Idiopathic pulmonary fibrosis (IPF) is an interstitial lung disease (ILD) characterized by chronic and progressive fibrosis of the lung parenchyma without known causes. It is a common form of ILD with a poor prognosis [1]. The overall mean survival is 4 years [2], and the median survival is only 3–4 months when acute exacerbation occurs [3]. In a pooled analysis, the outcome of patients with IPF who have respiratory failure requiring mechanical ventilation is poor, with a 3-month mortality of 94% [4]. The 2011 international guideline suggests not using mechanical ventilation in patients with IPF and respiratory failure who are not candidates for lung transplantation [5], consistent with subsequent reports on mechanical ventilation [3,6]. However, most previous studies reporting high mortality rates of patients with IPF on mechanical ventilation were conducted at a single center with a small sample size between the 1980s and the early 2000s [4,5]. Recently, short-term mortality of patients on mechanical ventilation has decreased over time, which implies an improved quality of intensive care unit (ICU) care [7,8]. The prognosis of patients with IPF admitted to the ICU has also improved [9]. In addition, the mortality of patients with IPF receiving mechanical ventilation has decreased, becoming lower than previously reported [10]. This discrepancy between the existing guideline and recent literature makes it difficult for clinicians to decide on the application of mechanical ventilation [11,12] or provide proper information to patients or their surrogates, particularly in situations where lung transplantation is not an option. Thus, the aim of this study was to investigate changes in the use of mechanical ventilation in patients with IPF and their outcomes over time.

During the study period, 2,536 patients with pulmonary fibrosis were treated with mechanical ventilation. Approximately half of them were excluded because of concomitant CTDs or exposure to injurious agents. A total of 1,227 patients were included in the final analysis (Fig. 1). The annual number of patients (prevalent cases) with IPF who received mechanical ventilation increased from 55 in 2011 to 189 in 2019 (p < 0.001). The annual number of patients with IPF also increased from 1,874 in 2011 to 6,316 in 2019 (p < 0.001). However, the annual rate of patients with IPF receiving mechanical ventilation among total patients over the study period did not change significantly (p = 0.143). Overall, 3.5% of patients with IPF were treated with mechanical ventilation (Fig. 2).

In this study, we investigated patients with IPF treated with mechanical ventilation between 2011 and 2019 in Korea. The number of patients with IPF treated with mechanical ventilation increased over time. Several changes in baseline characteristics including age and comorbidities were observed. However, their ratios to all patients with IPF and their mortality did not change significantly.

Recent studies on large cohorts using administrative data from the United States have revealed that hospital mortality of patients with IPF requiring mechanical ventilation ranges from 39.4 to 55.7%, much lower than those reported in previous studies. These studies have also revealed that hospital mortality has declined over time [10,14,15]. These aforementioned results are in contrast to the high hospital mortality demonstrated in our study (nearly 70%). In previous studies, patients who had ILD other than IPF, secondary pulmonary fibrosis, or who underwent surgery were included. In our study, we used both the ICD-10 and rare and intractable disease codes to identify only patients with IPF. As rare and intractable disease code was closely related to financial concerns of patients, health service providers, and government agencies, the registration of the code was performed more strictly. The diagnostic validity was higher in rare and intractable disease claim data than in general disease claim data [16]. In addition, we excluded patients with secondary pulmonary fibrosis or ILD and those who underwent surgery. Acute respiratory failure in other ILDs has demonstrated a better prognosis than that in IPF [17–19]. In patients with IPF who are mechanically ventilated, postoperative respiratory failure is associated with a lower hospital mortality [20]. Therefore, mortality rates of our study and previous studies seem to be different due to difference in the study population. Indeed, our results are similar to those of a study using clinical data of patients with IPF between 2015 and 2017 in Korea [21]. Furthermore, in contrast to previous studies, our study observed no significant temporal changes in mortality among patients with IPF. Acute exacerbation is the most common reason for mechanical ventilation in patients with IPF [22–24]. While guidelines recommend steroid administration for acute exacerbation of IPF, the quality of evidence is very low [3,5]. Recent trials evaluating the efficacy of agents other than steroids have also failed [25,26]. Given the absence of advancement in treatment for acute exacerbation of IPF, the stagnation in prognosis is not surprising.

In our study, the 90-day mortality declined during the study period, although the overall mortality rate remained high at 85.3%. Approximately half of the patients who survived in the hospital died within approximately two months. In pooled analysis of past data, the hospital mortality was 87% and the 3-month mortality was 94% [4]. When compared to pooled analysis results, the 90-day mortality demonstrated less improvement and the difference between hospital and 90-day mortalities widened in our study. This suggests that even if patients were discharged alive, they would still receive life-sustaining treatment rather than recover and return to their daily lives.

In the multivariable analysis, the year of mechanical ventilation was not associated with mortality. Age was the only independent risk factor for increased mortality. Comorbidity appeared to be associated with increased mortality, although such association was not statistically significant. This again demonstrated that there was little improvement in the prognosis of patients with IPF with respiratory failure over time.

The annual rate of patients with IPF receiving mechanical ventilation did not change significantly over the study period. This might reflect the disease course of IPF, wherein a certain proportion of patients would experience respiratory failure annually.

Strengths of our study were the utilization of NHIS data from the last 9 years and the rigorous case definition. These strengths enabled a large sample size, detection of annual changes, and precise selection of patients with IPF on mechanical ventilation.

This study had several limitations. First, this study was not free of inherent weaknesses resulting from the utilization of administrative data, including selection bias and accuracy of diagnosis codes [27]. However, selection bias would be low because the NHIS provides universal coverage to 97% of Koreans [13]. The requirement of rare and intractable disease codes for the case definition could compensate for the shortcomings of the ICD-10 codes and improve the accuracy of diagnosis. Second, the diagnostic criteria for IPF were revised twice (in 2011 and 2018) during the study period [5,28]. Patient characteristics such as chest computed tomography findings and frequency of patients who underwent surgical lung biopsy might differ. Third, information on causes of respiratory failure in patients requiring mechanical ventilation was not provided. Several studies have demonstrated that the most common cause of respiratory failure is acute exacerbation, followed by respiratory tract infection, and that the prognosis is worse in patients with acute exacerbation than in those with respiratory tract infection [22–24]. Another study has also reported that the prognosis of patients with ILD and infection is better than that of patients with acute exacerbation. In addition, the prognosis of patients with ILD exacerbation triggered by infection is better than that of patients with idiopathic or non-infection-trigged exacerbation [29]. For a more accurate analysis, the annual hospital mortality rate needs to be adjusted to the cause of respiratory failure each year. Fourth, other clinically important variables including general ICU prognostic measures and IPF-related parameters such as the extent of fibrosis on chest computed tomography, pulmonary function test results, and blood gas analysis were not provided. They are presumed to be closely associated with clinical outcomes. However, they could not be extracted from the NHIS database. We did not provide data on medications usage, particularly antifibrotic agents. The use of antifibrotic agents has recently increased in patients with IPF and improved the prognosis [30,31]. Moreover, antifibrotic agents are associated with a reduced risk of acute exacerbation and mortality after exacerbation [32]. Therefore, the prevalence and prognosis of patients with IPF requiring mechanical ventilation could be affected. In Korea, health insurance coverage for the antifibrotic drug, pirfenidone, was commenced in 2015. However, hospital, 30-day, and 90-day mortality rates of patients with IPF placed on mechanical ventilation did not differ significantly between 2011–2014 and 2015–2019. Short-term mortality did not change meaningfully from 2015 to 2019 either (data not displayed). Thus, despite an overall increase in the use of antifibrotic agents, mortality rates of patients with IPF receiving mechanical ventilation in Korea did not change significantly. Fifth, instead of utilizing individual comorbidities for mortality adjustments, we relied on the CCI. However, it is worth noting that within the CCI, both congestive heart failure and peptic ulcer disease are assigned a score of 1. Nevertheless, intuitively, congestive heart failure should have a more detrimental impact on mechanically ventilated patients than peptic ulcer disease. The increase in number of patients with dementia or metastatic solid tumor might have a negative impact on prognosis. Moreover, by treating CCI ≥ 3 as a single variable, we were unable to discern changes in the distribution of scores within 3 or higher range or their potential impact on mortality. The decline in the prevalence of peptic ulcer disease warrants further consideration. In 2011, 2015, and 2022 guidelines, suggestions for antiacid treatments such as proton pump inhibitors (PPIs), changed slightly [5,6,33]. Therefore, the use of PPIs might have changed over the study period. The requirement of a diagnostic code of peptic ulcer disease for prescribing PPIs introduces the possibility that changes in prescribing behavior might have influenced the recorded frequency of peptic ulcer disease irrespective of the true prevalence of the condition. In this case, the CCI score might also have been affected. Sixth, this study’s results cannot be applied to patients who have undergone lung transplantation or placed on extracorporeal life support. Considering the increasing number of lung transplantation cases in Korea, the prognosis for all patients with IPF who developed respiratory failure might have improved over time [34]. Finally, we did not provide information on conditions of patients who were discharged alive, such as performance status, dependency on mechanical ventilation, or residency in the nursing care facility.

In summary, in patients with IPF treated with mechanical ventilation, hospital and 30-day mortality rates did not change significantly during the study period. Although the 90-day mortality rate decreased over time, it remained high. Despite improvements in ICU care and ventilator management in the general population, the prognosis of patients with IPF receiving mechanical ventilation was not much improved. Given the unfavorable prognosis and the limited medical resources, the use of mechanical ventilation in this population should be approached cautiously.