The diaphragm is the primary muscle of respiration. It is a thin, dome-shaped muscle, comprising a costal and crural portion and a non-contracting central tendon. Contraction of the diaphragm, mediated by the phrenic nerve, generates the negative pressure that drives respiration. This contraction leads to caudal movement of the diaphragm’s domes while causing shortening and thickening at the zone of apposition (ZOA) where the diaphragm attaches to the chest wall and rib cage [1]. Despite its essential role in respiration, the diaphragm’s contribution to respiratory disease is often overlooked. Due to the technical and pragmatic issues related to traditional diaphragm imaging, clinical evaluation of diaphragmatic function is limited. Static imaging, such as computed tomography or x-ray, can demonstrate bilateral or hemi-diaphragm elevation but cannot provide insight into dynamic function. Fluoroscopy, long considered the gold standard to clinically assess diaphragmatic function, has several limitations including poor specificity (given the volitional component), limited ability to detect unilateral weakness (as opposed to paralysis), and limited ability to detect bilateral involvement [2]. Also, fluoroscopic sniff testing is very challenging to perform or interpret in patients who remain dependent on low level of positive pressure ventilation.

Point of care ultrasound (POCUS) has grown in popularity due to its expanding applications across multiple specialties. DU is a subset of POCUS that has been an area of active study as a tool to safely, reliably, and quickly characterize diaphragm function. Due to its availability, reproducibility, and safety, POCUS naturally overcomes many of the limitations of traditional modes of diaphragm imaging. Various measures to evaluate diaphragmatic function via ultrasound have been developed, allowing for a more sophisticated interpretation of diaphragmatic function. The most studied methods of ultrasonographic assessment of diaphragmatic function are excursion and thickening fraction.

Studies have demonstrated that DU competency is obtainable with reasonable effort, even for those with limited ultrasound experience. Acceptable competence has been reported in as few as 10 repetitions when under expert guidance [3,4,5]. Excellent inter- and intra-observer variability have been observed among both novice and experienced sonologists (Tables 1 and 2). When compared to gold-standard fluoroscopic measurements of excursion, DU has demonstrated high correlation, statistically non-significant differences, and in some cases, superiority [6,7,8]. Correlation with phrenic nerve stimulation has also been found [9]. With these technical and pragmatic aspects in mind, DU represents an attractive alternative and arguably a replacement of fluoroscopy for routine diaphragm assessment.

The accessibility of DU has allowed for evaluation of the diaphragm in clinical scenarios where it has not previously been feasible. In the inpatient setting, DU can be utilized to evaluate diaphragm dysfunction for patients with acute respiratory failure; its presence in the context of other respiratory illnesses may have prognostic value, including associations with non-invasive ventilation failure, extubation success, length of stay, and mortality [10,11,12,13]. In the outpatient setting, patients with chronic respiratory diseases such as asthma, chronic obstructive pulmonary disease, and interstitial lung disease have been found to have significant diaphragmatic dysfunction compared to controls, with DU demonstrating correlation with spirometry [14,15,16]. Patients with obstructive sleep apnea have been found to have differences in diaphragm thickness and DU may have utility as a screening modality [17].

DU is an emerging skill which has the potential to be integrated into a multitude of clinical settings. This review will focus on the practical aspects of performing this modality and an updated exploration of its use in weaning from mechanical ventilation, where its utility is becoming increasingly established and routine. As the use of POCUS and specifically DU becomes more widespread, it is important for clinicians, especially those treating respiratory disorders, to understand this tool and its applications.

DU is performed bilaterally in two locations: the subcostal space and the lateral chest wall. A subcostal approach is utilized to visualize the dome of the diaphragm, where cranio-caudal motion is most pronounced, and to measure diaphragmatic excursion. Visualization of the right hemidiaphragm is nearly always successful due to the presence of the liver, which is an excellent acoustic window. On the left, the spleen serves as a smaller acoustic window, and the presence of gastric structures may limit visualization. Imaging of one hemidiaphragm will often suffice unless unilateral dysfunction is suspected. A lateral chest wall approach is utilized to measure diaphragmatic muscle thickness and thickening fraction at the ZOA, where the diaphragm attaches to the rib cage.

A curvilinear or phased array probe may be utilized for measurement of diaphragmatic excursion using a subcostal approach. The probe is positioned below the rib cage in the mid-clavicular to anterior-axillary line with the probe marker oriented to the patient’s right side. The dome of the diaphragm will appear as a bright, thick, and curved hyperechoic line deep to the liver/spleen. It will move towards the probe on inspiration and away from the probe on expiration. Once visualization of the diaphragm is confirmed, diaphragmatic excursion can be interrogated in B-mode or M-mode, with the M-mode line positioned perpendicular to the diaphragm, where the most pronounced cranio-caudal motion is visualized. Markers are positioned to measure the vertical distance between the points of maximal cranial excursion (peak) and maximum caudal excursion (trough). The distance is measured from leading edge to leading edge with markers at the superior aspect of the tracing at both the peak and trough. (Fig. 1, Supplement 1).

Normal diaphragm excursion measured via subcostal approach in M-mode. Marker placed to measure distance from maximal cranial excursion to maximal caudal excursion. Image obtained by Dr. Cameron Baston

A high frequency probe may be utilized for measurement of diaphragmatic thickening at the ZOA using a lateral approach. The probe is placed at the anterior to mid-axillary line between the 8th and 11th intercostal space (Fig. 2). The probe marker is oriented cranially, and the beam is directed perpendicular to the chest wall. At the ZOA, the diaphragm appears as a three-layered structure, with two thick hyperechoic lines representing pleural and peritoneal linings bordering a hypoechoic muscular layer. A thin, hyperechoic line within the hypoechoic muscular layer represents the fibrous layer of the diaphragm. No current consensus exists regarding where along the ZOA thickness should be measured [18]. However, in our practice, the most pronounced thickening is often visualized at the most cranial portion of the ZOA. Diaphragm thickening can also be measured in both B- and M-modes. Using M-mode can improve accuracy for less experienced sonologists [o5]. Maximum inspiration is identified by the point of maximal thickness of the hypoechoic muscular layer, while maximum expiration is identified by the thinnest point during the respiratory cycle. When M-mode is employed, the M-mode line is positioned perpendicular to the diaphragm’s fibers at the point where the most dramatic change in thickness is seen. In both M-mode and B-mode, the perpendicular distance between the pleural and peritoneal linings is measured (Fig. 3, Supplement 2).

Probe Positioning showing the probe held at the Zone of Apposition on the patient’s right side. Image obtained by Dr. Cameron Baston

Normal diaphragm thickening measured at the zone of apposition in M-mode. Diaphragm thickening fraction (DTF) = (thickness at end inspiration—thickness at end expiration)/thickness at end expiration × 100. In this image, DTF = (0.56—0.38)/0.38 × 100 = 47%. Image obtained by Dr. Cameron Baston

Diaphragmatic dysfunction may be considered in patients with signs or symptoms of respiratory muscle weakness; symptoms that are seen particularly in patients with diaphragm weakness include orthopnea and dyspnea on exertion [19, 20]. Individuals with unilateral diaphragmatic paralysis may be asymptomatic but can manifest with respiratory symptoms, especially if they have a concomitant disease impacting their respiratory function. Pulmonary function tests may be revealing for a pattern of restriction on spirometry (decreased forced vital capacity [FVC] and total lung capacity) with normal forced expiratory volume in one second. A measurement of FVC and vital capacity (VC) in both upright and supine position is helpful in detecting diaphragm weakness specifically: in the absence of normal diaphragmatic strength, there is unopposed displacement of abdominal contents into the thoracic cavity that result in a drop in measured VC [21, 22]. A reduction in VC of 10–20% in the supine position for those with unilateral diaphragm weakness and a reduction in VC of 30–50% in the supine position for those with bilateral diaphragm weakness has been demonstrated [23].

While pulmonary function testing can be helpful in spontaneously breathing patients, it is less applicable in evaluating diaphragm function for patients who require mechanical ventilatory support. Ventilator-induced diaphragm dysfunction is a cause of diaphragm weakness in ventilated patients, with diaphragm atrophy occurring even after just one day of mechanical ventilation. One study that demonstrated a decrease in diaphragm thickness of approximately 6% per day of mechanical ventilation [24,25,26]. This phenomenon is felt to be related to diaphragm muscle inactivity during invasive mechanical ventilation, leading to muscle fiber injury and muscle atrophy [27, 28]. For these patients, diaphragm ultrasound can be both a feasible and reliable bedside method of detecting diaphragm weakness. Various measurements of diaphragm function (including diaphragm thickening fraction [DTF] and diaphragmatic excursion [DE]) obtained via ultrasonographic assessment have been demonstrated to be useful in predicting successful weaning from mechanical ventilation [4, 29,30,31,32].

Both DTF and DE have been studied as methods to detect diaphragm weakness. The equation for DTF is as follows: \(DTF=\left(\frac}_}-}_}}}_}}\right)\) [29, 31, 33]. The resultant number is a percentage; a DTF < 20% suggests diaphragm paralysis (Fig. 4). DTF of greater than or equal to 30% has been shown to predict success in weaning from mechanical ventilation with sensitivity of 88% and specificity of 71% [29]. DE has also been shown to be effective in predicting successful extubation with a cutoff of 10 mm excursion demonstrating sensitivity of approximately 79% and specificity of approximately 71% [30]. DE should not be used in isolation, especially in mechanically ventilated patients as there will be excursion purely due to receiving positive pressure ventilation.

Diaphragm paralysis measured at the zone of apposition in M-mode. Diaphragm thickening fraction (DTF) = (thickness at end inspiration—thickness at end expiration)/thickness at end expiration × 100. In this image, DTF = (0.39–0.36)/0.36 × 100 = 8%. Image obtained by Dr. Cameron Baston

Investigations into application of DU in ventilator weaning have expanded over the past several years (Table 3). Amongst the general population of invasively mechanically ventilated patients in the intensive care unit, both DTF and DE have been studied as tools to predict successful weaning from mechanical ventilation. Lalwani et al. performed a prospective observational study to compare percent change in diaphragm thickness to rapid shallow breathing index (RSBI) during spontaneous breathing trial (SBT) as predictors of successful extubation [34]. Patients had to have at least three days of invasive mechanical ventilation to be included in the study. Authors found that utilizing a DTF cutoff of > 29.71% resulted in high sensitivity (93.33%) and specificity (66.67%) of predicting extubation success compared to utilizing an RSBI, which had similar sensitivity but worse specificity (22.22%). Rabuske et al. performed a prospective, observational study to compare DU during SBT (DTF and DE) with modified lung ultrasound score (mLUS) to predict extubation success amongst patients in the general intensive care unit who had been on invasive mechanical ventilation for at least 24 h [35]. They found that DTF was the only measurement to show statistically significant association with extubation outcome, with mean DTF 42.1% ± 31.7% for patients who were successfully extubated and DTF 20.9% ± 24.2% for patients who failed extubation (p = 0.042).

Another area of interest has been the utility of diaphragm ultrasound in predicting extubation success specifically in patients recovering from acute respiratory distress syndrome (ARDS) [12, 13]. Gagliardi et al. performed an observational pilot study to determine the value of DE, inspiratory time, and velocity of diaphragm contraction in predicting extubation success amongst a cohort of patients mechanically ventilated as a result of COVID-19-related ARDS [12]. Their study calculated diaphragm acceleration from the velocity measurement and found that acceleration of diaphragm contraction was correlated with weaning outcome. Liu et al. tested a combination of lung and DU in predicting successful extubation after ARDS [13]. They found that combined measurements (such as lung ultrasound score [LUS] plus DTF) were more predictive of extubation success compared to any individual measurement. Their study incorporated diaphragmatic-rapid shallow breathing index (D-RSBI), a value that substitutes DE in place of tidal volume (VT) into the RSBI equation such that D-RSBI = RR/DE. They found that the combination of D-RSBI and LUS was the most highly predictive of extubation success, with AUC approximately 83%.

Other authors have also looked at diaphragm specific RSBI calculations in determining weaning outcomes [36,37,38]. Song et al. examined the diaphragmatic RSBI and investigated if substituting DTF in place of DE would yield a differential predictive value in weaning success [36]. They compared RSBI, DE-RSBI, and DTF-RSBI amongst a prospective cohort of patients on a spontaneous breathing trial (SBT). They found that the AUC for DTF-RSBI was superior to that of DE-RSBI (approximately 86% compared to approximately 81%, respectively) and that both measures were more accurate than traditional RSBI. Saravanan et al. compared the single measurements of RSBI, DE, and DTF to diaphragm specific RSBI calculations (both DE-RSBI and DTF RSBI) [37]. They found that both DE and DTF had high specificity for predicting success compared to RSBI alone and that DE-RSBI and DTF-RSBI (AUC approximately 81% and 66%) were more predictive than RSBI alone (AUC approximately 42%). However, Fossat et al. also studied DE-RSBI as a predictive measure for weaning success and found AUCs below 50% [38]. One potential reason for this discrepancy in findings could be related to differential definitions of weaning success: Fossat had a stricter definition of extubation success, measured extubation success at 72 h and 7 days compared to Saravanan who measured extubation success at 48 h.

There has also been interest in the utilization of DU as a clinical decision-making tool to reduce time to extubation. McCool et al. conducted a randomized controlled trial of 44 patients in the medical intensive care unit who had been on mechanical ventilation for more than 48 h and were ready to undergo SBT and randomized them to usual care or usual care with ultrasound assessment of diaphragm function, specifically the percent change of diaphragm thickness between end-expiration and end-inspiration [39]. They found that for subjects with a normal percent change in diaphragm thickness, time from ultrasound to extubation was shorter than in those patients with noted diaphragm dysfunction. Interestingly, the total number of days on a ventilator for these patients was the same between both groups. This study also assessed clinician perceived utility of DU as between “slightly useful” and “somewhat useful”.

Selecting the appropriate clinical scenario for the application of DU represents an evolving science. The evidence presented above and in the tables represents the current understanding of the diagnostic performance under difference circumstances. Currently, the authors do not believe that the evidence dictates applying this technique as a universal screening prior to extubation, as the prevalence of clinically significant diaphragmatic weakness in the general respiratory failure population is low enough that the time investment and diagnostic accuracy would not provide benefit. In selected populations with a higher pretest probability of diaphragmatic weakness, however, this can meaningfully inform clinical decision making around timing of extubation and choice of non-invasive positive pressure ventilation post extubation. For example, in the post-thoracic surgery population (e.g. lung transplant, heart transplant, etc.), or the autoimmune syndrome patient (i.e. with neuritis symptoms), there is justification for an additional evaluation prior to extubation. Similarly, if clinical symptoms (shoe tying orthopnea, etc.) were present prior to intubation, this examination can add clinical information. Finally, in the population who are showing signs of recovery after respiratory failure, but the decision has to be made about timing of tracheostomy for an anticipated prolonged weaning process, knowledge of whether diaphragmatic atrophy and weakness are present can help prognosticate timing for definitive ventilator wean. That said, there remain many specific applications that have not yet been explored, which are highlighted below in the following section.

Recent studies have extended assessment beyond the diaphragm to include other respiratory muscles, including those involved in cough since an effective cough has been shown to be associated with better extubation outcomes [40,41,42]. Qiu et al. investigated the role of pre-extubation abdominal muscle ultrasound in predicting extubation success [43]. They found that the cough thickening fraction of the internal oblique muscles had the strongest predictive value of extubation success although cough thickening fraction of the rectus abdominis was also associated with extubation success. Schreiber et al. performed a cohort study comparing abdominal muscle thickness and thickening fraction during cough and expiration in healthy subjects compared to ventilated patients both prior to and during an SBT [44]. They found that reduced abdominal muscle thickening during cough was associated with higher risk of extubation failure among ventilated patients who had passed an SBT. Interestingly, they noted that a summative measure of cough thickening fraction of three major abdominal muscles (transversus abdominis, rectus abdominis, and internal obliques) had stronger predictive discrimination for extubation failure compared to DTF as measured 5 min into the SBT (AUC 82% for abdominal muscles compared to AUC 70% for diaphragm).