Ultrasonography to Access Diaphragm Dysfunction and Predict the Success of Mechanical Ventilation Weaning in Critical Care

Weaning failure is common in mechanically ventilated patients, and whether ultrasound (US) can predict weaning outcome remains controversial. This review aims to evaluate the diaphragmatic function measured by US as a predictor of weaning outcome.

There are numerous factors such as respiratory and cardiac dysfunction, poor nutritional status, psychological issues, decreased muscle strength associated with weaning failure. 7Some of them are patients' related, while others may occur due to ICU care.The development of muscle weakness related to sepsis, multiorgan dysfunction syndrome, medications, bed rest, immobilization, and MV, named ICU-acquired weakness, contribute to difficulty in liberation from MV. 8 Physician's prediction for successful weaning has low accuracy, with positive (PPV) and negative predictive values (NPV) of only 50 and 67%, respectively. 9urrent guidelines 10 recommend the implementation of a spontaneous breathing trial (SBT) as a tool to predict weaning outcome.However, approximately 20% of all mechanically ventilated patients fail their first attempt to wean following a successful SBT, 11,12 since SBT monitoring is insensitive to detect early signs of load-capacity imbalance (the common pathophysiology of weaning failure).Several other parameters have been used extensively in clinical practice to predict weaning failure, such minute ventilation, vital capacity, maximum peak inspiratory pressure, airway occlusion pressure 0.1 seconds and rapid shallow breathing index (RSBI, ie, respiratory frequency/tidal volume), but none proved to be accurate. 13,14he diaphragm, the major respiratory muscle, is responsible for approximately 60-80% of the workload 15 with an excursion of 1-2 cm, while during the forced breathing its amplitude is up to 7-11 cm. 4 MV has been proved to induce several diaphragmatic abnormalities, leading to atrophy and contractile dysfunction of diaphragm (VIDD) 16 that is associated with poor prognosis at time of liberation from MV.As a result, diaphragmatic dysfunction remains one of the main causes of difficulty or failure in weaning, with a prevalence of around 30%. 17,18 Hence, an early diagnosis of diaphragmatic dysfunction before extubation is imperative to avoid weaning failure.
Point-of-care ultrasonography is emerging as an important bedside tool to enable expeditious decision-making in critically ill patients.Opposed to invasive methods to access diaphragmatic function, US is noninvasive, easily available at the bedside, and allows repeated measurements.It has been reported as an effective method to provide an estimation of respiratory effort during the weaning process in critically ill patients, to predict extubation success and to detect and monitor VIDD. 19However, routine evaluation of the functional status of diaphragm is still poorly applied in daily practice.
There are two diaphragm sonographic predictors of weaning outcome: the diaphragmatic excursion (DE), which measures the distance that the diaphragm is able to move during the respiratory cycle, and the diaphragm thickening fraction (DTF), which reflects variation in the thickness of the diaphragm during a respiratory effort. 20These US measurements can be used to define diaphragmatic dysfunction, although its definition varies widely.
The aims of this narrative review are to summarize the technique of ultrasonography in the evaluation of diaphragmatic function and to assess its utility and accuracy for predicting weaning outcomes in critically ill patients on invasive MV.

Search Strategy
The authors performed a search in PubMed to identify potentially relevant articles, using a preplanned systematic comprehensive and reproducible search strategy with the terms: ("Diaphragmatic ultrasonography" or "Diaphragmatic sonograph*" or "Diaphragm ultrasound" or "Diaphragm ultrasonography" or "Diaphragmatic excursion") combined with ("Ventilator Weaning"[Mesh] or "ventilator weaning" or "extubation success" or "extubation outcome") with no publication data restrictions applied.The search covered all relevant articles published until March 2023.

Study Selection
Title, abstract and full-text articles were screened in a standardized manner to assess their eligibility.The inclusion criteria were: 1) language: articles published in English, Spanish, or Portuguese; 2) type of study: experimental and systematic review articles, published as original studies in peer-reviewed journals, restricted to human studies; 3) population: critically ill adult patients under invasive MV admitted to the ICU and candidates for ventilator weaning; 4) intervention: diaphragm thickness and excursion measured by ultrasound during the weaning process; and 5) predefined outcomes: the accuracy of diaphragm ultrasound to predict weaning outcome.The exclusion criteria were: 1) case reports, opinion papers, editorials; and studies available only as abstract; 2) pediatric studies; 3) studies performed in settings other than critical care (ie, patients ventilated for elective surgery); 4) unusual diaphragm ultrasound methods (ie, indirect measures of diaphragm function).

Data Extraction
After a primary screening of studies, full articles of interest were reviewed and the information to be included in this literature review was extracted.Extracted data included: first author, year of publication, country, population size, ultrasound assessment technique, diaphragm thickness or excursion, and main results.

Ultrasound Assessment for Diaphragmatic Evaluation in ICU Patients
Ultrasonographic examination of the diaphragm can be achieved by two different acoustic windows. 20,21irst, by the subcostal area, between the mid-clavicular and anterior axillary lines, using liver or spleen as acoustic windows.A low frequency (2-5 MHz) ultrasound transducer in the longitudinal plane, can be used to identify diaphragm as a hyperechoic line.The B-mode is initially used to obtain the best approach and select the exploration line, then the M-mode is used to show movements and measure diaphragmatic excursion, that appears in a waveform (Figure 1).Inspiration is identified as an upward curvature of the tracing while, expiration is identified as downward curvature.
The values of diaphragmatic excursion in healthy individuals, performed in spontaneous breathing, were reported to be 1.8 AE 0.3, 7.0 AE 0.6, and 2.9 AE 0.6 cm for males, and 1.6 AE 0.3, 5.7 AE 1.0, and 2.6 AE 0.5 cm for females, during quiet, deep breathing, and voluntary sniffing, respectively. 22he role of excursion in the functional evaluation of diaphragm contractile activity during patienttriggered MV is far less clear.Under MV, excursion measured with M-mode represents the diaphragmatic contraction plus the pressure applied by the ventilator with no distinction between the active (muscle) and passive (ventilator) forces. 23,24In the case of mechanically ventilated patients, evaluation of diaphragmatic motion can be used at the time of weaning from MV during a SBT.Interestingly, the same diaphragmatic excursion values (1.8 cm) were found in ventilated patients who succeeded in the weaning trial. 17he second possible approach is at the zone of apposition of the diaphragm to the rib cage, between the 8th and 10th intercostal space in the anteroaxillary and mid-axillary lines at 0.5-2 cm below the costophrenic sinus.A high-frequency linear transducer (≥10 MHz) should be placed directed perpendicularly to the diaphragm at a depth of 1.5-3 cm.In this area, the diaphragm is observed as a structure made of three distinct layers (Figure 2): a nonechogenic central layer bordered by two echogenic layers, the peritoneum and the diaphragmatic pleurae.
This approach is used to assess thickness of the diaphragm, usually in M-mode (Figure 2C).In spontaneously breathing healthy patients, the normal thickness of the diaphragm is 1.7 AE 0.2 mm increasing to 4.5 AE 0.9 mm, while relaxing and when breath holding at total lung capacity, respectively. 25Diaphragm thickness measured at end inspiration correlates with maximal inspiratory pressure 26 and the change in diaphragm thickness during respiration is strongly related to lung volume. 27n clinical practice, thickening reflects the magnitude of diaphragm effort in spontaneously breathing patients and it can also be used during noninvasive MV (NIMV) and to predict extubation failure or success during a SBT.By measuring the muscle thickness at the end of inspiration (DT-end inspiration) and at the end of expiration (DT-end expiration), the DTF can be calculated as [(DT-end inspiration À DT-end expiration)/(DT-end expiration Â 100)]. 23

Study Identification and Selection
The process of literature search and selection is shown in Fig. 3.The initial database search yielded 61 references.Screening of the titles and abstracts yielded 36 studies, 10 of which were excluded following full text review and the remaining 26 studies were included in the final analysis.

Baseline Characteristics of Included Studies
The 20 original studies and 6 systematic reviews included in this narrative review took place between 2014 and 2023.Most of them were carried out in polyvalent ICUs, but 2 studies 28,29 were performed in respiratory ICUs.
The results are presented in Tables 1 and 2, which summarize the most relevant findings regarding ultrasonographic indices to assess diaphragm contractile function in adults.All original studies included were prospective cohort studies (Table 1) except for one which was retrospective. 43   main findings of the six systematic reviews and metaanalysis included.
Diaphragmatic dysfunction was assessed by DE in six studies, DTF in five and both in nine studies.Seven studies compared diaphragmatic US with other methods that have been previously used to predict successful extubation in the ICU: rapid shallow breathing index (RSBI) 33,34,36,37 and Lung US score 11,34,37 in four and three studies, respectively.
Regarding weaning protocols, all the studies assessed patient readiness to be weaned in order to perform a SBT, which was either performed with low pressure support (inspiratory pressure = 5-8 cmH 2 O and expiratory pressure = 0-5 cmH 2 O) or as T-piece trials.Diaphragm US was mainly performed during a SBT, but three studies also assessed diaphragm function before and after SBT and after extubation. 12,35,40he time interval between diaphragm US and extubation was poorly described in the majority of studies. 3,29,31,33,38In one study, a 24-to 48-hour interval between diaphragm US and extubation was tolerated. 43he definition of "weaning failure" is not standard, covering one or more of the following items in the first 48-72 hours after extubation: need for reintubation, need for non-scheduled post-extubation NIMV, tracheostomy requirement, death, and/or SBT failure.
The minimal duration of MV before inclusion in the studies ranged from 24 to 72 hours.One study included patients with complicated weaning (who had failed previous attempts of SBT) 3 and one study included patients at high risk of extubation failure. 42ome studies focused exclusively on patients with chronic obstructive pulmonary disease (COPD) 5 or COVID-19 41 or patients whose intubation was due to respiratory causes. 28,29,36,39This aspect could compromise the applicability of the results due to patient selection.Exclusion criteria mostly included conditions affecting diaphragm function or assessment.The review data indicates a satisfactory diagnostic accuracy in predicting extubation outcome.Diaphragmatic US is a promising tool for predicting reintubation within 48 hours of extubation.
Llamas-Álvarez et al, 2017 45 Spain 1071 The review detected a significantly higher specificity for DTF and higher sensitivity for DE in studies with applicability concerns.The AUC were 0.87 for DTF.The pooled sensitivity for DE were 75%.Lung and diaphragm US can provide valuable information for predicting weaning outcome, but taken alone, they may not perform as well as individual studies suggest.Its accuracy may vary depending on the patient subpopulation.
Mahmoodpoor et al, 2022 19 Iran 1114 The pooled sensitivities for DTF and DE were 0.89 and 0.79, specificities were 0.81 and 0.69, respectively.The pooled sensitivity and specificity for RSBI were 0.74 and 0.73, respectively.The meta-analysis showed the superiority of the diagnostic accuracy of DTF in comparison to the DE and the RSBI.Among the different US methods for prediction of weaning outcome, DTF seems to be most accurate.
Qian et al, 2018 46 China 436 The DE or DTF had a pooled sensitivity of 0.85 and a pooled specificity of 0.74 for predicting weaning success.Both DE and DTF showed good diagnostic performance to predict weaning outcomes in spite of limitations included high heterogeneity among the studies.DD was found to be a predictor of weaning failure.
Diaphragmatic US could be a useful tool to detect DD, to predict extubation success or failure, to monitor respiratory workload and to assess atrophy in patients under MV.

Predicting Value of DE and DTF on Weaning Outcome
To predict weaning outcome from MV either DE or DTF measurements performed during weaning process or around STB were employed as the test index.Sensitivity and specificity of both indices found in each study are shown in Tables 1 and 2 summarizes the pooled sensitivity and specificity of all studies.
In the studies that used DE ≥10 mm as a cutoff point to predict extubation success, sensitivity ranged from 69 to 97.1% and specificity from 62 to 85%.One study reported a high specificity (92.5%) to predict failure of extubation with a cutoff of DE <10 mm, but with low sensitivity (30%). 12DE was significantly lower in patients who fail compared with the successfully extubated group. 12,46TF was measured in 12 studies.In the studies that used a cutoff point of DTF ≥30%, sensitivity ranged from 68.1 to 94.1% and specificity from 61.5 to 100%.Qian et al further showed that DTF in the successfully weaning group was significantly higher than in failure group. 46ptimal cutoff to diagnose diaphragmatic dysfunction ranged from 10 to 14 mm for DE and from 30 to 36% for DTF during SBT. 47In Eltrabili's study, cutoff value to predict successful weaning was DE >10 mm and DTC >30.7%, with a sensitivity of 94 and 94.1%, a specificity of 85 and 100%, and an area under the curve (AUC) of 0.85 and 0.97, respectively. 33n the systematic reviews, despite the remarkable heterogeneity among the studies, the diagnostic performance, evaluated by the AUC for DE and DTF was 0.82-0.859and 0.82-0.87,respectively, suggesting a high level of overall accuracy diagnostic performance.

Discussion
The diaphragm is a fundamental respiratory muscle whose dysfunction is common in critically ill patients. 48Demoule et al 49 found that VIDD occurs in 64% of the patients on the first day of ICU admission.This was confirmed by Schepens et al 50 that also observed that diaphragm atrophy develops rapidly, within the first 24 hours of MV initiation in adults.
In recent years, ultrasonography emerged as a new method for assessment of diaphragm function, prevailing over other techniques.Diaphragmatic ultrasound is a non-invasive, cost-effective, safe, and easyto-perform technique, thus representing an attractive and suitable diagnostic tool for ICU patients. 16It allows a morphological and functional evaluation of the diaphragm in real time and can be repeated overtime at the bedside. 4redicting the optimal time for extubation is challenging, especially in patients with diaphragm dysfunction that has been increasingly recognized as the primary reason for difficult weaning or weaning failure from MV, and measurement of diaphragm function using US has the potential to predict weaning's outcome from MV. 20 McCool et al showed that incorporating information of diaphragm US into usual ICU care allowed clinicians to identify patients with a normally functioning diaphragm and decreased the time from US to extubation. 39ome recent studies have reported that reduced DE and DTF were associated with weaning difficulties in patients who are mechanically ventilated, 3,11,31 while others found opposite results. 12,41,42The most recent studies suggest that diaphragm US enables the prediction of extubation outcome. 28,32,35,45ccording to some authors, diaphragmatic movement correlates well with transdiaphragmatic pressure.Measurement of the DE could, therefore, be an important tool to evaluate the respiratory endurance of a patient and, by extension, predict successful extubation. 17,35,40,43Furthermore, Flevari et al concluded that this index may also be a reliable tool to assess patients with difficult and prolonged weaning, in whom the diaphragm has some degree of atrophy due to prolonged MV. 3 Kim et al found that DE of <10 mm or paradoxical movement during SBT identified patients at a higher likelihood of extubation failure. 17Similar results were found in the study of Farghaly et al 29 According to this author, a cut-off of >10.5 mm for DE at the time of SBT predicts successful extubation with a sensitivity of 87.5% and a specificity of 71.5%. 29owever, Carrie et al showed discordant results in their study, concluding that, although a decrease in DE values may be associated with an unfavorable weaning outcome, DE measured is not an accurate index by itself to predict weaning failure. 31This study has the particularity of using the maximal and not the mean DE as the US measurement.Nevertheless, mean values of maximal DE were significantly higher in patients who succeeded at their first weaning attempt (4.1 AE 2.1 vs 3 AE 1.8 cm, P = .04).Another study reported that, although DE was not statistically different between the success and failure groups, ΔDE (30-10 minutes during SBT) was higher in failure group than in the success group (1.07 AE 0.64 mm vs 3.33 AE 3.17 mm, P < .05). 5 The difference can be attributed to the timing of the US measurements (DE at 0, 10, and 30 minutes after initiation of SBT).
On the contrary, some available data suggest a lower sensitivity and specificity for DE as compared with the DTF in predicting weaning outcome. 4mbrello et al believe that DTF rather than DE is a reliable index of respiratory effort and active contraction of the diaphragm during MV, and reported a significantly higher DTF in the weaning success group, compared with the failure group. 21This study evaluated both indices during assisted breathing and concluded that DE should be limited to patients on SBT.DiNino et al found that DTF ≥30% had a PPV for extubation success of 91% in patients undergoing SBT with low levels of pressure support (PS of Δ5/5). 32McCool et al reported similar results for PPV, NPV, and AUC for DTF ≥30%. 39By contrast, Vivier et al found that TDF was not useful to distinguish between patients who were and were not successfully extubated. 42ifferent from the prior studies, they included patients under prolonged MV (at least 1 week), older patients (aged >65 years) and at high risk for re-intubation.Furthermore, they studied both hemidiaphragms and found unilateral dysfunction in 140 of 160 patients. 40ccording to three systematic reviews DTF is more accurate than DE to predict weaning outcome. 4,23,42lamas-Álvarez et al, based on 19 studies, showed a significantly higher specificity for DTF and higher sensitivity for DE. 45 Recently, Mahmoodpoor et al also showed a higher diagnostic accuracy of DTF compared with DE and RSBI. 19owever, some studies found that DE has higher sensitivity and specificity than DTF. 28,30,43This can be explained by the ventilator mode at the timing of US since, in these studies, measurements were made during a SBT without positive-pressure ventilation.The data obtained from measurements made during positive-pressure ventilation would affect the measured DE that is derived from adding the patient's effort to the pressure generated by the ventilator.Therefore, DTF is suitable to estimate the diaphragm function in patients under MV, while DE should be reserved to cases in the absence of the respiratory support, as the downward displacement of the muscle may reflect passive insufflation by the ventilator.
Nevertheless, two studies demonstrated that both indices are useful to predict successful extubation. 4,34eaning is also affected by non-diaphragm-related factors.Le Neindre et al concluded that diaphragm US predicts extubation failure with high specificity (0.84 for DTF and 0.82 for DE, respectively). 44owever, sensitivity was low (0.70 for DTF and 0.71 for DE) because absence of diaphragm dysfunction does not imply no risk of extubation failure.Therefore, a single diaphragmatic index may not be a perfect predictor.This supports several studies that emphasize the interest of combining diaphragmatic US with other traditional parameters to predict weaning outcome. 11,34,37

Conclusions
Diaphragm US is a novel method for measuring diaphragmatic function in mechanically ventilated patients and an attracting and a promising tool to predict weaning outcome.
The increased routine use of ultrasonography in the ICUs as a fast, inexpensive and noninvasive test is expected to lead to a timely identification of critically ill patients at risk of weaning failure.
Both ultrasonography indices, DE and DTF, showed a good predictive ability for successful liberation from MV in different populations, however DTF seems to be most accurate method to estimate the diaphragm function in patients undergoing MV, while DE should be reserved to cases in the absence of the breathing support (T-piece or low PS).Optimal cutoffs ranged from 1.0 to 1.4 cm for DE and 30-36% for DTF.
This technique may be a reliable and helpful tool to predict extubation outcome, however due to significant heterogeneities among studies, clinicians should be aware of its utility and limitations.
Additional high-quality randomized controlled trials are needed to standardize sonographic diagnostic criteria for diaphragmatic dysfunction and its use in daily clinical decisions in the weaning process.

Figure 1 .
Figure 1.Ultrasonographic assessment of diaphragm displacement.A, Ultrasonographic view of the normal diaphragm in the region of the liver dome, with B-mode in the upper part and M-mode in the lower part.B, Anatomical structures that can be identified in B-mode scanning.C, Anatomical structures that can be identified in M-mode scanning.D, Probe placement to explore the diaphragm in the region of the liver dome.[Image reproduced with permission of the rights holder.]

Figure 2 .
Figure 2. Ultrasonographic assessment of diaphragm thickness.A, Ultrasonographic view of the normal diaphragm in the zone of apposition, with B-mode in the upper part and M-mode in the lower part.B, Anatomical structures that can be identified in B-mode scanning.C, Anatomical structures that can be identified in M-mode scanning.D, Probe placement to explore the diaphragm in the zone of apposition.The distance identified by plus signs 1 in A and C is end-inspiratory thickness, whereas the distance between plus signs 2 in the same panels is the end-expiratory thickness.[Image reproduced with permission of the rights holder.]

Figure 3 .
Figure 3. PRISMA flow diagram for study identification and selection with reasons for exclusion.

Table 1 .
Characteristics of Included Studies

Table 1 .
Continued ΔDE 30-5 , the variation of DE between 5 and 30 minutes of SBT; AC, assist control mode of MV; AUC, area under the ROC curve; COPD, chronic obstructive pulmonary disease; DE, diaphragm excursion; DTF, diaphragm thickening fraction; DT ins , diaphragmatic thickness at end inspiration; FG, failure group extubation; LD, left diaphragm; LUS, lung ultrasound score; MV, mechanical ventilation; NPV, negative predictive value; NR, not reported; PPV, positive predictive value; PS, pressure support trial; RD, right diaphragm; RSBI, rapid shallow breathing index; SBT, spontaneous breathing trial; SG, success group extubation; TPIA, time to peak inspiratory amplitude diaphragm; TT, T-piece; US, ultrasound.a Age is expressed according to data extracted from each study as mean AE standard deviation or median (interquartile range).

Table 2 .
Characteristics of Included Systematic Reviews and Meta-Analysis