Incidence and impact of extra-pulmonary organ failures on hospital mortality in acute exacerbation of idiopathic pulmonary fibrosis

To evaluate hospital mortality and associated risk factors for acute exacerbations of idiopathic pulmonary fibrosis (AEIPF). Emphases were put on incidence and impact of extra-pulmonary organ failures. Patients diagnosed with AEIPF from July 2014 to September 2018 were enrolled. Clinical data were collected. Acute physiology and chronic health evaluation II (APACHE II) and simplified acute physiological score II (SAPS II) were calculated. Extra-pulmonary organ failures were diagnosed upon criteria of sequential organ failure assessment (SOFA). Forty-five patients with AEIPF were included. Eighteen patients (40.0%) developed extra-pulmonary organ failures, and 25 patients (55.6%) died during hospitalization. Serum C-reactive protein (CRP) (p = 0.001), SAPS II (p = 0.004), SOFA (p = 0.001) were higher, whereas arterial oxygen pressure (PaO2)/ fractional inspired oxygen (FiO2) (p = 0.001) was lower in non-survivors than survivors. More non-survivors developed extra-pulmonary organ failures than survivors (p = 0.002). After adjustment, elevated serum CRP (OR 1.038, p = 0.049) and extra-pulmonary organ failure (OR 13.126, p = 0.016) were independent predictors of hospital mortality in AEIPF. AEIPF had high hospital mortality and occurrence of extra-pulmonary organ failure was common. Elevated serum CRP and extra-pulmonary organ failure had predictive values for mortality.

Idiopathic pulmonary fibrosis (IPF) is a chronic, progressive and fatal fibrosing interstitial pneumonia with unknown etiology 1 . The clinical course of IPF is highly heterogeneous and unpredictable. For some patients, the disease remains stable or progresses slowly over years. But a small amount of patients may develop sudden exacerbations of respiratory function impairment, referred to as acute exacerbations of idiopathic pulmonary fibrosis (AEIPF) 2 , resulting in refractory hypoxemia and respiratory failure. AEIPF is the leading cause of death in IPF, with the hospital mortality up to 60% 2,3 .
AEIPF shared similar pathophysiological characteristics and clinical need with another severe condition, acute respiratory distress syndrome (ARDS) 4 . A majority of patients with AEIPF needed admission to intensive care unit (ICU) for severe complications and multiple organ failures. AEIPF was also considered as an inflammation associated disease 5 . Despite the high hospital mortality rates, the prognostic factors for short-term mortality in patients with AEIPF remained uncertain. Few studies described the conditions of extra-pulmonary organs in patients with AEIPF. Therefore, we conducted this retrospective study in an interstitial lung disease center in China, to investigate the hospital mortality and associated risk factors of hospital mortality in patients with AEIPF. Acute physiology and chronic health evaluation II (APACHE II) 6 , simplified acute physiological score II (SAPS II) 7 and sequential organ failure assessment (SOFA) system 8 were used to evaluate the organ conditions and assess the incidence and impact of extra-pulmonary organ failures on hospital mortality in AEIPF.  (Fig. 2). High-dose corticosteroids (0.5-1 g/d methylprednisolone or its equivalent) for 3-5 days was used in 26 patients (57.8%) after diagnosis. Non-invasive ventilation (NIV) was used in 25 patients (55.6%), and 2 of them were

Comparison of clinical characteristics between survivors and non-survivors
As presented in  Fig. 3 and Fig. 4).

Discussion
The study showed a majority ( www.nature.com/scientificreports/ mortality in AEIPF. These findings implicated an association between systemic inflammation and poor outcome in AEIPF. The outcome of AEIPF was poor, and no effective treatment options were available 3,[9][10][11] . A 10-year observational study showed that hospital mortality of AEIPF was 56.9% and 3-month mortality was 63.8% 11 . Saydain et al. reported that 61% of patients with AEIPF died during hospitalization and 92% of survivors died within 2 months after discharge 9 . The study from Song et al. also showed that hospital mortality in AEIPF was approximately 50.0% 3 , and in patients who were admitted to ICU, it exceeded 90% 12 . Various potential therapies had been reported in patients with AEIPF 13,14 . Unfortunately, these studies were mostly with small sample sizes and uncontrolled. Still, controversies remained over the use of corticosteroids and immunosuppressive agents in patients with AEIPF. A study from Japan demonstrated that early use of high-dose steroids could significantly increase survival rate 15 . But study from Papiris et al. showed corticosteroids therapy had no obvious advantage to 1-year survival 16 . International guidelines made a weak recommendation for the use of corticosteroids in AEIPF 1 . Results from clinical trials of nintedanib and pirfenidone suggested that IPF therapies may be helpful to prevent the development of AEIPF, but these data were from observational study [17][18][19] . A well-designed randomized clinical trial is still needed. In the current study, CRP was identified as an independent risk factor of hospital mortality in AEIPF, which was consistent with previous studies 3 . Viral infection was reported to cause exacerbation of established Table 2. Logistic regression analysis of risk factors of hospital mortality in AEIPF patients. Age, sex, PaO 2 / FiO 2, CRP, presence of extra-pulmonary organ failure and SAPS II score were included in the multivariate logistic regression analysis. AEIPF = acute exacerbation of idiopathic pulmonary fibrosis; PaO 2 = arterial oxygen tension; FiO 2 = fractional inspired oxygen; WBC = white blood cell; CRP = C-reactive protein; APACHE II = Acute physiology and chronic health evaluation II; SAPS II = simplified acute physiological score II; SOFA II = sequential organ failure assessment.  Previous studies had noted that extrapulmonary organ failures had increased the mortality risk of ARDS 23 . Sepsis syndrome/ multiple organ failures (MOFs) were the main cause of ARDS deaths 24 . Besides, higher APACHE II scores were reported to be associated with increased mortality in ARDS 25 . Similarly, this study showed that the presence of extra-pulmonary organ failure was a prognostic factor of AEIPF. It seemed that, during AEIPF, inflammation in lung triggered enhanced systemic inflammatory response leading to MOFs, and then worsened the clinical outcomes. The values of life supportive care and anti-inflammation therapy were also to be elucidated.
Notably, in our cohort, a majority of non-survivors (22/25, 88.0%) received NIV, and only 2 of them were performed IMV. Whether to withhold life sustaining therapies in patients with AEIPF is indeed a question. The available data suggested poor outcome for IPF patients who required NIV or IMV, although they temporarily corrected PaO 2 /FiO 2 and reduced respiratory rate 26,27 . The hospital mortality of AEIPF who required MV was about 90% 28,29 . A systematic review comprising 9 single-center studies reported that 87.0% of IPF patients who had received IMV died in hospital 30 . NIV seemed more promising. Rush et al. revealed AEIPF patients who underwent NIV had a lower mortality rate (30.9%) compared to those who underwent IMV (51.6%) 31 . An early application of NIV when the general conditions of patients were less severe and the prevention of ventilatorassociated pneumonia would benefit the survival 4 . However, ventilator-induced lung injury (VILI), which was an independent predictor for NIV failure 32 , should be acknowledged. In addition, excessive spontaneous effort and deflation injury resulted from NIV also caused lung injury 33 .
Several studies evaluated clinical predictors of mortality in patients with AEIPF, but a universally accepted predicted model did not exist. Akira et al. reported worse survival was associated with diffuse pattern on HRCT 34 . Lower forced vital capacity (FVC) and diffusion capacity for carbon monoxide (DLCO) before AE was observed in non-survivors compared to survivors 35 . Several prognostic biomarkers identified to be related with survival included increased Krebs Von den Lungen-6 (KL-6) 11 , higher hyaluronan 36 and anti-heat shock protein 70 autoantibodies 37 . Recently, a staging system with four parameters, serum lactate dehydrogenase (LDH) level, serum KL-6, PaO 2 /FiO 2 , and total extent of abnormal findings on HRCT, was proposed to predict prognosis 11 . Prognostic model which considers pathologic physiology of various organs are acquired to provide guidance on early intervention.
This study had several limitations. First, the data was retrospectively collected from a single-center study. The selection bias was evitable and the small sample size weakened the power of the study. Second, the pulmonary function tests data were incomplete partly because of the critical illness so we could not assess the association between the last pulmonary function data before AE with the mortality.

conclusion
The outcome of AEIPF remained poor with a high in-hospital mortality and occurrence of extra-pulmonary organ failure was common. Elevated CRP and extra-pulmonary organ failure were independently associated with hospital mortality. Further investigations were still needed to clarify the association between systemic inflammation and AEIPF.

Data collection
Clinical information was collected, including demographics, smoking history, clinical symptoms, duration from diagnosis of IPF to the occurrence of AE, treatment course and lengths of stay in RICU and hospital. The laboratory data included PaO 2 / FiO 2 , blood routine, CRP, erythrocyte sedimentation rate (ESR), LDH and lymphocyte subset etc. APACHE II score 6 and SAPS II 7 in the first 24 h after admission were calculated. The presence of extra-pulmonary organ failures (coagulation, liver, cardiovascular, central nervous system and renal) were defined according to criteria of SOFA system 8 . Chest HRCTs of all patients in supine position were performed at admission. The newly appeared parenchymal abnormalities were classified into peripheral, multifocal and diffuse patterns based on the distributions 34 . Treatment information included usages and courses of corticosteroids, acetylcysteine, pirfenidone/nintedanib and respiratory support. All patients were administrated with oxygen therapy firstly. NIV was initiated when respiratory rate was more than 30 breaths/min or PaO 2 /FiO 2 was less than 200mmHg 26 . The primary endpoint was hospital mortality. The length of stay in hospital was calculated from the date of admission to the date of discharge or death.
Statistical analysis. Data was presented as mean with standard deviation for continuous variables or frequencies with percentages for categorical variables. T-test or the Mann-Whitney U-test was used for continuous variables. Chi-square test was used for categorical variables. Binary logistic regression analysis was used to evaluate the association between variables and mortality. Survival analysis was estimated using Kaplan-Meier curves and was compared between groups using the log-rank test. For continuous variables, the median was used as cut-off value. P value less than 0.05 was considered statistically significant. All statistical analysis was conducted through IBM SPSS Statistics version 23.0 (SPSS, Inc., Chicago, IL, USA) and Prism 8.0 (GraphPad, Inc., La Jolla, CA, USA).

Ethical considerations.
The patient data for this study were recorded by the authors. This study was approved by Ethics Committee of Nanjing Drum Tower Hospital, Clinical College of Nanjing Medical School, Jiangsu, China. Written informed consents were obtained from all subjects and the study was performed in accordance with the relevant guidelines/regulations.

Data availability
Data can be submitted by the corresponding author in case of a request.