Palliative endoscopic treatment of malignant central airway obstruction

Abstract

Purpose

To examine the outcome of palliative endoscopic treatment of malignant central airway obstruction (CAO) and identify predictors for Days Alive and Out of Hospital (DAOH), overall survival and treatment related complications.

Methods

Consecutive adult patients treated endoscopically for malignant CAO at Aarhus University Hospital from 2012 to 2022 were included in the study. Statistical analyses were carried out to identify predictors for DAOH, survival and complications.

Results

127 consecutive patients met the inclusion criteria. The majority of patients were categorised with stage IV lung cancer, the majority being males, with a median age of 67 years. The endoscopic interventions were mainly tumour debulking combined with airway stent insertion or tumour debulking alone. The complication rate was 21.0% and the mortality rate was 3.9%. In total, 89.8% of the patients experienced symptom relief following surgery, and the majority (92.1%) were discharged from hospital within two days after intervention. Mean survival time following intervention was 144 days, mean DAOH₃₀ was 20.8 and mean DAOH₃₆₅ was 157. Survival was associated with comorbidity, type of intervention, preoperative respiratory support and postoperative oncologic treatment. A high preoperative ASA-score, preoperative respiratory support, urgency of intervention, female gender and insertion of airway stent were predictors for a poorer DAOH-outcome.

Conclusion

Endoscopic palliative treatment of malignant CAO is generally feasible and safe, offering symptom relief in most cases. The method is considered an effective measure for short to median term palliation of respiratory distress.

Introduction

Central airway obstruction (CAO) is a complex and potentially life-threatening condition, associated with considerable morbidity and mortality. The primary clinical manifestations of CAO include dyspnoea, reduced functional capacity, post-stenotic pneumonia, and respiratory failure [1, 2].

Lung cancer is the most prevalent cause of malignant central airway obstruction (CAO) [3–6], with approximately 20–30% of lung cancer patients developing CAO at some point during the course of their illness [7]. At the time of diagnosis, 13% of these patients exhibit radiological signs of CAO [8].

Conventional oncologic treatments for malignant CAO often exhibit a delayed onset of action [9] and post-radiation oedema may occur in the early phase following radiotherapy, potentially further compromising the airway. Consequently, therapeutic endoscopic surgery may be utilised to restore airway patency, typically within a palliative context [10, 11]. This approach may offer immediate symptom relief and serve as bridging to subsequent oncological treatment [4, 12].

Endoscopic procedures encompass both flexible and rigid endoscopy, in conjunction with various surgical techniques such as mechanical debulking, electrocautery, balloon dilation, and stent insertion. Especially suitable for extrinsic compression or rapid growing tumours [13], the endoscopic insertion of stents, is often effective. However, endoscopic intervention in malignant CAO is associated with potentially lethal risks comprising bleeding, infection and perforation [11, 14, 15].

The aim of this study was to examine the outcome of palliative endoscopic treatment of malignant CAO and identify predictors for Days Alive and Out of Hospital (DAOH), overall survival and treatment related complications in a large population of consecutive patients endoscopically treated for malignant CAO.

Methods

Study design and patient population

A retrospective analysis was conducted at Aarhus University Hospital, Denmark, including all patients diagnosed with malignant CAO who underwent endoscopic tracheobronchial treatment between January 2012 and December 2022. Follow-up was completed through December 2023. Patients treated with open surgery, including tracheotomy were not included in the cohort.

Information pertinent to the study was gathered from the Electronic Medical Record, covering demographic variables, disease characteristics, oncological treatment, urgency of intervention, symptoms, duration of admission and use of respiratory support, i.e. continuous positive airway pressure (CPAP), non-invasive ventilation (NIV), oxygen therapy, and intubation. CAO was characterized based on the Freitag classification [16]. Morbidity was reflected by ASA score and Charlson comorbidity index (CCI).

DAOH was applied to assess overall effectiveness of the intervention. DAOH [17–19] is a metric representing the number of days a patient is alive and not hospitalized within a specified period. It reflects the effectiveness of treatments or interventions as it emphasizes not only survival but also the quality of life by highlighting the time patients spend outside the hospital, thereby indicating their recovery and overall well-being. DAOH was calculated for 30 days (DAOH₃₀) and 365 days (DAOH₃₆₅) after the index intervention. The day of intervention was defined as day 0.

Complications within the first 30 days following an intervention were registered, i.e. during the procedure, during the primary admission or following discharge. These were graded using the Clavien-Dindo classification [20].

Average days from intervention to discharge from hospital was registered. Surgical efficacy on symptoms prior to discharge was registered.

Patient data were anonymously handled in accordance with regional research permit (1-45-70-16-23).

Interventional technique

Indications for therapeutic bronchoscopy included dyspnoea, recurrent post-obstructive pneumonia or atelectasis, stridor, fistula, impending asphyxiation and need for respiratory support in patients with malignant CAO. All patients underwent a pre-operative CT scan, delineating the stenosis.

All endoscopic procedures were performed under general anaesthesia. A scouting flexible bronchoscopy was performed in order to assess the pathology, followed by interventional bronchoscopy using rigid bronchoscope. Intraluminal masses were debulked and removed via the rigid endoscope, followed by monopolar cauterisation for haemostasis and, in some cases, stent insertion. In instances of an instable airway, extraluminal tumour compression or a high likelihood of bleeding, airway perforation or rapid tumour growth, a silicone stent (Novatech, La Ciotat, France) was inserted to maintain airway patency. Tracheobronchial, carinal or bronchial bifurcation involvement prompted the utilization of Y-shaped stents. Stents were endoscopically inserted using a rigid bronchoscope as per procedure described by Dumon [21].

Urgent interventions were defined as those conducted within 24 h of referral, while semi-urgent interventions were defined as those performed within seven days, and elective interventions more than seven days following the referral.

Outcomes

The primary outcomes were DAOH₃₀ and DAOH₃₆₅. Overall survival and procedure related complications within the first 30 postoperative days were secondary outcomes.

Statistical analysis

Patient data were registered in REDCap and processed using R software.

Unadjusted survival following index intervention was calculated using the Kaplan-Meier method. Survival curves were compared using the log-rank test. Univariate and multivariable Cox proportional hazard regression analyses were performed to calculate unadjusted and adjusted hazard rate ratio (HR) for independent prognostic factors and their impact on the risk of death.

Unadjusted means of DAOH₃₀ and DAOH₃₆₅ were calculated with a 95% confidence interval (CI). Means were compared using the Welch two-sample t-test.

A p-value < 0.05 was considered to be statistically significant.

Results

Patient characteristics

Clinical characteristics and procedural details are shown in Table 1.

Table 1.

Demographic data

Variablea	n = 127
Age (years)	64.5 (13.1)
BMI	24.0 (6.0)
Male Sex	62 (48.8%)
CCI	7.4 (2.4)
ASA
IV	61 (48.0%)
III	59 (46.5%)
Unknown	7 (5.5%)
Excessive Alcohol Consumption b	7 (5.5%)
Smoking
Never Smoked	28 (22.1%)
Ex-smoker	74 (58.3%)
Smoker	25 (19.7%)
Reported Degree of Dyspnoea
None	16 (12.6%)
Mild-moderate	59 (46.5%)
Severe	52 (40.9%)
Respiratory Status
Respiratory Support	40 (31.5%)
Intubated	8 (6.3%)
Urgency of Index Intervention
Elective	42 (33.1%)
Semi-urgent	57 (44.9%)
Urgent	28 (22.1%)
Level of Stenosis
Trachea	19 (15.0%)
Tracheobronchial	26 (20.5%)
Unilateral bronchi	69 (54.3%)
Bilateral bronchi	13 (10.2%)
Freitag Classification
Intraluminal	108 (85.0%)
Extern Compression	19 (15.0%)
Index Intervention
Stent Insertion	76 (59.8%)
Debulking & Dilation	51 (40.2%)
Malignancy
Lung	95 (74.8%)
NSCLC	80 (63.0%)
SCLC	5 (3.9%)
Other Primary Lung Cancer	10 (7.9%)
Oesophagus	9 (7.1%)
Thyroid	3 (2.4%)
Metastases	18 (14.2%)
Unclassified	2 (1.6%)
Postoperative Oncologic Therapy
Chemoradiotherapy	62 (48.8%)
Immunotherapyc	8 (6.3%)
None	57 (44.9%)

^aContinuous data are presented as mean (SD) and categorical data as numbers (percentage)

^bRecommendations: less than 10 standard drinks per week

^cImmunotherapy was introduced in Denmark in 2017

The majority of patients had histologically confirmed primary lung cancer (n = 95, 74.8%). 89.8% of patients experienced relief of symptoms after the index intervention, and the majority (92.1%) were discharged from hospital within two days after intervention.

92 patients (92.0% of patients with a valid description of the stenosis) had 71% obstruction of the airway cross-sectional area (Myer-Cotton Grade III or IV [22]), of whom 62 patients (62.2%) had no detectable lumen (complete stenosis). A valid description of stenosis severity was not found in 27 cases (21.3%).

Procedure characteristics

During the observational period, 127 patients underwent a total of 181 interventional bronchoscopies. 94 patients (74 0.0%) underwent one intervention, 24 underwent two (18.9%), five underwent three (3.9%), one underwent four (0.8%), one underwent five (0.8%), none underwent six and two underwent seven interventions (1.6%).

In total, 92 silicone stents were placed. Of these, 72 (78.3%) were Dumon I-silicone stents, and 20 (21.7%) were Dumon Y-silicone stents. Tumour debulking and dilation was performed in 89 cases.

DAOH

Mean DAOH₃₀ for the entire population was 20.8 (95% CI: 19.0-22.6). As shown in Table 2, DAOH₃₀ was significantly lower in patients with ASA IV, p = 0.03, and patients requiring respiratory support (oxygen therapy, NIV, CPAP or intubation), p = 0.01, prior to the index intervention. Patients requiring an urgent intervention had a significantly lower mean DAOH₃₀, p = 0.02, compared to electively treated patients. Patients receiving an airway stent during index intervention also exhibited a significantly lower mean DAOH₃₀, p = 0.006.

Table 2.

Predictors for DAOH₃₀ and DAOH₃₆₅

	DAOH30			DAOH365
	Mean (95% CI)	p-value		Mean (95% CI)	p-value
Sex
Male	23.1 (21.0-25.2)	Ref.		163 (131–194)	Ref.
Female	18.6 (15.8–21.4)	0.07		152 (117–187)	< 0.001
BMI
≥ 18.5	21.4 (19.5–23.3)	Ref.		166 (142–191)	Ref.
< 18.5	16.3 (10.5–22.0)	0.07		87.3 (29.9–145)	0.03
Malignancy
Primary Lung Tumours	20.7 (18.6–22.7)	Ref.		163 (136–189)	Ref.
Others	22.7 (19.1–26.3)	0.5		150 (100–199)	0.6
ASA
IV	18.6 (16.0-21.2)	Ref.		119 (88.3–150)	Ref.
III	22.6 (20.1–25.1)	0.03		192 (158–226)	0.002
Smoking
Never Smoked	22.8 (18.9–26.6)	Ref.		201 (148–255)	Ref.
Currently or previously	20.2 (18.2–22.3)	0.3		145 (119–170)	0.08
Respiratory Status
No respiratory Support	23.5 (21.6–25.3)	Ref.		188 (159–217)	Ref.
Respiratory Support	16.4 (13.1–19.7)	0.01		106 (70.5–141)	< 0.001
Freitag Classification
Intraluminal	21.1 (19.2–23.0)	Ref.		161 (137–186)	Ref.
Extern Compression	19.3 (14.0-24.5)	0.5		133 (63.6–203)	1
Index Intervention
Stent Insertion	19.0 (16.6–21.4)	Ref.		128 (99.6–157)	Ref.
Debulking & Dilation	23.4 (20.9–26.0)	0.006		200 (163–237)	0.3
Urgency of Intervention
Elective	23.6 (20.8–26.5)	Ref.		197 (155–239)	Ref.
Urgent or Semi-urgent	19.4 (17.2–21.6)	0.02		137 (110–165)	0.002
Level of Stenosis
Trachea	19.6 (14.4–24.7)	Ref.		158 (88.3–228)	Ref.
Tracheobronchial	19.4 (15.4–23.3)	0.8		145 (89.7–200)	0.8
Unilateral bronchi	21.8 (19.4–24.2)	0.9		174 (143–205)	0.6
Bilateral bronchi	20.2 (13.5–26.8)	0.09		91.2 (29.5–153)	0.09
Chemoradiotherapy after Intervention
Yes	24.1 (22.6–25.6)	Ref.		199 (169–229)	Ref.
No	17.7 (14.6–20.7)	0.005		117 (84.2–150)	< 0.001
Immunotherapy after Intervention
Yes	23.0 (16.1–29.9)	Ref.		241 (155–327)	Ref.
No	20.6 (18.7–22.4)	0.3		149 (125–173)	0.06

Mean DAOH₃₆₅ for the entire population was 157 (95% CI: 134–180). In line with DAOH₃₀, DAOH₃₆₅ was negatively impacted by ASA IV, p = 0.002, preoperative respiratory support, p < 0.001, urgent intervention, p = 0.002, BMI < 18.5, p = 0.03, absence of postoperative oncologic treatment, p < 0.001, and female gender, p < 0.001.

Survival

Median survival following the index intervention was 144 days, with only 22.1% being alive one year later. Mortality during the first 30 days after intervention was 16.5% which was mainly associated with pre-operative respiratory support, p = 0.001, stent insertion, p = 0.03, not having unilateral bronchial involvement, p = 0.008, and ineligibility for postoperative chemoradiotherapy, p < 0.001 (Table 3).

Table 3.

Predictors for death within 30 days after intervention

	Alive 30 days			Alive < 30 days								p-value
	n = 106			n = 21
Baseline Characteristics
Age 65 years	64 (60.4%)			10 (47.6%)								0.3
BMI 18.5	10 (9.4%)			5 (23.8%)								0.06
CCI 8	52 (49.1%)			14 (66.7%)								0.1
Female Sex	51 (48.1%)			14 (66.7%)								0.1
Primary Lung Tumours	82 (77.4%)			14 (66.7%)								0.3
Preoperative ASA III	51 (48.1%)			8 (38.1%)								0.4
Current or Previous Smoker	82 (77.4%)			17 (81.0%)								0.7
Respiratory Support Needed	35 (33.0%)			15 (71.4%)								0.001
Non-invasive Tumour Growth	14 (13.2%)			5 (23.8%)								0.2
Index Intervention
Stent Insertion	59 (55.7%)			17 (81.0%)								0.03
Urgent or Semi-urgent	69 (65.1%)			16 (76.2%)								0.3
Level of Stenosis
Tracheal	13 (12.3%)			6 (28.6%)								0.06
Tracheobronchial	20 18.9%			6 (28.6%)								0.3
Unilateral bronchi	64 (60.4%)			5 (23.8%)								0.008
Bilateral bronchi	9 (8.5%)			4 (19.1%)								0.2
Oncological Treatment after Intervention
Chemoradiotherapy	59 (55.7%)			3 (14.3%)								< 0.001
Immunotherapy	12 (11.3%)			0 (0.0%)								0.1

The impact of specific variables on time to death are depicted as hazard rates (Table 4). Only CCI, stent insertion, bilateral bronchial tumour involvement, and absence of post therapy CRT remained significant when adjusted for relevant confounders.

Table 4.

Hazards rate: survival

	Univariate Cox Regression			Multivariate Cox Regression
	HR (95% CI)	p-value		HR (95% CI)	p-value
Baseline Characteristics
Age (years)	1.01 (1.00-1.03)	0.06		1.00 (0.98–1.02)	0.8
BMI	0.99 (0.95-1-03)	0.6		1.03 (0.98–1.08)	0.2
CCI	1.17 (1.08–1.27)	0.001		1.14 (1.05–1.24)	0.002
Female Sex	1.01 (0.69–1.45)	1		0.94 (0.64–1.38)	0.8
Primary Lung Tumours	0.88 (0.57–1.37)	0.6		0.87 (0.55–1.36)	0.5
Preoperative ASA IV	1.82 (1.23–2.63)	0.002		1.33 (0.88-2.00)	0.2
Current or Previous Smoker	1.37 (0.87–2.15)	0.2		1.49 (0.92–2.42)	0.1
Respiratory Support Needed	1.97 (1.34–2.88)	< 0.001		1.27 (0.81-2.00)	0.3
Index Intervention
Stent Insertion	2.00 (1.35–2.94)	< 0.001		1.54 (1.03–2.33)	0.03
Urgent or Semi-urgent	2.08 (1.36–3.19)	< 0.001		1.30 (0.78–2.15)	0.3
Level of Stenosis
Tracheobronchial	1.10 (0.59–2.06)	0.8		1.47 (0.71–3.02)	0.4
Unilateral bronchi	0.83 (0.49–1.43)	0.5		0.85 (0.48–1.51)	0.6
Bilateral bronchi	1.64 (0.78–3.42)	0.2		2.93 (1.11–7.74)	0.03
Oncological Treatment after Intervention
Chemoradiotherapy	0.76 (0.52–1.10)	0.1		0.48 (0.32–0.72)	< 0.001
Immunotherapy	0.47 (0.23–0.97)	0.04		0.53 (0.25–1.10)	0.09

Multivariate HR: Adjusted for significant co-variables

Complications

Across 92 silicone stent insertions and 89 tumour debulking and/or dilation procedures, a total of 38 complications were recorded, resulting in an overall complication rate of 21.0%. One patient experienced complications associated with airway stenting on four separate occasions, while another patient experienced complications on two occasions. Additionally, 32 patients encountered complications related to a single intervention. The reported complications included 12 cases of pneumonia, 12 cases of stent-related complications (migration, occlusion), seven cases of respiratory insufficiency, four cases of bleeding, two of mediastinitis, two of emphysema, and two of deep vein thrombosis with lung embolism.

Reintervention was necessary in 14 cases due to stent migration, stent occlusion, or bleeding. Seven patients experienced a fatal outcome, with six deaths due to pneumonia and one due to respiratory insufficiency, giving a total mortality rate of 3.9%. Reintervention was performed in 10.9% of stent insertions and 4.5% of debulking/dilation procedures. None of the tested variables were significantly associated with complications. Based on the Clavien-Dindo classification, 12 cases were classified as grade 2 (requiring iv pharmacological treatment), 14 cases as grade 3b (requiring reintervention under general anaesthesia), and five cases as grade 4 (requiring ICU management).

Discussion

In this report, we present retrospective outcome data on therapeutic bronchoscopic procedures addressing malignant CAO. Our findings reveal comparable outcomes, both in terms of overall survival and complication rates, to those reported in the literature [10].

The primary indication for bronchoscopic intervention in malignant CAO is palliation, as most patients have incurable, advanced disease at the time of intervention. However, selection of patients can be challenging as the most marginalised individuals, and those with the greatest risk of complications, may achieve the greatest short term benefit [4]. The decision to proceed with endoscopic surgery necessitates a multidisciplinary assessment involving surgeons, radiologists and oncologists. However, in many cases, there is no effective alternative to endoscopic intervention, as conventional oncologic treatments frequently demonstrate a delayed onset of action and holds a risk of short-term exacerbation due to post-radiation oedema [9]. Accordingly, the patients in our cohort were indeed marginalized, and endoscopic intervention was considered a “last resort” intervention. Indeed, for some patients the intervention could be considered a “just before death therapy” as the postoperative survival was short. Nevertheless, the efficacy of endoscopic intervention was generally favourable in terms of alleviation of symptoms, short perioperative hospital stay and rate of complications. This prompts the question of determining the appropriate timing for offering this type of palliative intervention versus opting for non-interventional palliative pharmacological management. This represents a significant medical dilemma. Regarding survival and rate of complications, our data are comparable with previous reports.

In present study, the majority of interventions (66.9%) were performed in an urgent or semi-urgent setting where the risk of impeding asphyxiation was present. Half of our patients were classified as ASA IV, and more than 40% were ASA III, often presenting with severe airway obstruction (48% presented with Myer-Cotton grade IV, i.e., “no detectable lumen”). Accordingly, the patients in our cohort were indeed marginalized, and endoscopic intervention was considered a “last resort” intervention. Nevertheless, the efficacy of endoscopic intervention was generally favourable in terms of alleviation of symptoms, short perioperative hospital stay and rate of complications. Regarding survival and rate of complications, our data are comparable with previous reports [10].

To our knowledge, this is the first study to present DAOH in a cohort of patients receiving palliative endoscopic treatment of malignant end stage CAO. Combining hospitalizations and mortality following an intervention, DAOH is a patient-centred outcome measure, suitable for evaluation of the effect of surgery in particular [18, 19]. Thus, we found it to be a highly relevant measure in this group of patients.

In our cohort, we found a mean DAOH₃₀ of 20.8, meaning that the average patient was hospitalized or dead for 9.2 days within the first 30 days following the index intervention. Mean DAOH₃₆₅ was 157, meaning that the average patient was in hospital or dead for 208 days during the first year following primary endoscopic surgery. For reference, mean DAOH₃₆₅ has been reported to be 357 in patients undergoing trans-oral robotic surgery for oral pharyngeal cancers [23], 359 in patients undergoing video-assisted lobectomy for non-small cell lung cancer [24] and 356 in patients with oral cavity squamous cell carcinoma [25], . Inherently, as these procedures are performed with curative intent, direct comparison with our reports on DAOH₃₆₅ for a palliative intervention is limited.

As expected, DAOH varied among subgroups, with the stent group exhibiting the lowest outcomes in both DAOH₃₀ and DAOH₃₆₅. This makes sense, as stent insertion was predominantly performed in patients with an instable airway, i.e. due to major tumour bulk, external compression and disruption of the airway caused by an infiltrating tumour. Similarly, poor performance status and advanced disease, as indicated by a high ASA class, urgency of intervention, need for preoperative respiratory support, and ineligibility for chemoradiotherapy, were negative predictors of DAOH.

Not surprisingly, the same measures were predictors of overall survival, both on short term and long term. As one could expect in this cohort of patients with end stage disease, long tern survival was low, although short term mortality was modest, indicated by a 30-day mortality rate of 17%.

Generally, two pivotal observations can be made from these findings. One; should we adopt a more aggressive approach to endoscopic tumour resections, intervening at a stage when the general physical condition of the patient is better, potentially reducing complications and contributing to a prolonged period of symptom-free existence? And two; should we exercise more restraint and caution in the treatment of patients with a high risk of an unfavourable outcome, i.e., patients with ASA IV or receiving respiratory support? A study by Marchinoi et al. [26] supports the former, comparing standard oncologic treatment, i.e. chemoradiotherapy, for malignant CAO caused by NSCLC to integrated treatment (interventional endoscopy in combination standard treatment). The study found a favourable effect on overall survival, symptom free survival and hospitalisation in the integrated arm. Interestingly, endoscopic treatment was performed as an active cancer treatment modality i.e., not solely as a palliative measure and therefore also offered to asymptomatic patients and only to those with a performance status ≤ 2. Regarding the latter, our findings indicate a favourable outcome with alleviation of symptoms in the vast majority of patients, albeit maybe only on a short-term basis.

The ethical challenges of conducting randomised trials, particularly when dealing with obviously symptomatic patients who may experience immediate benefits from airway-opening therapy, present inherent difficulties. Nevertheless, exploring the optimal timing and aggressiveness of endoscopic interventions in a carefully designed prospective study could provide valuable insight.

The rate of complications within 30 days post intervention was 21.0%, ranging from mild to severe, according to the Clavien Dindo scoring system. However, the majority of complications being amiable for treatment. This aligns with findings presented previously [10, 12, 27]. Ost, et al. [10] described similar complication rates of 25% within 30 days of surgery, and a 30-day mortality rate of 16% for ASA ≥ II in a cohort of patients with malignant CAO. Mohan et al. [12] reported the overall complication rate for a similar cohort of patients to be 18.1%, only specifying that half of these occurred in the first 48 h following the intervention.

Several limitations were inherent in the design of our study, primarily due to its retrospective nature and the associated constraints typical of such a design. The study’s modest size diminishes the power of the analysed associations. Reliable subjective or objective data on the efficacy of the procedure was not registered as no standardised follow-up at the ENT department was planned.

Nevertheless, our results add to the current body of knowledge on patients treated endoscopically for end stage malignant CAO.

Fig. 1.

(a): Survival based on preoperative ASA classification. (b): Impact of preoperative need for respiratory support on survival

Conclusion

In summary, bronchoscopic intervention is a relevant and effective palliative treatment for CAO attributable to both pulmonary- and extra-pulmonary metastatic malignancies. Notably, this intervention remains relevant even in advanced stages of the disease. The immediate efficacy of treatment underscores its impact, providing prompt relief to patients facing respiratory compromise. As we navigate the complexities of end-stage malignancies, the principle of offering timely and impactful interventions remains paramount.

Abbreviations

ASA

American Society of Anaesthesiologists

CAO

Central Airway Obstruction

CI

Confidence Interval

CCI

Charlson Comorbidity Index

CPAP

Continuous Positive Airway Pressure

DAOH

Days Alive and Out of Hospital

HR

Hazard Rate Ratio

NIV

Non-Invasive Ventilation

Author contributions

MB and SH collected data, wrote the main manuscript text and prepared figures under guidance of TK. DGL, SG, SS and UP contributed with data collection and critical revision of the manuscript. The manuscript has been reviewed and approved by all authors.

Funding

Not applicable.

Data availability

The full dataset generated and analysed during the current study is not publicly available due to legal reasons. The granted approval only allows for publication of anonymised data.

Declarations

Ethics approval and consent to participate

The study was authorised by the ethics committee of the Central Denmark Region. Permit 1-45-70-16-23. This approval waivered the need for consent for participation and publication from included patients.

Consent for publication

The study was authorised by the ethics committee of the Central Denmark Region. Permit 1-45-70-16-23. This approval waivered the need for consent for participation and publication from included patients.

Competing interests

The authors declare no competing interests.