Intensive Care Med https://doi.org/10.1007/s00134-020-06228-7 NARRATIVE REVIEW Use of nasal high flow oxygen during acute respiratory failure Jean‑Damien Ricard1,2* , Oriol Roca3,4, Virginie Lemiale5, Amanda Corley6,7, Jens Braunlich8,9, Peter Jones10,11, Byung Ju Kang12, François Lellouche13, Stefano Nava14, Nuttapol Rittayamai15, Giulia Spoletini16,17, Samir Jaber18 and Gonzalo Hernandez19 © 2020 Springer-Verlag GmbH Germany, part of Springer Nature Abstract Nasal high flow (NHF) has gained popularity among intensivists to manage patients with acute respiratory failure. An important literature has accompanied this evolution. In this review, an international panel of experts assessed poten‑ tial benefits of NHF in different areas of acute respiratory failure management. Analyses of the physiological effects of NHF indicate flow-dependent improvement in various respiratory function parameters. These beneficial effects allow some patients with severe acute hypoxemic respiratory failure to avoid intubation and improve their outcome. They require close monitoring to not delay intubation. Such a delay may worsen outcome. The ROX index may help clini‑ cians decide when to intubate. In immunocompromised patients, NHF reduces the need for intubation but does not impact mortality. Beneficial physiological effects of NHF have also been reported in patients with chronic respiratory failure, suggesting a possible indication in acute hypercapnic respiratory failure. When intubation is required, NHF can be used to pre-oxygenate patients either alone or in combination with non-invasive ventilation (NIV). Similarly, NHF reduces reintubation alone in low-risk patients and in combination with NIV in high-risk patients. NHF may be used in the emergency department in patients who would not be offered intubation and can be better tolerated than NIV. Keywords: Acute respiratory failure, ARDS, Intubation, High flow oxygen, Nasal canula, Palliative care Introduction other than acute respiratory failure have emerged. In this narrative review, we have aimed to summarize the avail- In the past years, nasal high flow has gained an impor- able data and address the different clinical scenarios in tant popularity among intensivists to manage patients which NHF can be used, highlighting areas where further with acute respiratory failure, filling a gap in the venti- research is required to confirm or not the potential for latory support escalation between facemask oxygen and NHF to improve patient outcome. noninvasive or invasive mechanical ventilation. Interest- ingly, use of NHF was widely and rapidly adopted in ICUs Nasal High Flow (NHF) in acute hypoxemic before in-depth knowledge of its physiological effects respiratory failure (AHRF) and evidence of its efficacy were published. Since, a great Physiological effects of NHF amount of literature has been published and indications A unique feature of NHF is its ability to comfortably deliver high flows of warmed humidified gas, 20–70 L/ min, with a F ­iO2 range of 0.21–1.0. Physiological *Correspondence: jean‑damien.ricard@aphp.fr 1 responses to NHF therapy include increases in airway Medico‑surgical ICU, Assistance Publique ‑ Hôpitaux de Paris, DMU ESPRIT, Médecine Intensive Réanimation, Hôpital Louis Mourier, pressure, end-expiratory lung volume (EELV), and oxy- 92700 Colombes, France genation which are probably optimal with higher flows Full author information is available at the end of the article (60–70 L/min), while the effects on dead-space washout, work of breathing, and respiratory rate may be obtained with intermediate flows (20–45 L/min) [1–3] (Fig. 1). Take‑home message Most of this evidence relates to patients with hypoxemic In every segment of acute hypoxemic respiratory failure manage‑ respiratory failure and, to date, few physiological data ment (first-line ventilatory support, preoxygenation, post-extuba‑ have been obtained in hypercapnic populations [4]. tion, palliative care), evidence suggests that nasal high flow has the As NHF rate increases, so does EELV [2] (ESM Fig. 1). potential to benefit the patient. Beneficial effects also seem to be seen in patients with chronic respiratory disease. This relationship is linear with each litre of increased flow resulting in 0.7% increase in EELV [5]. A strong cor- relation exists between increases in airway pressure and hypoxemic respiratory failure, while lower flows may EELV (r = 0.7) [5], suggesting that this mechanism is be efficient to reduce the respiratory rate and work of responsible for the improvements in oxygenation seen breathing, especially in hypercapnic patients (see dedi- with NHF, particularly at higher flows [2], through alveo- cated section). lar recruitment (ESM Fig. 2). Inspiratory effort decreases with increasing flow rates, Avoidance of intubation and Mauri et al. [2] suggest it is best supported by flows Since the introduction of NHF in the armamentarium of 60 L/min. However these, and other authors, also rec- of respiratory failure management, a number of obser- ognised that important reductions in work of breath- vational studies have shown that respiratory status of ing can still be achieved with lower rates of 20–45 L/ patients with AHRF is significantly improved with NHF min [1, 2] (ESM Fig. 3). As flow increases, respiratory in comparison with standard oxygen [7–11]. Although rate decreases [2, 4], along with minute ventilation [2, they suggested that intubation may have been avoided 4], without any resultant hypercapnia. In other words, in some patients, design of these studies precluded any alveolar ventilation (minute ventilation − dead-space formal demonstration of intubation rate reduction with ventilation) remains stable whereas minute ventilation NHF. The characteristics of NHF and the related physi- decreases. The main mechanism to explain the stability ological effects account for the superiority of NHF in in ­PaCO2 is the washout of physiological dead space. terms of comfort, tolerance, alleviation of respiratory Importantly, higher flows are well tolerated by hypox- distress and improved oxygenation. The Florali study emic patients, whose comfort was not adversely affected [12] showed in a post hoc subgroup (patients with by higher gas flows [6]. a ­PaO2/FiO2 < 200) analysis that intubation rate was As NHF is used for respiratory support in a wide range lower in patients treated with NHF compared to those of conditions, clinicians should titrate flow to patient’s treated with NIV or standard oxygen. This clinically response to therapy. ESM Table 1 provides some guid- relevant difference translated into a reduction in hos- ance on achieving optimal flow rates for patients with pital and 90-day mortality rate in favour of patients that hypoxemic respiratory failure. If tolerated, the highest received NHF as first-line therapy. Several meta-anal- flow as a starting point seems optimal in patients with ysis [13], but not all [14], also found that use of NHF Fig. 1 Schematic representation of the physiologic effects of Nasal High Flow (NHF) and possible impact of the flow. Increase in airway pres‑ sure and ­FiO2 improve oxygenation by different mechanisms and may be optimal at higher flows. Most of dead-space wash-out-related effects (increased ­CO2 clearance, decrease respiratory drive, respiratory rate and effort to breathe) may be obtained for lower flows. All these physiological effects probably explain the improved comfort in patients with respiratory failure and possibly the outcomes. NHF nasal high flow, Paw airway pres‑ sure; FiO2 fraction of inspired oxygen, EELV end-expiratory lung volume, RR respiratory rate, VE minute volume, WOB work of breathing was associated with a reduction in intubation rate. A 108.6 and number of quadrants involved in the chest concern with some is the inclusion of studies assess- X-ray was 3 [7]. ing patients with very different intubation risks and Hence, in the absence of criteria for immediate intu- required information size was not reached [14]. Intu- bation, patients presenting with ARDS can be offered a bation rates in patients with AHRF admitted to the NHF trial. Those with a low P ­ aO2/FiO2 (< 100) will be at ICU range between 30 and 40%. These figures are only greater risk of intubation. Hence, a close monitoring will 1–2% in the ED as those needing immediate intubation be required to anticipate when intubation needs to be or NIV for AHRF were excluded. Thus, the choice of performed in a timely fashion. (see below for the preoxy- primary outcome selected in certain settings may not genation strategy in such patients). have been the most relevant. In the ED, three of five RCT of NHF compared to conventional oxygen therapy Anticipating outcome with NHF reported intubation rates, with no studies finding a dif- This close monitoring mentioned above should include ference between the groups [15–17]. A meta-analysis of several variables (oxygenation, thoracoabdominal asyn- the studies with either undifferentiated respiratory dis- chrony, need for vasopressors, SOFA score or disease tress [15, 17, 18] or cardiogenic pulmonary oedema [16] severity) that have been shown to be associated with found only 7/571 patients required intubation (1.2%), NHF failure [7, 9–11]. This association is not, how- with no difference in intubation rates: RR 0.69 (0.12, ever, directly predictive of the outcome of NHF. More 4.12), p = 0.68 [19]. A single RCT comparing NHF to recently, the ROX index, defined as the ratio of SpO2/ NIV for undifferentiated respiratory failure found that FiO2 over RR, has been described and prospectively vali- NHF was non-inferior to NIV for intubation rates [20]. dated to predict success and failure of NHF in pneumo- It, therefore, appears that contrary to the ICU, delayed nia patients with AHRF [22, 23]. The rationale behind intubation is a rare event in the ED, so much larger this index is that it combines in a single value, the best studies would be required to demonstrate any benefit describers of patients’ respiratory status. Patients with with respect to avoiding intubation. a ROX index greater than or equal to 4.88 after 2, 6 and 12 h of NHF therapy were less likely to be intubated. In contrast, patients with a ROX index below 2.85, 3.47 and Severe AHRF 3.85 after 2, 6 and 12 h of HFNC therapy, respectively, In patients with AHRF, NHF is usually initiated in case were more likely to fail. In addition, patients who failed of persistence of hypoxemia and/or respiratory dis- presented a smaller increase in the ROX index between tress despite conventional oxygen therapy. Hence, lev- 2 and 12 h and 6 and 12 h, which highlights the fact that els of hypoxemia of these patients vary considerably, dynamic assessment of the ROX index may also be help- from mild- to-moderate to severe hypoxemia ­(PaO2/ ful to identify those patients who are more likely to fail. FiO2 < 200). Although a noticeable proportion of them This dynamic assessment could be especially useful in is successfully treated with NHF, the lower boundary in classifying those patients who have ROX index values terms of P­ aO2/FiO2 has not been established. between the boundaries of failure and success facilitating In the Florali study [12], among the 80% of patients the day-to-day clinical decision process of NHF patients with a ­PaO2/FiO2 below 200, this ratio was in the range and minimizing the risk of delaying a needed intubation of 120 and many had bilateral chest X-ray involvement. (see Fig. 2 and ESM for detail). Of note however, the ROX They would have probably qualified as ARDS patients. was validated in patients with pneumonia-related AHRF. They clearly presented the clinical, radiographical and Although severe pneumonia is the most frequent cause biological features of ARDS. It has indeed been con- of AHRF, to what extent the same cut-offs values for the vincingly shown by Garcia-de-Acilu et al. that hypox- ROX will have the same performances in other etiologies emic patients with bilateral infiltrates treated with of AHRF remains to be shown. NHF present the same levels of plasma biomarkers of epithelial and endothelial injury and biomarkers of Hazards associated with NHF inflammation as ARDS patients who undergo mechani- As mentioned earlier, between 30 and 40% of patients cal ventilation [21]. Importantly, half of these patients with AHRF initially treated with NHF will require inva- treated with NHF did not require intubation. Similar sive mechanical ventilation finally. Hence, the principal findings were reported by Messika et al. In their cohort hazard associated with NHF is the delay in intubat- of 45 patients treated with NHF, all of whom met all the ing these patients in which NHF is insufficient. One Berlin criteria other than PEEP, success rate was 60%. study has shown that delayed intubation is associated Of note, worst recorded P ­ aO2/FiO2 for the cohort was with increased mortality in patients with community- acquired pneumonia. This observation was initially Fig. 2 Suggested algorithm using the ROX index to help with intubation decision. Because the index includes in a single value three relevant respiratory parameters, the overall philosophy of the index is that if its value is increasing, the patient’s respiratory status is improving. For each time- point, there are three possibilities: (1) the patient’s ROX index is below the cut-off value, we suggest considering intubation of the patient; (2) the index is between the lower and the higher cut-off value, we suggest increasing the level of NHF (increase flow to its maximum and F­ iO2 to 1) and re-evaluate after 30 min; (3) finally, if the index is above the upper boundary, we suggest pursuing NHF and close monitoring of the patient. Of note, this algorithm will require a formal validation by a RCT comparing standard of care and application of the algorithm in terms of safety and efficacy (timing of intubation) made in patients with AHRF receiving NIV [24] and has monitoring as described above of patients under NHF is been shown to occur also in patients treated with NHF. mandatory to enable early detection of clinical deterio- Rello et al. found that patients who failed NHF in H1N1 ration, and thereby prevent evitable delay in intubation. pneumonia-related ARDS had a 27% mortality rate whereas this figure was 20% in those immediately intu- NHF in immunocompromised patients bated, suggesting that in some instances at least, a delay In the past years, several studies have assessed non- in intubation might be detrimental [10]. Kang et al. invasive management of AHRF in immunocompromised further explored this hazard. They found very different patients (ESM Table 2). Most of them were retrospective. mortality rates in patients treated with NHF between Comparators (NIV or standard oxygen or NIV + NHF), those intubated within 48 h of therapy and those intu- primary endpoint (intubation or mortality) as well as bated after. This landmark study clearly illustrated the type of immunosuppression differed between studies, risk of a delayed intubation [25]. Several reasons may making direct comparisons difficult [9, 27–33]. None- explain these findings. Leaving a patient breathe spon- theless, taken together, these studies suggested that NHF taneously for too long may worsen the nature and the might reduce intubation rate, particularly when com- extent of the initial injury, a concept termed patient pared to NIV. They also showed that use of HNFC was self-inflicted lung injury. In addition, oxygen supple- possibly associated with a lower mortality rate and was mentation through NHF may rapidly normalize oxy- not harmful. gen saturation leading to the misjudgment that the Those results were confirmed in an international mul- patient is improving whereas in fact, the underlying ticentre observational cohort study [34] in which patients disturbances (such as ventilation perfusion mismatch admitted to ICU who received either NHF, NIV, standard or alveolar hypoventilation requiring inspiratory pres- oxygen or NHF + NIV were analysed. Intubation rate was sure support and positive-end expiratory pressure) are similar for each oxygenation strategy. However, the mul- insufficiently corrected [26]. Patients may hence gradu- tivariate analysis indicated a trend towards lower intuba- ally and “silently” deteriorate to the point of respiratory tion rate with NHF. muscle fatigue, cardiac dysfunction, and organ failure, These observational data required a prospective multi- thus contributing to a worse prognosis. Hence, close centre RCT to confirm the trends their results suggested. Such a trial was recently performed in 776 immunocom- changes increase tidal volume, decrease respiratory rate promised patients who either received standard oxygen and end-expiratory lung volume [4, 39, 41, 42] which or NHF [32]. Primary endpoint was D-28 mortality and result in a measurable reduction in work of breathing secondary endpoints included intubation rate. NIV was [23]. not used according to the results of a previous study [32]. A direct comparison of respiratory parameters under Patients were mostly admitted for pneumonia. Mortality three conditions (oxygen, NHF and NIV) was conducted rates did not differ between NHF (35, 6%) and standard by Longhini et al. [43]. Diaphragmatic sonography and oxygen (36, 1%) (p = 0, 94), as well as intubation rates ventilatory parameters rose during interruptions of NIV (respectively, 38, 7% and 43.8%, p = 0, 17). The mortality trials (30 min). P ­ aCO2 remained stable irrespective of the rates for patients who required intubation did not dif- modality of oxygen administration. However, respiratory fer either. In a predefined subgroup of patients (oxygen rate and diaphragm contractile activation (inspiratory flow > 9 l/min, duration between ICU admission and ran- thickening fraction) increased with standard oxygen, but domization > 3 days or use of vasopressors), primary and not with NHF. The latter was associated with improved secondary endpoints did not differ. In this study, greater comfort. This study, therefore, suggested that NHF might ­PaO2/FiO2 obtained with NHF did not translate into bet- lead to a greater unload of the diaphragm than standard ter outcome. oxygen. More recently Dumas et al. analysed the daily probabil- In stable COPD patients, studies have consist- ity of intubation according to the patient’s characteristics ently shown that NHF reduces hypercapnia despite a the day before [33]. Using a propensity score, authors decreased respiratory rate [41, 42, 44]. This is related to did not find any impact of the oxygenation device on the a flow- and leakage-dependent clearance of upper air- probability of intubation the following day. In this post way gases that reduces ­CO2 rebreathing [44, 45]. These hoc joint analysis of three previously published studies, findings were confirmed by Bräunlich et al. in the clinical only a small proportion of patients received NHF. setting [46]. They found that after 6 weeks, reduction of How can data from non-immunocompromised hypercapnia and improvement in quality of life obtained patients indicating more positive effects of NHF be rec- with NHF were non-inferior to those obtained with NIV. onciled with the more neutral data coming from immu- These results may be, however, limited to patients with- nocompromised patients? One possible explanation is out overlap syndrome (i.e. association of COPD and that patients’ underlying disease and/or the precipitat- obstructive sleep apnea), as recently demonstrated [47]. ing factor leading to AHRF in immunocompromised A summary of these studies is detailed in ESM Table 3. patients require more time for recovery. Consequently, these patients may have a longer duration of oxygen Strategies for NHF in hypercapnic respiratory failure dependency and may require more invasive procedures. NIV is the ventilatory modality of first choice in hyper- In addition, comfort was not improved in these patients. capnic ARF [48]. However poor tolerance often limits its This may be related to other sources of discomfort (side success. effects of chemotherapy) than respiratory ones in those NHF has been suggested as complementary therapy patients that were not relieved by NHF. For all those rea- during breaks off NIV [43, 49], or as an alternative to NIV sons, the nature of the ventilatory support may not have [50] or controlled oxygen therapy [51] in mild respiratory such impact in this particular setting. acidosis. Recent meta-analyses focusing on immunocom- By overcoming the drawbacks of standard oxygen dur- promised patients [35, 36] have included the stud- ing breaks off NIV [43], NHF could reduce the time spent ies presented here and all conclude that use of NHF is on ventilator and improve work of breathing during associated with a decreased intubation rate but with no breaks [50]. A single RCT assessed NHF in this scenario, impact on mortality. on a mixed population with 22 out of 54 hypercapnic subjects [49], while not reducing the time spent on NIV, NHF in hypercapnic respiratory failure NHF improved comfort, control of dyspnea and respira- Physiological effects in stable COPD tory rate compared to standard oxygen. The main limita- As illustrated above, there is a strong rationale for the use tions of this study, i.e. small sample size and possibility to of NHF in patients with chronic respiratory failure [37]. override the criteria for NIV sessions, are expected to be Comparable to patients with AHRF, NHF may reduce overcome by the High-flow ACRF trial [52]. This ongo- the entrainment of room air during high breathing ing large-scale RCT will assess ventilator-free days when efforts, such as during exercise [38]. Early physiologi- comparing NHF to standard oxygen during breaks off cal studies revealed an increase in mean airway pressure NIV. [39], and a decrease in dead-space ventilation [40]. These NHF has been proposed as an alternative to NIV in Hence, NHF has also been evaluated as a means to pre- mild-to-moderate respiratory acidosis [51]. An ongoing vent desaturation during intubation [59]. Compared non-inferiority multi-center trial will further explore this to standard oxygen, results of preoxygenation with strategy by assessing short-term effects on gas exchange NFH seem consistent in patients with mild-to-moder- and treatment failure [53]. ate hypoxemia [57, 60]. Paradoxically, in patients with These and future studies will be needed to better assess more severe hypoxemia, NHF failed to prove superior the role of NHF in hypercapnic ARF, to understand than standard oxygen to prevent profound desaturation which strategies are better to pursue. during intubation [61, 62]. Of note however, subopti- mal preoxygenation procedures (absence of jaw thrust NHF and preoxygenation which is essential to promote apneic oxygenation) may A significant proportion of patients admitted to the explain these discrepancies [63]. This illustrates the ICU for AHRF will fail a noninvasive ventilatory strat- importance of the two components of the “preoxygena- egy and, thus, require tracheal intubation [54]. In these tion process”: preoxygenation and apneic oxygenation. patients, this procedure is at even greater risk of com- Intuitively, because of the greater levels of positive plications, mainly oxygen desaturation [55]. Studies pressure applied, NIV may be seen superior to NHF have clearly shown that standard preoxygenation, even during the spontaneously breathing phase and NHF when properly performed, is insufficient to ensure satis- may prove more beneficial during apnea and laryngos- factory oxygenation levels in critically ill patients. This copy because it can be left in place. This hypothesis was has led clinicians to investigate other modes of pre- tested in a recent RCT that compared NIV and NHF to oxygenation. NIV was convincingly shown to be supe- ensure preoxygenation of patients with AHRF requir- rior to standard oxygen [56], although a confirmatory ing tracheal intubation in the ICU [64]. Primary out- trial failed to confirm NIV’s superiority [57]. Because a come measure was incidence of profound desaturation growing number of patients with AHRF are treated with (SpO2 < 80%). There was no difference between the two NHF as first-line ventilatory support [58], the device is strategies in this incidence (23% with NIV v. 27% with often in place when the decision to intubate is taken. HFNC) in the overall study patients. In patients with Fig. 3 Suggested algorithm for deciding how to preoxygenate a patient with AHRF who requires tracheal intubation. NIV: non-invasive ventila‑ tion; NHF: nasal high flow; PEEP: positive-end expiratory pressure; PSV: pressure support ventilation. *The MACOCHA score can be used; **low dose vasopressor may be also started in unstable patients; ***RSI: rapid sequence induction; ****jaw thrust is essential to ensure patent upper airway and efficient apneic oxygenation moderate-to-severe hypoxemia ­(PaO2/FiO2 ≤ 200 mm with more than 12 h of mechanical ventilation, NHF sig- Hg), desaturation < 80% occurred less frequently after nificantly reduce post-extubation respiratory failure and preoxygenation with NIV than with nasal high-flow reintubation rate after 24 h in comparison with standard (24% v. 35% respectively, p = 0·046). Serious adverse oxygen. No such difference was found in another study events did not differ between the two groups. Finally, in uncomplicated patients after abdominal surgery [69]. based on the above, the combination of both NIV and A probably lower risk for respiratory complications in NHF to prevent desaturation during intubation was the OPERA trial explains this discrepancy [69]. Recent compared to NIV alone. Significantly less desaturations meta-analyses confirm the superiority of NHF over con- occurred with the combination [65]. Based on all these ventional oxygen therapy in terms of post-extubation res- results, we can suggest an algorithm to guide clinicians piratory failure and reintubation rate [75]. in their choice of preoxygenation technique (see Fig. 3). In high-risk patients, two large studies found that NHF was non-inferior to NIV [68, 72]: one in a general popula- NHF in the postextubation tion [72], the other cardiothoracic surgery patients [68]. Preventing reintubation is the cornerstone of patient Discomfort limited the effective time under NIV in these management after extubation because of the complica- studies to 14 and 6.6 h, respectively, during the first 24 h. tions associated with reintubation (prolonged mechanical This was particularly the case after cardiothoracic sur- ventilation and ICU stay, ventilator-associated pneumo- gery, where discomfort was significantly more frequent nia…). It has been suggested that NIV should not be used with NIV than with NHF. in patients with established post-extubation respiratory A recent trial compared NHF to the combination of failure [48], after a safety concern was raised secondary NHF and NIV in high-risk patients [76]. Patients received to the risk of delaying reintubation. However, NIV has a preventive therapy for a minimum of 48 h. Very similar preventive role in patients at high-risk for failure [48, 66]. to previous results [72], time spent under NIV was 13 h Estimation of risk for failure generates controversy, during the first 24 h highlighting the fact that NIV toler- as no score predicting failure in a general population of ance may be a limitation to this strategy in this subset of intubated patients has been validated and risks depend patients. Nonetheless, authors found that the combina- largely on patient population. ESM Table 4 displays tion of NHF and NIV led to less reintubation that NHF these different risk factors. In surgical patients, expected alone. Interestingly, reintubation rate at day 7 was signifi- reintubation rate is low because of the short duration of cantly lower with NIV than with NHF alone in patients mechanical ventilation, and reasons for failing extubation with ­PaCO2 > 45 mmHg before extubation, but no such are specific (e.g. anaesthetic factors, airway patency, sur- difference was found in patients with ­PaCO2 < 45 mmHg. gical complications) [67]. Thus, studies in these patients In high-risk patients, the combination of NIV and NHF are usually powered for primary outcomes different therefore seems to benefit more patients with pre-extu- from the reintubation rate, and include composite end- bation hypercapnia. Duration of preventive therapy and points or have a non-inferiority design [68, 69]. In medi- date of assessment of reintubation are important vari- cal patients, many risk factors have been described, but ables to take into account. most of them with a low level of evidence [70]. In addi- tion, subjective definitions for these risks (those related NHF in the ED and palliative care to cough and secretions management for example) limit Several RCTs compare NHF to standard oxygen via their reproducibility in clinical trials. Traditional defi- low-flow nasal prongs or facemask in ED for adults with nition of patients at high-risk required at least one risk moderate-to-severe respiratory distress in different factor, but recent trials including a large number of those conditions [15–18]. A systematic review of these RCTs risk factors, suggest that the risk for failing extubation [15–18] found no difference in mortality (RR 1.20; 95% should be better defined [71, 72]. This is supported by CI 0.58, 2.48, p = 0.62) or treatment failure (RR 1.49; 95% results of multivariable models, reporting more than one CI 0.33, 6.82, p = 0.60) compared to standard oxygen. independently associated high-risk factor to reintuba- However, more patients did not tolerate NHF than stand- tion [70], with a potential summation effect, and previous ard oxygen (RR 6.81; 95% CI 1.18, 39.19, p = 0.03) [19]. data reporting a synergistic interaction between high-risk These data are limited as rates of intubation (1.2%) and factors for reintubation [73]. These issues are worth con- NIV (3.1%) were low, despite enrolment of patients with sidering when analyzing the data. AHRF in the included studies. This is because patients Several studies have compared NHF to standard oxy- needing immediate NIV or intubation were excluded gen to prevent reintubation according to a given risk for from the studies. Standard care in ED is usually effec- failure. In a large heterogeneous population of patients tive in avoiding mechanical ventilation in the remain- at low risk [74], including complicated surgical patients ing patients and studies may have included patients with ceilings of care. So, any differences between NHF and Institute for Medical Research, University of Leeds, Leeds, UK. 18 Saint Eloi ICU, Montpellier University Hospital and PhyMedExp, INSERM, CNRS, 34000 Mont‑ standard oxygen for these outcomes are likely to be small. pellier, France. 19 Intensive Care Medicine, University Hospital Virgen de la Compared to NIV for undifferentiated respiratory failure, Salud, Toledo, Spain. there was no difference between NHF for any outcome in Acknowledgements a single RCT [20]. In contrast, a small RCT in palliative The authors thank Camille Le Breton for her support in drawing Figure n.3. patients with limitations on invasive ventilation found dyspnea was reduced with NHF compared to NIV [77]. Compliance with ethical standards NHF may have a role in some ED patients who would not Conflicts of interest be offered intubation or when NIV is not well tolerated. JDR received travel expenses and accommodation coverage from More studies are required to confirm this. Fisher&Paykel Healthcare to attend scientific meetings. Fisher&Paykel Health‑ care provided support for the ongoing High Flow ACRF trial (NCT03406572) NIV is recommended in palliative care [48]. However, but took no part in design or conduct of the study. OR received speaker fees patient discomfort and inability to eat or communicate from Air Liquide. His institution received consultancy fees from Hamilton may limit its use. NHF may be an alternative method for Medical. VL has no conflict of interest to declare concerning this topic. JB received travel expenses, lecture fees and accommodation coverage from providing oxygen and respiratory support in this setting. TNI medical AG and Fisher & Paykel Healthcare. TNI medical AG provided Many studies have reported the benefit of NHF by allevi- support for several NHF trials. AC’s employer, on her behalf, has received travel ating dyspnea in patients with advanced cancer and do- expenses, lecture fees and accommodation from Fisher & Paykel Healthcare to attend scientific meetings. Fisher & Paykel Healthcare has also provided an not-intubate patients with hypoxemic respiratory failure unrestricted grant to support investigator driven research, but has had no part [78–81]. Furthermore, NHF might not be inferior to NIV in study design, conduct, analysis or reporting of the studies. PJ received com‑ in palliation of dyspnea in patients with terminal cancer petitive grant funding from the A + Trust (4926) and Greenlane Research and Education Fund (12/15/4086) for the HOTER RCT (ACTRN12610000960411) on [82]. However, further studies are needed to evaluate the HFNO in the ED. F&P Healthcare provided equipment for the HOTER RCT for benefits of NHF in terms of patient comfort and quality which PJ was the principal investigator, but took no part in design, conduct, of life in palliative care. analysis or reporting of the study. PJ has no other personal financial or profes‑ sional to declare. BJK: has no conflict of interest to declare. FL: Fisher&Paykel Healthcare provided support for the development of an application VentilO Concluding remarks on optimization of protective mechanical ventilation. SN: Fisher&Paykel Healthcare provided support for the study # NCT03759457. NR received travel In a small 10-year span, NHF has found its place in the expenses and lecture fees from Fisher&Paykel Healthcare. GS: Fisher&Paykel armamentarium of AHRF treatment. In every segment of Healthcare provided support for the study # NCT03965832 for which GS AHRF management (first-line ventilatory support, pre- is the principal investigator. SJ received personal fees as consultant from Fisher-Paykel. GH received travel expenses coverage and lecture fees from oxygenation, post-extubation, palliative care), evidence Fisher&Paykel. suggests benefits to the patient. Interestingly, there are still aspects that further need investigation. These include Publisher’s Note improving comfort, exploring the potential for even Springer Nature remains neutral with regard to jurisdictional claims in pub‑ higher flows, and addressing the issue of weaning. lished maps and institutional affiliations. Electronic supplementary material Received: 12 May 2020 Accepted: 21 August 2020 The online version of this article (https​://doi.org/10.1007/s0013​4-020-06228​-7) contains supplementary material, which is available to authorized users. Author details References 1 Medico‑surgical ICU, Assistance Publique ‑ Hôpitaux de Paris, DMU ESPRIT, 1. Delorme M, Bouchard P-A, Simon M et al (2017) Effects of high-flow nasal Médecine Intensive Réanimation, Hôpital Louis Mourier, 92700 Colombes, cannula on the work of breathing in patients recovering from acute France. 2 Université de Paris, IAME, U1137, Inserm, 75018 Paris, France. 3 Critical respiratory failure. Crit Care Med 45:1981–1988 Care Department, Vall d’Hebron University Hospital, Vall d’Hebron Research 2. Mauri T, Alban L, Turrini C et al (2017) Optimum support by high-flow Institute, Universitat Autònoma de Barcelona, Barcelona, Spain. 4 Ciber nasal cannula in acute hypoxemic respiratory failure: effects of increasing Enfermedades Respiratorias (CIBERES), Instituto de Salud Carlos III, Madrid, flow rates. Intensive Care Med 43:1453–1463 Spain. 5 Medical ICU Saint Louis Hospital, APHP, Paris, France. 6 Critical Care 3. Mauri T, Turrini C, Eronia N et al (2017) Physiologic effects of high-flow Research Group, The Prince Charles Hospital, Chermside, Australia. 7 University nasal cannula in acute hypoxemic respiratory failure. Am J Respir Crit Care of Queensland, Brisbane, QLD, Australia. 8 Department of Respiratory Medi‑ Med 195:1207–1215 cine, University of Leipzig, Liebigstraße 20, 04103 Leipzig, Germany. 9 Klinikum 4. Bräunlich J, Köhler M, Wirtz H (2016) Nasal highflow improves ventilation Emden, Bolardusstrasse 20, 26721 Emden, Germany. 10 School of Medicine, in patients with COPD. Int J Chron Obstruct Pulmon Dis 11:1077–1085 University of Auckland, Auckland, New Zealand. 11 Department of Emergency 5. Corley A, Caruana LR, Barnett AG et al (2011) Oxygen delivery through Medicine, Auckland City Hospital, Auckland, New Zealand. 12 Department high-flow nasal cannulae increase end-expiratory lung volume and of Internal Medicine, Ulsan University Hospital, University of Ulsan College reduce respiratory rate in post-cardiac surgical patients. Br J Anaesth of Medicine, Ulsan, Korea. 13 Quebec Heart and Lung Institute, Laval University, 107:998–1004 Québec City, QC, Canada. 14 Department of Clinical, Integrated, and Experi‑ 6. Mauri T, Galazzi A, Binda F et al (2018) Impact of flow and temperature on mental Medicine (DIMES), Respiratory and Critical Care, Sant’Orsola Malpighi patient comfort during respiratory support by high-flow nasal cannula. Hospital, Bologna, Italy. 15 Division of Respiratory Diseases and Tuberculosis, Crit Care 22:120 Department of Medicine, Faculty of Medicine Siriraj Hospital, Mahidol Uni‑ 7. Messika J, Ben Ahmed K, Gaudry S et al (2015) Use of high-flow nasal can‑ versity, Bangkok, Thailand. 16 Department of Respiratory Medicine, St James’s nula oxygen therapy in subjects with ARDS: a 1-year observational study. University Hospital, Leeds Teaching Hospital NHS Trust, Leeds, UK. 17 Leeds Respir Care 60:162–169 8. Roca O, Pérez-Terán P, Masclans JR et al (2013) Patients with New York with acute respiratory failure: an observational cohort study. Ann Inten‑ Heart Association class III heart failure may benefit with high flow nasal sive Care 6:45 cannula supportive therapy: high flow nasal cannula in heart failure. J Crit 29. Frat J-P, Ragot S, Girault C et al (2016) Effect of non-invasive oxygenation Care 28:741–746 strategies in immunocompromised patients with severe acute respira‑ 9. Roca O, de Acilu MG, Caralt B et al (2015) Humidified high flow nasal can‑ tory failure: a post hoc analysis of a randomised trial. Lancet Respir Med nula supportive therapy improves outcomes in lung transplant recipients 4:646–652 readmitted to the intensive care unit because of acute respiratory failure. 30. Tu G, He H, Yin K et al (2017) High-flow nasal cannula versus noninvasive Transplantation 99:1092–1098 ventilation for treatment of acute hypoxemic respiratory failure in renal 10. Rello J, Pérez M, Roca O et al (2012) High-flow nasal therapy in adults with transplant recipients. Transpl Proc 49:1325–1330 severe acute respiratory infection: a cohort study in patients with 2009 31. Lemiale V, Resche-Rigon M, Mokart D et al (2017) High-flow nasal cannula influenza A/H1N1v. J Crit Care 27:434–439 oxygenation in immunocompromised patients with acute hypoxemic 11. Sztrymf B, Messika J, Bertrand F et al (2011) Beneficial effects of humidi‑ respiratory failure: a groupe de recherche respiratoire en réanimation fied high flow nasal oxygen in critical care patients: a prospective pilot onco-hématologique study. Crit Care Med 45:e274–e280 study. Intensive Care Med 37:1780–1786 32. Azoulay E, Lemiale V, Mokart D et al (2018) Effect of high-flow nasal 12. Frat JP, Thille AW, Mercat A et al (2015) High-flow oxygen through oxygen vs standard oxygen on 28-day mortality in immunocompromised nasal cannula in acute hypoxemic respiratory failure. N Engl J Med patients with acute respiratory failure: the HIGH randomized clinical trial. 372(23):2185–2196 JAMA 320:2099–2107 13. Rochwerg B, Granton D, Wang DX et al (2019) High flow nasal cannula 33. Dumas G, Chevret S, Lemiale V et al (2018) Oxygenation/non-invasive compared with conventional oxygen therapy for acute hypoxemic ventilation strategy and risk for intubation in immunocompromised respiratory failure: a systematic review and meta-analysis. Intensive Care patients with hypoxemic acute respiratory failure. Oncotarget Med 45:563–572 9:33682–33693 14. Monro-Somerville T, Sim M, Ruddy J et al (2017) The effect of high-flow 34. Azoulay E, Pickkers P, Soares M et al (2017) Acute hypoxemic respira‑ nasal cannula oxygen therapy on mortality and intubation rate in acute tory failure in immunocompromised patients: the Efraim multinational respiratory failure: a systematic review and meta-analysis. Crit Care Med prospective cohort study. Intensive Care Med 43:1808–1819 45:e449–e456 35. Cortegiani A, Crimi C, Sanfilippo F et al (2019) High flow nasal therapy 15. Jones PG, Kamona S, Doran O et al (2016) Randomized controlled trial of in immunocompromised patients with acute respiratory failure: a humidified high-flow nasal oxygen for acute respiratory distress in the systematic review and meta-analysis. J Crit Care 50:250–256 emergency department: the HOT-ER study. Respir Care 61:291–299 36. Wang Y, Ni Y, Sun J, Liang Z (2020) Use of high-flow nasal cannula for 16. Makdee O, Monsomboon A, Surabenjawong U et al (2017) High-flow immunocompromise and acute respiratory failure: a systematic review nasal cannula versus conventional oxygen therapy in emergency and meta-analysis. J Emerg Med 58:413–423 department patients with cardiogenic pulmonary edema: a randomized 37. Pisani L, Vega ML (2017) Use of nasal high flow in stable COPD: ration‑ controlled trial. Ann Emerg Med 70:465–472.e2 ale and physiology. COPD J Chron Obstruct Pulmon Dis 14:346–350 17. Bell N, Hutchinson CL, Green TC et al (2015) Randomised control trial of 38. Ritchie JE, Williams AB, Gerard C, Hockey H (2011) Evaluation of a humidified high flow nasal cannulae versus standard oxygen in the emer‑ humidified nasal high-flow oxygen system, using oxygraphy, capnog‑ gency department: high flow oxygen randomised control trial. Emerg raphy and measurement of upper airway pressures. Anaesth Intensive Med Australas 27:537–541 Care 39:1103–1110 18. Rittayamai N, Tscheikuna J, Praphruetkit N, Kijpinyochai S (2015) Use of 39. Bräunlich J, Beyer D, Mai D et al (2013) Effects of nasal high flow on high-flow nasal cannula for acute dyspnea and hypoxemia in the emer‑ ventilation in volunteers, COPD and idiopathic pulmonary fibrosis gency department. Respir Care 60:1377–1382 patients. Respiration 85:319–325 19. Tinelli V, Cabrini L, Fominskiy E et al (2019) High flow nasal cannula 40. Biselli P, Fricke K, Grote L et al (2018) Reductions in dead space ventila‑ oxygen vs. conventional oxygen therapy and noninvasive ventilation in tion with nasal high flow depend on physiological dead space volume: emergency department patients: a systematic review and meta-analysis. metabolic hood measurements during sleep in patients with COPD J Emerg Med 57:322–328 and controls. Eur Respir J 51:1702251 20. Doshi P, Whittle JS, Bublewicz M et al (2018) High-velocity nasal insuffla‑ 41. Fraser JF, Spooner AJ, Dunster KR et al (2016) Nasal high flow oxygen tion in the treatment of respiratory failure: a randomized clinical trial. Ann therapy in patients with COPD reduces respiratory rate and tissue car‑ Emerg Med 72:73–83.e5 bon dioxide while increasing tidal and end-expiratory lung volumes: a 21. García-de-Acilu M, Marin-Corral J, Vázquez A et al (2017) Hypoxemic randomised crossover trial. Thorax 71:759–761 patients with bilateral infiltrates treated with high-flow nasal cannula 42. Pisani L, Fasano L, Corcione N et al (2017) Change in pulmonary present a similar pattern of biomarkers of inflammation and injury to mechanics and the effect on breathing pattern of high flow oxygen acute respiratory distress syndrome patients. Crit Care Med 45:1845–1853 therapy in stable hypercapnic COPD. Thorax 72:373–375 22. Roca O, Messika J, Caralt B et al (2016) Predicting success of high-flow 43. Longhini F, Pisani L, Lungu R et al (2019) High-flow oxygen therapy nasal cannula in pneumonia patients with hypoxemic respiratory failure: after noninvasive ventilation interruption in patients recovering from the utility of the ROX index. J Crit Care 35:200–205 hypercapnic acute respiratory failure: a physiological crossover trial. 23. Roca O, Caralt B, Messika J et al (2019) An index combining respiratory Crit Care Med 47:e506–e511 rate and oxygenation to predict outcome of nasal high-flow therapy. Am 44. Bräunlich J, Mauersberger F, Wirtz H (2018) Effectiveness of nasal high‑ J Respir Crit Care Med 199:1368–1376 flow in hypercapnic COPD patients is flow and leakage dependent. 24. Bellani G, Laffey JG, Pham T et al (2017) Noninvasive ventilation of BMC Pulm Med 18:14 patients with acute respiratory distress syndrome. insights from the LUNG 45. Möller W, Feng S, Domanski U et al (2017) Nasal high flow reduces SAFE study. Am J Respir Crit Care Med 195:67–77 dead space. J Appl Physiol 122:191–197 25. Kang BJ, Koh Y, Lim C-M et al (2015) Failure of high-flow nasal cannula 46. Bräunlich J, Dellweg D, Bastian A et al (2019) Nasal high-flow versus therapy may delay intubation and increase mortality. Intensive Care Med noninvasive ventilation in patients with chronic hypercapnic COPD. 41:623–632 COPD 14:1411–1421 26. Nedel WL, Deutschendorf C, Moraes Rodrigues Filho E (2017) High-flow 47. Pisani L, Betti S, Biglia C et al (2020) Effects of high-flow nasal cannula nasal cannula in critically ill subjects with or at risk for respiratory failure: a in patients with persistent hypercapnia after an acute COPD exacerba‑ systematic review and meta-analysis. Respir Care 62:123–132 tion: a prospective pilot study. BMC Pulm Med 20:12 27. Mokart D, Geay C, Chow-Chine L et al (2015) High-flow oxygen therapy 48. Rochwerg B, Brochard L, Elliott MW et al (2017) Official ERS/ATS clini‑ in cancer patients with acute respiratory failure. Intensive Care Med cal practice guidelines: noninvasive ventilation for acute respiratory 41:2008–2010 failure. Eur Respir J 50:1602426 28. Coudroy R, Jamet A, Petua P et al (2016) High-flow nasal cannula oxygen 49. Spoletini G, Mega C, Pisani L et al (2018) High-flow nasal therapy vs therapy versus noninvasive ventilation in immunocompromised patients standard oxygen during breaks off noninvasive ventilation for acute respiratory failure: a pilot randomized controlled trial. J Crit Care 66. Esteban A, Frutos-Vivar F, Ferguson ND et al (2004) Noninvasive positive- 48:418–425 pressure ventilation for respiratory failure after extubation. N Engl J Med 50. Lee MK, Choi J, Park B et al (2018) High flow nasal cannulae oxygen 350:2452–2460 therapy in acute-moderate hypercapnic respiratory failure. Clin Respir J 67. Canet J, Gallart L, Gomar C et al (2010) Prediction of postoperative pul‑ 12:2046–2056 monary complications in a population-based surgical cohort. Anesthesi‑ 51. Pilcher J, Eastlake L, Richards M et al (2017) Physiological effects of ology 113:1338–1350 titrated oxygen via nasal high-flow cannulae in COPD exacerbations: a 68. Stéphan F, Barrucand B, Petit P et al (2015) High-flow nasal oxygen vs randomized controlled cross-over trial. Respirology 22:1149–1155 noninvasive positive airway pressure in hypoxemic patients after cardio‑ 52. Ricard J-D, Dib F, Esposito-Farese M et al (2018) Comparison of high thoracic surgery: a randomized clinical trial. JAMA 313:2331–2339 flow nasal cannula oxygen and conventional oxygen therapy on ven‑ 69. Futier E, Paugam-Burtz C, Godet T et al (2016) Effect of early postextuba‑ tilatory support duration during acute-on-chronic respiratory failure: tion high-flow nasal cannula vs conventional oxygen therapy on hypox‑ study protocol of a multicentre, randomised, controlled trial. The aemia in patients after major abdominal surgery: a French multicentre “HIGH-FLOW ACRF” study. BMJ Open 8:e022983 randomised controlled trial (OPERA). Intensive Care Med 42:1888–1898 53. Cortegiani A, Longhini F, Carlucci A et al (2019) High-flow nasal 70. Thille AW, Boissier F, Ben Ghezala H et al (2015) Risk factors for and predic‑ therapy versus noninvasive ventilation in COPD patients with mild-to- tion by caregivers of extubation failure in ICU patients: a prospective moderate hypercapnic acute respiratory failure: study protocol for a study. Crit Care Med 43:613–620 noninferiority randomized clinical trial. Trials 20:450 71. Ferrer M, Valencia M, Nicolas JM et al (2006) Early noninvasive ventilation 54. Ricard JD, Gaborieau B, Bernier J et al (2019) Use of high flow nasal can‑ averts extubation failure in patients at risk: a randomized trial. Am J Respir nula for preoxygenation and apneic oxygenation during intubation. Ann Crit Care Med 173:164–170 Transl Med 7:S380 72. Hernández G, Vaquero C, Colinas L et al (2016) Effect of postextubation 55. Ricard JD (2016) Hazards of intubation in the ICU: role of nasal high flow high-flow nasal cannula vs noninvasive ventilation on reintubation and oxygen therapy for preoxygenation and apneic oxygenation to prevent postextubation respiratory failure in high-risk patients: a randomized desaturation. Minerva Anestesiol 82:1098–1106 clinical trial. JAMA 316:1565–1574 56. Baillard C, Fosse J-P, Sebbane M et al (2006) Noninvasive ventilation 73. Salam A, Tilluckdharry L, Amoateng-Adjepong Y, Manthous CA (2004) improves preoxygenation before intubation of hypoxic patients. Am J Neurologic status, cough, secretions and extubation outcomes. Intensive Respir Crit Care Med 174:171–177 Care Med 30:1334–1339 57. Baillard C, Prat G, Jung B et al (2018) Effect of preoxygenation using 74. Hernández G, Vaquero C, González P et al (2016) Effect of postextubation non-invasive ventilation before intubation on subsequent organ high-flow nasal cannula vs conventional oxygen therapy on reintubation failures in hypoxaemic patients: a randomised clinical trial. Br J Anaesth in low-risk patients: a randomized clinical trial. JAMA 315:1354–1361 120:361–367 75. Zhu Y, Yin H, Zhang R et al (2019) High-flow nasal cannula oxygen 58. Papazian L, Corley A, Hess D et al (2016) Use of high-flow nasal can‑ therapy versus conventional oxygen therapy in patients after planned nula oxygenation in ICU adults: a narrative review. Intensive Care Med extubation: a systematic review and meta-analysis. Crit Care 23:180 42:1336–1349 76. Thille AW, Muller G, Gacouin A et al (2019) Effect of postextubation high- 59. Miguel-Montanes R, Hajage D, Messika J et al (2015) Use of high-flow flow nasal oxygen with noninvasive ventilation vs high-flow nasal oxygen nasal cannula oxygen therapy to prevent desaturation during tracheal alone on reintubation among patients at high risk of extubation failure: a intubation of intensive care patients with mild-to-moderate hypoxemia. randomized clinical trial. JAMA 322:1465–1475 Crit Care Med 43:574–583 77. Ruangsomboon O, Dorongthom T, Chakorn T et al (2020) High-flow nasal 60. Guitton C, Ehrmann S, Volteau C et al (2019) Nasal high-flow preoxygena‑ cannula versus conventional oxygen therapy in relieving dyspnea in tion for endotracheal intubation in the critically ill patient: a randomized emergency palliative patients with do-not-intubate status: a randomized clinical trial. Intensive Care Med 45:447–458 crossover study. Ann Emerg Med 75:615–626 61. Vourc’hVourc’h M, Asfar P, Volteau C et al (2015) High-flow nasal cannula 78. Epstein AS, Hartridge-Lambert SK, Ramaker JS et al (2011) Humidified oxygen during endotracheal intubation in hypoxemic patients: a rand‑ high-flow nasal oxygen utilization in patients with cancer at Memorial omized controlled clinical trial. Intensive Care Med 41:1538–1548 Sloan-Kettering Cancer Center. J Palliat Med 14:835–839 62. Simon M, Wachs C, Braune S et al (2016) High-flow nasal cannula versus 79. Peters SG, Holets SR, Gay PC (2013) Nasal high flow oxygen therapy in bag-valve-mask for preoxygenation before intubation in subjects with do-not-intubate patients with hypoxemic respiratory distress. Respir Care hypoxemic respiratory failure. Respir Care 61:1160–1167 58:597–600 63. De Jong A, Jaber S (2016) Apneic oxygenation for intubation in the criti‑ 80. Koyauchi T, Hasegawa H, Kanata K et al (2018) Efficacy and tolerability of cally ill. Let’s not give up! Am J Respir Crit Care Med 193:230–232 high-flow nasal cannula oxygen therapy for hypoxemic respiratory failure 64. Frat J-P, Ricard J-D, Quenot J-P et al (2019) Non-invasive ventilation versus in patients with interstitial lung disease with do-not-intubate orders: a high-flow nasal cannula oxygen therapy with apnoeic oxygenation for retrospective single-center study. Respiration 96:323–329 preoxygenation before intubation of patients with acute hypoxaemic 81. Zemach S, Helviz Y, Shitrit M et al (2019) The use of high-flow nasal can‑ respiratory failure: a randomised, multicentre, open-label trial. Lancet nula oxygen outside the ICU. Respir Care 64:1333–1342 Respir Med 7:303–312 82. Hui D, Morgado M, Chisholm G et al (2013) High-flow oxygen and bilevel 65. Jaber S, Monnin M, Girard M et al (2016) Apnoeic oxygenation via high- positive airway pressure for persistent dyspnea in patients with advanced flow nasal cannula oxygen combined with non-invasive ventilation cancer: a phase II randomized trial. J Pain Symptom Manage 46:463–473 preoxygenation for intubation in hypoxaemic patients in the intensive care unit: the single-centre, blinded, randomised controlled OPTINIV trial. Intensive Care Med 42:1877–1887