Trends and management of acute respiratory failure in hospitalized patients: a multicenter retrospective study in South Korea

Won Jin Yang; Yong Jun Choi; Kyung Soo Chung; Ji Soo Choi; Bo Mi Jung; Jae Hwa Cho

doi:10.4266/acc.004728

Articles

Page Path: HOME > Acute Crit Care > Volume 40(2); 2025 > Article

Original Article Epidemiology Trends and management of acute respiratory failure in hospitalized patients: a multicenter retrospective study in South Korea: Acute and Critical Care 2025;40(2):171-185.
DOI: https://doi.org/10.4266/acc.004728
Published online: May 28, 2025

Won Jin Yang^1,*

, Yong Jun Choi^1,*

, Kyung Soo Chung²

, Ji Soo Choi³

, Bo Mi Jung¹

, Jae Hwa Cho¹

¹Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Gangnam Severance Hospital, Yonsei University College of Medicine, Seoul, Korea

²Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Severance Hospital, Yonsei University College of Medicine, Seoul, Korea

³Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Yongin Severance Hospital, Yonsei University College of Medicine, Seoul, Korea

Corresponding author: Jae Hwa Cho Division of Pulmonary and Critical Care Medicine, Department of Internal Medicine, Gangnam Severance Hospital, Yonsei University College of Medicine, 211 Eonju-ro, Gangnam-gu, Seoul 06273, Korea Tel: +82-2-2019-3455, E-mail: jhcho66@yuhs.ac

*These authors contributed equally to this study as co-first authors.

• Received: December 23, 2024 • Revised: April 28, 2025 • Accepted: May 2, 2025

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

890 Views
56 Download

prev next

Full Article

Download PDF

Abstract
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
KEY MESSAGES
NOTES
SUPPLEMENTARY MATERIALS
REFERENCES

Abstract

Background
Acute respiratory failure (ARF) is the leading cause of hospitalization and is associated with in-hospital mortality. This study aimed to elucidate the epidemiology and clinical outcomes of ARF.
Methods
We retrospectively screened patients admitted to three hospitals in South Korea between January 2018 and December 2022. We included individuals aged 18 years, diagnosed with either type 1 ARF (arterial oxygen partial pressure [PaO₂] <60 mm Hg) or type 2 ARF (arterial carbon dioxide partial pressure (PaCO₂) >45 mm Hg) with a pH of <7.35, or diagnosed with the combined-type ARF.
Results
Among the 768,700 hospitalized patients, 33,278 (4.3%) developed ARF. The most common cause of ARF was sepsis (15,757 patients, 47.3%), and the most frequent comorbidity was malignancy (15,403 patients, 43.6%). Among ARF patients, 15,671 (47.1%) required intensive care unit transfer, while 8,980 (27.0%) experienced in-hospital mortality. Over 5 years, the proportion of ARF patients aged 80 years and older has shown a consistent annual increase (coefficient, 0.085 and P_trend <0.001). Concurrently, the in-hospital mortality rate exhibited an upward trend, increasing from 25.5% in 2018 to 29.3% in 2022 (coefficient, 1.017 and P_trend<0.001). Among the respiratory support methods used for patients with ARF over the 5-year period, high-flow nasal cannula usage steadily increased (coefficient, 4.137 and P_trend<0.001), whereas the use of invasive mechanical ventilation declined (coefficient, –0.983 and P_trend<0.001).
Conclusions
ARF frequency and in-hospital mortality rates are increasing, driven by various etiologies. Despite these trends, research on the epidemiology and individualized treatments for older patients is limited, highlighting the need for nationwide prospective multicenter studies.
Key Words: epidemiology; intensive care units; Korea; mortality; respiratory failure; respiratory insufficiency

INTRODUCTION

Acute respiratory failure (ARF) is a severe condition characterized by the inability to maintain normal oxygen delivery to tissues or carbon dioxide removal from tissues [1]. ARF is traditionally defined by an arterial oxygen partial pressure (PaO₂) of <60 mm Hg (type 1), an arterial carbon dioxide partial pressure (PaCO₂) of >45 mm Hg with a pH of <7.35 (type 2), or a combination of both. ARF is associated with increased healthcare resource utilization, high in-hospital mortality rates, and significant ethical considerations regarding end-of-life decisions. It is also a common reason for admission and transfer to the intensive care unit (ICU) [2,3]. In the United States, 3,213,030 patients were discharged due to ARF in 2017, with an incidence rate of 1,275 cases per 100,000 person-years [4].

ARF treatment is divided into the management of the underlying cause and the provision of supportive oxygen therapy [5]. Supportive oxygen therapy for ARF follows a phased escalation strategy based on various ventilatory and non-ventilatory interventions. The primary goals of treatment are to stabilize the patient while addressing the underlying causes of respiratory failure and preventing or mitigating potential lung damage from interventions such as mechanical ventilation and to reverse the causes of acute respiratory injury [6]. Respiratory support options include conventional oxygen therapy, high-flow nasal cannula (HFNC), non-invasive ventilation (NIV), invasive mechanical ventilation (IMV), and extracorporeal membrane oxygenation (ECMO) [7-13]. However, clinical data on the use of these devices and their associated trends are lacking. Furthermore, large-scale research investigating the frequency and characteristics of type 1, type 2, and combined-type ARF in Korea is limited [1]. This multicenter study aimed to investigate the frequency, causes, and current management practices of ARF.

MATERIALS AND METHODS

Ethics

This study was approved by the Institutional Review Board of Gangnam Severance Hospital (No. 2023-0928-003). As this was a retrospective study that utilized de-identified data, the requirement for informed consent was waived.

Study Design and Data Source

From January 2018 to December 2022, all patients admitted to three hospitals in South Korea (Severance Hospital: 2,499 beds, Gangnam Severance Hospital: 766 beds, and Yongin Severance Hospital: 658 beds) were screened through a retrospective review of electronic medical records (EMRs) using the Severance Clinical Research Analysis Portal (SCRAP) at three hospitals. Patients aged 18 years who were diagnosed with type 1 ARF (PaO₂ <60 mm Hg), type 2 ARF (PaCO₂ >45 mm Hg with a pH of <7.35, or combined-type ARF; whose type 2 ARF developed during the progression of type 1 ARF or vice versa; or whose type 1 ARF developed during the progression of type 2 ARF were included in the study [10]. Conversely, patients who were previously diagnosed with chronic respiratory disease (International Classification of Diseases, Tenth Revision [ICD-10] code J96.1 or J96.9) prior to admission and long-term ventilator support, NIV, or oxygen therapy from previous admissions were excluded (Figure 1).

Data Collection and Outcomes

ICD-10 codes, sex, age, and laboratory results (hemoglobin, white blood cell, platelet, total protein, albumin, aspartate transaminase, alanine transaminase, total bilirubin, sodium, potassium, chloride, and arterial blood gas levels) of patients were extracted from the EMR database. Comorbidities were assessed using the Charlson Comorbidity Index (CCI) [14].

Seventeen disease groups were identified as common causes of ARF [15]. The ICD-10 codes for each condition are as follows: J43.0x-J44.x, excluding J43.0, for chronic obstructive pulmonary disease; J45-J46 for asthma; G10-G14, G20-G41, G70-G83, and G90-G99 for neuromuscular disease; I60-I64 and I69 for stroke; C32-C33, C42-C43, and C70-C72 for neurologic tumor; M40-M54 for musculoskeletal disease; S00-S99 and T00-T99 for trauma; I20-I25 for ischemic heart disease; and I09.9, I11.0, I13.0, I13.2, I25.5, I42.0, I42.5-I42.9, P29.0, I40, and I50 for heart failure. Lung lesions such as pneumonia, pulmonary edema, atelectasis, pneumothorax, pulmonary embolism, and pleural effusion were defined based on the radiologists’ interpretation of the results of all radiologic evaluations performed during the hospital stay. Sepsis was defined based on the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3) criteria and using the algorithm described by Shashikumar et al. [16,17], which involves a ≥2-point change in the Sequential Organ Failure Assessment (SOFA) score occurring 48 hours before or 24 hours following the onset of infection [18]. The time of infection was defined as the administration of antibiotics between 24 hours before and 72 hours after blood culture acquisition [18]. Acute respiratory distress syndrome (ARDS) was defined according to the Berlin definition [19]. The Berlin definition includes the following criteria: onset within 1 week of a known clinical insult or worsening respiratory symptoms; bilateral opacities on chest imaging not explained by other conditions; respiratory failure not fully attributed to cardiac failure or fluid overload, requiring objective assessment if no risk factors are present; and impaired oxygenation, defined as a PaO₂/FiO₂ ratio of ≤300 mm Hg with a positive end-expiratory pressure or continuous positive airway pressure of ≥5 cm H₂O.

The SOFA scores were calculated based on six subcategories [20]. For the respiratory SOFA score, when only SpO₂ data were available, a nonlinear imputation strategy proposed by Brown et al. [21] was used to estimate the PaO₂/FiO₂ ratio from the SpO₂/FiO₂ ratio, indirectly deriving the PaO₂ value.

Statistical Analysis

Categorical variables were expressed as frequencies (percentages). Continuous variables were expressed as the mean±standard deviation for normally distributed variables and as the median (interquartile range) for non-normally distributed variables. The normality of continuous variables was confirmed using the Shapiro-Wilk test. Outcomes associated with hospital admissions were assessed at the level of individual admissions, while adjusting for the clustering of multiple admissions within the same patient. P-values for non-normally distributed continuous variables were calculated using a clustered Kruskal-Wallis test with bootstrap sampling. For normally distributed continuous variables and categorical outcomes, Generalized Estimating Equations were used to adjust for within-patient correlation due to repeated admissions. The trend curve was generated using linear regression analysis, and the P for trend was calculated using the proportion trend test. All analyses were performed using R software version 4.4.0 (R Foundation). A P-value of <0.05 was considered significant.

RESULTS

Characteristics of Patients with ARF

Among 768,700 hospitalized patients aged >18 years, 33,278 (4.3%) were diagnosed with ARF (Figure 1). Based on the arterial blood gas analysis results, 13,659 patients (41.0%) had type 1 ARF, 10,459 patients (31.4%) had type 2 ARF, and 9,160 patients (27.5%) had combined-type ARF. The median age of patients with ARF was 68 years (interquartile range, 58–77). Type 1 ARF was significantly more prevalent in older patients compared with type 2 ARF (70 years [59–78] vs. 66 years [57–75], P<0.001) (Table 1). Males (60.0%) were more prevalent than females, accounting for 60.0 % of cases. This trend was consistent across all ARF types (Table 1), with a particularly higher proportion of males observed in type 2 ARF compared to type 1 and combined ARF (61.6%, 58.7%, and 60.1%, respectively; P<0.001). The median body mass index of ARF patients was 23.2 kg/m² (20.5–25.8). Patients with type 2 ARF had a significantly higher BMI compared to those with type 1 and combined ARF (23.6 kg/m² [21.1–26.2], 22.8 kg/m² [20.2–25.4], and 23.1 kg/m² [20.2–26.0], respectively; P<0.001).

In terms of comorbidities assessed using the CCI, patients with type 1 ARF had significantly higher scores compared to those with type 2 and combined ARF (4.0 [2.0–8.0], 3.0 [1.0–5.0], and 4.0 [2.0–8.0], respectively; P<0.001). The most common comorbidities were malignancy, diabetes mellitus (with or without complications), and chronic pulmonary disease, accounting for 46.3%, 31.9%, and 26.6%, respectively. Most comorbidities were more prevalent in patients with type 1 ARF compared to type 2 ARF (Table 1). However, peripheral vascular disease was significantly more common in type 2 ARF patients than in type 1 ARF patients (15.9% vs. 14.0%, P <0.001) (Table 1).

The combined ARF group was further subdivided based on whether the initial presentation was as type 1 ARF, type 2 ARF, or simultaneous onset of both types. The results of this subgroup analysis are presented in Supplementary Tables 1-5.

Causes of ARF

The causes of ARF are presented in Table 2. Of all cases, 20,848 (62.6%) were complicated by more than one underlying cause (Figure 2A). Combined type was caused by more complex factors compared to type 1 or type 2, with a higher number of associated causes (3.0 [1.0–4.0], 2.0 [1.0–4.0], and 1.0 [1.0–3.0], respectively) (Table 2). The predominant causes of all ARF were sepsis (15,757 cases, 47.3%), pleural effusion (8,501 cases, 25.5%), pneumonia (8,298 cases, 24.9%), atelectasis (7,205 cases, 21.7%), and ischemic heart disease (6,911 cases, 20.8%) (Table 2). Similarly, in type 1 ARF, the leading causes were sepsis (7,197 cases, 52.7%), pleural effusion (4,143 cases, 30.3%), pneumonia (4,075 cases, 29.8%), atelectasis (3,372 cases, 24.7%), and ischemic heart disease (2,769 cases, 20.3%) (Table 2). For type 2 ARF, the most frequent causes were sepsis (3,066 cases, 29.3%), ischemic heart disease (2,063 cases, 19.7%), pleural effusion (1,611 cases, 15.4%), trauma (1,567 cases, 15.0%), and atelectasis (1,516 cases, 14.5%). In the combined type, sepsis (5,494 cases, 60.0%), pneumonia (2,974 cases, 32.5%), pleural effusion (2,747 cases, 30.0%), atelectasis (2,317 cases, 25.3%), and ischemic heart disease (2,079 cases, 22.7%) (Table 2).

Biomarker Values of Patients with ARF

Table 3 presents the differences in biomarker values according to the ARF type. The platelet (type 1 ARF vs. type 2 ARF; 181.0 ×10³/μl [107.0–258.5] vs. 173.0 ×10³/μl [111.0–233.0], P<0.001), creatinine (0.9 [0.6–1.4] mg/dl vs. 0.8 [0.6–1.1] mg/dl, P<0.001), and protein (5.9 g/dl [5.2–6.6] vs. 5.6 g/dl [4.9–6.3], P<0.001) levels were higher in type 1 ARF compared with type 2 ARF. By contrast, the hemoglobin (10.4 g/dl [8.8–12.2] vs. 11.3 g/dl [9.6–12.9], P<0.001), white blood cell (9.7 ×10³/μl [6.6–13.8] vs. 11.2 ×10³/μl [8.0–15.0], P<0.001), sodium (137.0 mmol/L [134.0–140.0] vs. 139.0 mmol/L [136.0–141.0], P<0.001), potassium (4.1 mmol/L [3.6–4.5] vs. 4.2 mmol/L [3.8–4.6], P<0.001), and albumin (3.2 g/dl [2.8–3.7] vs. 3.3 g/dl [2.8–3.7], P<0.001) levels were significantly higher in type 2 ARF compared with type 1. The C-reactive protein levels were significantly elevated in type 1 ARF (69.2 mg/L [19.2–145.8] vs. 16.3 mg/L [2.5–66.4], P<0.001), whereas the chloride (105.0 mmol/L [102.0–108.0] vs. 102.0 mmol/L [98.0–106.0], P<0.001) and creatine kinase (143.0 U/L [75.0–312.0] vs. 88.0 U/L [47.0–203.0], P<0.001) levels were significantly higher in type 2 ARF. These findings suggest distinct differences in inflammatory responses and metabolic profiles between the two ARF types.

Proportion of ICU Transfer, In-Hospital Mortality, and Multiorgan Failure in Patients with ARF

Over 5 years, 55,833 patients required ICU admission, of whom 15,671 (28.1%) were admitted due to ARF. Among all patients with ARF, 15,671 (47.1%) were transferred to the ICU (Table 4). ICU transfer rates were highest in patients with the combined type (71.2%), followed by type 2 (53.2%) and type 1 (26.3%) (Table 4). The in-hospital mortality rate was highest in patients with the combined type (48.2%), followed by type 1 (20.9%) and type 2 (16.4%).

The baseline total SOFA score showed a significant difference among the ARF types (Table 4). Patients with the combined type had the highest baseline SOFA score (9.0 [5.0–13.0]), followed by type 2 (6.0 [3.0–10.0]) and type 1 (5.0 [3.0–8.0]). In all six SOFA categories, patients with the combined type (type C) exhibited a higher proportion of severe cases (indicated by higher scores) compared with those with type 1 or type 2 ARF. However, type 1 and type 2 ARF showed distinct and characteristic distributions within the individual categories. In the respiratory system, a significantly higher proportion of patients with type 2 ARF initially required ventilator support (≥3 points) compared with those with type 1 ARF (≥3 points in respiratory system: type 1 vs. type 2, 1,055 patients [7.8%] vs. 2,432 patients [23.3%]; P<0.001). In the central nervous system, a significantly higher proportion of patients with type 2 ARF obtained a Glasgow Coma Scale score of ≤12 (≥2 points in the central nervous system) compared with those with type 1 ARF (3,597 [26.3%] vs. 5,309 [50.7%], P<0.001). In the cardiovascular system, patients with type 2 ARF more frequently used vasopressors at baseline (score ≥3 points) than those with type 1 ARF (3,022 [22.1%] vs. 3,838 [36.7%], P<0.001). By contrast, liver and renal injuries were more prevalent in patients with type 1 ARF than in those with type 2 ARF. For liver injury, a higher proportion of patients with type 1 ARF obtained a liver SOFA score of ≥1 (4,063 patients, 29.7%) compared with those with type 2 ARF (2,830, 27.1%) (P<0.001). Similarly, renal injury (SOFA score ≥1) was more prevalent in type 1 ARF (5,648, 41.4%) than in those with type 2 ARF (3,363, 32.2%) (P<0.001).

Respiratory Support in ARF

Table 5 summarizes the respiratory support provided during the course of ARF. Conventional oxygen therapy was the most commonly used form of supportive care, administered to 31,356 ARF patients (94.2%), regardless of ARF type. A HFNC was used in 8,089 patients with ARF (24.3%), with the highest usage observed in those with the combined-type ARF (3,921 patients, 42.8%). HFNC was more frequently used in type 1 ARF than in type 2 ARF (2,853 patients [20.9%] vs. 1,315 patients [12.6%], P<0.001). NIV was more frequently used in 625 patients with ARF (1.9%), with the highest usage observed in those with the combined-type ARF (435 patients [4.7%]). NIV was more frequently used in patients with type 2 ARF than in those with type 1 ARF (142 patients [1.4%] vs. 48 patients [0.4%], P<0.001).

IMV was used in 11,810 patients with ARF (35.5%), with the highest usage observed in those with the combined-type ARF (5,905 patients, 64.5%). IMV was significantly more frequently used in patients with type 2 ARF than in those with type 1 ARF (4,220 patients [40.3%] vs. 1,685 patients [12.3%], P<0.001). ECMO was more commonly used in 673 patients with ARF (2.0%), with the highest usage observed in those with the combined-type ARF (480 [5.2%]). ECMO was also more frequently used in type 2 ARF than those with type 1 ARF (70 [0.5%] vs.123 [1.2%], P<0.001).

Annual Trend of Prevalence and Mortality in Patients with ARF

Figure 2 and Supplementary Table 6 present the annual trends in ARF prevalence. The annual prevalence of total ARF among all hospitalized patients over the past 5 years ranged from 3.9% to 4.6% (P for trend [P_trend]<0.001) (Supplementary Table 6), with a marked increase to over 4.4% during the coronavirus disease 2019 (COVID-19) pandemic period (2020–2022) (Figure 2A). The prevalence of ARF among annual hospitalized patients showed an average increase of 0.175% per year. A consistent annual rise was observed in the proportion of patients aged ≥80 years across all ARF categories, with a statistically significant trend (total ARF-coefficient 0.085, P_trend<0.001; type 1 ARF-coefficient 0.032, P_trend<0.001; type 2 ARF-coefficient 0.023, P_trend<0.001; and combined ARF-coefficient 0.030, P_trend<0.001) (Figure 2A). The prevalence of ARF among all ICU patients showed a similar annual trend (Figure 2B). In-hospital mortality also showed an annual increasing trend, rising from 25.5% in 2018 to 29.3% in 2022 (coefficient 1.017 and P_trend<0.001) (Figure 3A, Supplementary Table 6). Even after excluding patients with do-not-resuscitate (DNR) status, the rate increased from 10.4% to 13.4% over the same period (coefficient 0.873, P_trend<0.001) (Figure 3A). The annual increase in mortality rate was more pronounced in type 1 ARF, increasing from 17.3% to 23.0% (coefficient 1.401, P_trend<0.001) (Figure 3A). However, after excluding patients with DNR status, the rate of increase was attenuated, from 1.55% to 3.30% (coefficient 0.456, P_trend<0.001). The mortality trend differed by age group, showing a consistent decline in in-hospital mortality among young patients aged <40 years with ARF (coefficient –1.579, P_trend=0.013) (Figure 3B). The trends in mortality according to the type of respiratory support are presented in Figure 3C and Supplementary Table 6, with most support types showing an increase in mortality over the 5-year period.

Annual Trend in the Causes of ARF and Respiratory Support Usage

The major causes of ARF, including sepsis (coefficient 0.796, P_trend<0.001), pneumonia (coefficient 0.728, P_trend<0.001), and heart failure (coefficient 1.961, P_trend<0.001), showed an annual increasing trend (Figure 4A, Supplementary Table 6). By contrast, the proportion of respiratory failure caused by trauma has decreased. The primary causes of ARF vary by age group. An increasing trend in sepsis and pneumonia was observed among ARF patients aged ≥60 years. Specifically, in sepsis patients, the coefficient was 0.757 (P_trend=0.004) among those aged 60–79 years and 1.502 (P_trend=0.001) among those aged ≥80 years. In pneumonia patients, the corresponding coefficients were 0.612 (P_trend=0.005) and 1.415 (P_trend<0.001), respectively. No significant trend was identified in patients aged <60 years (Figure 4B, Supplementary Table 6). Similarly, heart failure exhibited a marked increase over the past 5 years with advancing age. The coefficients for heart failure were 0.742 (P_trend= 0.111) in patients aged <40 years, 1.278 (P_trend<0.001) in those aged 40–59 years, 1.960 (P_trend<0.001) in those aged 60–79 years, and 2.467 (P_trend<0.001) in those aged ≥80 years.

Figure 5 illustrates the annual changes in respiratory support for ARF. The HFNC usage steadily increased over the past 5 years across all types—type 1, type 2, and combined types (total ARF-coefficient 4.137, P_trend<0.001; type 1 ARF-coefficient 4.441, P_trend<0.001; type 2 ARF-coefficient 1.773, P_trend<0.001; and combined ARF-coefficient 5.639, P_trend<0.001) (Figure 5A, Supplementary Table 6). In contrast, the use of IMV showed a declining trend across all types (total ARF-coefficient –0.983, P_trend<0.001; type 1 ARF-coefficient –0.391, P_trend=0.046; type 2 ARF-coefficient –2.649, P_trend<0.001; and combined ARF-coefficient –2.038, P_trend<0.001) (Figure 5A, Supplementary Table 6). A similar trend was observed in both DNR and non-DNR patient groups (Figure 5B and 5C, respectively). In particular, in type 2 ARF, although the use of HFNC significantly increased (coefficient 1.773, P_trend<0.001) (Figure 5B) and the use of IMV significantly decreased (coefficient –2.649, P_trend<0.001) (Figure 5B), there was no significant change in mortality (coefficient –0.246, P_trend=0.409) (Figure 3A).

DISCUSSION

This study presents the frequency, causes, and management practices of ARF within the multicenter cohort in South Korea over a 5-year period. ARF accounted for 4.3% of all hospital admissions and 28.1% of all ICU admissions. The etiology of ARF is complex, with most patients having multiple contributing factors rather than a single cause. The most common etiology was sepsis, which accounted for 47.3% of all ARF cases. Among ARF patients, the most common underlying condition was a tumor. In addition, the study highlighted several significant trends in ARF that warrant attention.

First, the proportion of elderly patients among ARF cases has been consistently increasing. As South Korea becomes an aging society, the growing elderly population has led to healthcare challenges. Among these, respiratory failure is one of the most significant issues. The aging of the ARF patient population may negatively impact ARF outcomes and is likely closely associated with the increase in patients with DNR. The poor outcomes of ARF in elderly patients are well-established [22-25]. Recently, there have been reports focusing on specific mechanisms of ARF in elderly patients. Brown et al. [23] identified key features of lung injury in the elderly, including diminished immune response, alterations in the structural components of the airway and vasculature, changes in lung function and respiratory mechanics (increased compliance and decreased small airway diameter). They reported that these factors may lead to an increased risk of mechanical injury. Therefore, as the number of elderly ARF patients increases, individualized severity assessments and treatment strategies tailored to age and patient condition become essential.

The second key point is that sepsis continues leading cause of ARF, with its proportion showing a consistent upward trend, particularly in elderly patients. Sepsis can directly or indirectly cause respiratory failure, a key criterion for evaluating multiorgan failure using the SOFA score, highlighting the interrelated nature of sepsis and respiratory failure [20]. Sepsis has been consistently reported to have a high prevalence, with an annual increase in incidence [26]. Until the early 2010s, many studies reported a steady decline in sepsis-related mortality [26-28]. However, since the latter half of the 2010s, some studies have reported a rebound in mortality rates [29,30]. This trend may be partly attributed to the adoption of the Sepsis-3 definition, which redefined sepsis, but it also suggests a potential increase in the severity of complications like respiratory failure in patients with sepsis. Prest et al. reported that the mortality rate of pulmonary sepsis increased in both sexes and all age groups [30]. In our study, sepsis and pneumonia were also found to be steadily increasing causes of ARF, especially among patients aged ≥60 years. These conditions are likely major contributors to the increased in-hospital mortality associated with ARF. Further research is warranted to address this issue.

Third, the proportion of cardiogenic has rapidly increased. In respiratory failure, heart failure can influence pulmonary congestion, oxygen delivery, and carbon dioxide transport, potentially leading to type 1 and type 2 ARF [31]. The rising prevalence of heart failure is closely linked to the aging population [32]. In this study, heart failure rapidly became one of the primary causes of ARF within just 5 years. This highlights the importance of early cardiac evaluation in the treatment of ARF, underscoring its significance in clinical practice.

Fourth, the annual increase in HFNC use for ARF supportive care was also significant. HFNC reduces intubation rates in patients with hypoxemic respiratory failure [33]. Recent studies have reported that HFNC may benefit patients with hypercapnic respiratory failure by facilitating the removal of carbon dioxide from the upper airway dead space [33]. Despite its potential, the clinical efficacy of HFNC compared with NIV remains controversial, and it is not yet recommended in clinical guidelines. Interestingly, the increasing use of HFNC in type 2 ARF observed in this study suggests that clinicians in real-world practice are progressively favoring HFNC over NIV for the management of hypercapnic respiratory failure. Moreover, the observed reduction in ICU transfer and mechanical ventilation rates in the past 5 years, without an increase in in-hospital mortality. However, the efficacy of HFNC in the management of type 2 respiratory failure has not been clearly established, and further studies are needed to address this issue.

In addition, laboratory data also highlighted clear distinctions between type 1 and type 2 ARF. Elevated C-reactive protein in type 1 ARF suggests an association with inflammatory response, typical of conditions like pneumonia or ARDS [34,35]. In contrast, type 2 ARF showed relatively stable inflammatory markers but subtle electrolyte imbalances, reflecting hypercapnic compensation [36]. These laboratory profiles likely reflect not only the intrinsic mechanisms of each ARF type, but also the impact of prevalent underlying conditions, such as COPD and asthma. Furthermore, the characteristics of the ARF discussed in this study varied by year, ARF type, and center. The COVID-19 pandemic is thought to have significantly affected the etiology and clinical outcomes of ARF during the study period. Therefore, long-term, large-scale, multicenter studies are essential to further understand the characteristics of ARF and improve its treatment.

This multicenter study, involving hospital care institutions from various regions, included data from more than 30,000 patients, which is a significant strength. However, this study has several limitations. First, the retrospective design may have introduced bias during data collection and analysis. As the study relied on EMRs, it may be limited by missing data, recording errors, and variability in data quality. We performed some data processing to address missing values—for example, when PaO₂ was unavailable for calculating the PaO₂/FiO₂ ratio used in the SOFA score, we estimated PaO₂ based on SpO₂ values. However, this imputation process may have introduced potential bias. Second, the COVID-19 pandemic may have influenced the results, particularly in terms of etiology, clinical outcomes, and ICU transfer patterns. Third, as an observational study, it cannot provide direct evidence of the effects of interventions, which requires careful interpretation of the findings.

In conclusion, ARF is a significant medical issue that affects hospitalized patients and has the potential to become a major societal concern, particularly in the aging population. The frequency of ARF and in-hospital mortality have been steadily increasing, with various etiologies contributing to this trend. However, epidemiological studies and research on individualized treatment for elderly patients are lacking. Therefore, nationwide prospective multicenter studies are required to address these issues.

KEY MESSAGES

▪ Acute respiratory failure remains a growing medical and societal concern, particularly in aging populations.

▪ Its frequency and in-hospital mortality rates are increasing, driven by various etiologies.

▪ Despite these trends, research on the epidemiology and individualized treatments for older patients is limited, highlighting the need for nationwide prospective multicenter studies.

NOTES

CONFLICT OF INTEREST

No potential conflict of interest relevant to this article was reported.

FUNDING

None.

ACKNOWLEDGMENTS

Data was obtained from the Severance Clinical Research Analysis Portal (SCRAP) at three hospitals. The authors would like to thank the SCRAP service support team for their excellent support.

AUTHOR CONTRIBUTIONS

Conceptualization: WJY, YJC, JHC. Methodology: WJY, YJC, JHC. Formal analysis: WJY, YJC, JHC. Data curation: WJY, YJC, JHC. Visualization: WJY, YJC, JHC. Project administration: WJY, YJC, BMJ, JHC. Funding acquisition: WJY, YJC, JHC. Writing - original draft: WJY, YJC, KSJ, JSC, JHC. Writing - review & editing: WJY, YJC, KSJ, JSC, JHC. All authors read and agreed to the published version of the manuscript.

SUPPLEMENTARY MATERIALS

Supplementary materials can be found via https://doi.org/10.4266/acc.004728.

Supplementary Table 1.

Baseline characteristics of combined ARF group

acc-004728-Supplementary-Table-1.pdf

Supplementary Table 2.

Causes of acute respiratory failure in combined ARF

acc-004728-Supplementary-Table-2.pdf

Supplementary Table 3.

Biomarker levels in patients with combined ARF

acc-004728-Supplementary-Table-3.pdf

Supplementary Table 4.

Clinical outcomes and multiorgan failure in patients with combined ARF

acc-004728-Supplementary-Table-4.pdf

Supplementary Table 5.

Respiratory support used during hospitalization in patients with combined ARF

acc-004728-Supplementary-Table-5.pdf

Supplementary Table 6.

Coefficient and P-value of trend curve

acc-004728-Supplementary-Table-6.pdf

Figure 1.

Flowchart of the patient selection process. ABGA: arterial blood gas analysis; ICD-10: International Classification of Diseases, Tenth Revision; NIV: non-invasive ventilation.

Figure 2.

Annual prevalence of acute respiratory failure in hospitalized and intensive care unit (ICU)–admitted patients. The prevalence of acute respiratory failure (ARF) among all hospitalized (A) and ICU–admitted (B) patients, respectively. Both graphs present the overall data, stratified by age group. The dotted lines represent the observed changes over the 5-year period, whereas the solid lines indicate the trend curves fitted using logistic regression. The coefficients and P–values for each trend curve are presented in Supplementary Table 6.

Figure 3.

Annual trends in the major causes of acute respiratory failure. The annual trends in in-hospital mortality (excluding individuals with do-not-resuscitate [DNR] status), do-not-intubate cases, DNR cases, and intensive care unit (ICU) transfer cases stratified by acute respiratory failure (ARF) type (A) and age (B). (C) The annual trends in in-hospital mortality according to different types of respiratory support. The dotted lines represent the observed changes over the 5-year period, whereas the solid lines indicate the trend curves fitted using logistic regression. The coefficients and P-values for each trend curve are presented in Supplementary Table 6. COT: conventional oxygen therapy; NIV: non-invasive ventilation; ECMO: extracorporeal membrane oxygenation; HFNC: high-flow nasal cannula; IMV: invasive mechanical ventilation.

Figure 4.

Annual trends in the major causes of acute respiratory failure. The annual trends in the causes of acute respiratory failure (ARF) stratified by ARF type (A) and age (B) group, respectively. The dotted lines represent the observed changes over the 5-year period, whereas the solid lines indicate the trend curves fitted using logistic regression. The coefficients and P–values for each trend curve are presented in Supplementary Table 6.

Figure 5.

Annual trends in respiratory support use for acute respiratory failure. The annual trends in respiratory support use among patients with acute respiratory (A), do-not-resuscitate (DNR; B), and non-DNR (C) acute respiratory failure (ARF) status, respectively. The dotted lines represent the observed changes over the 5-year period, whereas the solid lines indicate the trend curves fitted using logistic regression. The coefficients and P-values for each trend curve are presented in Supplementary Table 6. COT: conventional oxygen therapy; NIV: non-invasive ventilation; ECMO: extracorporeal membrane oxygenation; HFNC: high-flow nasal cannula; IMV: invasive mechanical ventilation.

Table 1.

Baseline characteristics of patients with acute respiratory failure

Variable	Total (n=33,278)	ARF type			P-value
Variable	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Center A	8,724 (26.2)	2,682 (19.6)	3,448 (33.0)	2,594 (28.3)	<0.001
Center B	21,557 (64.8)	9,820 (71.9)	6,014 (57.5)	5,723 (62.5)
Center C	2,997 (9.0)	1,157 (8.5)	997 (9.5)	843 (9.2)
Age (yr)	68 (58–77)	70 (59–78)	66 (57–75)	68 (58–78)	<0.001
Sex (male)	19,963 (60.0)	8,017 (58.7)	6,438 (61.6)	5,508 (60.1)	<0.001
Height (cm)	163.2 (156.3–170.0)	163.0 (156.0–169.0)	164.0 (157.0–170.0)	163.2 (156.2–170.0)	<0.001
Weight (kg)	61.0 (52.8–70.0)	60.0 (52.0–68.5)	63.0 (54.2–72.0)	61.0 (52.1–70.5)	<0.001
BMI (kg/m²)	23.2 (20.5–25.8)	22.8 (20.2–25.4)	23.6 (21.1–26.2)	23.1 (20.2–26.0)	<0.001
Baseline ABGA
pH	7.3 (7.3–7.4)	7.4 (7.4–7.5)	7.3 (7.3–7.3)	7.3 (7.2–7.4)	<0.001
PaCO₂ (mm Hg)	45.1 (33.4–50.0)	33.8 (28.9–39.0)	49.2 (46.7–54.7)	46.6 (36.4–55.0)	<0.001
PaO₂ (mm Hg)	57.9 (50.5–115.0)	53.3 (46.7–57.1)	154.7 (100.5–206.4)	55.1 (46.3–75.1)	<0.001
HCO₃^– (mmol/L)	23.7 (20.6–26.0)	22.3 (19.5–25.2)	24.7 (23.1–26.2)	23.6 (20.3–26.5)	<0.001
CCI (score)	3.0 (2.0–8.0)	4.0 (2.0–8.0)	3.0 (1.0–5.0)	4.0 (2.0–8.0)	<0.001
Myocardial infarction	1,946 (5.8)	771 (5.6)	489 (4.7)	686 (7.5)	<0.001
Congestive heart failure	7,270 (21.8)	3,112 (22.8)	1,802 (17.2)	2,356 (25.7)	<0.001
Peripheral vascular disease	5,709 (17.2)	1,915 (14.0)	1,666 (15.9)	2,128 (23.2)	<0.001
Cerebrovascular disease	6,692 (20.1)	2,861 (20.9)	1,924 (18.4)	1,907 (20.8)	<0.001
Dementia	2,608 (7.8)	1,364 (10.0)	531 (5.1)	713 (7.8)	<0.001
Chronic pulmonary disease	8,867 (26.6)	3,712 (27.2)	2,309 (22.1)	2,846 (31.1)	<0.001
Rheumatic disease	707 (2.1)	324 (2.4)	148 (1.4)	235 (2.6)	<0.001
Peptic ulcer disease	5,488 (16.5)	2,524 (18.5)	1,319 (12.6)	1,645 (18.0)	<0.001
Mild liver disease	4,350 (13.1)	2,034 (14.9)	1,148 (11.0)	1,168 (12.8)	<0.001
Diabetes without chronic complication	6,398 (19.2)	2,876 (21.1)	1,845 (17.6)	1,677 (18.3)	<0.001
Diabetes with chronic complication	4,212 (12.7)	2,113 (15.5)	918 (8.8)	1,181 (12.9)	<0.001
Paraplegia and hemiplegia	1,730 (5.2)	749 (5.5)	448 (4.3)	533 (5.8)	<0.001
Renal disease	6,472 (19.4)	2,932 (21.5)	1,432 (13.7)	2,108 (23.0)	<0.001
Any malignancy^a)	15,403 (46.3)	6,823 (50.0)	4,997 (47.8)	3,583 (39.1)	<0.001
Moderate or severe liver disease	1,493 (4.5)	672 (4.9)	263 (2.5)	558 (6.1)	<0.001
Metastatic solid tumor	6,387 (19.2)	3,352 (24.5)	1,433 (13.7)	1,602 (17.5)	<0.001
AIDS/HIV	31 (0.1)	18 (0.1)	3 (0.0)	10 (0.1)	0.050

Values are presented as number (%) or median (interquartile range). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

ARF: acute respiratory failure; BMI: body mass index; ABGA: arterial blood gas analysis, PaCO₂: arterial carbon dioxide partial pressure; PaO₂: arterial oxygen partial pressure; HCO₃^–: bicarbonate; CCI: Charlson Comorbidity Index; AIDS: acquired immunodeficiency syndrome; HIV: human immunodeficiency virus.

^a)Including lymphoma and leukemia, except malignant neoplasm of the skin.

Table 2.

Causes of acute respiratory failure

Variable	Total (n=33,278)	ARF type			P-value
Variable	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Number of associated ARF causes	2.0 (1.0–4.0)	2.0 (1.0–4.0)	1.0 (1.0–3.0)	3.0 (1.0–4.0)	<0.001
COPD	3,682 (11.1)	1,407 (10.3)	1,065 (10.2)	1,210 (13.2)	<0.001
Asthma	2,090 (6.3)	855 (6.3)	550 (5.3)	685 (7.5)	<0.001
Neuromuscular disease	4,840 (14.5)	1,924 (14.1)	1,444 (13.8)	1,472 (16.1)	<0.001
Stroke	3,707 (11.1)	1,521 (11.1)	1,091 (10.4)	1,095 (12.0)	0.029
Neurologic tumor	923 (2.8)	319 (2.3)	394 (3.8)	210 (2.3)	<0.001
Musculoskeletal disease	2,842 (8.5)	1,263 (9.2)	976 (9.3)	603 (6.6)	<0.001
Sepsis	15,757 (47.3)	7,197 (52.7)	3,066 (29.3)	5,494 (60.0)	<0.001
Trauma	5,538 (16.6)	2,348 (17.2)	1,567 (15.0)	1,623 (17.7)	<0.001
Pneumonia	8,298 (24.9)	4,075 (29.8)	1,249 (11.9)	2,974 (32.5)	<0.001
Pulmonary edema	4,362 (13.1)	1,867 (13.7)	1,040 (9.9)	1,455 (15.9)	<0.001
Atelectasis	7,205 (21.7)	3,372 (24.7)	1,516 (14.5)	2,317 (25.3)	<0.001
Pneumothorax	670 (2.0)	224 (1.6)	201 (1.9)	245 (2.7)	<0.001
Pulmonary thromboembolism	729 (2.2)	432 (3.2)	106 (1.0)	191 (2.1)	<0.001
Pleural effusion	8,501 (25.5)	4,143 (30.3)	1,611 (15.4)	2,747 (30.0)	<0.001
Acute respiratory distress syndrome	1,837 (5.5)	238 (1.7)	582 (5.6)	1,017 (11.1)	<0.001
Heart failure	5,828 (17.5)	2,491 (18.2)	1,502 (14.4)	1,835 (20.0)	<0.001
Ischemic heart disease	6,911 (20.8)	2,769 (20.3)	2,063 (19.7)	2,079 (22.7)	<0.001

Values are presented as median (interquartile range) or number (%). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

ARF: acute respiratory failure; COPD: chronic obstructive pulmonary disease; PaO₂: arterial oxygen partial pressure; PaCO₂: arterial carbon dioxide partial pressure.

Table 3.

Biomarker levels in patients with acute respiratory failure

Variable	Total (n=33,278)	ARF type			P-value
Variable	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Hemoglobin (g/dl)	10.7 (9.0–12.5)	10.4 (8.8–12.2)	11.3 (9.6–12.9)	10.4 (8.8–12.3)	<0.001
WBC (10³/μl)	10.6 (7.2–14.8)	9.7 (6.6–13.8)	11.2 (8.0–15.0)	11.2 (7.4–15.7)	<0.001
Platelet (10³/ μl)	173.0 (102.0–245.0)	181.0 (107.0–258.5)	173.0 (111.0–233.0)	162.0 (86.0–241.0)	<0.001
Creatinine (mg/dl)	0.9 (0.6–1.3)	0.9 (0.6–1.4)	0.8 (0.6–1.1)	0.9 (0.7–1.5)	<0.001
Sodium (mmol/L)	138.0 (134.5–140.7)	137.0 (134.0–140.0)	139.0 (136.0–141.0)	138.0 (134.0–141.0)	<0.001
Potassium (mmol/L)	4.1 (3.7–4.6)	4.1 (3.6–4.5)	4.2 (3.8–4.6)	4.2 (3.7–4.7)	<0.001
Chloride (mmol/L)	104.0 (99.0–107.0)	102.0 (98.0–106.0)	105.0 (102.0–108.0)	103.0 (99.0–107.0)	<0.001
Total protein (g/dl)	5.7 (5.0–6.4)	5.9 (5.2–6.6)	5.6 (4.9–6.3)	5.7 (4.9–6.4)	<0.001
Albumin (g/dl)	3.2 (2.7–3.7)	3.2 (2.8–3.7)	3.3 (2.8–3.7)	3.1 (2.6–3.6)	<0.001
AST (IU/L)	33.0 (22.0–60.0)	31.0 (21.0–53.0)	32.0 (22.0–59.0)	37.0 (24.0–73.0)	<0.001
ALT (IU/L)	20.0 (13.0–39.0)	20.0 (13.0–35.0)	20.0 (13.0–40.0)	21.0 (13.0–44.0)	<0.001
Total bilirubin (mg/dl)	0.8 (0.5–1.2)	0.8 (0.5–1.2)	0.8 (0.5–1.1)	0.7 (0.5–1.3)	0.415
CRP (mg/dl)	45.8 (7.9–125.4)	69.2 (19.2–145.8)	16.3 (2.5–66.4)	47.4 (7.6–132.8)	<0.001

Values are presented as median (interquartile range). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

ARF: acute respiratory failure; WBC: white blood cell; AST: aspartate transaminase; ALT: alanine transaminase; CRP: C-reactive protein; PaO₂: arterial oxygen partial pressure; PaCO₂: arterial carbon dioxide partial pressure.

Table 4.

Clinical outcomes and multiorgan failure in patients with ARF

	Total (n=33,278)	ARF type			P-value
	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Transfer to ICU	15,671 (47.1)	3,594 (26.3)	5,559 (53.2)	6,518 (71.2)	<0.001
In–hospital mortality	8,980 (27.0)	2,850 (20.9)	1,716 (16.4)	4,414 (48.2)	<0.001
Total LOS (day)	13.2 (6.9–26.2)	13.3 (6.9–24.8)	10.8 (6.0–18.8)	18.8 (8.8–39.2)	<0.001
Long–term MV	1,190 (3.6)	113 (0.8)	254 (2.4)	823 (9.0)	<0.001
Baseline SOFA score	6.0 (4.0–10.0)	5.0 (3.0–8.0)	6.0 (3.0–10.0)	9.0 (5.0–13.0)	<0.001
Respiratory score					<0.001
0	2,715 (8.2)	615 (4.5)	1,896 (18.1)	204 (2.2)
1	1,830 (5.5)	176 (1.3)	1,491 (14.3)	163 (1.8)
2	21,562 (64.8)	11,813 (86.5)	4,640 (44.4)	5,109 (55.8)
3	3,368 (10.1)	444 (3.3)	1,525 (14.6)	1,399 (15.3)
4	3,803 (11.4)	611 (4.5)	907 (8.7)	2,285 (24.9)
Coagulation score					<0.001
0	18,216 (54.7)	7,917 (58.0)	5,864 (56.1)	4,435 (48.4)
1	5,714 (17.2)	2,271 (16.6)	1,898 (18.1)	1,545 (16.9)
2	4,355 (13.1)	1,666 (12.2)	1,298 (12.4)	1,391 (15.2)
3	2,229 (6.7)	936 (6.9)	442 (4.2)	851 (9.3)
4	2,764 (8.3)	869 (6.4)	957 (9.2)	938 (10.2)
Neurologic score					<0.001
0	17,155 (51.6)	9,054 (66.3)	4,752 (45.4)	3,349 (36.6)
1	1,942 (5.8)	1,008 (7.4)	398 (3.8)	536 (5.9)
2	5,588 (16.8)	1,945 (14.2)	2,101 (20.1)	1,542 (16.8)
3	3,806 (11.4)	1,021 (7.5)	1,426 (13.6)	1,359 (14.8)
4	4,787 (14.4)	631 (4.6)	1,782 (17.0)	2,374 (25.9)
Liver score					<0.001
0	23,271 (69.9)	9,596 (70.3)	7,629 (72.9)	6,046 (66.0)
1	5,330 (16.0)	2,050 (15.0)	1,699 (16.2)	1,581 (17.3)
2	3,462 (10.4)	1,445 (10.6)	896 (8.6)	1,121 (12.2)
3	678 (2.0)	328 (2.4)	139 (1.3)	211 (2.3)
4	537 (1.6)	240 (1.8)	96 (0.9)	201 (2.2)
Cardiovascular score					<0.001
0	15,251 (45.8)	7,594 (55.6)	4,820 (46.1)	2,837 (31.0)
1	6,276 (18.9)	2,960 (21.7)	1,761 (16.8)	1,555 (17.0)
2	182 (0.5)	83 (0.6)	40 (0.4)	59 (0.6)
3	8,818 (26.5)	2,655 (19.4)	2,776 (26.5)	3,387 (37.0)
4	2,751 (8.3)	367 (2.7)	1,062 (10.2)	1,322 (14.4)
Renal score					<0.001
0	19,893 (59.8)	8,011 (58.6)	7,096 (67.8)	4,786 (52.2)
1	4,701 (14.1)	1,894 (13.9)	1,195 (11.4)	1,612 (17.6)
2	2,049 (6.2)	884 (6.5)	406 (3.9)	759 (8.3)
3	3,704 (11.1)	1,661 (12.2)	976 (9.3)	1,067 (11.6)
4	2,931 (8.8)	1,209 (8.9)	786 (7.5)	936 (10.2)

Values are presented as number (%) or median (interquartile range). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

ARF: acute respiratory failure; ICU: intensive care unit; LOS: length of stay; MV: mechanical ventilator; SOFA: Sequential Organ Failure Assessment; PaO₂: arterial oxygen partial pressure; PaCO₂: arterial carbon dioxide partial pressure.

Table 5.

Respiratory support used during hospitalization for acute respiratory failure

Variable	Total (n=33,278)	ARF type			P-value
Variable	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Conventional oxygen therapy	31,356 (94.2)	12,664 (92.7)	9,747 (93.2)	8,945 (97.7)	<0.001
High-flow nasal cannula	8,089 (24.3)	2,853 (20.9)	1,315 (12.6)	3,921 (42.8)	<0.001
Non-invasive ventilation	625 (1.9)	48 (0.4)	142 (1.4)	435 (4.7)	<0.001
Invasive mechanical ventilation	11,810 (35.5)	1,685 (12.3)	4,220 (40.3)	5,905 (64.5)	<0.001
Extracorporeal membrane oxygenation	673 (2.0)	70 (0.5)	123 (1.2)	480 (5.2)	<0.001
Do-not-resuscitate patients	6,670 (20.0)	3,151 (23.1)	1,193 (11.4)	2,326 (25.4)	<0.001
Do-not-intubate patients	4,817 (14.5)	2,517 (18.4)	812 (7.8)	1,488 (16.2)	<0.001

Values are presented as number (%). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

ARF: acute respiratory failure; PaO₂: arterial oxygen partial pressure; PaCO₂: arterial carbon dioxide partial pressure.

REFERENCES

1. Choi YJ, Cho JH. Current status of treatment of acute respiratory failure in Korea. J Korean Med Assoc 2022;65:124-9.Article PDF
2. Vincent JL, Sakr Y, Ranieri VM. Epidemiology and outcome of acute respiratory failure in intensive care unit patients. Crit Care Med 2003;31(4 Suppl):S296-9.Article PubMed
3. Cartin-Ceba R, Kojicic M, Li G, Kor DJ, Poulose J, Herasevich V, et al. Epidemiology of critical care syndromes, organ failures, and life-support interventions in a suburban US community. Chest 2011;140:1447-55.Article PubMed
4. Kempker JA, Abril MK, Chen Y, Kramer MR, Waller LA, Martin GS. The epidemiology of respiratory failure in the United States 2002-2017: a serial cross-sectional study. Crit Care Explor 2020;2:e0128. Article PubMed PMC
5. Scala R, Heunks L. Highlights in acute respiratory failure. Eur Respir Rev 2018;27:180008.Article PubMed PMC
6. Narendra DK, Hess DR, Sessler CN, Belete HM, Guntupalli KK, Khusid F, et al. Update in management of severe hypoxemic respiratory failure. Chest 2017;152:867-79.Article PubMed
7. See KC, Sahagun J, Taculod J. Patient characteristics and outcomes associated with adherence to the low PEEP/FIO₂ table for acute respiratory distress syndrome. Sci Rep 2021;11:14619.Article PubMed PMC PDF
8. Qaseem A, Etxeandia-Ikobaltzeta I, Fitterman N, Williams JW Jr, Kansagara D; Clinical Guidelines Committee of the American College of Physicians. et al. Appropriate use of high-flow nasal oxygen in hospitalized patients for initial or postextubation management of acute respiratory failure: a clinical guideline from the American College of Physicians. Ann Intern Med 2021;174:977-84.Article PubMed
9. Rochwerg B, Brochard L, Elliott MW, Hess D, Hill NS, Nava S, et al. Official ERS/ATS clinical practice guidelines: noninvasive ventilation for acute respiratory failure. Eur Respir J 2017;50:1602426.Article PubMed
10. See KC. Approach to acute respiratory failure for frontline clinicians. Singapore Med J 2022;63:740-5.Article PubMed PMC
11. Glenardi G, Chriestya F, Oetoro BJ, Mangkuliguna G, Natalia N. Comparison of high-flow nasal oxygen therapy and noninvasive ventilation in COVID-19 patients: a systematic review and meta-analysis. Acute Crit Care 2022;37:71-83.Article PubMed PMC PDF
12. Park S. Treatment of acute respiratory failure: high-flow nasal cannula. J Korean Med Assoc 2022;65:131-43.Article PDF
13. Koh SO, Kim JR. Complications during ventilatory support in patients with acute respiratory failure. Yonsei Med J 1994;35:142-8.Article PubMed
14. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373-83.Article PubMed
15. Vincent JL, de Mendonça A, Cantraine F, Moreno R, Takala J, Suter PM, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on “sepsis-related problems” of the European Society of Intensive Care Medicine. Crit Care Med 1998;26:1793-800.Article PubMed
16. Shashikumar SP, Stanley MD, Sadiq I, Li Q, Holder A, Clifford GD, et al. Early sepsis detection in critical care patients using multiscale blood pressure and heart rate dynamics. J Electrocardiol 2017;50:739-40.Article PubMed PMC
17. Shashikumar SP, Li Q, Clifford GD, Nemati S. Multiscale network representation of physiological time series for early prediction of sepsis. Physiol Meas 2017;38:2235-48.Article PubMed PMC PDF
18. Fleuren LM, Klausch TL, Zwager CL, Schoonmade LJ, Guo T, Roggeveen LF, et al. Machine learning for the prediction of sepsis: a systematic review and meta-analysis of diagnostic test accuracy. Intensive Care Med 2020;46:383-400.Article PubMed PMC PDF
19. Ferguson ND, Fan E, Camporota L, Antonelli M, Anzueto A, Beale R, et al. The Berlin definition of ARDS: an expanded rationale, justification, and supplementary material. Intensive Care Med 2012;38:1573-82.Article PubMed PDF
20. Singer M, Deutschman CS, Seymour CW, Shankar-Hari M, Annane D, Bauer M, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 2016;315:801-10.Article PubMed PMC
21. Brown SM, Duggal A, Hou PC, Tidswell M, Khan A, Exline M, et al. Nonlinear imputation of PaO₂/FIO₂ from SpO₂/FIO₂ among mechanically ventilated patients in the ICU: a prospective, observational study. Crit Care Med 2017;45:1317-24.Article PubMed PMC
22. Baek MS, Chung CR, Kim HJ, Cho WH, Cho YJ, Park S, et al. Age is major factor for predicting survival in patients with acute respiratory failure on extracorporeal membrane oxygenation: a Korean multicenter study. J Thorac Dis 2018;10:1406-17.Article PubMed PMC
23. Brown R, McKelvey MC, Ryan S, Creane S, Linden D, Kidney JC, et al. The impact of aging in acute respiratory distress syndrome: a clinical and mechanistic overview. Front Med (Lausanne) 2020;7:589553.Article PubMed PMC
24. Polok K, Fronczek J, Guidet B, Artigas A, De Lange DW, Fjølner J, et al. Outcomes of patients aged ≥80 years with respiratory failure initially treated with non-invasive ventilation in European intensive care units before and during COVID-19 pandemic. Ann Intensive Care 2023;13:82.Article PubMed PMC PDF
25. Rodriguez Lima DR, Anzueta Duarte JH, Rubio Ramos C, Otálora González L, Pinilla Rojas DI, Gómez Cortés LA, et al. Risk factors for in-hospital mortality in older patients with acute respiratory distress syndrome due to COVID-19: a retrospective cohort study. BMC Geriatr 2024;24:878.Article PubMed PMC
26. Rhee C, Klompas M. Sepsis trends: increasing incidence and decreasing mortality, or changing denominator? J Thorac Dis 2020;12(Suppl 1):S89-100.Article PubMed PMC
27. Herrán-Monge R, Muriel-Bombín A, García-García MM, Merino-García PA, Cítores-González R, Fernández-Ratero JA, et al. Mortality reduction and long-term compliance with surviving sepsis campaign: a nationwide multicenter study. Shock 2016;45:598-606.Article PubMed
28. Lorencio Cárdenas C, Yébenes JC, Vela E, Clèries M, Sirvent JM, Fuster-Bertolín C, et al. Trends in mortality in septic patients according to the different organ failure during 15 years. Crit Care 2022;26:302.Article PubMed PMC
29. Bauer M, Gerlach H, Vogelmann T, Preissing F, Stiefel J, Adam D. Mortality in sepsis and septic shock in Europe, North America and Australia between 2009 and 2019: results from a systematic review and meta-analysis. Crit Care 2020;24:239.Article PubMed PMC PDF
30. Prest J, Nguyen T, Rajah T, Prest AB, Sathananthan M, Jeganathan N. Sepsis-related mortality rates and trends based on site of infection. Crit Care Explor 2022;4:e0775. Article PubMed PMC
31. Cundrle I Jr, Olson LJ, Johnson BD. Pulmonary limitations in heart failure. Clin Chest Med 2019;40:439-48.Article PubMed PMC
32. Lee CJ, Lee H, Yoon M, Chun KH, Kong MG, Jung MH, et al. Heart failure statistics 2024 update: a report from the Korean Society of Heart Failure. Int J Heart Fail 2024;6:56-69.Article PubMed PMC PDF
33. Park S. High-flow nasal cannula for respiratory failure in adult patients. Acute Crit Care 2021;36:275-85.Article PubMed PMC PDF
34. Giamarellos-Bourboulis EJ, Siampanos A, Bolanou A, Doulou S, Kakavoulis N, Tsiakos K, et al. Clarithromycin for early anti-inflammatory responses in community-acquired pneumonia in Greece (ACCESS): a randomised, double-blind, placebo-controlled trial. Lancet Respir Med 2024;12:294-304.Article PubMed
35. Matthay MA, Zemans RL, Zimmerman GA, Arabi YM, Beitler JR, Mercat A, et al. Acute respiratory distress syndrome. Nat Rev Dis Primers 2019;5:18.Article PubMed PMC PDF
36. Shi T, Feng L. Blood biomarkers associated with acute type II respiratory failure in COPD: a meta-analysis. Clin Respir J 2022;16:75-83.Article PubMed PMC PDF

Figure & Data

References

Citations

Citations to this article as recorded by

PubReader
ePub Link

Cite

CITE: export Copy Format; Close

Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:

RIS — For EndNote, ProCite, RefWorks, and most other reference management software
BibTeX — For JabRef, BibDesk, and other BibTeX-specific software

Include:

Citation for the content below
Citation and abstract for the content below

Trends and management of acute respiratory failure in hospitalized patients: a multicenter retrospective study in South Korea

Acute Crit Care. 2025;40(2):171-185. Published online May 28, 2025

DOI: https://doi.org/10.4266/acc.004728

XML Download

Figure

Related articles

Trends and management of acute respiratory failure in hospitalized patients: a multicenter retrospective study in South Korea

Figure 1. Flowchart of the patient selection process. ABGA: arterial blood gas analysis; ICD-10: International Classification of Diseases, Tenth Revision; NIV: non-invasive ventilation.

Figure 2. Annual prevalence of acute respiratory failure in hospitalized and intensive care unit (ICU)–admitted patients. The prevalence of acute respiratory failure (ARF) among all hospitalized (A) and ICU–admitted (B) patients, respectively. Both graphs present the overall data, stratified by age group. The dotted lines represent the observed changes over the 5-year period, whereas the solid lines indicate the trend curves fitted using logistic regression. The coefficients and P–values for each trend curve are presented in Supplementary Table 6.

Figure 3. Annual trends in the major causes of acute respiratory failure. The annual trends in in-hospital mortality (excluding individuals with do-not-resuscitate [DNR] status), do-not-intubate cases, DNR cases, and intensive care unit (ICU) transfer cases stratified by acute respiratory failure (ARF) type (A) and age (B). (C) The annual trends in in-hospital mortality according to different types of respiratory support. The dotted lines represent the observed changes over the 5-year period, whereas the solid lines indicate the trend curves fitted using logistic regression. The coefficients and P-values for each trend curve are presented in Supplementary Table 6. COT: conventional oxygen therapy; NIV: non-invasive ventilation; ECMO: extracorporeal membrane oxygenation; HFNC: high-flow nasal cannula; IMV: invasive mechanical ventilation.

Figure 4. Annual trends in the major causes of acute respiratory failure. The annual trends in the causes of acute respiratory failure (ARF) stratified by ARF type (A) and age (B) group, respectively. The dotted lines represent the observed changes over the 5-year period, whereas the solid lines indicate the trend curves fitted using logistic regression. The coefficients and P–values for each trend curve are presented in Supplementary Table 6.

Figure 5. Annual trends in respiratory support use for acute respiratory failure. The annual trends in respiratory support use among patients with acute respiratory (A), do-not-resuscitate (DNR; B), and non-DNR (C) acute respiratory failure (ARF) status, respectively. The dotted lines represent the observed changes over the 5-year period, whereas the solid lines indicate the trend curves fitted using logistic regression. The coefficients and P-values for each trend curve are presented in Supplementary Table 6. COT: conventional oxygen therapy; NIV: non-invasive ventilation; ECMO: extracorporeal membrane oxygenation; HFNC: high-flow nasal cannula; IMV: invasive mechanical ventilation.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Trends and management of acute respiratory failure in hospitalized patients: a multicenter retrospective study in South Korea

Variable	Total (n=33,278)	ARF type			P-value
Variable	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Center A	8,724 (26.2)	2,682 (19.6)	3,448 (33.0)	2,594 (28.3)	<0.001
Center B	21,557 (64.8)	9,820 (71.9)	6,014 (57.5)	5,723 (62.5)
Center C	2,997 (9.0)	1,157 (8.5)	997 (9.5)	843 (9.2)
Age (yr)	68 (58–77)	70 (59–78)	66 (57–75)	68 (58–78)	<0.001
Sex (male)	19,963 (60.0)	8,017 (58.7)	6,438 (61.6)	5,508 (60.1)	<0.001
Height (cm)	163.2 (156.3–170.0)	163.0 (156.0–169.0)	164.0 (157.0–170.0)	163.2 (156.2–170.0)	<0.001
Weight (kg)	61.0 (52.8–70.0)	60.0 (52.0–68.5)	63.0 (54.2–72.0)	61.0 (52.1–70.5)	<0.001
BMI (kg/m²)	23.2 (20.5–25.8)	22.8 (20.2–25.4)	23.6 (21.1–26.2)	23.1 (20.2–26.0)	<0.001
Baseline ABGA
pH	7.3 (7.3–7.4)	7.4 (7.4–7.5)	7.3 (7.3–7.3)	7.3 (7.2–7.4)	<0.001
PaCO₂ (mm Hg)	45.1 (33.4–50.0)	33.8 (28.9–39.0)	49.2 (46.7–54.7)	46.6 (36.4–55.0)	<0.001
PaO₂ (mm Hg)	57.9 (50.5–115.0)	53.3 (46.7–57.1)	154.7 (100.5–206.4)	55.1 (46.3–75.1)	<0.001
HCO₃^– (mmol/L)	23.7 (20.6–26.0)	22.3 (19.5–25.2)	24.7 (23.1–26.2)	23.6 (20.3–26.5)	<0.001
CCI (score)	3.0 (2.0–8.0)	4.0 (2.0–8.0)	3.0 (1.0–5.0)	4.0 (2.0–8.0)	<0.001
Myocardial infarction	1,946 (5.8)	771 (5.6)	489 (4.7)	686 (7.5)	<0.001
Congestive heart failure	7,270 (21.8)	3,112 (22.8)	1,802 (17.2)	2,356 (25.7)	<0.001
Peripheral vascular disease	5,709 (17.2)	1,915 (14.0)	1,666 (15.9)	2,128 (23.2)	<0.001
Cerebrovascular disease	6,692 (20.1)	2,861 (20.9)	1,924 (18.4)	1,907 (20.8)	<0.001
Dementia	2,608 (7.8)	1,364 (10.0)	531 (5.1)	713 (7.8)	<0.001
Chronic pulmonary disease	8,867 (26.6)	3,712 (27.2)	2,309 (22.1)	2,846 (31.1)	<0.001
Rheumatic disease	707 (2.1)	324 (2.4)	148 (1.4)	235 (2.6)	<0.001
Peptic ulcer disease	5,488 (16.5)	2,524 (18.5)	1,319 (12.6)	1,645 (18.0)	<0.001
Mild liver disease	4,350 (13.1)	2,034 (14.9)	1,148 (11.0)	1,168 (12.8)	<0.001
Diabetes without chronic complication	6,398 (19.2)	2,876 (21.1)	1,845 (17.6)	1,677 (18.3)	<0.001
Diabetes with chronic complication	4,212 (12.7)	2,113 (15.5)	918 (8.8)	1,181 (12.9)	<0.001
Paraplegia and hemiplegia	1,730 (5.2)	749 (5.5)	448 (4.3)	533 (5.8)	<0.001
Renal disease	6,472 (19.4)	2,932 (21.5)	1,432 (13.7)	2,108 (23.0)	<0.001
Any malignancy^a)	15,403 (46.3)	6,823 (50.0)	4,997 (47.8)	3,583 (39.1)	<0.001
Moderate or severe liver disease	1,493 (4.5)	672 (4.9)	263 (2.5)	558 (6.1)	<0.001
Metastatic solid tumor	6,387 (19.2)	3,352 (24.5)	1,433 (13.7)	1,602 (17.5)	<0.001
AIDS/HIV	31 (0.1)	18 (0.1)	3 (0.0)	10 (0.1)	0.050

Variable	Total (n=33,278)	ARF type			P-value
Variable	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Number of associated ARF causes	2.0 (1.0–4.0)	2.0 (1.0–4.0)	1.0 (1.0–3.0)	3.0 (1.0–4.0)	<0.001
COPD	3,682 (11.1)	1,407 (10.3)	1,065 (10.2)	1,210 (13.2)	<0.001
Asthma	2,090 (6.3)	855 (6.3)	550 (5.3)	685 (7.5)	<0.001
Neuromuscular disease	4,840 (14.5)	1,924 (14.1)	1,444 (13.8)	1,472 (16.1)	<0.001
Stroke	3,707 (11.1)	1,521 (11.1)	1,091 (10.4)	1,095 (12.0)	0.029
Neurologic tumor	923 (2.8)	319 (2.3)	394 (3.8)	210 (2.3)	<0.001
Musculoskeletal disease	2,842 (8.5)	1,263 (9.2)	976 (9.3)	603 (6.6)	<0.001
Sepsis	15,757 (47.3)	7,197 (52.7)	3,066 (29.3)	5,494 (60.0)	<0.001
Trauma	5,538 (16.6)	2,348 (17.2)	1,567 (15.0)	1,623 (17.7)	<0.001
Pneumonia	8,298 (24.9)	4,075 (29.8)	1,249 (11.9)	2,974 (32.5)	<0.001
Pulmonary edema	4,362 (13.1)	1,867 (13.7)	1,040 (9.9)	1,455 (15.9)	<0.001
Atelectasis	7,205 (21.7)	3,372 (24.7)	1,516 (14.5)	2,317 (25.3)	<0.001
Pneumothorax	670 (2.0)	224 (1.6)	201 (1.9)	245 (2.7)	<0.001
Pulmonary thromboembolism	729 (2.2)	432 (3.2)	106 (1.0)	191 (2.1)	<0.001
Pleural effusion	8,501 (25.5)	4,143 (30.3)	1,611 (15.4)	2,747 (30.0)	<0.001
Acute respiratory distress syndrome	1,837 (5.5)	238 (1.7)	582 (5.6)	1,017 (11.1)	<0.001
Heart failure	5,828 (17.5)	2,491 (18.2)	1,502 (14.4)	1,835 (20.0)	<0.001
Ischemic heart disease	6,911 (20.8)	2,769 (20.3)	2,063 (19.7)	2,079 (22.7)	<0.001

Variable	Total (n=33,278)	ARF type			P-value
Variable	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Hemoglobin (g/dl)	10.7 (9.0–12.5)	10.4 (8.8–12.2)	11.3 (9.6–12.9)	10.4 (8.8–12.3)	<0.001
WBC (10³/μl)	10.6 (7.2–14.8)	9.7 (6.6–13.8)	11.2 (8.0–15.0)	11.2 (7.4–15.7)	<0.001
Platelet (10³/ μl)	173.0 (102.0–245.0)	181.0 (107.0–258.5)	173.0 (111.0–233.0)	162.0 (86.0–241.0)	<0.001
Creatinine (mg/dl)	0.9 (0.6–1.3)	0.9 (0.6–1.4)	0.8 (0.6–1.1)	0.9 (0.7–1.5)	<0.001
Sodium (mmol/L)	138.0 (134.5–140.7)	137.0 (134.0–140.0)	139.0 (136.0–141.0)	138.0 (134.0–141.0)	<0.001
Potassium (mmol/L)	4.1 (3.7–4.6)	4.1 (3.6–4.5)	4.2 (3.8–4.6)	4.2 (3.7–4.7)	<0.001
Chloride (mmol/L)	104.0 (99.0–107.0)	102.0 (98.0–106.0)	105.0 (102.0–108.0)	103.0 (99.0–107.0)	<0.001
Total protein (g/dl)	5.7 (5.0–6.4)	5.9 (5.2–6.6)	5.6 (4.9–6.3)	5.7 (4.9–6.4)	<0.001
Albumin (g/dl)	3.2 (2.7–3.7)	3.2 (2.8–3.7)	3.3 (2.8–3.7)	3.1 (2.6–3.6)	<0.001
AST (IU/L)	33.0 (22.0–60.0)	31.0 (21.0–53.0)	32.0 (22.0–59.0)	37.0 (24.0–73.0)	<0.001
ALT (IU/L)	20.0 (13.0–39.0)	20.0 (13.0–35.0)	20.0 (13.0–40.0)	21.0 (13.0–44.0)	<0.001
Total bilirubin (mg/dl)	0.8 (0.5–1.2)	0.8 (0.5–1.2)	0.8 (0.5–1.1)	0.7 (0.5–1.3)	0.415
CRP (mg/dl)	45.8 (7.9–125.4)	69.2 (19.2–145.8)	16.3 (2.5–66.4)	47.4 (7.6–132.8)	<0.001

	Total (n=33,278)	ARF type			P-value
	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Transfer to ICU	15,671 (47.1)	3,594 (26.3)	5,559 (53.2)	6,518 (71.2)	<0.001
In–hospital mortality	8,980 (27.0)	2,850 (20.9)	1,716 (16.4)	4,414 (48.2)	<0.001
Total LOS (day)	13.2 (6.9–26.2)	13.3 (6.9–24.8)	10.8 (6.0–18.8)	18.8 (8.8–39.2)	<0.001
Long–term MV	1,190 (3.6)	113 (0.8)	254 (2.4)	823 (9.0)	<0.001
Baseline SOFA score	6.0 (4.0–10.0)	5.0 (3.0–8.0)	6.0 (3.0–10.0)	9.0 (5.0–13.0)	<0.001
Respiratory score					<0.001
0	2,715 (8.2)	615 (4.5)	1,896 (18.1)	204 (2.2)
1	1,830 (5.5)	176 (1.3)	1,491 (14.3)	163 (1.8)
2	21,562 (64.8)	11,813 (86.5)	4,640 (44.4)	5,109 (55.8)
3	3,368 (10.1)	444 (3.3)	1,525 (14.6)	1,399 (15.3)
4	3,803 (11.4)	611 (4.5)	907 (8.7)	2,285 (24.9)
Coagulation score					<0.001
0	18,216 (54.7)	7,917 (58.0)	5,864 (56.1)	4,435 (48.4)
1	5,714 (17.2)	2,271 (16.6)	1,898 (18.1)	1,545 (16.9)
2	4,355 (13.1)	1,666 (12.2)	1,298 (12.4)	1,391 (15.2)
3	2,229 (6.7)	936 (6.9)	442 (4.2)	851 (9.3)
4	2,764 (8.3)	869 (6.4)	957 (9.2)	938 (10.2)
Neurologic score					<0.001
0	17,155 (51.6)	9,054 (66.3)	4,752 (45.4)	3,349 (36.6)
1	1,942 (5.8)	1,008 (7.4)	398 (3.8)	536 (5.9)
2	5,588 (16.8)	1,945 (14.2)	2,101 (20.1)	1,542 (16.8)
3	3,806 (11.4)	1,021 (7.5)	1,426 (13.6)	1,359 (14.8)
4	4,787 (14.4)	631 (4.6)	1,782 (17.0)	2,374 (25.9)
Liver score					<0.001
0	23,271 (69.9)	9,596 (70.3)	7,629 (72.9)	6,046 (66.0)
1	5,330 (16.0)	2,050 (15.0)	1,699 (16.2)	1,581 (17.3)
2	3,462 (10.4)	1,445 (10.6)	896 (8.6)	1,121 (12.2)
3	678 (2.0)	328 (2.4)	139 (1.3)	211 (2.3)
4	537 (1.6)	240 (1.8)	96 (0.9)	201 (2.2)
Cardiovascular score					<0.001
0	15,251 (45.8)	7,594 (55.6)	4,820 (46.1)	2,837 (31.0)
1	6,276 (18.9)	2,960 (21.7)	1,761 (16.8)	1,555 (17.0)
2	182 (0.5)	83 (0.6)	40 (0.4)	59 (0.6)
3	8,818 (26.5)	2,655 (19.4)	2,776 (26.5)	3,387 (37.0)
4	2,751 (8.3)	367 (2.7)	1,062 (10.2)	1,322 (14.4)
Renal score					<0.001
0	19,893 (59.8)	8,011 (58.6)	7,096 (67.8)	4,786 (52.2)
1	4,701 (14.1)	1,894 (13.9)	1,195 (11.4)	1,612 (17.6)
2	2,049 (6.2)	884 (6.5)	406 (3.9)	759 (8.3)
3	3,704 (11.1)	1,661 (12.2)	976 (9.3)	1,067 (11.6)
4	2,931 (8.8)	1,209 (8.9)	786 (7.5)	936 (10.2)

Variable	Total (n=33,278)	ARF type			P-value
Variable	Total (n=33,278)	Type 1 (n=13,659)	Type 2 (n=10,459)	Combined (n=9,160)	P-value
Conventional oxygen therapy	31,356 (94.2)	12,664 (92.7)	9,747 (93.2)	8,945 (97.7)	<0.001
High-flow nasal cannula	8,089 (24.3)	2,853 (20.9)	1,315 (12.6)	3,921 (42.8)	<0.001
Non-invasive ventilation	625 (1.9)	48 (0.4)	142 (1.4)	435 (4.7)	<0.001
Invasive mechanical ventilation	11,810 (35.5)	1,685 (12.3)	4,220 (40.3)	5,905 (64.5)	<0.001
Extracorporeal membrane oxygenation	673 (2.0)	70 (0.5)	123 (1.2)	480 (5.2)	<0.001
Do-not-resuscitate patients	6,670 (20.0)	3,151 (23.1)	1,193 (11.4)	2,326 (25.4)	<0.001
Do-not-intubate patients	4,817 (14.5)	2,517 (18.4)	812 (7.8)	1,488 (16.2)	<0.001

Table 1. Baseline characteristics of patients with acute respiratory failure

Values are presented as number (%) or median (interquartile range). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

Including lymphoma and leukemia, except malignant neoplasm of the skin.

Table 2. Causes of acute respiratory failure

Values are presented as median (interquartile range) or number (%). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

ARF: acute respiratory failure; COPD: chronic obstructive pulmonary disease; PaO₂: arterial oxygen partial pressure; PaCO₂: arterial carbon dioxide partial pressure.

Table 3. Biomarker levels in patients with acute respiratory failure

Values are presented as median (interquartile range). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

Table 4. Clinical outcomes and multiorgan failure in patients with ARF

Values are presented as number (%) or median (interquartile range). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

Table 5. Respiratory support used during hospitalization for acute respiratory failure

Values are presented as number (%). Type 1: PaO₂ <60 mm Hg; Type 2: PaCO₂ >45 mm Hg with a pH <7.35.

ARF: acute respiratory failure; PaO₂: arterial oxygen partial pressure; PaCO₂: arterial carbon dioxide partial pressure.