Abstract
INTRODUCTION
A frequent cause of weaning and extubation failure in critically ill mechanically ventilated patients is diaphragm muscle dysfunction. Ultrasound (US) evaluation of the diaphragm yields important data regarding its thickness (diaphragm thickening fraction [TFdi]) and its movement or excursion (diaphragmatic dynamics) that reveal the presence of diaphragmatic dysfunction.
METHODS
Cross-sectional study, which included patients older than 18 years with invasive mechanical ventilation with an expected duration of more than 48 h, in a tertiary referral center in Colombia. The excursion of the diaphragm, inspiratory and expiratory thickness, and TFdi were evaluated by US. Prevalence and use of medications were evaluated, and the association with failure in ventilatory weaning and extubation was analyzed.
RESULTS
Sixty-one patients were included. The median age and APACHE IV score were 62.42 years and 78.23, respectively. The prevalence of diaphragmatic dysfunction (assessed by excursion and TFdi) was 40.98%. The sensibility, specificity, positive predictive value, and negative predictive value for TFdi < 20% was 86%, 24%, 75%, and 40%, respectively, with an area under the receiver operating characteristic (ROC) curve of 0.6. The ultrasonographic analysis of excursion of the diaphragm, inspiratory and expiratory thickness, and TFdi (>20%) allow in its set and with normal values, predict success or failure for the extubation with an area under the ROC curve of 0.87.
CONCLUSION
Diaphragmatic dynamics and thickness parameters together assessed by ultrasonography could predict the success of extubation in critically ill patients in Colombia, based on the finding of diaphragmatic dysfunction.
Introduction
Mechanical ventilation (MV) is a therapeutic tool for daily use in intensive care units (ICU), whose objective is to maintain adequate gas exchange, improving alveolar ventilation in critically ill patients with respiratory failure, shock, postresuscitation status, among others.1,2 Once the patient overcomes the harmful condition that motivated the need for ventilatory support, weaning or reduction of the ventilatory support parameters should be programmed and disconnection of mechanical ventilation should be evaluated. 2 However, some patients may not tolerate the decrease in ventilatory support or weaning from the ventilator, and it is often necessary to reintubate the patient and restart ventilatory therapy.3,4 These cases account for 30% to 40%, leading to prolonged ventilation, increased ICU stay, worse prognosis, and up to 12% higher mortality compared to patients who could be extubated.2–4
A frequent cause of failure in weaning and extubation is dysfunction of the diaphragm muscle, which is the main breathing muscle (increasing mortality from 10% to 50%, according to some studies).4,5 Many factors have a negative impact on the function of this structure, for example, prolonged ventilatory support, sepsis, muscle relaxants, malnutrition, among others, which are very frequent in critically ill patients.6–8
There are currently several methods to assess diaphragmatic dysfunction (DD), such as phrenic nerve stimulation, electromyography, and ultrasound (US), the latter being a noninvasive, inexpensive, and easily accessible method.5,6 The US evaluation of the diaphragm yields important data on this muscle, for example, its thickness and its movement or excursion (diaphragmatic dynamics) that reveal the presence of dysfunction of the diaphragm.5,6 There is literature that addresses the issue of ultrasonographic evaluation of the diaphragm in critically ill patients, however, the results of the measurements, the sensitivity and specificity of this technique are very heterogeneous.5–9
In Colombia, the potential of this tool in predicting the success of extubation in critically ill patients has not been thoroughly evaluated. For this reason, the following study aims to evaluate the characteristics of the diaphragm by ultrasonography to predict the success of extubation in a critically ill Colombian population.
Additionally, we consider that it is a topic of the utmost importance, since it will allow, at the bedside of the patient, to determine which patients present dysfunction in the diaphragm, correlate them with associated risk factors and propose a comprehensive management to lead to the successful disconnection of the patient from the ventilator.
Methods
Study Design and Population
Prospective cross-sectional study, which included patients older than 18 years with invasive MV with an expected duration of more than 48 h, in a tertiary referral center in the city of Bucaramanga, Colombia, for 1 year (Clínica FOSCAL). Patients who were admitted to the ICU from a home care program and who were on chronic MV, pregnant women, patients with a history of unilateral diaphragmatic paralysis, individuals on MV under palliative management in whom extubation was not considered, and individuals with polyneuropathies, such as Guillain Barré syndrome, myasthenia gravis, among others, were excluded. The sample was calculated based on the average annual population received in the hospital, obtaining 95% confidence intervals (95% CI) of at least 60 patients (power of 0.8 and alpha of 0.05).
Data Collection and Definition of Operational Variables
Sociodemographic data such as age and gender were collected; clinical variables such as length of hospital stay, duration of intubation, hemoglobin, hematocrit, arterial blood gases, definitive diagnosis, medications used, and diaphragmatic parameters were collected. The Acute Physiology and Chronic Health Evaluation (APACHE) IV score was also calculated.
DD is defined as a diaphragmatic excursion of <10 mm and a diaphragm thickening fraction (TFdi) < 20% (to talk about a clinically relevant DD). 10 This parameter was evaluated with M-mode. The TFdi was calculated using the following formula: ((end- inspiratory diaphragm thickness (tdi) - end-expiratory diaphragm thickness)×100)) / end-expiratory diaphragm thickness). 11
Likewise, failure in ventilatory weaning was defined as when the patient did not tolerate a spontaneous ventilation mode for a period of <12 h, in which the need to switch back to a controlled ventilatory mode (by pressure or volume, according to the criteria of the intensivist in charge of the patient) was observed. Extubation failure was defined as the need for orotracheal reintubation during the first 48 h after a scheduled extubation. The criteria for extubation of a patient were at the discretion of the medical specialist in charge of the patient.
Description of the Ultrasonographic Technique
The US evaluation was performed by a radiologist. The transducer was positioned under the right and left costal ridge with clavicular or anterior axillary midline, directing it cranially, medially, and dorsally. Diaphragmatic excursion was evaluated by this technique and was performed with a convex probe or a micro-convex probe. A 3.5 to 5 MHz probe was used, it was placed below the subcostal space with mid-clavicular line, making the sound beam travel perpendicular to the posterior third of the diaphragm.
To evaluate the variation of diaphragmatic thickness, the US transducer was positioned between the 8th to 10th intercostal space with the axillary midline in the place known as the zone of costal apposition. Here, either B-mode or M-mode was used, the latter evaluating the variation of thickness over time. The movement mode was used to evaluate diaphragmatic excursion (distance measured in centimeters) and diaphragmatic thickness (distance measured in millimeters).
For diaphragmatic thickness, the probe (10-12 MHz) was positioned perpendicularly over the 9th to 10th intercostal spaces, with anterior axillary line, in the right hemidiaphragm, as this offered a simpler and more accessible measurement thanks to the hepatic acoustic window. The diaphragmatic thickness fraction was obtained at the end of inspiration and at the end of expiration. For its calculation, the thickness value at the end of inspiration and at the end of expiration is subtracted and this result is divided by the thickness at the end of expiration, this value is multiplied by 100.
Timing of Ultrasonographic Evaluation
Patients who were candidates for participation in this study were identified according to the study entry criteria. Active communication was maintained with the intensivists present in the ICU. The US assessment of the diaphragm was performed once the ventilatory weaning process was initiated with spontaneous breathing test or spontaneous mode. The latter process was defined independently by the intensivist in charge of the patient.
Cases in which extubation was successful, cases in which weaning from ventilation failed and those that, after extubation, required reintubation during the first 48 h of the first extubation were recorded.
Statistical Analysis
All statistical analyses were performed using STATA version 14 (StataCorp LP). Nominal and ordinal variables were expressed as percentages, while discrete and continuous variables were expressed as median and interquartile range, since they did not have a normal distribution. To evaluate the performance of the diaphragmatic parameters, the result obtained was compared with the extubation success, from which the values and 95% CI of sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated, and area under the curve (AUC) was calculated. Additionally, the relative risk (RR) and odds ratio (OR) of variables associated with extubation success was calculated. The P value < .05 was defined as statistical significance.
Ethical Considerations
Endorsement was obtained from the ethics committee of the participating institution (Act 003-21). Taking into account that the diagnostic method used was US, the patients were not exposed to ionizing radiation, nor were they subjected to invasive interventions. The study was conducted following the ethical guidelines established by the last update made to the Declaration of Helsinki at the 64th General Assembly, Fortaleza, Brazil, October 2013 12 ; and to the ethical considerations as set forth in Resolution 8430 of 1993 of the Colombian Ministry of Health. 13 Therefore, the present study was classified as minimal risk research. Both the patient and family members were informed of their right both to participate and to withdraw from the research at any time during the investigation, therefore, informed consent was obtained from each participant.
Results
Baseline and Clinical Characteristics of the Study Population
Sixty-one patients were included. The 60.66% (n = 37) of the patients were male. The median age was 62.42 years. The severity of the patients’ clinical picture according to APACHE IV scores had a median of 78.2 points and a predicted mortality of 39.66%. The most frequent pathology was COVID-19 disease in 34.43%, followed by traumatic or surgical causes (31.15%) and infectious disease due to causes other than COVID-19 (18.03%). Within these groups, the main diagnosis was acute respiratory distress syndrome in 44.26% of the patients (Table 1).
Sociodemographic, clinical, and laboratory characteristics of the study population (N = 61).
Abbreviations: APACHE, Acute Physiology and Chronic Health Evaluation; IQR, interquartile range.
The clinical ventilation and oxygenation characteristics of the patients at the time of examination showed that the median Pa/FiO2 and PO2 were 225.84 and 92.98 mm Hg, respectively. The median length of stay and duration of orotracheal intubation were 25.34 and 13.29 days, respectively (Table 1). Of the patients, 60.66% required management with more than 2 analgesic and sedative drugs: opioids were the most commonly used group of drugs in 98.36%, followed by benzodiazepines 73.77%, propofol (26.23%) and dexmedetomidine (11.48%). Neuromuscular relaxant infusion was used in 34% of patients (Table 2).
Distribution of the frequency and type of drugs used in the population evaluated.
Abbreviation: NMR, neuromuscular relaxant.
Diaphragmatic Assessment by Ultrasonography
The overall results of the diaphragmatic parameters are summarized in Table 3. The prevalence of DD (alteration of diaphragm excursion and TFdi < 20%) was 16.39% (Table 4). Failure of ventilatory weaning was 27.87%; and failure of extubation was 16.39%. And 11.48% of extubated patients required noninvasive MV, and it was not considered as extubation failure.
Results of diaphragmatic parameters evaluated by ultrasonography.
Prevalence ofDD, according to evaluated parameters and defined cut-off scores.
Abbreviations: CI, confidence interval; DD, diaphragmatic dysfunction.
Variables Associated With Extubation Success
This analysis showed that the highest OR for successful extubation was a TFdi > 20% (OR 2.73) and a TFdi >30% (OR 3.6). However, with a cut-off point of 35% TFdi only showed an OR of 1.97.
Performance of Diaphragmatic Dynamics and Thickness as Predictor of Extubation Success
When comparing the cases of successful extubation with the cut-off scores obtained and the rest of the US variables, it was observed that increasing the cut-off point of TFdi slightly improves sensitivity (90%), with changes in specificity with TFdi 30% and better AUC. The other diaphragmatic US variables showed a sensitivity between 82% and 86%, specificity between 13% and 40%, and AUC <0.60 (Figure 1).
Distribution of the area under the curve for the parameters: (A) TFdi <20%; (B) TFdi <25%; (C) TFdi <30%; (D) TFdi <35%. Abbreviations: ROC, receiver operating characteristic; TFdi, diaphragm thickening fraction.
By building a model with all US variables, showing that having normal values (TFdi, diaphragmatic thickness, and excursion together), it allows adequate discrimination to predict extubation success or failure, with an AUC of 0.8794 (Figure 2).
Area under the curve model by combining all parameters of diaphragmatic dynamics and thickness evaluated by ultrasound (TFdi, diaphragmatic thickness, and diaphragmatic excursion together). Abbreviations: ROC, receiver operating characteristic; TFdi, diaphragm thickening fraction.
Discussion
The use of US in critical care has been on the rise, given its easy availability, portability, and minimal risk to the patient; in addition, it provides real-time information, serving not only as a monitoring tool, but also as a therapeutic element.14,15 One of the most crucial moments in the management of the critically ill patient is the time of weaning from ventilatory support and extubation, since it is a reflection of the resolution of the underlying disease that generated the critical condition, assuming the risk of extubation failure and the need for reintubation, which is associated with increased morbidity and mortality.16,17
In this study, a prevalence of DD was found to be approximately 20%. There is a significant variability in this prevalence reported in the literature, which ranges from 23% to 36%. Kim et al 10 conducted a prospective cross-sectional study with 82 patients in Korea, where they evaluated the prevalence of DD by means of US parameters (diaphragmatic excursion <10 mm), evidencing a prevalence of 29%, documented during the spontaneous breathing test. 10 In the United States, McCool et al 18 conducted a clinical trial with 32 patients, showing a prevalence of DD of 43.7%, using TFdi <30% as the cut-off score, which was strongly associated with extubation time compared to the control group (23.2 ± 35.2 vs 57.3 ± 52.0 h; P = .046). 18 In Colombia, one of the few studies carried out was that of Varón et al, 19 which included 84 patients and found a prevalence of DD of only 1.2%. 19 Around the world, our results are comparable. However, at the national level, data are almost nonexistent. For this reason, it is difficult to hypothesize possible causal relationships with precision. In addition, our study presents emerging etiologies that should be studied separately to avoid biases in the generalization of results. Although most of the literature on diagrammatic US evaluates TFdi and diaphragmatic excursion (variables analyzed in this study), there are other parameters that have recently gained importance, such as diaphragmatic contraction velocity, which in the study by Varón et al, 19 was also used for the diagnosis of DD and could be related to the low prevalence reported. Diaphragm excursion divided by inspiration time has shown acceptable discriminatory power in predicting extubation success.19,20
Of the total population studied in our study, 27.87% failed during ventilatory support weaning tests, 16.39% of those extubated required reintubations before 48 h (extubation failure), and 32.79% required percutaneous tracheostomy. And 29.5% had failure both at destination and at extubation with normal values of diaphragm excursion and TFdi, which correlates with extra diaphragmatic factors that can lead to failure to release mechanical ventilatory support. In the study by Vivier et al, 21 DD assessed by US was similar in those who succeeded or failed extubation, and effective cough was found to be the only independent factor associated with extubation failure, 21 a parameter that was not evaluated in our study and should be considered in future studies.
Different values were obtained for successful and unsuccessful extubation. Of the patients with DD detected by US, 18% failed to wean from ventilatory support and failed extubation. This variation in reported percentages is determined by several factors such as steroid use, hospital stay, and infections, as found in our final analysis. Our final model evaluated the sensitivity, specificity, positive and negative predictive value of the variables measured with US to determine the success of extubation when the patient was on spontaneous invasive ventilation. The sensitivity of diaphragmatic excursion > 10 mm to determine extubation success was 86%, with a specificity of only 25%. The cut-off point of a TFdi of 25% to 30% to define extubation success, had a sensitivity of 86% to 90% and a PPV of 75% to 69%, respectively.
When contrasting these findings with those reported in the literature at a global level, it became evident that in the study by DiNino et al, 15 who conducted a clinical trial with the aim of evaluating the predictive capacity of diaphragmatic parameters on extubation success in MV patients, they demonstrated that a TFdi of > 30% has a sensitivity of 88% for successful extubation with a high positive predictive value of 91%. However, unlike our findings, it has a higher specificity and NPV. 15 A recent systematic review that analyzed 28 studies, of which 16 studies were included in a meta-analysis, found that the sensitivity and specificity of TFdi was 70% and 84%, respectively, and for diaphragm excursion, the sensitivity was 71% and specificity 80%. 22 However, the study published by Gok et al, 23 in which 62 patients from Turkey were evaluated with the same target line of research, showed that the sensitivity for TFdi (cut-off score was 27.5%) was 67.5%, with a specificity of 66.6%, a PPV of 95% and a low NPV (only 24%). Although the specificity obtained was higher than in our research, it is not the best and these findings are partially similar to our work with a similar number of patients. Additionally, diaphragmatic excursion in this study had a fair sensitivity (69%), with a specificity, PPV, and NPV of 67.5%, 96%, and 27.8%, respectively. 23
Although the sensitivity and PPV reported in our work for TFdi and diaphragmatic excursion is similar to that found in the literature, the specificity is lower, which may be explained by the heterogeneity of the patients, the latter given the high prevalence in our work of patients with COVID-19 with steroid use (34%), long-term infusion of muscle relaxants and prolonged ICU stay, recognized factors associated with deconditioning, myopathy and also related to failure of weaning and extubation.2–4 In the final model of the AUC with the US measurements of the diaphragm, where the variables are evaluated together, we obtained an AUC of 0.87, which shows that these measurements, having normal reference values, together and not as isolated data, improve the predictive and discriminative ability to predict success in extubation of the patient. We obtained an AUC of 0.64 for TFdi with a cut-off point of 30%, less than that reported in the work of McCool et al 18 where an AUC for TFdi alone of 0.9 was found for extubation success within 24 h and maintained up to 48 h. However, there is no clarity in that study on the type of population selected, or the type of drugs administered, days of stay as a strong associated outcome or days of MV and it has a very small number of patients in relation to our research, which may not be comparable. 18
Normal TFdi values in the literature vary from 29% to 36%, a cut-off of 30% has a reported sensitivity of 88%, specificity of 71%, and AUC of 0.79, being the combined use (TFdi and diaphragmatic excursion), relevant parameters when evaluating the suspension of MV.11,24,25 This study, with a cut-off point of TFdi of 30%, obtained a good sensitivity and PPV, but with low specificity, which allows us to suggest that a single variable should not be taken as a single parameter and a predictive model should be constructed that also evaluates excursion and diaphragm thickness, which increases the predictive capacity from 0.60 to 0.87.
Paradoxically, ARDS proved to be a protective factor for failed extubation with an RR of 0.37. There is no evidence that explains pathophysiologically that this disease can protect the critically ill patient from a failed extubation; probably, this finding is related to the nonperformance of ventilatory support weaning and scheduled extubation in ventilated patients with ARDS due to prolonged intubation, since in these cases, it is preferred to proceed to perform a percutaneous tracheostomy (in patients with intubation of more than 15 days). Therefore, at the moment of initiating the complex process of weaning from ventilatory support, the assessment of diaphragmatic activity by US should be complete, gathering the previously mentioned measurements as a whole, but it should not be the only thing to take into account, since there are several causes of extubation failure (eg, ventricular dysfunction, pulmonary edema, pleural effusion, neurological deterioration, physical deconditioning, etc), so it should be integrated with other clinical and laboratory variables. 26
Keeping in mind only US measurements of the diaphragm for weaning from ventilatory support could lead to ignoring other factors that are associated with extubation failure and thus to adopt inappropriate behaviors in our patients. Ideally, clinical, physiological, and laboratory measures, such as rapid shallow breathing rate, pulmonary compliance, muscle strength, Glasgow coma scale, hematocrit, renal function, should be evaluated together with US measurements of both the lung, diaphragm and heart, 27 to increase the discriminatory power to predict success or failure of extubation and to guide the necessary therapeutic measures.25,26 However, a major barrier that currently exists is the lack of consensus on a common definition of DD, based on ultrasonographic parameters, so the results of similar studies should be interpreted with caution and rigor.
Strengths and Limitations of the Study
As strengths of this work, highlights the potential for the application of US in daily respiratory monitoring at the patient's bedside, which is necessary for diagnosis and to guide treatment. US is a tool that should not be missing in the daily work of the intensivist, and this requires not only the availability of equipment, but also education, training, and periodic certifications. Additionally, it poses a new challenge and opens the debate on the management of Acute respiratory distress syndrome (ARDS) secondary to SARS-Cov-2 infection and encourages readers to carry out specific studies for this pathology.28,29
The limitations of this study include, the limited sample, which does not allow the difference between risk factors to be robustly detected; the heterogeneity of the population studied, the dependence on the operator performing the US and the lack of concordance evaluation among those who took the US can be recognized. In addition, US should not be the only method to be evaluated during weaning from ventilatory support, as there may be other variables that can negatively influence the success of extubation, such as infections, steroids, use of muscle relaxants, and we must keep these and other variables in mind when initiating the ventilatory weaning plan.
Conclusions
Diaphragmatic dynamics and thickness parameters together assessed by ultrasonography could predict the success of extubation in critically ill patients in Colombia, based on the finding of DD. The performance trend of these parameters needs to be evaluated in depth, as it could have a substantial impact on the outcome of the mechanically ventilated patient.
Footnotes
Author Contribution(s)
DECLARATION OF CONFLICTING INTERESTS
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
FUNDING
The authors received no financial support for the research, authorship, and/or publication of this article.
