Diaphragmatic ultrasound findings correlate with dyspnea, exercise tolerance, health-related quality of life and lung function in patients with fibrotic interstitial lung disease

Dyspnea and exercise intolerance characterize fibrotic interstitial lung diseases (FILD) [1, 2]. Both are determinants of impaired health-related quality of life (HRQoL) observed in FILD patients [1, 3, 4]. Dyspnea and exercise intolerance in FILD have several underlying mechanisms [5]. Among them is respiratory muscle dysfunction [6, 7], which is usually difficult to diagnose [6].

In ILD, increased lung elastic recoil overloads the respiratory muscles [5, 8]. In addition, many other factors present in ILD, such as chronic hypoxemia, corticosteroids, systemic inflammation, physical inactivity and malnutrition may promote muscle dysfunction [6]. However, the few studies that have evaluated diaphragmatic function in ILD report conflicting results. While some authors have found preserved inspiratory muscle strength in ILD [5, 8‐10], Walterspachen et al. [7] found reduced diaphragmatic strength, assessed with non-volitional measures, in a subset of ILD subjects with severe disease.

The diaphragm is the main inspiratory muscle being responsible for 60 to 70% of the tidal volume during quiet breathing [11, 12]. Due to its location and size, its function can be assessed by ultrasound (US). Initially, standardized techniques for mobility and diaphragmatic thickening measurements were reported in healthy volunteers [13‐17]. Later, thoracic ultrasound (TUS) focusing on the diaphragm was employed in patients with respiratory disorders such as asthma [18], chronic obstructive pulmonary disease (COPD) [19], cystic fibrosis [20] and respiratory failure [21‐23]. Recently, TUS was applied in FILD patients [24] and results have shown decreased diaphragmatic mobility during deep breathing that is associated with reduced lung volumes [24].

Improvement of HRQoL, including relief of dyspnea and exercise intolerance, is the cornerstone of the care of FILD patients, most of whom have irreversible lung diseases [25, 26]. In cases where TUS findings were correlated with dyspnea, exercise tolerance, HRQoL and lung function, US could be used to propose and monitor interventions such as rehabilitation and to assess disease progression [27].

Firstly, we hypothesized that diaphragmatic mobility and thickness, assessed by TUS, would be associated with dyspnea, exercise tolerance, HRQoL and lung function in FILD cases. Secondly, we hypothesized that a proportion of FILD cases would present with reduced diaphragmatic thickening and they would have different clinical and functional characteristics from FILD cases without reduced thickening.

Demographic and clinical data of FILD cases and healthy controls are depicted in Table 1. The diaphragmatic mobility and thickness were measured for all subjects. The classification of FILD patients and the current use of steroids are presented in Table 1. Two patients were on long-term oxygen therapy. Respiratory muscle strength was similar between FILD cases and healthy controls (Table 1). The prevalence of comorbidities was similar between FILD cases and healthy controls (Additional file 1: Methods and Results). The scores of SGRQ revealed a reduction in HRQoL in all domains (Additional file 1: Table SA1, Results) in FILD cases when compared to reference values [39].

FILD cases walked less than predicted and presented peripheral oxygen desaturation, increased heart rate, dyspnea and leg fatigue at the end of the 6MWT (Additional file 1: Table SA2).

Diaphragmatic mobility during QB was similar between FILD and control groups (p = 0.95). However, during DB, diaphragmatic mobility was lower in the FILD cases when compared to healthy controls (p <  0.01). During QB, at FRC, the diaphragm of FILD cases was significantly thicker than the healthy controls (p = 0.01). But, after a maximal DB, at TLC, the diaphragm of FILD cases was significantly thinner than the healthy controls (p <  0.01), resulting in a lower TF in the FILD cases (p <  0.01). (Table 2 and Additional file 2: Figure S1).

During DB, diaphragmatic mobility and thickness correlated with lung function (FVC, FEV_1, TLC and DLCO), exercise tolerance and HRQoL but negatively correlated with resting dyspnea. During DB, lesser diaphragmatic mobility and thickness correlated with more resting dyspnea, more desaturation and dyspnea at the end of the 6MWT; quality of life is worse (mainly respiratory symptoms and activity domain of SGRQ). (Table 3 and Additional file 3: Figure S2 and Additional file 4: Figure S3). However, nor corticosteroid use, nor a specific group of FILD diagnoses were associated with diaphragmatic mobility and thickness in FILD cases.

For the healthy controls, the 95% confidence interval for TF during DB was 101 to 354%. To define the FILD cases with reduced diaphragmatic thickening, the choice of TF < 101% represents the values below which only 5% of the healthy controls’ values fall (5th percentile). Seventy percent of FILD cases presented reduced diaphragmatic thickening (Table 4). FILD cases with reduced diaphragmatic thickening had lower lung volumes (FVC and FEV₁), higher resting dyspnea, worse HRQoL (activity and total domains of SGRQ), higher desaturation and dyspnea after the 6MWT (Table 4). Age, sex, BMI and corticosteroid were similar among FILD cases with and without reduced diaphragmatic thickening.

The novel findings of this study are that in FILD cases compared to healthy adults, lower deep breathing diaphragmatic mobility and thickening correlated with increased dyspnea, decreased exercise tolerance, worse HRQoL and worse lung function. Most FILD cases (75%) presented reduced diaphragmatic thickening and these patients had higher dyspnea, higher desaturation, worse HRQoL and lung function than FILD cases without reduced diaphragmatic thickening. In addition, FILD cases presented a thicker diaphragm at rest compared to healthy controls. Using US, we assessed the diaphragm function of a heterogeneous sample of FILD patients. Findings of higher dyspnea, exercise intolerance, worse HRQoL and lung function in FILD cases with reduced diaphragmatic mobility and thickening are novel and provide further evidence that diaphragm function affects FILD patients clinically.

Few studies explored the relationship between diaphragmatic function, dyspnea, HRQoL and exercise tolerance in FILD. In our study, FILD cases with lower DB diaphragmatic mobility and TF had more resting and exertional dyspnea and worse HRQoL. Further, the lower the DB diaphragmatic mobility and the TF, the greater the desaturation at the end of the 6MWT. The reduced diaphragmatic mobility and thickening in FILD cases may compromise the overloaded inspiratory muscles [40, 41] eliciting early onset breathlessness [5, 41] resulting in poor HRQoL [1, 3]. In addition, impairment in diaphragm function may hinder ventilation throughout exercise causing additional mismatch on the ventilation to perfusion ratio, leading to desaturation similar to that shown in diaphragm paralysis [42].

The relationship between muscle function, dyspnea and HRQoL has been explored in sarcoidosis [43, 44] lung cancer patients [45] and in COPD [46]. Interestingly, in COPD, Paulin et al. [46] used TUS to assess diaphragm function and showed that COPD patients with lower diaphragmatic mobility also had lower 6MWT performance and higher exertional dyspnea [46]. Furthermore, air trapping influenced the diaphragmatic mobility in COPD. The authors hypothesize that COPD patients with air trapping imposed a mechanical disadvantage of the inspiratory muscles due to their shortened lengths, which was accentuated during exertion and elicited breathlessness [46].

In our study, diaphragmatic mobility and TF was associated with dyspnea and desaturation, but did not influence the 6MWT distance covered by FILD cases.

To the best of our knowledge, only two studies used the TUS to investigate the diaphragmatic function in ILD. He et al. [47] showed that diaphragmatic mobility, but not thickness, was similar between a small sample of idiopathic pulmonary fibrosis patients and healthy controls. In contrast, we previously recorded that FILD patients presented reduced diaphragmatic mobility and TF during DB when compared to healthy controls [24]. Diaphragmatic mobility during DB was associated with FVC, FEV₁ and TLC in FILD [24].

The present study reinforced those findings. We found a strong correlation between diaphragmatic mobility and TF with lung function. Although expected and intuitive, these correlations are not consistent findings even in healthy adults. While some studies found a linear relationship between diaphragmatic mobility and lung volumes in healthy volunteers [48, 49], another study showed that diaphragmatic mobility measured during QB and DB or inspiratory capacity maneuvers poorly correlated with lung volumes [50]. Recently, in 210 healthy volunteers, Boussuges et al. [14] found only a weak correlation between diaphragmatic mobility and lung volumes. This controversy may be explained by the fact that inspiratory lung volumes are determined by diaphragmatic mobility, and also by recruitment of extra diaphragmatic muscles [51] and thoracoabdominal compliance [14]. These factors also influence the thickening of diaphragm. In healthy subjects, thickening of diaphragm is greatest in the zone of apposition, during a primarily abdominal breath [52] and is affected by the diaphragm curvature [53].

In previous studies in COPD, both diaphragmatic mobility [54] and thickening [55] have been correlated with air trapping and airway obstruction. In our study, decreased diaphragmatic mobility seen in FILD patients during DB may reflect the reduced lung volume. The reduced diaphragmatic thickening may be due to an intrinsic muscle dysfunction associated with several causal factors (inflammatory and oxidative stress, corticosteroid use, physical inactivity and malnutrition) usually present in ILDs [6, 7]. We cannot rule out that the curvature of the diaphragm may have affected diaphragmatic thickening. The lower lung volumes in ILDs decrease the radius of the diaphragmatic curvature, which may explain our finding of increased diaphragmatic thickness in patients with FILD. Increased diaphragmatic thickness could also represent muscle hypertrophy consistent with a training effect, and has been showed in other respiratory diseases such as chronic asthma [18] and cystic fibrosis [20].

Respiratory muscle strength was similar between the two FILD subgroups (with and without reduced diaphragmatic thickening). Respiratory muscle function is often neglected in ILD and the few studies that have addressed it report conflicting results. Preserved inspiratory strength in ILD has been described [8‐10], but recently reduced diaphragmatic strength in a sample of patients with more severe ILD was shown [7]. These conflicting results may reflect a “set of opposing forces” in ILD. ILD subjects have a mechanical advantage of inspiratory force generation at lower operating lung volumes and a “training effect” on the diaphragm because of the lung stiffness and increased elastic recoil [6]. However, several factors (hypoxemia, inflammatory status, corticosteroids, physical inactivity and malnutrition) may be harmful for the respiratory muscles in ILD [6]. Overall, the impaired diaphragmatic function (reduction of mobility and thickening) may represent the harmful effects on respiratory muscles, while “training effect” maintains preserved inspiratory strength in these ILD patients. Further, the heterogeneity and severity of the underlying ILD may be a confounding factor in the interpretation of respiratory muscle assessment.

In our study, corticosteroid was not associated with impaired diaphragmatic function in FILD cases possibly because only 5 patients (16,6%) were actually using a dosage of prednisone higher than 20 mg/day which is recognized as myotoxic. Corticosteroid use has been associated with mitochondrial dysfunction and oxidative stress leading to corticosteroid-induced myopathy [56]. However, the occurrence of corticosteroid-induced myopathy may be influenced by the dosage. Unfortunately, we did not investigate prospectively before and under steroid treatment for FILD. Our findings suggest that TUS may be used to monitor variables that are clinically relevant and prognostic determinants in FILD. In addition, identifying impairment of diaphragm function may alert physicians to avoid or minimize the use of myotoxic drugs, such as corticosteroids. Characterization of diaphragm function with TUS could further suggest targeted actions to improve function, such as rehabilitation, a recognized intervention to improve exercise tolerance and HRQoL and to decrease dyspnea.

Our study has some limitations. First, we studied a heterogeneous sample of FILD patients, which could be a confounding factor. However, heterogeneity is usual among FILD and gives a pragmatic character to the present study. Secondly, lung volumes were not measured concurrently with diaphragm thickness and mobility. Normalizing the mobility and thickness for lung volumes measured during the protocol could add to our understanding about the relationship between diaphragm function and lung volume restriction. Thirdly, TUS was assessed only on the right side. Fourthly, only one observer performed the TUS. However, we used M-mode to measure diaphragm mobility and standardized technique to evaluate diaphragm thickness, and both have shown to have a high reproducibility. A fifty consideration was that respiratory muscle strength was similar, which might be a reflection of a volitional measurement of overall respiratory muscle strength. Perhaps TUS discloses diaphragm function impairment before strength reduction, or the activity of global inspiratory muscle may compensate for diaphragmatic weakness. A sixth consideration was the performance of multiple comparisons in our study, which present the risk of erroneously finding a significant difference by chance. However, our study aimed to understand whether diaphragmatic ultrasound findings would correlate with clinically relevant parameters for ILD subjects, such as dyspnea and exercise intolerance that are determinants of their impaired quality of life (HRQoL). Lastly, our healthy controls did not have TLC measurements, but we have no reasons to believe that they could have any restriction or hyperinflation.

Compared with healthy controls, FILD cases had higher diaphragmatic thickness at rest, but lower diaphragmatic mobility and thickening fraction during deep breathings that correlate to dyspnea, exercise tolerance, HRQoL and lung function. FILD cases with diaphragmatic reduced thickening have higher dyspnea, less exercise tolerance, lower HRQoL and lung function than FILD cases without reduced thickening.

Diaphragmatic US in FILD cases may be a useful tool to investigate and monitor diaphragm function, to propose and monitor interventions such as rehabilitation and use of pharmacological treatments such as corticosteroids.