The main finding of the present study is that the combination of cardiopulmonary ultrasound parameters, notably the LUS score and RVD, with the ILD-GAP score improves the prediction of prognosis in patients with ILD compared to the ILD-GAP score alone. This shows that cardiopulmonary ultrasound is a useful method that aids risk stratification in patients with ILD. Aside from the severity of pulmonary fibrosis, the cardiac impact of ILD right ventricular remodeling and dysfunction are also important in the prognosis of ILD.
RV remodeling and dysfunction correlated with poor outcome in patients with ILD
Patients with ILD may develop pulmonary hypertension (PH) at the end-stage of the disease. Pulmonary fibrosis-induced destruction of pulmonary vessels, hypoxic pulmonary vasoconstriction and pulmonary vascular remodeling induced by vasoactive compounds can all lead to PH [17]. The RV is highly sensitive to pressure overload [18] and the gradual increase of SPAP in patients with ILD causes compensatory enlargement and hypertrophy of the RV, which can lead to dysfunction, right heart failure and death.
Using multivariate analysis, we showed that RVD and RVGLS were more related to the outcome of patients with ILD than the SPAP parameter. This finding is in agreement with the study on IPF patients by Rivera-Lebron et al. [19] and other previous studies that examined PH [20]. Andersen et al. showed that ILD patients with PH had markedly higher mortality than those without PH [13]. Our study also found that the death group had a higher proportion of PH cases, but no significant correlation between SPAP and prognosis.
In our study, only RVGLS was found to correlate with the prognosis of patients with ILD, but not the traditional right ventricular functional parameters (i.e. TAPSE, RVFAC). This might be because such parameters do not show abnormalities in the early stage of right ventricular remodeling. Nevertheless, a previous study showed that RVGLS could reveal early impairment of right ventricular function, even in IPF patients without pulmonary hypertension [21].
Although our study initially found a correlation between RVGLS and the risk of death in patients with ILD, this correlation was not confirmed by the multivariate Cox proportional hazard regression model. This might be due to the fact that right ventricular function was relatively preserved among the patients enrolled in the present study. The median value of the RVGLS was -21.5 %, which was significantly higher than the -20 % value that is recommended by the ASE guidelines [16]. This may be the reason why RVGLS was not found to be related to the prognosis of patients with ILD in our multivariate analysis.
The RV parameter that correlated with prognosis in patients with ILD was RVD. Bax et al. showed in a recent study that the ratio between the RV and the left ventricle on CT pulmonary angiograms predicted mortality in ILD, rather than echocardiography parameters such as TAPSE or RVFAC [22]. These findings demonstrated the importance of morphological changes in the RV, especially when SPAP cannot be obtained. A change in RV shape and volume can be the first sign of RV dysfunction or volume overload [23].
ILD LUS score correlated with poor outcome in patients with ILD
A large number of previous studies have demonstrated the accuracy of LUS in evaluating the severity of pulmonary fibrosis in patients with ILD [6-8, 24]. Severe lesions of ILD are often present in the peripheral area of the subpleural lung, which facilitates the evaluation of the severity of pulmonary fibrosis by LUS [25]. Previous studies conducted by our team revealed a certain correlation between the LUS score and PFT parameters (DLCO%, FVC%), the ILD-GAP score, and even some echocardiographic parameters, including SPAP and RVGLS, in patients with ILD [11, 26]. The present study further found the LUS score to be an independent predictor of death in patients with ILD.
LUS has been shown to be useful in the evaluation of all stages of ILD. In the early stage of ILD, when pulmonary parenchyma injury and alveolitis manifests, as well as when increased fibrous components in the alveolar septum, thickened alveolar wall, and interlobular septal are observed, numerous B-line artifacts or white lung indicators can be detected by LUS when the subpleural lung tissue is involved. When the progressive aggravation of alveolar inflammatory injury leads to the destruction of the alveolar wall, and the appearance of subpleural cysts and a honeycomb, LUS will show blurred pleura lines [1, 10]. Recently, Gargani et al. analyzed the relationship between B lines and prognosis in patients with systemic sclerosis. Their results showed that B lines were an independent predictor of ILD deterioration [27]. Although their study ignored the importance of pleural lines in evaluating lung involvement, they still found a predictive value of B-lines. All of these findings indicate that LUS is a useful tool for evaluating lung pathology in patients with ILD.
Clinical implications
The ILD-GAP score currently used for risk assessment of ILD patients largely depends on the FVC and DLCO. Prior to disease development, the normal pulmonary function of a person can range from 80 % to 120 % of values predicted from age, sex and height. A FVC threshold of 75 % would therefore represent a decline of between 5 % and 45 % depending on premorbid values, a 9-fold variability. It is therefore necessary to use HRCT to refine the assessment and identify patients in whom measured values are misleading [28]. However, HRCT is not feasible for dynamic monitoring in the intensive care unit or for regular follow-up sessions in chronic patients. LUS on the other hand, is much more suitable for dynamic monitoring and can be easily repeated without exposing patients to ionizing radiation. The ILD-GAP score combined with the LUS score doesn’t only evaluate the severity of pulmonary fibrosis, but can also monitor patient reaction after treatment, which greatly facilitates the management of critically ill patients and the long-term follow-up of outpatients.
In clinical practice, close attention is paid to lung lesions in patients with ILD, but their right ventricular remodeling is largely ignored. In our cohort study, we found that RVD can add additional predictive value to the ILD-GAP score. RVD is also an easily obtainable parameter. Although we failed to obtain RVGLS in some patients, we found that using RVGLS in combination with RVD could not improve prediction efficiency. Therefore, when evaluating the RV in patients with ILD, if RVGLS cannot be obtained, the dilatation of RVD is considered a sign of poor prognosis.
Limitations
Various limitations of the current work should be noted. This study is a retrospective single-center study with a small sample size, which limits the generalizability of the conclusions drawn. The endpoint occurrence (all-cause death) was determined by telephone follow-up in almost all cases, which can be inaccurate due to recall bias of the patients’ families. A large-scale prospective multicenter study is needed to confirm the results of this work in the future.
Standardized methods for evaluating patients with ILD by LUS are lacking, both for pulmonary fibrosis, and RV function and structure. The present study used the LUS score, which was developed by Buda et al. to evaluate pulmonary fibrosis [10]. For RV function, previous studies have shown that RVGLS might be a sensitive parameter [16]. However, since no three-dimensional echocardiographic images were collected in the current work, we were unable to evaluate the correlation between RV volume [23] and prognosis of patients with ILD. Nevertheless, we still found a good correlation between the structural parameter RVD and the risk of death in patients with ILD.
In all, this study preliminarily explored the prognostic value of cardiopulmonary ultrasound parameters in patients with ILD. This question has been rarely studied in the past, and based on the current and encouraging results, we will perform large-scale prospective studies in the future.