Introduction

Interstitial lung disease (ILD) includes several chronic lung diseases, with an estimated incidence of around 30 per 100 000 individuals in France1. Regardless of the aetiology, the disease course is marked by the occurrence of acute exacerbation (AE), and may lead to hypoxemic acute respiratory failure (ARF), with an in-hospital mortality rate of around 50%2.

Various oxygenation techniques are now available for the management of patients with severe hypoxemia3, but the optimal strategy for these patients has yet to be defined. Moreover, patients with acute exacerbation of interstitial lung disease (AE-ILD) are prone to patient self-inflicted lung injury (P-SILI)4 and ventilator-induced lung injury (VILI)5 during acute respiratory distress syndrome (ARDS)6. Whenever possible, non-invasive ventilation (NIV) is preferred to invasive mechanical ventilation7, due to its reduced morbidity and mortality. Some teams use NIV in this indication8, despite an absence of consensus on the subject. In view of its physiological effects, the use of high-flow nasal oxygen therapy (HFNO) in AE-ILD has been the subject of several retrospective series indicating better tolerance than NIV9,10,11,12. A small, prospective, randomized study comparing NIV and HFNO in the management of AE-ILD showed an improvement in comfort and a reduction in the number of days of mechanical ventilation when HFNO was used, with no significant difference between groups in terms of mortality13. A recent meta-analysis of these studies concluded that NIV and HFNO could represent relevant alternatives to mechanical ventilation in the management of AE-ILD, with no difference in mortality or intubation between the two techniques, despite a better oxygenation with NIV14.

Little is known about the optimal oxygenation strategy in AE-ILD, in particular since the development of HFNO. Therefore, we conducted a nationwide survey among respiratory and intensive care unit (ICU) physicians in France to assess their real-life practices regarding oxygenation strategies in the management of AE-ILD. Based on the responses obtained, we intend to assess the feasibility of a French multicenter randomized clinical trial, with the aim of answering the question of the optimal oxygenation strategy in the management of AE-ILD.

Methods

The study was conducted from February 2024 to June 2024. Physicians in respiratory wards and medical ICUs in France were contacted and asked to fill an electronic survey on the use of different oxygenation strategies in AE-ILD. Given the surveys nature regarding professional practices in healthcare providers (exception II-2, Article R1121-1), no ethical approval was necessary15.

Survey

The survey was created and administered using the SurveyMonkey® online platform. Data were collected on the ventilatory management of patients with AE-ILD who were admitted to respiratory wards and medical ICUs in France. Data collection was anonymous.

The whole spectrum of chronic diffuse ILD was included, not just AE of idiopathic pulmonary fibrosis. Physicians were asked about the oxygenation strategies they used to manage patients with AE-ILD: face mask (FM), HFNO, continuous positive airway pressure (CPAP) or NIV.

The survey was reviewed and validated by the Groupe Assistance Ventilatoire et Oxygène (GAVO2) of the Société de Pneumologie de langue française (SPLF) and by the Société de Réanimation de langue française (SRLF). The final survey included 11 questions, 6 of which were multiple-choice questions, 3 with Yes/No answers and 4 with the option of including free text to supplement the answer.

Participating physicians

In order to best identify the physicians involved in the management of patients with AE-ILD, we first contacted all university professors registered with the French Collège des Enseignants de Pneumologie, as well as the heads of all ICUs in France. As a rule, these physicians were ICU physicians and respiratory physicians working on wards that had a respiratory ICU or not. Each physician contacted was asked to forward the survey to other physicians in their region who were involved in the management of patients with AE-ILD. Physicians were surveyed, with only one response per ward, in order to obtain a representative description of the different practices.

Results

The survey was sent to 127 respiratory physicians and 143 ICU physicians. A total of 119 responses (44.1%) were obtained and analysed. Eight physicians only answered the first question and their data were excluded from the analysis. Among the remaining 111 physicians, 45 were respiratory physicians and 66 were ICU physicians. The majority of the physicians, 81/111 (73.0%), worked in tertiary centres. Among the 45 respiratory physicians, 25 (55.6%) worked on a ward that had a respiratory ICU.

The number of patients with AE-ILD managed per year was: less than five patients for 15.3% of physicians, between five and ten patients for 38.7% of physicians, between ten and twenty patients for 26.2% of physicians and more than twenty patients for 19.8% of physicians. Among these latter, 17/20 patients were managed in a tertiary centre.

The majority of patients with AE-ILD were managed in an ICU setting. However, 30.6% of physicians reported that patients were managed in a conventional respiratory ward, in particular when the patient was not eligible for ICU admission. Outside an ICU setting, 57.8% of respiratory physicians reported that they managed patients with AE-ILD using HFNO with a flow rate of at least 30 L/min.

The first-line oxygenation strategy was HFNO for 59.5% of physicians, followed by conventional oxygen therapy with a FM for 24.3% of physicians (Fig. 1). Face masks were used by the majority of respiratory physicians, in respiratory wards or respiratory ICUs. CPAP was used by only one physician. NIV was used as a first-line therapy by only 17/111 of physicians (15.3%), in combination with HFNO for 15 of them, and with a face mask for 2 of them. Forty-five physicians (40.5%) reported that they never used NIV in this indication. In the event of failure of the first-line oxygenation strategy, NIV was more frequently used, in combination with HFNO for 57 physicians (51.8%), in particular in cases of hypercapnia or as a bridge to mechanical ventilation. NIV was almost exclusively used in medical or respiratory ICUs. Only 4/47 (8.5%) of physicians working on a respiratory ward reported using NIV outside an ICU setting. Seventeen ICU physicians reported that they never used NIV in a context of AE-ILD.

Fig. 1
figure 1

First-line oxygenation strategy used in patients with acute exacerbation of interstitial lung disease without do-not intubate order according to physician expertise (global (n = 111), > 5 patients/year (n = 94), > 20 patients/year (n = 22)). HFNO, high flow nasal oxygen; CPAP, continuous positive airway pressure; NIV, non-invasive ventilation; FM, face mask.

Full size image

When analysing the responses of physicians who treat more than 5 patients with AE-ILD per year (n = 94), we find that HFNO is used most widely (n = 56, 59.6%), followed by face mask (n = 23, 24.5%), a combination of NIV/HFNO (n = 12, 12.8%), a combination of NVI/face mask (n = 2, 0.02%) and, lastly, CPAP (n = 1, 0.01%).

Physicians (only 108/111) were also asked whether they would take part in a national multicentre randomized study to assess the potential impact of the first-line oxygenation strategy on the prognosis of patients with AE-ILD: 91/108 physicians (84.3%) answered yes. Nevertheless, the accompanying comments underlined the limitations of such an investigation in terms of potential recruitment, given the small number of patients with this disease. For 90/108 (83.3%) physicians, the proposal of a randomization arm including NIV did not constitute an obstacle to participating in such a study, unlike the remaining 18 physicians (10 ICU physicians and 8 respiratory physicians) who considered it would be difficult to propose the use of NIV in this indication.

Discussion

This national survey has provided an overview of the ventilatory management of AE-ILD in France in 2024. Our results show a large predominance of HFNO as a first-line oxygenation strategy in this indication, while the use of NIV remains in the minority, regardless of the wards involved.

This type of study about AE-ILD is fairly rare. We only found one international practice survey evaluating the management of ILD8. Our response rate (44.1%) is satisfactory for this type of online survey16. The participation in our survey may have been limited by the rarity of the disease evaluated and the extent of the panel of physicians contacted, since the involvement or not of physicians in the management of AE-ILD was not a criterion prior to sending out the survey. In the survey on AE-ILD carried out among an international panel of respiratory physicians, the authors selected the physicians for questioning on the basis of their interest in this disease8. Nevertheless, the response rate to this international survey was not reported.

The majority of the physicians in our study were ICU physicians (59.5%), which was to be expected given the often-severe ARF associated with AE-ILD and the limited number of ICUs associated with respiratory wards in France (22 respiratory ICUs listed in 2023 according to the SPLF). The survey highlighted the significant use of HFNO in ARF by respiratory physicians outside an ICU setting (57.8%), including some patients with AE-ILD. The development of HFNO for hypoxemic ARF outside an ICU setting appears to be a recent change in clinical practice which, according to the ward concerned, was probably made possible by the conditions of the respiratory management of SARS-CoV-2 viral pneumonia during the COVID-19 pandemic17.

While almost 20% of participating physicians, mostly in tertiary centers, reported managing more than 20 patients with AE-ILD per year, most centers managed between 5 and 10 patients with AE-ILD per year. The clinical experience of our centers to manage AE-ILD can therefore be considered less than that of the international survey, which reported a median number of 18 patients with AE-ILD per year per center8.

The respiratory management between the two studies was also different. In our study, HFNO was the first-line oxygenation strategy for 59.5% of physicians, compared to 15.3% for NIV, although 51.8% of physicians used NIV as a second-line strategy to avoid or limit the need for intubation. These results were not impact by the clinical experience of the physicians. In the international survey, NIV was used by 74% of physicians worldwide and 68% in Europe, while HFNO was also used by the majority worldwide (81% of physicians) and in Europe (68%) without any data about clinical context (first-line oxygenation therapy or do not intubate orders for example)8. Therefore, NIV was less widely used by the physicians in our study. Possible factors to explain this difference could be the period when the two surveys were carried out: 2017 for Kreuter et al.8 and 2024 for ours. Indeed, in addition to the undoubted impact of the results of the FLORALI study to use HNFO in hypoxemic ARF18, we were able to highlight a change in clinical practices over time in a retrospective regional multicentre cohort, with a clear decrease in the use of NIV between 2010 and 2020 in the management of AE-ILD19. Another possible explanation lies in the fact that the first study was an international survey8 whereas our study was a nationwide survey.

Recent physiological data have suggested the benefits of NIV administered during ARDS using a helmet and high levels of positive expiratory pressure, in particular when inspiratory efforts are high and when patients are hypocapnic20. In a study comparing patients managed for ARDS or AE-ILD treated with HFNO followed by NIV, tidal volume was not increased during AE-ILD, unlike ARDS21. Dynamic compliance and mechanical power remained stable. These results suggest that NIV does not increase the risk of P-SILI in AE-ILD. In this study, patients with AE-ILD were severely hypoxemic (median PaO2/FiO2 of 108 mmHg (80;126) and hypocapnic (median PaCO2 31 mmHg (28;32)), presenting a phenotype close to that of patients described as “responders” to NIV in ARDS by Menga et al.20. Moreover, an updated median HACOR score of 7 (6;8) was found to be predictive of a low probability of NIV failure22.

Between ERS clinical practice guidelines about NIV for ARF in 201723 and those of ESICM about ARDS in 202324, we note a more balanced message regarding the role of NIV and CPAP in ARF, especially in light of data from the ventilatory management of COVID-19. It is because of this shift that we introduced the question of NIV and CPAP use into our survey, but our results showed a minority use of NIV and exceptional use of CPAP. Although ARDS and AE-ILD are similar in radiological and clinical terms25,26, the major difference between the two lies in the low recruitment potential of the AE-ILD lung and the deleterious effect of PEEP in this indication6,27, although the effects of these techniques on respiratory effort28,29 and oxygenation30 appear relevant to assess during AE-ILD .

Our survey highlights several strengths: the inclusion of both ICU and ward physicians, providing a more representative sample of all clinicians involved in respiratory support for patients with AE-ILD, the endorsement by professional societies, which likely improved recruitment and the up-to-date data presented in a field of fast-changing practices, particularly since COVID-19. Our study also has certain limitations. A response rate of 44%, although good, introduces a certain response bias, and it is obviously possible that only physicians who felt concerned by the issue and therefore had some knowledge of the subject responded. In fact, only 15.3% of survey participants treated fewer than 5 patients per year. We cannot therefore rule out the possibility that physicians who did not respond had a different approach to ventilatory care. We opted for a national survey, although the international exportability of the results remains to be discussed. Indeed, in the study by Kreuter et al.8, significant differences in respiratory care were found between continents. Finally, because of survey design constraints: the 11-question format limited depth, and the survey was not formally validated, although it was reviewed by expert societies.

Nevertheless, the optimal non-invasive respiratory support for the management of AE-ILD remains to be established. The results of our survey suggest that a majority of physicians would be interested in participating in a prospective multicentre randomized clinical trial, comparing several strategies, including a first-line strategy using NIV in this indication. Given the rarity of this disease and the potential difficulties of recruitment, a robust methodology will be required to enable the feasibility of such a prospective study and obtain relevant results.

Conclusion

This national multicentre “real-life” survey shows that, despite a lack of recommendations, HFNO has become the first-line oxygenation strategy in the management of AE-ILD, including outside an ICU setting. Despite the risks of P-SILI4, NIV remains used as a second-line technique, mainly in cases of underlying hypercapnia or a risk of HFNO failure to avoid endotracheal intubation and mechanical ventilation. Nevertheless, a multicentre prospective randomized clinical trial is now needed to assess the real impact of oxygenation strategies on the outcomes of patients with AE-ILD.