Abstract
BACKGROUND: Given the long ventilation times of patients with COVID-19 that can cause atrophy and contractile weakness of respiratory muscle fibers, assessment of changes at the bedside would be interesting. As such, the aim of this study was to determine the evolution of respiratory muscle thickness assessed by ultrasound.
METHODS: Adult (> 18 y old) patients admitted to the ICU who tested positive for SARS-CoV-2 and were ventilated for < 24 h were consecutively included. The first ultrasound examination (diaphragm, rectus abdominis, and lateral abdominal wall muscles) was performed within 24 h of intubation and regarded as baseline measurement. After that, each following day an additional examination was performed, for a maximum of 8 examinations per subject.
RESULTS: In total, 30 subjects were included, of which 11 showed ≥ 10% decrease in diaphragm thickness from baseline; 10 showed < 10% change, and 9 showed ≥ 10% increase from baseline. Symptom duration before intubation was highest in the decrease group (12 [11–14] d, P = .03). Total time ventilated within the first week was lowest in the increase group (156 [129–172] h, P = .03). Average initial diaphragm thickness was 1.4 (1.1–1.6) mm and did not differ from final average thickness (1.3 [1.1–1.5] mm, P = .54). The rectus abdominis did not show statistically significant changes, whereas lateral abdominal wall thickness decreased from 14 [10–16] mm at baseline to 11 [9–13] mm on the last day of mechanical ventilation (P = .08). Mixed-effect linear regression demonstrated an association of atrophy and neuromuscular-blocking agent (NMBA) use (P = .01).
CONCLUSIONS: In ventilated subjects with COVID-19, overall no change in diaphragm thickness was observed. Subjects with decreased or unchanged thickness had a longer ventilation time than those with increased thickness. NMBA use was associated with decreased thickness. Rectus muscle thickness did not change over time, whereas lateral abdominal muscle thickness decreased but this change was not statistically significant.
- ultrasound
- COVID-19
- diaphragm
- rectus abdominis
- lateral abdominal wall muscles
- respiratory muscles
- atrophy
- mechanical ventilation
Introduction
Mechanical ventilation unloads the respiratory muscles and can cause atrophy and contractile weakness of muscle fibers.1-4 This is crucial, as studies have shown that loss of thickness has important implications for clinical outcomes. This includes duration of mechanical ventilation.3,4 Monitoring these changes is, therefore, of potential interest for clinicians. It could identify patients at risk for remaining mechanically ventilated and potentially help adapt therapeutic goals accordingly.5 In light of the current COVID-19 pandemic, this is becoming more relevant than ever, with patients being ventilated for extended periods, creating conditions favoring respiratory muscle atrophy and dysfunction.6,7
A recent histological study in deceased subjects with COVID-19 demonstrated infiltration of the diaphragm by the virus and development of muscle fibrosis.8 Whether this also results in additional loss of muscle mass of the diaphragm and other respiratory muscles is unclear. Ultrasound offers an inexpensive, noninvasive possibility to assess muscle mass by measuring changes of diaphragm thickness over time.9-11 As such, the aim of this study was to determine the evolution of respiratory muscle thickness assessed by ultrasound as has been previously described in non–COVID-19 subjects.4,12
QUICK LOOK
Current Knowledge
Histological data confirm the presence of viral infiltration in the diaphragm by COVID-19 and development of muscle fibrosis. Whether this results in loss of muscle mass remains unknown.
What This Paper Contributes to Our Knowledge
Whereas on average no change in diaphragm thickness was seen, subjects with unchanged or decreased thickness had longer ventilation time than those with increased thickness. Rectus muscle thickness did not change over time, whereas lateral abdominal muscle thickness fell, but this change was not statistically significant.
Methods
This prospective observational study was conducted in an academic ICU (Amsterdam UMC, location VUmc, Amsterdam, the Netherlands) with daily ultrasound examination being standard care. The protocol was approved by the local ethics board (Medisch Ethische Toetsings Commissie, study number 2017.056), registered in the Netherlands Trial Register (Registration ID: NL9053), and performed according to guidelines in the Declaration of Helsinki. Informed consent was approved as an opt-out procedure. Subjects were followed up to 28 d after ICU discharge or until mortality. Standards of Reporting of Observational Studies in Epidemiology guidelines were followed.
Subjects
The study population consisted of adult (> 18 y old) patients admitted to the ICU who tested positive for SARS-CoV-2 and were ventilated for < 24 h. Subjects were consecutively included from November 2, 2020–January 21, 2021. Exclusion criteria were past medical history of spinal cord injury or diaphragm paralysis. Even though not specifically selected for, no patients with significant chronic pulmonary pathology, neuromuscular disease, or mechanical ventilation for > 48 h in the past 3 months were included. If patients were extubated, transferred to a different medical center, or passed away within the first 72 h of mechanical ventilation, they were excluded as well (Fig. 1).
Flow chart.
The first ultrasound examination was performed within 24 h of intubation and regarded as baseline measurement. After that, each following day an additional examination was performed for a maximum of 8 examinations per subject. It was aimed to perform the measurements as closely to 24 h apart as possible, but measurements within the same shift (day, evening, or night) were accepted. Baseline characteristics (age, sex, body mass index) and symptom duration were recorded on the day of inclusion. Symptom duration was calculated from the first day of experiencing flu-like symptoms such as coughing, shortness of breath, fever, headache, digestive symptoms, or fatigue. At the time of each examination, clinical scores (Sequential Organ Failure Assessment [SOFA] and Acute Physiology and Chronic Health Evaluation II [APACHE II]), ventilator settings, sedatives, and neuromuscular-blocking agents (NMBAs) use were recorded. All subjects were treated with dexamethasone; treatment with monoclonal antibodies was not yet introduced in our ICU.13
Ultrasound Examination
Ultrasound images were acquired by 5 well-trained researchers (MHa, JS, MHe, LV, AL) with extensive ultrasound experience (> 2 y of almost daily ultrasound use) using a Sonosite Edge II ultrasound machine (Fujifilm Sonosite, Bothell, Washington). Reproducibility measurements were not performed as they have been performed in previous studies.4,14,15 Measurements were performed upon completion of the exam in the hospital’s picture archiving and communication system to minimize the time spent on the isolation unit. All images were acquired in supine position, or in case of the diaphragm sometimes in prone position, with a 4–12 MHz high-frequency linear transducer. Gain and depth settings were freely adjustable by the operator but aimed to create ideal images for measurement (ie, depth just below the diaphragm so that it does not leave the window during respiration and gain to create contrast between the muscle tissue and pleural/peritoneal lines).16 Clip length was adjusted to the patient’s breathing frequency to capture at least 3 full breathing cycles. Measurements included muscle thickness at end expiration (for the diaphragm), muscle thickness at end inspiration (for rectus abdominis and lateral wall muscles), and were performed in at least 3 cycles. The calculated average of these values was used as final value. During the first exam, the sites of examination were marked with a skin marker for the following examinations (Fig. 2).
Placement of ultrasound transducer for image acquisition. PL = pleura; D = diaphragm; PE = peritoneum; RA = rectus abdominis; EO = external oblique; IO = internal oblique; TA = transverse abdominis.
Diaphragm.
Images were acquired in the right zone of apposition between the 8th and 11th intercostal space perpendicular to the chest wall so that 3 hyperechogenic lines (pleura, fibrous layer, and peritoneum) were visible.10,14,17 Calipers were placed between the 2 outer hyperechogenic layers.
Rectus abdominis.
Rectus muscle was examined on the right side, perpendicular to the abdominal wall at its thickest portion in on the right side of the patient, slightly supraumbilical.18 Calipers were placed between the fascia surrounding the muscle. During examination, only minimal pressure was applied to prevent compression of the muscle.
Lateral abdominal muscles.
Lateral abdominal muscles were examined on the right side, perpendicular to the abdominal wall approximately between at the midpoint of arcus costalis and ilium, with assessment of total thickness of the most ventral and dorsal fascia in addition to measurement of individual muscle thickness.4,19 During examination, minimal pressure was applied to prevent compression of the muscle.
Statistical Analysis
No sample size calculation was carried out for this pilot study as no previous data on which to base this calculation exist in this patient group. At registration time, we aimed to include 40 subjects; however, due to external factors such as lack of enough trained personnel for daily measurements and decreasing admissions of COVID-19 cases, inclusion was stopped at 30 subjects. Analyses were performed on the entire cohort, and comparisons were made between 3 predefined subgroups. These were defined based on previous literature by the percentage change from the muscle baseline thickness on day 7 (≥ 10% increase, ≥ 10% decrease, and < 10% change).3 Missing data on muscle thickness during the study period were considered missing at random and were not imputed.
Baseline characteristics are presented for the entire cohort and per subgroup. They were tested for normality with the Shapiro-Wilk test, evaluation of histograms and quantile-quantile plots, and are presented as means ± SD, medians with interquartile range, or numbers (percentages) when appropriate. Differences in baseline characteristics between groups were tested with analysis of variance, Kruskal-Wallis test, or chi-square test when appropriate.
In addition, a mixed-model regression analysis was performed to determine the presence of a statistically relevant association between variables hypothesized to impact loss of diaphragm thickness. To this end, independent variables (SOFA, APACHE II, level of inspiratory support, level of PEEP, P0.1, hours of sedation, and hours of NMBA use) were analyzed in a univariate analysis with percent change of diaphragm thickness from baseline as dependent variable. The cutoff for inclusion in the multiple linear mixed-effects model analysis was set at P ≤ .05.
A P < .05 was regarded as statistically significant. Statistical analyses were performed using IBM SPSS Version 22 (IBM, Armonk, New York) and R statistical software Version 3.6 (R Foundation for Statistical Computing, Vienna, Austria).
Results
The study flow chart is presented in Fig. 1. Thirty subjects were included, of which 11 showed ≥ 10% decrease from baseline diaphragm thickness; 10 showed < 10% change, and 9 showed ≥ 10% increase from baseline. Of 212 examinations, images of the diaphragm could be acquired in 209 (99%) of cases, rectus abdominis 142 (67%), and lateral abdominal muscles 133 (62%). For the diaphragm, missing data were exclusively due to missed measurements, whereas for abdominal muscles prone ventilation was the main reason for missing data (151 missing data points, of which 142 [94%] were due to prone ventilation).
Regarding clinical characteristics summarized in Table 1, a statistically significant difference was found in symptom duration before intubation between groups, longest in the decrease group (12 [11–14] d, P = .03). The total time spent mechanically ventilated within the first week also differed significantly (P = .03). In addition, a statistically significant difference was seen in h of sedative use (P = .004), whereas duration of NMBA use was lower in the increase group, albeit without statistical significance. No difference was seen in clinical outcome on follow-up on day 28; however, h spent ventilated during the first week were lowest in the increase group (156 [129–172] h, P = .03).
Clinical Characteristics of Study Cohort and Subgroups According to Change of Diaphragm Thickness Over Time
Values of ultrasound measurements of the respiratory muscles are depicted in Table 2 and Table 3. Changes of diaphragm thickness over time per group are depicted in Fig. 3. Average initial diaphragm thickness was 1.4 (1.1–1.6) mm and did not differ from final average thickness (1.3 [1.1–1.5] mm, P = .54). A statistically significant difference on the last day was seen (P = .09). The group with the highest initial thickness (1.6 [1.4–1.9] mm) also demonstrated the lowest thickness on the last day of ultrasound (1.2 [1.1–1.4] mm). The rectus abdominis did not show statistically significant changes, whereas lateral abdominal wall thickness decreased from 14 (10–16) mm at baseline to 11 (9–13) mm on the last day of mechanical ventilation (P = .08). Of the latter, both external and internal oblique muscles showed a decrease from first to last day, whereas the transverse abdominis did not, albeit all without statistical significance.
Evolution of diaphragm thickness according to percent change over time.
Ultrasonographic Characteristics First Day and Last Day
Muscle Thickness Baseline Versus Last Day
After backward stepwise exclusion of statistically nonsignificant variables, the final mixed-effect linear regression model contained h NMBA use (P = .01) and PEEP (P = .041) (ESM 1, see related supplementary materials at http://www.rcjournal.com). There were no missing data for the covariates. Three data points were excluded due to missing of the dependent variable.
Discussion
The findings of interest in this study in adult mechanically ventilated subjects with COVID-19 with daily ultrasound examination of the respiratory muscles are (1) average diaphragm thickness did not change over time; (2) ventilation time was the shortest in the increase group; (3) NMBA use was correlated with loss, whereas PEEP with gain of diaphragm thickness; and (4) lateral abdominal muscle thickness decreased but this change was not statistically significant.
In light of viral expression and concomitant activation of fibrosis-related gene expression in diaphragm muscle fibers and prolonged ventilation in patients with COVID-19, we expected to observe a relevant decrease of overall diaphragm thickness in our cohort.1,8,20 Interestingly, this was not the case. As no histological data of muscle tissue are available, we can only speculate about the underlying reasons for this observation. First, inflammatory changes or increased respiratory effort–induced changes during the course of COVID-19 infection might lead to the formation of tissue edema, masking the loss of muscle mass due to fibrosis. Whereas this might be possible explanation, it should be noted that in a previous study the increases of respiratory muscle thickness originated from the fascia and membranes rather than muscle tissue itself.4 Second, subjects were examined during the first week of mechanical ventilation, and a longer follow-up might be necessary to appreciate the full effect of these changes. This is supported by the fact that specimens in the histological studies were taken from subjects that had passed away, with a median ventilation time of 12 d. Third, we hypothesize that the relatively small sample size with a roughly equal distribution of subjects with increase, no change, and decrease of diaphragm thickness could have also resulted in a lack of measurable difference overall.
Interestingly, the group of subjects that did show > 10% atrophy also had the highest initial diaphragm thickness. Whereas this is in line with a previous study, the underlying pathophysiology remains to be elucidated.3 Even though it could be seen as initial compensatory hypertrophy of functional muscle mass due to higher respiratory work load and its loss when this load is reduced through mechanical ventilation, the subacute time course of respiratory distress in both studies makes this unlikely. The accumulation and loss of tissue edema as described above seem like a more probable explanation. Nevertheless, these subjects still showed a longer time spent ventilated during the first week than those with increased thickness, highlighting the importance of monitoring these changes closely and elucidating the underlying causes.
With this in mind, our study found an association of decrease in diaphragm thickness with increased duration of NMBA use. Physiologically speaking, this is a plausible conclusion as diaphragm inactivity has been shown to rapidly result in myofibrillar disuse.1 As follows, an approach that encompasses diaphragm-protective ventilation next to lung-protective ventilation, that is, allowing sufficient diaphragm activity to prevent atrophy without overt activity causing damaging to itself or the lungs, is of great interest and poses an important challenge in this patient category.5,9,21 Whether such an approach has impact on clinical outcomes is yet to be elucidated in future studies.22 In addition, PEEP was positively associated with change of diaphragm thickness. We hypothesize that this effect is not based on hypertrophy but rather results from various confounding factors. A lower resting position of the diaphragm at end expiration, resulting in increased end-expiratory thickness through passive shortening could, for example, be at play.23
Whereas overall thickness of the rectus abdominis muscle remained constant throughout the course of mechanical ventilation, the overall thickness of the lateral abdominal wall muscles did not, albeit without statistical significance. Interestingly, the decreased total thickness is most likely attributable to decreases of the internal and external oblique muscles, as the transverse abdominis thickness did not change over time. This needs to be corroborated in a larger study. The lack of statistically significant changes could be explained by the difference in rate of atrophy of the lateral abdominal muscles.1,6,20 Nevertheless, these are interesting findings as they are in line with a previous study in non–COVID subjects, underlining the importance of not only considering loss of diaphragmatic muscle mass but also of the expiratory muscles.4 Loss of thickness and in turn potentially functionality could result in impaired airway clearance and result in longer hospital stay.4
Strengths and Limitations
This study has some important limitations and strengths. First, an important limitation is that no sample size calculation was carried out and only 30 subjects were included. Nevertheless, a strength of this study was that subjects were included consecutively, and only one patient meeting the inclusion criteria was excluded due to extreme body fat percentage hampering diaphragm visualization by ultrasound. In addition, measurements were performed daily without interruption, which is commonly not achieved. It should also be noted that results were comparable to a similar study in non–COVID 19 subjects.12 This increases the probability that we included a representative sample. Second, due to the frequent necessity for prone ventilation, measurements of the abdominal muscles could not always be performed, resulting in missing data. For the diaphragm, however, almost all attempted measurements were successful. Third, the cutoff at 10% change in thickness in either directions from baseline was derived from a previously published study.3 Whereas these categories had important clinical implications, they are not based on physiological or histological evidence for relevant atrophy or hypertrophy.
Conclusions
In ventilated subjects with COVID-19, overall no change in diaphragm thickness was observed. Subjects with decreased or unchanged thickness had a longer ventilation time than those with increased thickness. NMBA use was associated with decreased thickness. Rectus muscle thickness did not change over time, whereas lateral abdominal muscle thickness decreased but this change was not statistically significant.
Acknowledgments
We want to thank Professor Johannes Berkhof for his statistical advice during the project.
Footnotes
- Correspondence: Mark E Haaksma MD, Amsterdam UMC, location VU University Medical Centre Amsterdam, Postbox 7507 1007MB, Amsterdam, the Netherlands. E-mail: m.haaksma{at}amsterdamumc.nl
See the Related Editorial on Page 1489
The authors have disclosed no conflicts of interest.
Supplementary material related to this paper is available at http://www.rcjournal.com.
- Copyright © 2022 by Daedalus Enterprises
