Ultrasonographic assessment of parasternal intercostal muscles during mechanical ventilation

Together with the diaphragm, extra-diaphragmatic inspiratory muscles participate in the generation of the tidal volume. The intercostal muscles are composed of three thin layers of muscle fibers occupying each of the intercostal spaces. The external intercostals extend from the tubercles of the ribs dorsally to the costochondral junctions ventrally, and their fibers are oriented obliquely, downward, and forward, from the rib above to the rib below (Fig. 1, Panel A). Ventrally, between the sternum and the costochondral junctions, the fibers are those of the internal intercostal muscles; these are particularly thick in parasternal region of the rib cage, where they are conventionally called the parasternal intercostals. These anatomical structures are active during the inspiratory phase and interact with the diaphragm and other extra diaphragmatic inspiratory muscles [27]. In contrast, the internal intercostal muscles are directed obliquely, upward, and forward from the superior border of the rib and costal cartilage below to the base of the subcostal border of the rib. The innermost intercostal muscles are the deepest, separated from the internal intercostal muscles by a grouping of nerves and blood vessels known as the neurovascular bundle [15]. These internal intercostals are active during the expiration phase of respiratory cycle (Fig. 1b). The complex interactions among all these fiber groups have been investigated in few physiological humans’ studies. The inspiratory drive to the external intercostal muscles was assessed by De Troyer et al. [8], who showed how the third dorsal external intercostal was active earlier in inspiration. Moreover, the degree of activation of the external intercostal muscles during quiet breathing correlated with the magnitude of their mechanical advantage, which diminishes fourfold from the second to the fifth interspace [17]. The same authors, in a complex physiological study, showed how the parasternal intercostals have an inspiratory action on the lung even if with a pressure-generating ability less remarkable as compared to other extra diaphragmatic muscles [16]. Moreover, the study described how the fractional shortening of the parasternal intercostals decreased gradually from the second to the fifth interspace. Nevertheless, it remains unknown whether the neural drive to parasternal intercostal muscles is graded across the interspaces [28]. The spatial distribution of neural drive along rostro-caudal gradient has been demonstrated in a study in which the timing of activation and the discharge performance of single motor units activated during inspiration in the parasternal intercostals of the first to fifth interspaces were measured [29]. During respiratory loading, Sampson et al. [30] showed how the pattern of chest wall deformation was closely related to the activity and coordination of the various inspiratory intercostal muscles, and how the parasternal intercostals do not necessarily represent all inspiratory intercostals. In summary, the movement of the ribs depends on the relative amount of torque around the vertebral articulations acting on the two points of attachment of the muscle to the respective ribs: when external intercostal muscles contract, the torque acting on the lower rib is greater than that acting on the upper rib, with a net effect to raise the ribs. The reverse actually happens for the internal intercostals, which act to lower the ribs to which they attach. Eventually, the parasternal intercostals raise the ribs as their action is transmitted to the sternum (Fig. 1b).

Ultrasound imaging has been advocated as an interesting imaging modality as it has been shown to be accurate, reproducible and convenient tool for the assessment of linear dimensions, cross-sectional area, thickness, and indices of muscle architecture [31, 32]. In addition, ultrasound has already been used to assess the relationship between muscles and chest wall motion [33, 34]. An animal model demonstrated the ability of ultrasound to accurately estimate the cross-sectional area of the parasternal intercostal muscles, comparing it with the measurement directly performed on the specimen [35]. In healthy volunteers, ultrasound has been used to investigate parasternal intercostal muscles in the easily accessible anterior parasternal region [35, 36]. Conversely, in the lateral and in the posterior part of the intercostal space, the internal and external intercostal muscles often overlap, making the ultrasound detection of both muscle layers impossible. Therefore, the increasing muscle thickness of the external intercostal muscle during inspiration (an estimate of muscle recruitment) cannot be assessed in this location.

Using ultrasound imaging in healthy volunteers, Cala et al. [33] reported that the muscles move ventrally and straighten. Cala et al. [36] measured the inter-rib distance, i.e., the parasternal intercostal muscle thickness. The authors showed a ventral movement during tidal breathing without changes in inter-rib distance and thickness, thereby concluding that the parasternal intercostal muscle moves ventrally and more perpendicular to the ribs. The findings support an intercostal stabilizing function of the parasternal intercostal muscles. Wallbridge [24] postulated that the relationship between intercostal muscle sizes might reflect different roles and recruitment that depends on chest wall site. This hypothesis was based on weak correlation (r = 0.33) observed between ultrasound thickness of parasternal intercostals and severity of airflow obstruction as measured by spirometric parameters (i.e., FEV1% predicted) in stable COPD patients. Given the preliminary nature and the weak correlation of the results, this topic deserves to be explored more thoroughly. A recent study by Yoshida et al. [26] used ultrasound to assess whether human parasternal intercostal muscle thickness increased during vigorous breathing efforts (the maximal inspiratory level using an inspiratory resistance loading device). The images obtained showed an increase in muscle thickness only in the anterior portion of the intercostal space during maximal breathing, whereas no significant differences could be detected in the lateral and posterior portions. This remark suggest how the parasternal intercostal portion may be the best region to investigate the intercostal muscle, as in the lateral and posterior part of the intercostal space, internal intercostal muscles and external intercostal muscles overlap, and ultrasound cannot separate both muscle layers. Nevertheless, these observations must be investigated more thoroughly in mechanically ventilated patients. Regarding the reproducibility of ultrasound technique, a high interobserver reproducibility of parasternal intercostal end-expiratory and peak-inspiratory thickness has recently been confirmed [37] as previously reported [35, 36]. Moreover, intra-rater reliability of intercostal muscle thickness depended on site, ranging between 0.87 and 0.97 [24]. Relevant observations are summarized in Table 1.

The parasternal intercostal muscles can be assessed using a 10–15 MHz, linear ultrasonography transducer in M-mode, placed 3–5 cm laterally from the sternum, and oriented transversally in the sagittal plane between the 2nd and the 3rd rib [37, 38]. For instance, while the patient is lying at a 20° head-up position, the linear array transducer is positioned perpendicular to the anterior thorax surface in the longitudinal scan. Starting in B mode, the pleural line is easily detected as a part of the “bat sign” [39]. Just above the pleural line, the parasternal intercostal muscle can be identified as a three-layered biconcave structure, where two linear hyperechoic membranes running from the anterior and posterior aspects of the adjoining ribs, and a medial portion with muscle echotexture (Fig. 2) [24, 37]. Muscle thickness is measured between the inner and outermost hyperechogenic layer of the muscle fascial borders. Moreover, using the M-mode, the inspiratory contraction of the muscle can be detected, as described for the thickening fraction of the diaphragm [40]. In fact, during the inspiration phase, an increase in muscle thickness can be visualized as the muscle fiber contraction displaces the rib cage cranially and anteriorly, while muscle mass is maintained constant [27]. Parasternal intercostal muscle-thickening fraction (TFic) can be calculated as the ratio of the difference between end-expiratory and end-inspiratory thickness to end-expiratory thickness (TF = ((TH end inspiration – TH end expiration)/TH end expiration)) * 100), similarly to the diaphragm thickening fraction (TFdi). Furthermore, indirect qualitative information on muscle composition/structure can be recorded by measuring muscle echogenicity (image-grey-scale) [41]. Muscle echogenicity can define the region of interest for analysis performed using a standard histogram function widely available in many commercial software for image editing [42]. Higher muscle echogenicity (indicating poorer muscle quality) of parasternal intercostal muscle is negatively correlated with COPD severity [24]. However, since echogenicity measurements are influenced by observer-dependent factors more strongly than are other measures, a rigorous ultrasound setting must be selected before acquiring the image to standardize the methods. Some features regarding the probe orientation must also be considered when exploring parasternal muscles. In fact, it has been supposed that the parasternal intercostal muscle image quality progressively reduces as the angle of incidence of ultrasound changes by moving the transducer away from the parasagittal plane. In a pioneering study, Wait et al. [43] showed how the value of the misalignment angle—i.e., the angle at which the transducer no longer produces a readable signal—ranged from 1.6° to 8.4° for transducer frequencies between 10 and 1 MHz, respectively, during the evaluation of diaphragm movements. Such observations have never been tested for the evaluation of intercostal muscle. Since the ventral and dorsal edges of the parasternal intercostal muscles are not parallel in the majority of the images, the thickness measurements should be performed at the cranio-caudal midpoint between the ribs. There, the difference in the curvature between the two surfaces is least, and the lines flattest.

Even if only few studies have focused specifically on the intercostal muscles in mechanically ventilated patients, some recent papers investigated the effects of MV on extra diaphragmatic respiratory muscles (Table 1). Starting from animal observations, Capdevilla et al. [44] showed how MV in rabbits produces alterations in contractile properties of the 5th external intercostal muscle, increasing muscle fatigue and promoting atrophy of type II fibers. More specifically, Bernard et al. [45] showed a disruption of myofibrils in both the diaphragm and the 5th and 6th external intercostal muscles.

Direct extrapolation of these scarce results to human subjects may be questioned; however, a time-dependent relationship between the duration of ventilation and the subsequent development of weakness appears likely [37, 46, 47]. In this regard, whereas the effects of MV on the diaphragm have extensively been studied [10, 22], there is little evidence regarding its effects on the performance of other inspiratory muscles and the relative contribution of intercostal muscle dysfunction remains to be determined, especially in ventilator-dependent patients. Nakanishi et al. [48] showed how an excessive inspiratory support causes atrophy of the disused diaphragm and intercostals muscle and that excessive pressure might cause injury to respiratory muscles and structural changes with gradually increasing thickness. Recently, Dres [37] reported that the TFic was significantly associated with failure of a spontaneous breathing trial in mechanically ventilated patients and how this thickening was responsive to the level of ventilator assistance and significantly higher in patients with diaphragm dysfunction. More specifically, a TFic exceeding 8% identified patients with diaphragm dysfunction, and a value greater than 10% predicted weaning failure. In support of this observation, Umbrello et al. [38] showed similar results, describing a complex relationship between the diaphragm and parasternal intercostal muscles. The authors showed how diaphragm thickening fraction was higher (> 30%) and that of parasternal intercostal was lower (< 5%) in patients without as compared to patients with diaphragm dysfunction. In the presence of diaphragm dysfunction a low TFdi may reflect a low inspiratory effort, or an elevated effort with inspiratory work performed by extra-diaphragmatic muscles associated with low or elevated levels of Tfic, depending on level the mechanical respiratory support delivered by the ventilator. While we acknowledge that these are recent and preliminary observations, and more importantly given the monocentric and pilot nature of the studies, their consistency and strong physiological rationale led us to propose their use in clinical practice.

Based on the recent consistent findings on this topic [37, 49], we propose a practical flowchart during the weaning trial from MV, in which ultrasound muscular evaluation of the whole respiratory apparatus is taken into account. A similar approach has recently been suggested in an ABCD ultrasound flowchart [49]. Weaning failure has a multifactorial nature, which results from muscular dysfunction, excessive mechanical load, weaning-induced cardiovascular dysfunction, or reduced ability to clear airway secretions. Most physicians simply evaluate the patient’s ability to tolerate a spontaneous breathing trial (SBT) without distress to determine likely weaning success [50, 51]. We propose that once a SBT is started, a diaphragm evaluation exploring the TFdi should be considered in those patients who failed a termination of the weaning process within 24 h [1]. At the same time, we propose to evaluate the parasternal intercostal muscles (Fig. 3). This suggestion derives from considering that even when low contractile activity of the diaphragm is identified, two different scenarios may be present: the diaphragm evaluation may be in the normal range because of an overassisted machine setting, or this reduced activity may signify diaphragm dysfunction. In the latter case, the involvement of parasternal intercostals should be considered, as a significant thickening of these muscles may be a sign of muscle recruitment. Thus, a value of TFic less than 10%, associated with a TFdi greater than 20% is suggestive of a breath in which the diaphragm is the main inspiratory muscle and the extra diaphragmatic muscle is not “stressed”, thereby potentially leading to a successful weaning trial [37, 38]. Otherwise, it may be possible that, in the presence of diaphragm dysfunction, the intercostal muscles have become more active, which is a negative indicator, potentially predictive of weaning failure. The clinicians must consider that these thresholds have not been prospectively validated, and that all these measurements may be integrated with other commonly used predictive parameters of weaning success, such as the rapid shallow breathing index (RSBI), the negative inspiratory pressure (NIP), forced vital capacity (FVC) [52], gas exchange and the tidal changes in esophageal pressure (Δpes) [53]. These, however, are not always investigated, and each is characterized by its own specificity and sensitivity. While we are well aware that it is simplistic to think that a single parameter could predict overall weaning success or failure, we nevertheless think that the strong physiological rationale of parasternal intercostal ultrasound may be of help to the clinician after an adequate training. With the exception of chronic muscular disease [54], the contributions to respiratory failure from respiratory muscles other than the diaphragm, and the possible causative role of prior muscle inactivity, still are poorly understood. No study has specifically investigated the role of intercostal muscles during support by different respiratory devices, such as non-invasive ventilation.