Non-invasive respiratory devices, high-flow oxygen therapy by nasal cannula (HFNC) and non-invasive ventilation (NIV), have become useful tools for the management of acute respiratory failure (ARF), having demonstrated their usefulness in different entities.1,2 Although most of the evidence for the use of NIV is relegated to patients with chronic obstructive pulmonary disease (COPD) with severe exacerbation, and acute cardiogenic pulmonary edema (ACPE), NIV can be proposed as a first-line treatment in almost any etiology of ARF, especially when endotracheal intubation is considered inadequate.3,4

Nowadays, a significant proportion of patients admitted to the ICU are elderly, with severe comorbidities and presenting frailty.5 These factors affect their physical, physiological and cognitive reserve, making them especially vulnerable to a stressful situation. When chronic pathology is advanced or the frailty situation is severe, the quality of life is so deteriorated that they are not candidates for receiving certain types of organ support.5 Among the measures adopted to adapt the therapeutic effort, do not intubate (DNI) order is the most frequent in patients with ARF.6 In these patients, the use of NIV as ceiling therapy is frequent.7

The use of NIV in patients with DNI order aims to reverse the acute illness and to promote patient comfort.7 The importance of DNI order is determined by its high prevalence and its impact on the prognosis of patients. DNI order is applied to nearly a quarter of patients with ARF and its incidence has increased in recent years.8 The relationship between DNI order and a worse prognosis has been established in different observational studies, which showed a high mortality rate, both in-hospital and in the long term.9–13 Recently, a systematic review showed an in-hospital survival rate of 56%, and 32% at one year in this population.14 Despite this, there is still significant controversy over the use of NIV, since it could lead to the prolongation of a chronic and irreversible process.15

The primary objective of this study was to describe the prevalence of DNI order in a large sample of patients with ARF. The secondary objective was to analyze the relationship between the presence of DNI order and the patients’ prognosis.

We performed an observational and retrospective study of patients consecutively admitted for ARF to the ICU of an university hospital, in the period from January 1st, 1997, to December 31st, 2022. The study was approved by the Ethics Committee of the institution (Code EXT:10/23; approved on March 6, 2023).

During the analyzed period, 6495 patients with ARF treated with NIV were admitted; 493 patients were excluded because they had been admitted previously. Finally, 5972 patients were analyzed. During the first 24 h of ICU stay, a DNI order was determined in 1001 patients. In the following days, 294 additional patients were determined to have a DNI order, while in 20 patients for whom a DNI order had initially been determined, it was decided to change it (Fig. 1). Finally, 1275 (21.3%) had a DNI order, with a decrease in the incidence over the years (Fig. 2). The reasons for the DNI order are shown in Fig. S1.

Figure 1.

ICU admission flowchart. (ARF: acute respiratory failure; NIV: noninvasive ventilation; DNI: do-not-intubate).

(0.19MB).

Figure 2.

Evolution of DNI orders over time.

(0.24MB).

In our study, where we used NIV as respiratory support to prevent a fatal outcome and not as a palliative measure to minimize symptoms, the prevalence of DNI order in patients with ARF treated with NIV was high (21.3%). However, a progressive decrease was observed in the last decade, being 11.9% in recent years. The prevalence of DNI order in different studies is very variable, depending on the type of patients, the geographic area8 and the allocation of the patient in the hospital, with a lower rate in those admitted to the ICU compared to those admitted to intermediate care units or conventional wards.9–13 In a recent systematic review, 27% of patients with ARF had a DNI order, with a wide variability between studies (9–58%).14 The decrease observed in our series could be related to the characteristics of patients admitted to the ICU, especially the decrease in patients with exacerbation of hypercapnic chronic respiratory failure and the increase in hypoxemic ARF. This fact may be related to the greater current availability of NIV in conventional hospitalization wards, where mainly patients with chronic or less severe hypoxemic pathologies are admitted.4 On the other hand, the decrease in DNI order in patients with active cancer is evident in view of the availability of new medications that increase the survival and quality of life of these patients, including intubated patients.16 It remains to be clarified whether, over the years, new generations of intensive care physicians have changed their perspective on the management of patients with chronic diseases, or whether the worse prognosis of these patients (even those treated with NIV in the ICU) represents an impediment to the time to consult them and admit them to the ICU.

Defining an adequate therapeutic effort is a common practice in critical care areas. Of these, DNI order is one of the most frequent. Although the use of NIV in the treatment of advanced disease, both as support or to improve comfort, has been widely debated15,17 due to the possibility of increasing the agony of both the patient and the family, it has become a common practice in DNI order patients. In fact, in the systematic review carried out by Wilson et al., the establishment of DNI order has increased in recent years,8 unlike what occurred in our study.

The factors influencing the decision on a DNI order are multiple, and are mainly related to age, associated comorbidities, the patient’s baseline and frailty status, the prognosis of the underlying disease, and the expectations and wishes of the patient and family.8 Among the factors related to the patient, older age,9,12,18–20 higher comorbidity21,22 and lower baseline status,12 are more frequent in patients with DNI order. In our study, patients with DNI order were older, had a greater number of comorbidities and were frailer than patients eligible for “full” treatment. As in other studies, the pathologies most frequently associated with DNI order in our patients were chronic cardiorespiratory and neurological diseases.11,22 However, in recent years, advanced oncological pathology has become one of the fundamental causes for establishing DNI order. The use of NIV as ceiling therapy in haemato-oncological patients with ARF, despite the controversy over its use, is an effective tool that in many cases can solve the clinical problem that motivated hospital admission, without worsening the quality of life previously shown by the patient.12,16

In our study, neurological, hemodynamic and respiratory parameters did not show relevant differences between patients with and without DNI order. For this reason, the ventilatory parameters used during NIV were very similar. As a single exception, we had the highest level of maximum IPAP in patients with DNI order, which was related to a higher level of PaCO2 and lower pH. The duration of NIV was longer in patients with DNI order, since patients with this limitation of therapeutic effort were usually maintained on non-invasive support until clinical improvement or death. Duan et al., in a study that analyzed the duration of NIV, found that DNI order was an independent predictive factor for prolonged NIV.23 Complications related to NIV are frequent but generally not serious.24 In our series, patients with DNI order presented more frequent complications of NIV, which were related to the longer duration of this kind of respiratory support.

The prognosis of patients with ARF and DNI order is poor. In our study, in-hospital mortality rate was 57.9%. The different series that have addressed the use of NIV in this type of patients show very variable hospital mortality rates. Scarpazza et al. in a series of 62 patients with acute on chronic respiratory failure, showed a mortality of 11.3%.25 Lemyze et al., in a series of patients with acute on chronic respiratory failure and impaired consciousness, had an in-hospital mortality rate of 28%.26 However, most series show higher mortality. It was 44% in the series by Azoulay et al.,12 48.5% in the pneumonia series by Brambilla et al.,27 71.1% in the series by Bulow et al.,28 56.1% in Schetino et al.,11 74% in Fernandez et al.13 and 55% in Meduri et al.29

Long-term survival is very short. In the study by Bullow et al.28 with a 5-year follow-up, survival rate was 11%. Chu et al. showed a one-year survival rate of 29.7% in COPD patients.9

The recent COVID-19 pandemic has led to the need for in-patient and intermediate care unit treatment for many patients with ARF. DNI order and mortality rates have been highly variable. Aliberti et al.,30 in a multicenter study, showed a DNI order rate of 41.4%, while it was up to 50% in the case of Faraone et al.31 or Kofod et al.32 Van der Veer et al. published a series of 39 patients with DNI order, out of a total of 100, with an in-hospital mortality rate of 48.7%.19 However, other authors showed higher mortality rates. Bradley et al. reported a mortality rate of 70% at 30 days,33 while Holm et al.34 observed mortality rates of 81% in those with DNI order. These results are similar to those observed in Italy with mortality rates close to 90%.31,35 In a recent systematic review by Cammarota et al.36 evaluating the use of non-invasive respiratory devices in 31 observational studies with 6645 patients, the DNI order rate was 25.4% and in-hospital mortality was 83.6%.

The presence of DNI order has been identified as an independent risk factor for in-hospital mortality in both patients with COVID-19 pneumonia37 and non-COVID patients.27 In our study, tanto univariate analysis as adjusted analysis of multiple confounding variables showed that the presence of DNI order as an independent predictive factor for increased in-hospital mortality.

The relationship between the cause of ARF and the prognosis of patients with DNI order is not the same in all pathologies. In our study, the survival of patients with COPD, asthma and obesity-hypoventilation syndrome was higher than in other etiologies. This finding has also been described by other authors, with better survival rates in COPD and acute heart failure,10,11 than in patients with hypoxemic ARF11 secondary to cancer38 or interstitial lung disease.39

This study has important strengths. First, this is a study with a large sample size, with a wide variety of clinical diagnoses, encompassing patients with hypoxemic ARF and exacerbation of hypercapnic chronic respiratory failure, which presents a broad spectrum of diseases that condition the presence of respiratory failure. Second, multiple variables were evaluated and systematically collected. The limitations of the study are related to the observational and retrospective nature of its design. On the other hand, it is a single-center study, in an ICU with extensive experience in the management of patients with NIV, highly qualified personnel and multiple devices for the treatment of these patients, which can affect the generalization of the results. Therefore, the external validity of the study may be compromised, and the results obtained may not be extrapolated to other units with less means or experience. Third, while the determination of the DNI usually requires a unanimous criterion not only with the patient or the family, but with all the medical doctors in the ICU, in our study this decision was sometimes taken by the doctors on duty without being able to be endorsed by the rest of the medical staff. Despite the presence of these limitations, we believe that they do not invalidate the results.

The prevalence of DNI order was frequent in patients with ARF admitted to the ICU, being related to age, comorbidity, frailty and the underlying disease. NIV as a respiratory support measure to restore the previous state of health was an effective tool in these patients. The mortality of patients with DNI order was high, both in-hospital and in the long term, with patients with chronic respiratory disease being those who can benefit the most from NIV.

Andres Carrillo-Alcaraz: conception and design of the study, acquisition of data, analysis and interpretation of data, drafting the article, revising it critically for important intellectual content, final approval of the version to be submitted.

Miguel Guia: analysis and interpretation of data, drafting the article, revising it critically for important intellectual content, final approval of the version to be submitted.

Pilar Tornero-Yepez: acquisition of data, analysis and interpretation of data, drafting the article, revising it critically for important intellectual content, final approval of the version to be submitted.

Laura López-Gómez: acquisition of data, analysis and interpretation of data, revising the article critically for important intellectual content, final approval of the version to be submitted.

Nuria Alonso-Fernandez: acquisition of data, analysis and interpretation of data, revising the article critically for important intellectual content, final approval of the version to be submitted.

Juan Gervasio Martin-Lorenzo: analysis and interpretation of data, drafting the article, revising it critically for important intellectual content, final approval of the version to be submitted.

Juan Miguel Sanchez-Nieto: analysis and interpretation of data, drafting the article, revising it critically for important intellectual content, final approval of the version to be submitted.

No form of artificial intelligence was used.

None. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.