Respiratory Musculature Evaluated by Computed Tomography in the Setting of Prolonged Mechanical Ventilation

Abstract

BACKGROUND:

Diaphragm atrophy has been observed in subjects who undergo invasive mechanical ventilation. We propose a new method to assess for respiratory muscle (RM) changes in subjects who undergo invasive mechanical ventilation by assessing for changes in respiratory muscles through computed tomography (CT).

METHODS:

A retrospective case series study was conducted on subjects who underwent invasive mechanical ventilation and received at least 2 chest CT scans during admission. Exclusion criteria included history of chronic mechanical ventilation dependence and neuromuscular disease. Respiratory muscle cross-sectional area (CSA) was measured at the T6 vertebrae.

RESULTS:

Fourteen subjects were included: mean (± SD) age, BMI, and admission APACHE II scores were 54.0 y (± 14.9), 32.6 kg/m² (± 10.9), and 23.5 (± 6.0), respectively. Ten (71%) subjects were male. Mean length of time between CT chest scans was 7.5 d (± 3.3). Mean duration of invasive mechanical ventilation was 4.5 d (± 3.4). The percentage change in TM CSA among those who underwent invasive mechanical ventilation was 10.5% (± 6.1).

CONCLUSIONS:

We demonstrated that serial analysis of respiratory muscle CSA through CT chest scans can be a method to assess for respiratory muscle atrophy in subjects undergoing mechanical ventilation. Future prospective studies involving larger populations are needed to better understand how this method can be used to predict outcomes in mechanically ventilated patients.

Introduction

Invasive mechanical ventilation is a lifesaving intervention that is commonly used in ICUs across the United States. It has been previously reported that 30–40% of ICU patients undergo invasive mechanical ventilation at any time during their hospitalization.¹ Although invasive mechanical ventilation is a life-sustaining intervention in those with acute respiratory failure, there are inherent complications and increased costs when used for prolonged periods of time. For example, up to 30% of patients who undergo invasive mechanical ventilation have difficulty weaning from the ventilator, which is associated with increased risk of mortality, poor functional outcomes, and greatly increased health care costs.^2,3 Invasive mechanical ventilation has been estimated to cost $600–$1,500 per day, and those requiring invasive mechanical ventilation for more than 3 weeks account for over half of all ICU costs.⁴ Initiatives to minimize prolonged invasive mechanical ventilation have been an area of interest to many critical care specialists in an effort to minimize costs and pulmonary complications and to improve patient outcomes.

Several reversible causes have been thought to contribute to difficulty weaning including neuropsychological issues, critical illness neuromuscular abnormalities, metabolic derangements, and iatrogenic influences.⁵ Ventilator-induced diaphragm dysfunction, or weakness in the force-generating capacity of the primary respiratory muscle (RM), is a frequently cited cause of difficulty weaning.⁶ Various factors have thought to contribute to ventilator-induced diaphragm dysfunction including sepsis, severe multi-organ dysfunction, and pharmacologic agents.⁷ Perhaps one of the largest contributors to ventilator-induced diaphragm dysfunction has been hypothesized to occur secondary to invasive mechanical ventilation.⁸ Specifically, the ventilatory assistance provided by invasive mechanical ventilation is thought to suppress the patient’s own inspiratory drive, which has been shown to promote diaphragm atrophy.^8,9

There have been multiple attempts made to measure diaphragm atrophy, most recently through ultrasonography (US), as a predictor of outcomes in invasive mechanical ventilation patients with limited success.^10-14 Given lack of standardization with various US techniques utilized in measuring the diaphragm and influence of patient position throughout the examination, other methods of assessing ventilator-induced diaphragm dysfunction are needed.^14,15 Thus, as a proof of concept, we investigated the usefulness of measuring respiratory muscle cross-sectional area (CSA) via computed tomography (CT) chest scans as a technique to measure changes in respiratory musculature during prolonged invasive mechanical ventilation.

QUICK LOOK.

Current knowledge

Respiratory muscle atrophy and diaphragm dysfunction have been reported in patients undergoing prolonged invasive mechanical ventilation. Various methods including ultrasound have been utilized to quantitatively measure respiratory muscle atrophy with limited success.

What this paper contributes to our knowledge

We present a method to quantitatively measure respiratory muscle atrophy, observed in mechanically ventilated subjects through the use of computed tomography (CT) chest scans. Analysis of respiratory muscle cross-sectional area at the upper thoracic vertebral level on CT chest scans demonstrated reduction in subjects who underwent prolonged ventilation.

Methods

After institutional review board approval (IRB number 14062505) was obtained, a retrospective review was conducted on subjects admitted to an adult ICU who underwent invasive mechanical ventilation and received at least 2 chest CT scans: one scan on admission and the subsequent scan at least 3 d thereafter. Subjects were eligible for inclusion if the underlying cause for intubation had resolved before initiation of weaning. Exclusion criteria included history of recent cardiothoracic surgery, previous invasive mechanical ventilation dependence, tracheostomy prior to admission, or known neuromuscular disease.

Variables collected included demographic, treatment, and outcome data. Respiratory muscle CSA was calculated using sliceOmatic v5.0 software (TomoVision, Magog, Québec, Canada) at the sixth thoracic mid-vertebrae (Fig. 1). Respiratory muscles were highlighted by a human operator, and the CSA was then computer generated. Muscles included in analysis were the pectoralis minor and major, serratus anterior muscle, intercostal muscles, subscapularis muscle, erector spinae muscle, and trapezius muscle (Fig. 2). The CSA values for each subject was then divided by their respective height in meters squared to create an index that could be comparable across subjects. Two operators trained on the software analyzed each subject, and results were compared for inter-rater variability. The primary aim of the study was to investigate a new technique to assess respiratory muscle change in subjects undergoing mechanical ventilation. Descriptive statistics were performed using Stata/IC software (version 16.1, StataCorp, College Station, Texas).

Fig. 1.

Admission respiratory muscle cross-sectional area (CSA) of 97.3 cm² (A) compared to subsequent computed tomography scan 15 d later with respiratory muscle CSA of 67.52 cm² in the same subject (B).

Fig. 2.

Cross-sectional area of the respiratory muscle groups at the sixth thoracic vertebra, including pectoralis major and minor muscles (red), serratus anterior and intercostal muscles (orange), and subscapularis, trapezius, and erector spinae muscles (green).

Results

Fourteen subjects met inclusion criteria with demographic data available in Table 1. Most subjects were male (71%, n = 10) and Black (57%, n = 8), followed by white (28%, n = 4), and other race (15%, n = 2). Primary causes of intubation were acute hypoxic respiratory failure in 11 subjects (78%), circulatory compromise in 2 subjects (15%), and postoperative respiratory failure in 1 subject (7%). Sepsis was identified as the most common ICU admission diagnosis in 6 subjects (44%), followed by gastrointestinal hemorrhage in 2 subjects (14%), acute systolic heart failure in 2 subjects (14%), ARDS in 1 subject (7%), necrotizing pancreatitis in 1 subject (7%), diffuse alveolar hemorrhage in 1 subject (7%), and orthotopic liver transplantation in 1 subject (7%). Mean (± SD) duration of time between CT chest scans was 7.5 d (± 3.3). The mean (± SD) thoracic muscle (TM) CSA on initial and second CT chest scan was 78.7 cm²/m² (± 18.8) and 69.8 cm²/m² (±17.2), respectively. This represents a mean percent reduction in TM CSA (cm²/m²) of 10.5%. Secondary ICU outcomes can be found in Table 2. Total mean (± SD) duration of invasive mechanical ventilation was 4.5 d (± 3.4). Most subjects (79%, n = 11) were placed on pressure-regulated volume control ventilation. Mean time to invasive mechanical ventilation weaning was 29.5 h (± 17.3). Nine subjects (64%) failed to wean, and 2 (22%) underwent tracheostomy. There was one (11%) in-patient mortality, and 7 (50%) were successfully able to discharge to home.

Table 1.

Subject Demographics

Table 2.

Hospital Course and Outcomes

Discussion

We have demonstrated that serial analysis of respiratory muscle CSA measured via CT chest scans can provide a possible method to assess respiratory musculature in subjects undergoing invasive mechanical ventilation. To our knowledge this is the first report of using serial CT scans to directly measure changes in respiratory musculature in subjects undergoing invasive mechanical ventilation. Although other attempts have been made to measure this phenomenon, including phrenic nerve stimulation, airway pressure testing, and most notably US assessment of the diaphragm, our technique may provide an easy method to incorporate in studies investigating the impact that invasive mechanical ventilation has on respiratory musculature. Furthermore, this technique may provide a method to assess for frailty in critical care patients that undergo radiographic imaging of the chest.

Sarcopenia, defined as a state of generalized muscle wasting and reduced function, has been reported to be associated with poor outcomes in various critically ill, surgical, and malignancy cohorts.^16-18 The protocol to assess for generalized sarcopenia has been well described and includes assessing for the skeletal muscle CSA at the third lumbar vertebra.¹⁶ Assessing for generalized sarcopenia on ICU patients at the third lumbar level has demonstrated that those with sarcopenia have worse overall survival, decreased ventilator-free days, and increased risk of extubation failure.^19-21 Similarly, thoracic sarcopenia, or reduced muscle mass at the thoracic vertebral level, has recently emerged as a tool to predict outcomes of patients in cardiothoracic surgery cohorts. Fintelmann et al²² and Troschel et al²³ have both demonstrated that reduced thoracic skeletal muscle CSA was associated with worse postoperative outcomes and more cardiopulmonary complications following lung resection surgery. However, little is known about the utility of assessing for thoracic sarcopenia, or respiratory muscle atrophy, in critical ill patients who undergo invasive mechanical ventilation.

Within the last several years, there has been much research investigating the underlying pathophysiologic mechanisms driving ventilator-induced diaphragm dysfunction.²⁴ Ventilator-induced diaphragm dysfunction is thought to result from a complex interplay between treatable and untreatable causes that contribute increased oxidative stress, reduced protein synthesis, and mitochondrial dysfunction that leads to respiratory muscle disuse atrophy.^8,25,26 Key to assessing for ventilator-induced diaphragm dysfunction in human subjects is utilizing methods that can directly measure changes in respiratory musculature in patients undergoing invasive mechanical ventilation. Traditionally, phrenic nerve stimulation was considered the reference test for assessing for diaphragm dysfunction.²⁷ However, phrenic nerve stimulation is not practical in everyday practice and is often difficult to perform.

More recently, US has emerged as a popular method to assess for ventilator-induced diaphragm dysfunction related to its ease of use and the ability to record bedside measurements over multiple time points. Various techniques have been cited as methods to evaluate for ventilator-induced diaphragm dysfunction through US including measuring diaphragm thickness, and its percent change throughout respiration, or respiratory excursion via ultrasound B and M modes.²⁸ Multiple studies have reported the prevalence of ventilator-induced diaphragm dysfunction, as measured through US, throughout the weaning process to include between 30–70% of mechanically ventilated subjects.^29-31 To date, a majority of studies have used the diaphragm thickness as a measure of diaphragm dysfunction.^10,32-38 Grosu et al³² demonstrated in 7 mechanically ventilated subjects that the diaphragm thickness decreased at a rate of 6% per day. However, the relationship between changes in diaphragm thickness and extubation success is less understood; and multiple authors have argued that the diaphragm thickening fraction, as measured with US, best approximates diaphragmatic strength.^10,34,38

The above studies of diaphragm evaluation with US are limited by their small sample sizes; they were performed at a single institution, and the authors did not adhere to the same protocol to assess for changes in the diaphragm musculature. Furthermore, definitions of extubation failure used between studies were not in agreement, with some authors including subjects extubated for end-of-life reasons.^10,31 Recently, Vivier et al¹³ investigated the usefulness of diaphragm US measurements at time of extubation in a large cohort of 192 subjects in an attempt to mitigate these limitations and determined that diaphragmatic dysfunction present on US was not associated with increased risk of extubation failure. Thus, there is a continued need to find better predictors of extubation outcomes in patients who undergo invasive mechanical ventilation.

There are limitations to our study that need to be addressed. First, this proof-of-concept study was a retrospective study that comprised a small sample size of subjects admitted to an adult medical ICU that had 2 CT chest scans during their admission. There are numerous patients who require invasive mechanical ventilation; however, only a small minority of those admitted to our institution happened to have 2 CT chest scans available for analysis. Furthermore, as this was retrospective in design and most subjects were on invasive mechanical ventilation at time of CT chest scan, it was not feasible to control phase of respiration, which may have interfered with respiratory muscle measurements. There are risks associated with obtaining CT scans including radiation exposure and accidental extubation during transport. Given that this study was not sufficiently powered, we were not able to perform multivariate analysis to compare the change in respiratory muscle to other variables known to be associated with extubation failure. Future investigations should include mode of mechanical ventilation, pressure support levels, nutritional support, and other variables when assessing for changes in respiratory musculature. Finally, due to the retrospective nature of this study, subjects were not evaluated for respiratory muscle function using sniff tests or maximal inspiratory pressure determination to correlate CT findings of muscle wasting and measures of muscle function. It is unclear how respiratory muscle CSA would correlate with measures of respiratory function and strength, which should be addressed in future larger prospective studies. This study was a proof of concept that aimed to demonstrated that CT scanning of the chest can be a possible tool used to assess respiratory muscle CSA and this may have a direct relationship with weaning success and respiratory muscle function.

Conclusions

Overall, we have reported the first use of a technique that attempts to measure change in respiratory musculature through serial CT chest scans in subjects who are undergoing invasive mechanical ventilation. Because this is a small study, further research involving larger cohorts and more homogenous patient populations are needed to fully understand the role that CT scans can have in assessing respiratory muscle changes throughout invasive mechanical ventilation. Furthermore, future prospective studies comparing this technique with previously reported US techniques are needed to understand which may offer the best strategy for assessing respiratory muscle changes and correlate those changes to respiratory muscle function and weaning success.