Using Ultrasound to Study Respiratory Muscle Function in Critically Ill Patients

This is a prospective observational study designed to assess serial changes in quantitative respiratory parameters in patients on different modes of MV e.g. synchronized intermittent mandatory ventilation, controlled mechanical ventilation and pressure support ventilation. This strategy will compare serial changes in diaphragm muscle parameters in ventilator strategies that provide different level of spontaneous respiratory effort and targets for MV. The study will be carried out at Wake Forest Baptist Medical Center. The protocol has been approved by the Institutional Review Board. All adult (>18 years) patients admitted to an ICU will be screened. Written informed consent will be obtained from each patient or their surrogate.

Daily ultrasound studies will be done on the diaphragm muscle. The first study will be conducted within 48 hours of ICU admission. Serial exams will be repeated daily until the patients are extubated or transferred out of the ICU. A maximum of 7 ultrasound studies will be performed since the focus is to demonstrate the utility of early serial changes in measurable parameters on ultrasound. Images will be acquired on both sides during quiet breathing with the patient's arms relaxed by the side in a neutral position while lying with the head of bed at 30 degrees. Most ICU patients are in this position. Diaphragm muscle image at end expiration on each side will be acquired in B mode with a 6-13 MHz transducer in the mid axillary line using the intercostal approach. Diaphragm excursion will be measured bilaterally through anterior subcostal approach in the M mode through a 2-5 MHz transducer. Images will be acquired in accordance with the guidelines set forth in the AANEM position statement. Each ultrasound examination will be done at a time when it does not interfere with concurrent ongoing nursing \& medical care.

To determine the effect of the serial changes that occur in respiratory muscle thickness and echodensity on MV, linear regression models will be applied. The distribution of the outcome variables and the need for any transformation to approximate normality will be checked. The outcome variables are serial changes between first and last visit diaphragm muscle thickness and echodensity. The covariate of interest is the mode of MV. The confounding variables are neurological disease, acute lung injury, cumulative fluid balance, age, BMI, history of smoking, duration between the first and last ultrasound examination and the nutritional status (albumin, cumulative calorie intake and type of tube feeds). Since serial changes are the outcome variables, besides the potential confounders, baseline outcome measures will be also adjusted in the regression models. Model selection approaches will be used to select the confounding variables adjusted in the models. Furthermore, to consider the changes across different ultrasound visits, first the individual trajectories of the outcome variables will be displayed graphically. Then, the mixed effects model will be used to examine the associations. This model can handle repeated measures over time by modeling the covariance structure. The covariates will be the same as described previously for the linear regression model. Logistic regression models will then be used to determine if the serial changes in diaphragm muscle thickness and echodensity predict the total number of days on MV, success at SBTs and successful extubation.

Diaphragm ultrasound is a promising new method for evaluating the diaphragm during MV. The investigators feel that the study proposal is extremely relevant in its ability to explore a novel approach to assess the risk of extubation failure, very early on, in the course of MV based on change in muscle echotexture using a noninvasive bedside tool. The investigators propose that changes in diaphragm configuration associated with MV correlate to clinical outcomes as measured by days on MV and extubation failure. The early serial changes detected by ultrasound suggest a potential benefit of proactive early therapies designed to preserve respiratory muscle architecture to reduce days on MV and prevent extubation failure. The investigators further hypothesize that different ventilator modes may affect these changes differentially and these changes might be prevented by titrating ventilatory support to maintain normal levels of respiratory effort. If spontaneously breathing patients and patients on low level of ventilatory support show differential changes on diaphragm muscle parameters compared to patients on controlled modes and higher ventilatory support, clinical protocols might need to emphasize alternate strategies that focus on "muscle protective" or "muscle restorative" ventilation. Ultrasound can also be used to develop clinical paradigms that involve inspiratory muscle training in helping patients who have failed extubation in getting liberated from MV.

The goals are to use the pilot data generated through this grant to power and design a randomized controlled trial of a "muscle protective" or "muscle restorative" ventilation strategy and assess its effectiveness in reducing days on MV and extubation failure compared to current medical practice. This team approach will involve a novel collaboration of key members with expertise in muscle ultrasound, mechanical ventilation, mechanisms of muscle wasting related to acute lung injury and physiological effects of exercise on respiratory muscle hence representing critical input into developing clinical strategies to assess and reduce the burden of extubation failure. The investigators believe this approach is innovative in aiming to use a non-invasive bedside clinical tool to develop "muscle protective" and "muscle restorative" ventilator strategies with a focus on early respiratory rehabilitation in critically ill patients on MV.