Liberation from invasive mechanical ventilation is a pivotal phase in the management of critically ill patients. Both premature and unnecessarily delayed extubation have important consequences. Failed extubation may necessitate reintubation and is associated with prolonged intensive care unit (ICU) treatment and increased morbidity, whereas unnecessary continuation of invasive ventilation exposes patients to ventilator-associated complications, sedation, immobility, and respiratory-muscle deconditioning [1,2]. Consequently, identifying physiological readiness for sustained spontaneous breathing remains a fundamental component of contemporary critical care.

Current liberation strategies incorporate clinical improvement, adequate gas exchange, hemodynamic stability, assessment of airway protection, and performance during a spontaneous breathing trial (SBT) [1,2]. Conventional physiological indices, particularly the rapid shallow breathing index (RSBI), provide useful information but cannot completely characterize the complex interaction between respiratory drive, respiratory-muscle capacity, pulmonary mechanics, cardiac function, and airway competence that ultimately determines extubation outcome [3]. A patient may therefore successfully complete an SBT yet subsequently require reintubation.

The diaphragm is the principal inspiratory muscle and is particularly vulnerable during critical illness. Mechanical ventilation itself may contribute to rapid alterations in diaphragm structure and contractile activity. Goligher et al. demonstrated measurable changes in diaphragm thickness during mechanical ventilation, with more than 10% reduction in thickness occurring in 44% of their cohort during the first week of ventilation [4]. Earlier work similarly demonstrated that ultrasonographically identified diaphragm dysfunction was associated with substantially longer weaning and ventilation durations and a higher frequency of weaning failure [5].

Diaphragm ultrasound (DUS) has consequently emerged as a bedside technique capable of directly assessing respiratory-muscle structure and function without ionizing radiation or patient transportation. The principal parameters include diaphragmatic excursion (DE), diaphragm thickness at end-expiration and end-inspiration, and diaphragmatic thickening fraction (DTF). Initial prospective investigations demonstrated that DTF values around 30% could identify patients more likely to remain successfully extubated [6]. Subsequent studies evaluated DE, DTF and combinations of ultrasound measurements with conventional weaning indices, although diagnostic thresholds and performance have varied between populations and ventilatory conditions [7,8].

Systematic reviews subsequently strengthened the evidence supporting DUS. Meta-analytic assessments indicate that both DE and DTF possess clinically meaningful diagnostic performance for predicting successful liberation from mechanical ventilation, while also identifying substantial between-study heterogeneity [9,10]. These observations have shifted the clinical question from whether diaphragm ultrasonography contains prognostic information to how it should be standardized and integrated into multimodal extubation assessment.

The present review therefore evaluates current evidence regarding DUS for prediction of extubation and weaning outcomes in mechanically ventilated patients, summarizes the diagnostic performance of major sonographic parameters, examines its clinical utility relative to conventional indices, and identifies methodological priorities for future research.

Study Design A structured review of the available clinical evidence was undertaken to evaluate the diagnostic and prognostic utility of diaphragm ultrasonography in predicting successful liberation from invasive mechanical ventilation. The review focused on the relationship between ultrasound-derived measurements of diaphragmatic structure and function and clinically relevant weaning or extubation outcomes. Literature Search The literature search was constructed around PubMed/MEDLINE and Embase-indexed literature. Combinations of Medical Subject Headings and free-text terms were used, including “diaphragm ultrasound,” “diaphragmatic ultrasonography,” “diaphragmatic excursion,” “diaphragm thickening fraction,” “mechanical ventilation,” “ventilator weaning,” “spontaneous breathing trial,” “extubation,” “extubation failure,” “reintubation,” and “rapid shallow breathing index.” Reference lists of relevant systematic reviews and meta-analyses were additionally examined to identify influential primary studies. Eligibility Criteria Studies were considered eligible when they: 1. evaluated adult patients receiving invasive mechanical ventilation; 2. assessed diaphragm function using bedside ultrasonography during mechanical ventilation, an SBT, or immediately before planned extubation; 3. reported at least one quantitative diaphragm parameter, including DE, end-expiratory diaphragm thickness, inspiratory diaphragm thickness, DTF, or a diaphragm-derived composite index; 4. reported weaning success, extubation success, weaning failure, extubation failure, reintubation, or another directly related ventilator-liberation outcome; and 5. were available as peer-reviewed original investigations, randomized clinical trials, systematic reviews, or meta-analyses. Case reports, conference abstracts without adequate numerical information, animal studies, and reports without clinically interpretable diaphragm-ultrasound measurements were excluded from the principal evidence synthesis. Ultrasound Parameters The two principal parameters assessed were diaphragmatic excursion (DE) and diaphragmatic thickening fraction (DTF). DE represents the displacement of the diaphragm during inspiration and is generally measured using M-mode ultrasonography through a subcostal acoustic window. DTF represents the relative increase in diaphragm thickness during inspiration and was interpreted according to: DTF (%) = [(Thickness at end-inspiration − Thickness at end-expiration) / Thickness at end-expiration] × 100 Where reported, composite parameters incorporating respiratory frequency and ultrasound measurements, including diaphragmatic rapid shallow breathing indices, were also evaluated. Outcomes The primary outcome was the ability of diaphragm-ultrasound measurements to predict successful liberation from mechanical ventilation or successful extubation. Secondary outcomes included extubation failure, reintubation within 48–72 hours, weaning failure, diagnostic sensitivity and specificity, positive and negative predictive performance, receiver-operating-characteristic area under the curve (AUC), and comparative performance against conventional indices such as RSBI. Because definitions of successful weaning and extubation varied among published studies, the outcome definition employed by each individual investigation was retained when interpreting its findings. Evidence Extraction and Synthesis For eligible key studies, data were extracted regarding study design, sample size, clinical population, ultrasound parameter, measurement conditions, proposed threshold, extubation or weaning outcome, sensitivity, specificity, and AUC where available. Results were synthesized according to four clinically relevant domains: (1) characteristics and outcomes of representative primary studies; (2) performance of diaphragmatic excursion; (3) performance of diaphragm thickening fraction and ultrasound-derived indices; and (4) findings from systematic reviews, meta-analyses, and interventional evidence. Numerical estimates reported in the Results tables were retained from their corresponding published investigations rather than recalculated across heterogeneous populations. Particular attention was given to differences in SBT methodology, ventilatory support during ultrasound acquisition, patient positioning, ultrasound technique, outcome definition, and timing of assessment because these factors may influence diagnostic thresholds. Clinical Interpretation Diagnostic performance was interpreted within the broader clinical context of ventilator liberation. DUS was considered an adjunctive physiological assessment rather than a stand-alone indication for extubation. Accordingly, ultrasound findings were interpreted alongside SBT performance, gas exchange, hemodynamic stability, neurological status, cough effectiveness, secretion burden, airway patency, and conventional clinical assessment.

The available evidence consistently demonstrated an association between preserved diaphragmatic function and successful liberation from mechanical ventilation. Across individual prospective studies, diaphragmatic excursion (DE) and diaphragmatic thickening fraction (DTF) were generally greater among patients who were successfully weaned or extubated. However, considerable variability was observed in proposed diagnostic thresholds, reflecting differences in patient populations, spontaneous breathing trial (SBT) methodology, ventilatory support, ultrasound technique, and definitions of weaning or extubation success.

Evidence from Representative Clinical Studies

Representative prospective studies demonstrated successful liberation rates ranging from approximately 70% to 86%. Farghaly and Hasan evaluated 54 patients who had successfully completed an SBT and observed extubation failure in 14 patients (25.9%). DE, inspiratory and expiratory diaphragm thickness, and DTF were significantly greater among successfully extubated patients. A DE threshold of ≥10.5 mm demonstrated 87.5% sensitivity and 71.5% specificity, whereas DTF ≥34.2% provided 90.0% sensitivity and 64.3% specificity.

In a larger prospective cohort of 200 mechanically ventilated patients, 171 (85.5%) were successfully weaned and 29 (14.5%) experienced failure. DE >1.21 cm showed particularly favorable discrimination, with 94% sensitivity, 71% specificity, and an AUROC of 0.809. DTF >37% was less discriminative, with 80% sensitivity and 52% specificity.

Table 1. Representative clinical studies evaluating diaphragm ultrasound for prediction of ventilator liberation

*Study identification and complete bibliographic information are provided in the final reference list.

Overall, these findings indicate that diaphragm ultrasound identifies clinically meaningful differences between patients with successful and unsuccessful liberation. Nevertheless, the range of optimal thresholds—from approximately 10.5 to 12.5 mm for DE and approximately 24% to 37% for DTF—argues against applying a single universal cutoff across all ICU populations.

Diagnostic Performance of Diaphragmatic Excursion

DE was among the most consistently evaluated ultrasound parameters. Most studies demonstrated better diaphragmatic displacement among successfully liberated patients. Thresholds around 1.0–1.3 cm were commonly associated with favorable outcomes, although the precise threshold depended on measurement conditions.

The prospective study involving 200 patients demonstrated an AUROC of 0.809 for DE, compared with only 0.422 for conventional RSBI in that cohort. Using a DE cutoff >1.21 cm yielded sensitivity of 94% and specificity of 71%, suggesting that preserved excursion may be particularly useful for identifying patients physiologically capable of sustaining spontaneous ventilation.

Another prospective study involving 62 patients reported that DE at a cutoff of 1.25 cm predicted successful weaning with sensitivity of 97.1% and specificity of 82.1%. Similarly, Farghaly and Hasan reported sensitivity of 87.5% and specificity of 71.5% at a DE threshold ≥10.5 mm.

Table 2. Diagnostic performance of diaphragmatic excursion

The strongest aggregate evidence was provided by the 2023 systematic review and meta-analysis. Nineteen studies involving 1,204 patients contributed to quantitative synthesis. For DE, pooled sensitivity was 0.80 (95% CI: 0.77–0.83), pooled specificity was 0.80 (95% CI: 0.75–0.84), and the area under the summary receiver-operating-characteristic curve was 0.87. The diagnostic odds ratio (DOR) was 17.1 (95% CI: 10.2–28.6).

These pooled results support moderate-to-good discriminatory performance for DE but also indicate that approximately one-fifth of outcomes may be incorrectly classified if DE is used independently. Therefore, DE appears more appropriate as an adjunct to comprehensive extubation assessment than as an isolated decision threshold.

Diagnostic Performance of Diaphragmatic Thickening Fraction

DTF reflects active diaphragmatic contraction and is calculated from the proportional change in diaphragm thickness between expiration and inspiration. Several investigations found significantly higher DTF among patients successfully liberated from mechanical ventilation.

Although early clinical studies frequently proposed values around 30% as clinically useful thresholds, subsequent studies demonstrated considerable variation. Farghaly and Hasan reported that DTF ≥34.2% predicted successful extubation with 90% sensitivity and 64.3% specificity. In the 200-patient prospective study, DTF >37% demonstrated 80% sensitivity but only 52% specificity. Conversely, another prospective study identified an approximately 24% threshold with sensitivity of 93.5% and specificity of 94.7%.

Table 3. Diagnostic performance of diaphragmatic thickening fraction and conventional indices

In the 2023 meta-analysis, DTF demonstrated pooled sensitivity of 0.85 (95% CI: 0.82–0.87) and specificity of 0.75 (95% CI: 0.69–0.80). The summary AUROC was 0.87 and the DOR was 17.2 (95% CI: 9.16–32.3). After studies employing atypical cutoff values were excluded, sensitivity increased to 0.86 and specificity to 0.78, with the summary AUROC increasing to 0.90.

Thus, DTF appears to have somewhat greater sensitivity than DE in pooled analyses, whereas DE demonstrates slightly higher specificity. These complementary characteristics provide a rationale for evaluating the two parameters together rather than treating them as competing measurements.

Comparison with Conventional Weaning Parameters and Pooled Evidence

A major potential advantage of DUS is its ability to provide direct information regarding respiratory-muscle performance rather than relying exclusively on global breathing-pattern indices. Conventional RSBI remains clinically useful, but its discriminatory performance has been inconsistent in contemporary cohorts.

Farghaly and Hasan reported that RSBI <105 had sensitivity of 90% but specificity of only 18.7% for successful extubation. By comparison, DE ≥10.5 mm provided substantially higher specificity of 71.5%. Combining DE ≥10.5 mm with inspiratory diaphragm thickness ≥21 mm increased specificity to 100%, although sensitivity decreased to 64.9%.

Similarly, in the 200-patient prospective cohort, RSBI had an AUROC of 0.422 compared with 0.809 for DE and 0.654 for DTF. The findings illustrate the potential incremental value of direct diaphragmatic assessment, particularly when traditional indices produce equivocal results.

Table 4. Summary of pooled and comparative diagnostic evidence

The 2023 meta-analysis included 26 studies in qualitative synthesis and 19 studies comprising 1,204 patients in quantitative analysis. Of these, 908 patients contributed DE measurements and 945 contributed DTF measurements, with several studies evaluating both parameters.

Despite encouraging pooled accuracy, substantial methodological heterogeneity was evident. Patient position during ultrasound examination was identified as one contributor to heterogeneity. Differences in pressure-support versus T-tube SBTs did not produce significant differences in pooled sensitivity and specificity. Exclusion of atypical thresholds improved overall diagnostic performance, particularly for DTF.

An earlier systematic review and meta-analysis of diaphragm and lung ultrasound involving 19 studies and 1,071 participants similarly reported an area under the summary ROC curve of 0.87 for DTF and a diagnostic odds ratio of 21 (95% CI: 11–40). For DE, pooled sensitivity and specificity were both approximately 75%, with a diagnostic odds ratio of 10 (95% CI: 4–24).

Collectively, the available evidence demonstrates that DUS provides clinically relevant prognostic information during ventilator liberation. DTF may offer somewhat greater sensitivity for identifying patients likely to succeed, whereas DE provides comparable overall discrimination with potentially higher specificity in pooled analyses. The variability in thresholds across studies remains an important limitation and supports integration of DUS with SBT performance, respiratory pattern, airway assessment, neurological status, secretion burden, and other clinical determinants rather than reliance on a single sonographic cutoff.

Successful liberation from mechanical ventilation requires an appropriate balance between respiratory load and the patient’s capacity to sustain spontaneous ventilation. The present review demonstrates that diaphragm ultrasonography provides clinically relevant information regarding this balance and may complement conventional assessment of readiness for extubation. Across the available evidence, both diaphragmatic excursion (DE) and diaphragmatic thickening fraction (DTF) show useful diagnostic performance, although considerable heterogeneity in measurement technique and threshold selection remains an important barrier to universal implementation [11,12]. One of the major advantages of diaphragm ultrasound is its ability to assess respiratory-muscle function directly at the bedside. Conventional parameters such as respiratory rate, tidal volume and rapid shallow breathing index (RSBI) reflect the overall breathing pattern but do not specifically identify diaphragmatic dysfunction. This distinction is important because respiratory failure following extubation may occur despite apparently satisfactory conventional indices. Farghaly and Hasan demonstrated this limitation clearly: RSBI <105 had 90% sensitivity but only 18.7% specificity for successful extubation, whereas DE ≥10.5 mm provided 87.5% sensitivity and 71.5% specificity [13]. Combining diaphragm measurements further improved specificity, suggesting that ultrasound may provide information not captured by RSBI alone. The pooled evidence nevertheless requires careful interpretation. Parada-Gereda et al. analyzed 19 studies comprising 1,204 patients and reported an AUSROC of 0.87 for both DE and DTF. DE demonstrated pooled sensitivity and specificity of 80% and 80%, respectively, whereas DTF demonstrated sensitivity of 85% and specificity of 75% [11]. These results support good overall discrimination but are insufficient to justify ultrasound as an isolated extubation criterion. Earlier meta-analytic evidence similarly demonstrated clinically useful but imperfect diagnostic performance [12,14]. The inability of diaphragm ultrasound to predict every extubation failure is physiologically understandable. Successful extubation is not determined exclusively by diaphragmatic contractility. Upper-airway obstruction, impaired consciousness, ineffective cough, excessive respiratory secretions, cardiac dysfunction, unresolved pulmonary disease and neurological impairment may all cause extubation failure despite preserved diaphragm function. A 2021 diagnostic meta-analysis found moderate-to-high specificity but lower sensitivity and emphasized that absence of diaphragm dysfunction does not exclude extubation failure [12]. Accordingly, diaphragm ultrasound should be regarded as one component of a multidimensional liberation assessment. DTF has particular physiological appeal because it reflects inspiratory muscle contraction rather than diaphragm displacement alone. DiNino et al. demonstrated that diaphragm thickening measured by ultrasound could predict extubation success, helping establish DTF as a practical bedside marker [15]. Subsequent investigations have reported useful DTF thresholds ranging broadly from approximately 25% to 35% or higher. Samanta et al., for example, observed that DTF values around 25% predicted simple weaning with ROC AUC values ≥0.90 under several measurement conditions [16]. This variability emphasizes that a rigid universal cutoff should be avoided. DE is technically straightforward and has also demonstrated substantial prognostic utility. However, excursion may be influenced by ventilator assistance, breathing pattern, abdominal pressure and patient positioning. Moreover, diaphragmatic displacement does not necessarily represent active muscle contraction when positive-pressure assistance contributes to inspiratory movement. Consequently, measurement conditions should always accompany interpretation of DE. The clinical context may also modify diagnostic performance. In patients with abdominal sepsis, Theerawit et al. reported that DTF ≥30.7% predicted successful liberation with 94.1% sensitivity and 100% specificity, while DE ≥10.4 mm demonstrated 94% sensitivity and 85% specificity [17]. Such excellent performance in a selected population illustrates both the potential value of DUS and the difficulty of extrapolating individual thresholds to heterogeneous ICU populations. Measurement standardization is therefore central to future clinical adoption. A systematic methodological evaluation found substantial variation in ultrasound acquisition and analysis despite generally favorable measurement reliability. Diaphragm thickness demonstrated particularly high reliability, with reported inter-/intrarater performance reaching an ICC of approximately 0.98 in pooled methodological assessment [18]. Standardized definitions of anatomical landmarks, patient position, probe orientation, respiratory phase, ventilatory assistance and timing relative to the SBT would substantially improve comparability among studies. Clinical Utility In clinical practice, DUS may be most useful when conventional assessment produces uncertainty. A patient who successfully completes an SBT but has markedly reduced DTF or DE may warrant closer evaluation for respiratory-muscle weakness, whereas preserved diaphragm function may redirect attention toward pulmonary, cardiac, neurological or airway-related causes of difficult liberation. Serial assessment represents another potential application. Rather than relying solely on a single pre-extubation measurement, repeated ultrasound examinations can characterize changes in diaphragm thickness and function during mechanical ventilation. This is relevant because prolonged ventilation may be associated with diaphragm atrophy and altered contractile activity. Recognition of evolving dysfunction may permit earlier modification of ventilatory support, mobilization, nutritional strategies and respiratory-muscle rehabilitation. The practical advantages of ultrasound are considerable. It is portable, repeatable, radiation-free and can be performed at the bedside without interrupting mechanical ventilation. These features make it particularly attractive for critically ill patients who cannot be transported easily. Future Directions Future research should move beyond establishing whether DUS predicts extubation outcomes and determine how best to integrate it into decision-making algorithms. Large multicenter prospective studies should employ standardized ultrasound protocols and uniform definitions of weaning and extubation failure. Investigations should also evaluate predefined subgroups, including patients with sepsis, chronic obstructive pulmonary disease, obesity, prolonged mechanical ventilation and neurological disease. Integration of diaphragm measurements with lung ultrasound, cardiac ultrasound and conventional physiological variables may provide better discrimination than any single parameter. Composite models incorporating DTF, DE, lung aeration, RSBI, cough effectiveness and clinical variables are therefore a logical direction for future investigation. Automated image analysis and artificial intelligence may further reduce operator dependence by facilitating diaphragm border identification, measurement of thickness and excursion, and longitudinal comparison. However, such systems require external validation before routine clinical implementation. Importantly, recent meta-analytic evidence continues to identify substantial heterogeneity. Even where pooled diagnostic performance appears favorable, variability in cutoffs, patient positioning, SBT technique and outcome definitions limits direct translation into a universal bedside rule [11,12,19]. Future trials should therefore determine whether ultrasound-guided liberation strategies actually reduce reintubation, ventilation duration, ICU stay or mortality rather than focusing exclusively on diagnostic accuracy. LIMITATIONS Several limitations of the available evidence should be recognized. Most primary investigations are single-center observational studies with relatively small sample sizes. Definitions of weaning success, extubation success and failure vary between studies, as do ultrasound timing and ventilatory conditions during measurement. DE and DTF cutoff values are consequently heterogeneous. Ultrasound is also operator dependent, and inadequate visualization may occur in patients with obesity, surgical wounds, dressings or unfavorable anatomy. Finally, extubation is a multifactorial process; diaphragm performance cannot account for airway protection, secretion clearance, neurological status, cardiac reserve and all pulmonary causes of post-extubation deterioration [11,12,18,20].

Diaphragm ultrasound is a valuable non-invasive bedside technique for assessment of respiratory-muscle function during liberation from mechanical ventilation. Both diaphragmatic excursion and diaphragmatic thickening fraction demonstrate clinically useful predictive accuracy for successful weaning and extubation, with pooled evidence suggesting an overall area under the ROC curve of approximately 0.87. Nevertheless, substantial heterogeneity in measurement protocols, thresholds and patient populations prevents adoption of a single universal cutoff. DUS should therefore complement rather than replace spontaneous breathing trials and comprehensive clinical assessment. Its greatest potential may lie in identifying occult diaphragm dysfunction, evaluating patients with difficult or prolonged weaning, and contributing to multimodal prediction models. Standardized acquisition protocols and adequately powered multicenter studies are required to establish whether ultrasound-guided extubation strategies improve patient-centered clinical outcomes.

DECLARATIONS

Ethical approval: Not applicable, as this article is based on previously published literature and involved no direct recruitment of human participants.

Informed consent: Not applicable.

Funding: No specific funding was received for this work.

Conflict of interest: The authors declare no conflicts of interest.

Data availability: All data discussed in this review were obtained from published literature.

Author contributions: All authors contributed to the conceptualization, literature evaluation, interpretation of evidence, drafting, and critical revision of the manuscript. All authors reviewed and approved the final manuscript.