and sample size
We cannot rule out selection bias in this study, as one of the primary inclusion criteria (i.e., the physician’s personal assessment of the need for 48 hrs of mechanical ventilation) was subjective in nature. The likelihood that a patient would need mechanical ventilation for >48 hours was assessed after discussion between the investigator and the senior ICU physician. This discussion took into account findings from initial diagnostic work-up and the reaction to therapeutic interventions. Because we were unable to locate any previous studies that documented the incidence of significant muscle wasting in critically ill adult patients, we were unable to use previous data to determine the relevant sample size. However, we did note that Bloch et al. [14] reported significant muscle wasting (defined as a decrease in RFcsa of more than 10% at day 7 of an ICU stay) in 55% of patients who had undergone cardiothoracic surgery. Based on these findings, we estimated that 100 patients would provide the appropriate sample size for determining the incidence of significant muscle wasting. Given the possibility that many of enrolled patients might not fully complete the study protocol, we increased the number of participants to 175. To cover potential loss of follow-up, the number of participants was increased to 186.
Statistical analysis
Continuous demographic data and baseline patient characteristics are presented as means with standard deviations (SDs). Continuous outcome characteristics and measurements of RFcsa are presented as means and 95% confidence intervals (CIs) with SDs. Categorical characteristics are summarized using absolute counts and percentages. There were no missing RFcsa measurements, outcome characteristics, demographic, or baseline data for any of the 104 patients we evaluated.
All potential risk factors were examined using both descriptive statistics and forward stepwise selection in a logistic regression model. Risk factors were retained in the model if they had p-value <0.1. A logistic regression model that included change in RFcsa as a dependent binary variable and risk factors that were retained after forward selection as predictors were constructed to estimate odds ratios and 95% Wald CIs.
The relationships between patient characteristics and decreases in RFcsa were examined using Wilcoxon–Mann–Whitney U-tests. The relationship between a decrease in RFcsa and 28-day mortality was evaluated using Fisher’s exact test with an alpha level of 0.05. All statistical analyses were conducted using SAS 9.4 (SAS Institute, Cary NC).