Introduction

Interstitial lung disease (ILD) refers to a heterogeneous lung disease characterized by a combination of inflammation and fibrosis. In recent years, studies have identified effective antifibrotic therapies for many forms of ILD1,2,3. The success of these clinical trials has also expanded the understanding of ILD from the disease classification and aetiology to the disease behaviour and treatment response, followed by the coinage of terms describing pulmonary fibrosis, such as progressive pulmonary fibrosis (PPF), fibrotic interstitial lung disease (FILD), and progressive fibrosing interstitial lung diseases (PF-ILD)1,4. FILD is a general term that includes many diseases that manifest as pulmonary fibrosis and are often progressive and associated with high morbidity and early mortality. Among them, idiopathic pulmonary fibrosis (IPF) is a common disease that causes a chronic progressive pulmonary fibrosis with an unknown aetiology, but a larger group of FILD patients develop a progressive fibrosing phenotype during their clinical course5. These FILD types include idiopathic nonspecific interstitial pneumonia (iNSIP), connective tissue disease-associated ILD (CTD-ILD), interstitial pneumonia with autoimmune features (IPAF), fibrotic hypersensitivity pneumonitis (FHP), unclassifiable ILD (U-ILD) and others4,6,7.

Although each individual FILD is rare, collectively, they affect numerous patients6. Due to the incomplete knowledge of the underlying pathogenetic pathways and natural history, the diagnosis of fibrotic ILD is frequently considerably delayed, misdiagnoses occur, and antifibrotic therapy has limited effectiveness, which in turn imposes a substantial burden of disease. Research has shown that the time from symptom onset to death in PF-ILD patients is 61–80 months, but the time between the detection of progressive fibrosis and death is only approximately 30–45 months, which is similar to that of IPF8. The delay in the diagnosis of FILD accompanied by the delay in antifibrotic therapy may be the main reason for the poor prognosis of FILD. Currently, there are limited data regarding the frequency and prognosis of FILD patients other than IPF.

In this cohort study, we retrospectively analysed the baseline clinical data of patients who were diagnosed with fibrotic ILD other than IPF in the past eleven years to identify clinical characteristics and early prognostic risk factors for FILD.

Methods

Study design

We retrospectively reviewed the medical records of 381 patients who received a diagnosis of ILD at the First Affiliated Hospital of Ningbo University between April 2011 and September 2022. One hundred ninety-four patients diagnosed with non-IPF fibrotic ILD, including idiopathic nonspecific interstitial pneumonia (iNSIP), connective tissue disease (CTD)-associated ILDs, fibrotic hypersensitivity pneumonitis (FHP), microscopic polyangiitis (MPA)-associated ILD, and ILDs related to occupational exposures, were screened in our study. The CTDs included rheumatoid arthritis (RA), polymyositis/dermatomyositis (PM/DM), primary Sjogren’s syndrome (pSS), systemic sclerosis (SSc), systemic lupus erythematosus (SLE) and mixed connective tissue disease (MCTD). Each disease was diagnosed according to its respective criteria9,10,11,12,13,14,15,16,17,18,19. A flowchart of the patient screening and classification for our study is presented in Fig. 1.

The exclusion criteria were as follows: (1) patients diagnosed with IPF (n = 44); (2) ILD patients for whom chest CT or HRCT did not show clear pulmonary fibrosis (n = 104); and (3) ILD patients for whom chest CT or HRCT data were not available (n = 40). The cohort study was approved by the institutional review board of the First Affiliated Hospital of Ningbo University (IRB no: 2023-R102RS).

Fig. 1
figure 1

Flow chart of the study. ILD: interstitial lung disease; IPF: idiopathic pulmonary fibrosis; iNSIP: idiopathic nonspecific interstitial pneumonia; MPA-ILD: microscopic polyangiitis-associated ILD; CTD-ILD: connective tissue disease-associated ILD; IPAF: interstitial pneumonia with autoimmune features; FHP: fibrotic hypersensitivity pneumonitis; RA-ILD: rheumatoid arthritis-associated ILD; PM/DM-ILD: polymyositis/dermatomyositis-associated ILD; pSS-ILD: primary Sjögren’s syndrome-associated ILD; SSc-ILD: systemic sclerosis-associated ILD; SLE-ILD: systemic lupus erythematosus-associated ILD; MCTD-ILD: mixed connective tissue disease-associated ILD.

Full size image

Data collection

Clinical data, laboratory tests, chest CT/HRCT, pulmonary function tests, treatment regimens and outcomes were retrospectively extracted from electronic medical records and telephone interviews. These data included basic patient information (age, sex, medical history, diagnosis, smoking status, and occupational history), and pulmonary function test (PFT) data, including FVC, FVC% predicted, FEV1, FEV1/FVC, DLCO% predicted, and subsequent hospitalization or outpatient PFTs were also noted. Radiographic information, including all previous chest CT and HRCT scans, was collected. Chest HRCT images were independently evaluated by two thoracic radiologists. Echocardiographic estimates of systolic pulmonary artery pressure (SPAP) were also recorded.

Acute exacerbation of ILD (AE-ILD) was diagnosed based on the following slightly modified diagnostic criteria for AE-IPF20. The definition of AE-ILD was based on an acute worsening or the development of dyspnoea of less than 1 month in duration, chest HRCT showing new bilateral ground-glass opacities and/or consolidation superimposed on fibrosis, and deterioration that was not fully explained by cardiac failure or fluid overload20,21. The ILD patterns were classified according to the criteria of the 2013 ATS/ESR classification of idiopathic interstitial pneumonia (IIPs) and the usual interstitial pneumonia (UIP) pattern according to the criteria of the 2018 ATS/ERS/JRS/ALAT collaboration to develop a clinical practice guideline for the diagnosis and management of IPF9,22. Interobserver agreement for NSIP and UIP between radiologists was good (Cohen’s kappa 0.83). A combined pulmonary fibrosis and emphysema (CPFE) diagnosis was based on the 2022 ATS/ERS/JRS/ALAT guidelines23. The ILD-gender, age and physiologic variables (ILD-GAP) index was defined by data obtained at baseline evaluation24. The last follow-up was in December 2022, and the outcomes were defined as death from all causes.

Statistical analysis

Statistical analysis was performed using IBM SPSS 26 and GraphPad Prism 9 software. Landmark analysis was performed using SAS 9.3. Continuous variables are presented herein as the mean ± standard deviation (SD); Continuous, nonnormally distributed data are presented as the median with interquartile range (IQR); A t test or the Mann‒Whitney U test was used to assess variations in continuous variables between the NSIP pattern and the UIP pattern groups; for the two-group comparisons of binary data, the chi‒square test or Fisher’s exact test was used. We quantified radiologists’ agreement with Cohen’s kappa coefficient. A Kaplan–Meier survival analysis was performed to estimate the survival rates, and factors associated with survival were identified by log-rank tests. Additionally, landmark analyses were used to assess outcomes at 48 months and between 48 and 138 months in the UIP pattern and NSIP pattern groups. A multivariate Cox regression model was used to identify variables independently associated with death and variables associated with a given outcome at P < 0.1 in the univariate survival analysis. A P value < 0.05 was considered to indicate statistical significance.

Results

Demographic and clinical characteristics and outcome in fibrotic ILD

A total of 194 patients diagnosed with non-IPF fibrotic ILD (FILD) were reviewed, including 117 with connective tissue disease-related interstitial lung disease (CTD-ILD), 30 with microscopic polyangiitis-associated ILD (MPA-ILD), 23 with idiopathic nonspecific interstitial pneumonia (iNSIP), 16 with interstitial pneumonia with autoimmune features (IPAF), 7 with occupational ILD and 1 with fibrotic hypersensitivity pneumonitis (FHP) (Fig. 1). CTD-ILD was the most common subgroup of FILD in our study, including 35 RA-ILD, 31 PM/DM-ILD, 21 pSS-ILD, 15 SSc-ILD, 12 SLE-ILD and 3 MCTD-ILD (Fig. 1).

The demographics and main characteristics of FILD are summarized in Table 1. The study population was predominantly female (62.9%), with a mean age of 66.6 ± 13.3 years (IQR, 58.0–76.0). A total of 50 (25.8%) patients had a history of smoking. One hundred and twenty-one patients underwent PFTs at the initial presentation, and the most common manifestation was restrictive ventilation dysfunction with diffusion dysfunction, with a mean FVC% predicted and DLCO% predicted of 79.0 ± 18.8 and 62.0 ± 19.5, respectively. The ILD-GAP Stages I-IV were 72 (59.5%), 38 (31.4%), 8 (6.6%) and 3 (2.5%), respectively. The HRCT pattern was mainly nonspecific interstitial pneumonia (NSIP) (59.18%), followed by usual interstitial pneumonia (UIP) (26.53%). A total of 51 patients had the UIP HRCT pattern, including 33 with UIP, 16 with probable UIP, and 2 who were indeterminate for UIP. The median follow-up period was 46 months (IQR, 24.0-65.3), 32 patients experienced at least one AE of ILD during the follow-up process, and 41 (21.1%) patients died from all causes.

Table 1 Fibrotic ILD patients’ baseline characteristics.
Full size table

Comparison of demographic and clinical characteristics between the NSIP pattern and UIP pattern

The demographic and clinical characteristics of the NSIP and UIP groups are summarized in Table 2. The age at diagnosis of ILD (64.4 ± 13.5 vs. 70.0 ± 11.6, P = 0.001) and the proportion of males, smokers and AEs (13.5% vs. 27.5%, P = 0.026) were significantly higher in the UIP group than in the NSIP group. However, the FVC% predicted (81.3 ± 17.9 vs. 70.9 ± 20.8, P = 0.011) and DLCO% predicted (63.5 ± 19.8 vs. 52.0 ± 20.5, P = 0.012) were significantly lower in the UIP pattern group.

The ILD-GAP stages I-IV were 62 (70.5%), 22 (25.0%), 4 (4.5%) and 0 (0%) in the NSIP pattern group and 7 (24.1%), 16 (55.2%), 3 (10.3%) and 3 (10.3%) in the UIP pattern group, respectively.

Eighteen (13.5%) and 14 (27.5%) patients experienced at least one AE during the follow-up in the NSIP group and UIP group, respectively.

Table 2 Comparison between the NSIP pattern group and UIP pattern group in clinical baseline characteristics.
Full size table

Survival

The univariate and multivariate analyses of the risk of all-cause mortality in patients with FILD are summarized in Table 3. The median follow-up time of the 194 FILD patients was 46.0 (24.0-65.3) months, and all-cause mortality was observed in 41 (21.1%) patients during follow-up. The median survival time was 105 months, and the 1-, 3- and 5-year overall survival rates were 97.9%, 93.9% and 77.2% for patients with FILD, respectively. The Kaplan‒Meier survival analysis showed that the IPF group (hazard ratio (HR) 1.97, 95% confidence interval (CI) 1.05–3.70, P = 0.011) had a worse survival rate than the non-IPF FILD group (Fig. 2A). The subgroup analysis showed that CTD-ILD (HR 0.51 95% CI 0.26–0.98 P = 0.036) had a better survival than other types of FILD (Fig. 2B), with statistically significant differences.

In the univariate analysis, survival was found to be significantly shorter for the patients who had experienced AEs (HR 3.98 95% CI 1.88–8.43 P < 0.001) (Fig. 3), combined with lung cancer (HR 2.36 95% CI 0.70–7.95 P = 0.044), chest HRCT had honeycomb (HR 1.88 95% CI 0.95–3.74 P = 0.042), DLCO lower than 50% predicted (HR 2.49 95% CI 0.97–6.44 P = 0.050) and those with an SPAP ≥ 37 mmHg (HR = 2.06, 95% CI: 1.02–4.16, P = 0.019). However, treatment with corticosteroids (HR 0.28, 95% CI 0.15–0.53, P < 0.001) or a combination of corticosteroids and immunosuppressants (HR 0.26, 95% CI 0.14–0.49, P < 0.001) were found to prolong survival. In the multivariate Cox analysis, AEs of ILD (HR 3.72 95% CI 2.00-6.93 P < 0.001) and combined with lung cancer (HR 3.12 95% CI 1.27–7.65 P = 0.013) remained as significant factors that were independently associated with shorter survival, and receiving corticosteroids combined with immunosuppressants (HR 0.27, 95% CI 0.12–0.61, P = 0.002) was independently associated with prolonged survival. Although there was no statistically significant difference in overall survival between the UIP pattern and NSIP pattern on chest HRCT scan (Fig. 4A), landmark analyses showed that the survival of FILD patients with the UIP pattern (HR 3.07 95% CI 1.37–6.86 P = 0.006) was significantly shorter than that of FILD patients with the NSIP pattern 48 months after ILD diagnosis (Fig. 4B).

Table 3 Univariate and multivariate analyses of the risk of all-cause mortality in fibrotic ILD patients (from the time of ILD diagnosis).
Full size table
Fig. 2
figure 2

A Kaplan–Meier survival curves of the non-IPF FILD and IPF groups. B Kaplan–Meier survival curves of the CTD-ILD and other type FILD groups. IPF: idiopathic pulmonary fibrosis; FILD: fibrotic interstitial lung disease; CTD-ILD: connective tissue disease-associated interstitial lung disease.

Full size image
Fig. 3
figure 3

Kaplan–Meier survival curves of the patients who had at least one AE and those who did not have any AEs of their FILD.

Full size image
Fig. 4
figure 4

A Kaplan–Meier survival curves of patients with the UIP pattern and NSIP pattern on HRCT in the FILD group. B Landmark analyses (0 to 48 and 48 to 138 months) of patients with the UIP pattern and NSIP pattern on HRCT in the FILD group. UIP: usual interstitial pneumonia; NSIP: nonspecific interstitial pneumonia.

Full size image

Discussion

This retrospective study screened 194 non-IPF fibrotic ILD patients. CTD-ILD (60.3%) was the most common subgroup of FILD and was associated with better survival than other types of FILD in our study. The radiologic pattern of FILD was mainly NSIP, followed by UIP. The age at diagnosis of ILD and the proportion of males, smokers and AEs were significantly higher in the UIP pattern group than in the NSIP pattern group. Furthermore, landmark analyses showed that the survival of FILD patients with the UIP pattern was significantly shorter than that of patients with the NSIP pattern 48 months after ILD diagnosis. The multivariate Cox analysis showed that AEs and combined with lung cancer were significant factors that were independently associated with shorter survival of FILD patients, and receiving corticosteroids combined with immunosuppressants was associated with prolonged survival.

As reported previously24,25,26, the prognosis of CTD-ILD is better than that of other types of FILD patients, which may be related to the main pulmonary inflammation rather than to fibrosis and thus, the condition is likely to respond to corticosteroids and other immunosuppressive agents27. In addition, the majority of CTD diagnoses occur significantly earlier than the diagnosis of ILD, which can enable better management and early treatment of CTD-ILD and is also one of the reasons for the relatively good prognosis of CTD-ILD. In our present study, only DLCO < 50% predicted was considered associated with poor prognosis, while low FVC did not reach statistical significance. Previous studies have shown that low FVC and/or low DLCO are risk factors for poor prognosis or progression to progressive pulmonary fibrosis (PPF) in FILD26,28,29,30,31. Due to the measurement variability of DLCO within patients and varying techniques across pulmonary function laboratories, low FVC values may be more commonly used in clinical work to evaluate patient prognosis. However, approximately 20% of CPFE patients in our study had FVC values within the normal range, while DLCO was significantly reduced23. This suggests that DLCO is more valuable for predicting prognosis in CPFE patients than FVC.

Our present study also showed that patients with the UIP pattern on chest HRCT had a poorer survival rate than those with the NSIP pattern 48 months after ILD diagnosis, but there was no statistically significant difference in overall survival. Although the UIP pattern on chest HRCT is generally considered associated with worse outcomes, especially in patients with specific radiological features, such as honeycombing and bronchiectasis32,33, in fact, among patients with fibrotic lung disease, the extent of fibrosis often influences outcomes and prognosis more strongly than chest imaging patterns1,7,34. Recent studies show that a long asymptomatic phase often occurs before physiologic impairments develop in patients with IPF and other fibrotic ILDs35. The increased use of CT screening has led to more frequent recognition of interstitial lung abnormalities (ILAs), and some ILAs likely represent early UIP36,37. Most FILD of the NSIP pattern is considered primarily inflammatory, at least in its early stages, while the UIP pattern is mainly characterized by pulmonary fibrosis27. If inflammation in the lungs is not controlled in a timely manner in the early stages of the disease, it may lead to premature death of patients, such as MDA5-positive myositis. In addition, the use of glucocorticoids and immunosuppressants is also prone to secondary complications such as opportunistic infections. Therefore, patients may have a better prognosis in the early stages of UIP, but over time, the progression of pulmonary fibrosis is more likely to occur, leading to a worse prognosis in the later stages; this finding can also explain why significant crossover was observed between the UIP pattern and NSIP pattern in this study.

AE is a life-threatening event in patients with IPF, and the median survival time of these patients is only 3–4 months38. However, accumulating evidence indicates that patients with ILD other than IPF also develop AEs during their clinical course and have an extremely poor prognosis with high mortality38,39,40. Studies have shown that AEs of FILD are often reported in patients with a histological or radiological pattern of UIP38,41,42, which is consistent with our findings. Univariate and multivariate Cox regression analyses also indicated that an AE is an independent risk factor for poor prognosis in FILD. In microscopic polyangiitis-associated ILD, a lower FVC% predicted was an independent prognostic factor for patients to develop an AE41. In fibrotic HP, a lower DLCO value and radiologic UIP pattern at diagnosis were associated with the development of AEs43. Some possible common factors include the UIP pattern, low PFT, and extensive pulmonary fibrosis, which can lead to the occurrence of AEs in these patients. Once an AE event occurs, the survival rate will significantly decrease, so more attention should be given to FILD patients who have poor PFT, UIP pattern and extensive pulmonary fibrosis on chest HRCT, especially if the patient experiences an AE.

Our retrospective cohort study had several limitations. First, this was a single-centre study with a small sample size; therefore, it is possible that our findings would not be consistent with findings in larger patient populations. Second, due to the calculation of the survival time starting from the time of diagnosis of FILD, some patients seek medical attention at other hospitals in the early stage and lack chest HRCT scans in the early stage, leading to an inevitable time bias. Finally, pulmonary function tests were performed for some patients, and related variables were not analysed by multivariate survival analysis. The results of the multivariate analysis of risk factors for prognosis should be carefully interpreted. Thus, larger-scale prospective studies and clinical randomized controlled trials are needed for further evaluation.

Conclusions

In conclusion, CTD-ILD was the most common subgroup of FILD in our study and is associated with longer survival than other types of FILD. Patients with the UIP pattern are older, may have a poorer prognosis in the later stage, and have a higher proportion of AEs than patients with the NSIP pattern. AE and combined with lung cancer are independent risk factors for the prognosis of FILD patients, more attention should be given to FILD patients who have poor PFT, UIP pattern and extensive pulmonary fibrosis on chest HRCT, especially if the patient experiences an AE.