Role of High flow nasal cannula in managing Type I respiratory failure in patients with neuromuscular disorders

Abstract

Management of acute Type I respiratory failure (RF) in patients with neuromuscular disorders (NMD) and myopathies poses a clinical challenge in the acute care setting. There is a lack of studies which looked into the effectiveness of High flow nasal cannula (HFNC) in acutely ill NMD patients with Type I RF. This single centre, retrospective case-series study was conducted at the RICU of a tertiary care institute in the state of Rajasthan, India, between 1 January 2020 and 31 March 2025 .All patients with a diagnosis of NMD who were admitted in RICU with Type I respiratory failure were sequentially enrolled in the study. Along with demographic and clinical data of patients, arterial blood gas analysis, routine blood and radiological investigations were incorporated in the study. Patient data was traced till the patient was discharged from the hospital or expired. Risk factors for HFNC failure were evaluated. The patient data was analysed using R version 4.5.0. Risk factors found statistically significant on univariate analyses underwent logistic regression .Odds ratio was calculated for those factors which were significant on logistic regression. p value < 0.05 was taken to be significant. In the study period, 42 patients were enrolled, out of whom, 29 (69%) patients could be successfully managed on HFNC, while 13 (31%) patients had HFNC failure. On multiple logistic regression, exacerbation of restrictive lung disease in patients with kyphoscoliosis, serum albumin and Simplified Acute Physiology Score (SAPS) 3 scores were found to be independent risk factors for HFNC failure with Odds ratio (95% CI) of 4.1(2.1–6.1), 6.9(5.2–8.6) and 5.1(3.2-7) respectively. ROX > 4.88 at 12 h had a sensitivity of 80.3% and specificity of 69.23% and an accuracy of 76.2% of predicting HFNC success. This retrospective study shows that HFNC can be an essential adjunct in managing Type I respiratory failure in patients with neuromuscular disorders.

Introduction

Management of acute Type I respiratory failure (RF) in patients with neuromuscular disorders (NMD) and myopathies poses a clinical challenge in the acute care setting. In these patients partial pressure of arterial oxygen (PaO2) in ABG on room air is less than 60 mm Hg. Type I respiratory failure in these patients can be caused by aspiration pneumonia, community acquired pneumonia, atelectasis due to ineffective mucous clearing, pulmonary embolism and left ventricular failure. Non-invasive ventilation(NIV) in the form of BIPAP (bilevel positive airway pressure) or CPAP (continuous positive airway pressure) is often used in the management of patient of NMD with respiratory failure [1]. NIV is indicated in acutely ill patients with NMD even in the absence of hypercapnia to manage respiratory failure [2]. NIV ameliorates respiratory failure by unloading the inspiratory muscles of excessive work of breathing and increased ventilation by recruitment of alveolar units [1]. Intolerance of NIV is however a common occurrence, reported to be as high as 55.5% as in a nationwide study [3]. Poor tolerance can be due to facial skin abrasion, claustrophobia, gastric and/or colonic distension, or the accumulation of bronchial secretions due to the patient’s difficulty to cough [4]. High flow Nasal Cannula (HFNC) has emerged as an essential adjunct in management of Type I respiratory failure in recent times. The small, soft nasal interface and humidified delivery system of air with or without oxygen, clinically appears comfortable in patients who cannot tolerate NIV. High flow rates support alveolar recruitment and ventilation more than nasal prong or Hudson mask oxygen delivery and the nasal prong interface allow frequent suctioning and cleaning of pooled oral secretions. Whilst these are important treatment benefits in NMD patients with respiratory muscle weakness, dysphagia and a poor cough, use of HFNC carries failure risk [5]. There is a lack of studies which has looked into the effectiveness of HFNC in acutely ill NMD patients with Type I RF. To bridge this hiatus, our retrospective study was carried out in the RICU (Respiratory Intensive care unit) of a tertiary care hospital.

Methods

This single centre, retrospective case-series study was conducted at the RICU between 1 January 2020 and 31 March 2025 in a tertiary care institute in the state of Rajasthan, India. All participants provided consent forms permitting the use of their anonymized clinical data for research, analysis, and reporting. Each file was anonymised by assigning a unique de-identified code. The local Ethics Committee waived the requirement for ethical approval because the study was retrospective and not conducted as a formal research project (IRB min no:325 dated 11.12.2019). The study was conducted in compliance with the 1975 Declaration of Helsinki.

All patients with a diagnosis of NMD or myopathies who were admitted in RICU with Type I respiratory failure were sequentially enrolled in the study. HFNC was started at setting of 60% fio2 and 60 lt/min flow and titrated down to maintain spo2 > 92%.The patients’ age, gender, BMI, smoking status, type of NMD and comorbidities were among the demographic and clinical data gathered. Data regarding ABG, routine blood investigations including complete haemogram(CBC), Serum electrolytes (SE), Urea, creatinine, Liver function test at the time of admission were recorded. ABG was repeated at 2 h and again at 12 h and every 24 h. ABG was also repeated before declaring HFNC failure as per criteria and patient was put on NIV or Invasive Mechanical ventilation(IMV).

HFNC success was defined as the patient maintaining a saturation of 92% on HFNC and then could be gradually weaned off HFNC into low delivery oxygen systems. Criteria for HFNC failure was prespecified and followed in our RICU [5].

1. When a patient has acidosis (low blood pH) and refractory hypotension (systolic BP < 90 mmHg or mean BP < 65 mmHg).

2. If the patient’s mental status deteriorates (e.g., confusion, unresponsiveness) or there is a risk of aspiration (food or liquid entering the lungs), airway protection becomes essential.

3. Hypercapnia refers to elevated carbon dioxide levels in the blood, which can result in respiratory acidosis (pH < 7.35).

4. In hypoxemic respiratory failure, despite the maximal flow and FiO2 settings on the HFNC, the patient cannot maintain adequate oxygen saturation (SpO2 > 90%).

Findings of radiological investigations including Chest X Ray, or CT (computed tomography) Thorax if done were also included in our study. Fibreoptic bronchoscopy (FOB) was done in patients for microbiologial sampling in cases of pneumonia/ARDS if clinically indicated. FOB guided bronchial toileting was also done in patients who had poor cough reflex and had recurrent pooling of secretions. Clinical severity scores Simplified Acute Physiology Score (SAPS) 3 scores were calculated and correlated with the outcome of the patient like decrease in oxygen supplementation and requirement of NIV/IMV. Patient data was traced till the patient was discharged from the hospital or expired.

Muscular weakness was diagnosed when the Chest X ray of the patient was within normal limits and ultrasound of diaphragm muscles revealed decreased amplitude, thickness of diaphragm and shortening fraction of the muscles. Diaphragmatic values were measured on both right and left and their average were taken. Indices including shortening fraction < 20%, diaphragmatic excursion in quiet breathing < 15 mm and maximal inspiratory diaphragmatic excursion < 25 mm were taken to be indicators of diaphragmatic/muscular weakness [6].

Statistical analysis

The patient data was put into Microsoft Excel and analysed using R version 4.5.0.

Continuous variables were expressed in the form of mean ± standard deviation and categorical variables in the form of percentages. The statistical tests which have been used in the study were, t test for means, Chi-square test, Fisher’s exact test, and z proportion test. Risk factors found significant on univariate analyses underwent logistic regression and Odds ratio calculated for those factors which were significant on logistic regression. Receiver Operating Curve (ROC) analysis was done for ROX index done at 2 and 12 h with relation to HFNC failure.P value less than 0.05 was taken to be significant.

Results

Patients were subdivided on the basis of HFNC success or failure and various demographic, clinical and laboratory parameters were compared between these two groups. They are detailed in Table 1. Statistically significant p-values (<0.05) for parameters and their corresponding Odds ratios and 95% Confidence Intervals are shown in bold in Table 1.

Table 1.

Comparison of demographic, clinical and laboratory parameters between HFNC success and failure

Parameters	HFNC success group (n = 29)	HFNC failure group(n = 13)	Univariate analysis (statistical test used)	p value	Logistic regression p value	Odds ratio (95% CI)
Age, mean (sd) yrs	36(8.4)	39(7.9)	t test for means	0.56
Gender
Male (n,%)	16(55.2%)	8(61.5%)	Z proportion test	0.7
Female (n,%)	13(44.8%)	5(38.5%)	Z proportion test	0.7
Body mass index mean, (sd)kg/m2	22.5(4.2)	20.8(5.3)	t test for means	0.65
Diagnosis of Type I RF
Aspiration pneumonia(n,%)	6(20.7%)	0	Z proportion test	0.038	0.12
Community acquired pneumonia(n,%)	8(27.6%)	2(15.4%)	Z proportion test	0.19
Atelectasis due to mucous impaction(n,%)	4(13.8%)	1(7.7%)	Z proportion test	0.57
Muscular/diaphragmatic weakness	7(24.1%)	5(37.5%)	Z proportion test	1
Pulmonary embolism (n,%)	1(3.4%)	1(7.7%)	Z proportion test	0.55
Pulmonary oedema(n,%)	3(10.2%)	1(7.7%)	Z proportion test	0.79
Exacerbation of restrictive lung disease (n,%)	0	3(23.1%)	Z proportion test	0.007	0.04	4.1(2.1–6.1)
Types of NMD
Myasthenia gravis(n,%)	4(13.6%)	3(23.1%)	Z proportion test	0.46
Gullain bare syndrome(n,%)	11(37.4%)	2(15.4%)	Z proportion test	0.14
MND(n,%)	6(20.4%)	2(15.4)	Z proportion test	0.69
CIDP(n,%)	4(13.6%)	1(7.7%)	Z proportion test	0.57
Muscular Dystrophy(n,%)	3(10.2%)	3(23.1%)	Z proportion test	1
Sequelae of polio(n,%)	1(3.4%)	2(15.4%)	Z proportion test	0.1650
Comorbidties
Hypertension(n,%)	5(17%)	3(23.1%)	Z proportion test	0.66
Type II DM(n,%)	2(6.8%)	5(37.5%)	Z proportion test	0.01	0.08
Chronic Kidney disease(n,%)	1(3.4%)	3(23.1%)	Z proportion test	0.045	0.1
COPD(n,%)	0	1(7.7%)	Z proportion test	0.13
Bronchial asthma(n,%)	3(10.2%)	2(15.4%)	Z proportion test	0.6
Malignancy(n,%)	1(3.4%)	0	Z proportion test	0.5
Clinical parameters
Blood Pressure mmHg mean(sd)	103(21.7)	112(23.5)	t test for means	0.14
Heart rate beats/min mean(sd)	102(18.2)	98(15.7)	t test for means	0.49
Respiratory rate mean(sd)	28(6.9)	31(4.7)	t test for means	0.16
spO2% mean(sd)	89(8.4)	88(5.6)	t test for means	0.14
S.Albumin g/dl Mean(sd)	4.0(1.2)	3.1(1.1)	t test for means	0.03	0.02	6.9(5.2–8.6)
TLC(*1000/uL) Mean(sd)	13(5.1)	16(4.1)	t test for means	0.14
CRP (mg/dl) mean(sd)	53(11.2)	61(8.2)	t test for means	0.02	0.18
Procalcitonin (ng/ml) mean(sd)	0.9(0.2)	1.1(0.4)	t test for means	0.14
SAPS 3 mean(sd)	51(8.9)	58(10.2)	t test for means	0.03	0.04	5.1(3.2–7.2)
ABG parameters at admission
pH mean(sd)	7.41(0.8)	7.37(1.5)	t test for means	1
PaO2 mmHg mean(sd)	58(4.3)	57(6.8)	t test for means	0.56
PaCO2 mmHg mean(sd)	34(7.5)	37(6.1)	t test for means	0.21
PaO2:FiO2 mean(sd)	187(20.5)	181(16.8)	t test for means	0.36

In the study period, 42 patients were admitted in the RICU with a prior diagnosis of NMD and was put on HFNC for Type I respiratory failure. Out of this, 29 (69%) patients had successful weaning from HFNC to room air or low flow oxygen therapy while 13 (31%) patients had HFNC failure. Out of the 13 patients who had HFNC failure, 11(84.6%) required IMV while only 2 patients (15.4%) patients were successfully managed on NIV therapy. Six patients out of those 13 who failed HFNC expired during the course of hospital stay. Mean duration of HFNC therapy from initiation to be declared treatment failure was 12.6 ± 6.1 h. Out of the 13 patients who failed HFNC therapy, 9 (69.2%) failed within the first 12 h. ABG prior to be declared HFNC failure was done in 10/13 (76.9%) patients as in the other 3 patients, sudden desaturation lead to crash intubation before ABG analysis can be done. Mean (sd) mm of Hg of PaCO2 and PaO2 at the time of HFNC failure were 49.6(11.2) and 41(19.2) respectively. Change in PaCO2 and PaO2 of all patients at initiation and failure of HFNC were calculated and their mean (sd) change were 11.2(8.9) and 16(9.6) mm of Hg respectively. FOB guided Bronchoalveolar lavage (BAL) was done for microbiological sampling in 19/42 (45.2%) patients while 7 (16.6%) patients underwent bronchial toileting.

Flow of patients is depicted in Fig. 1. Functional or physical status of the patients prior to their admission in the RICU are given in Table 2. Thirteen (31%) of the patients had functional limitation prior to admission and were bed or wheelchair bound while 5(11.9%) were suffering from bulbar dysfunction and had undergone PEG tube insertion earlier to prevent aspiration. No statistically significant relationship was found between functional status of patients and HFNC success/failure.

Fig. 1.

Flowchart showing outcomes of patients

Table 2.

Relationship of Functional/Physical status of patients (n = 42)prior to admission in RICU with HFNC success/failure

Parameters	Total Number of patients	HFNC success (n,%)	HFNC failure (n,%)	p value	Statistical test used
Bed/Wheel chair bound	13	6(46.1%)	7(53.9%)	0.7	t test for proportions
Home NIV use	7	4(57.1%)	3(42.9%)	0.6	t test for proportions
Presence of bulbar dysfunction/dysphagia	5	3(60%)	2(40%)	0.53	t test for proportions
Presence of Kyphoscoliosis	4	3(75%)	1(25%)	0.16	t test for proportions
Presence of PEG tube	5	3(60%)	2(40%)	0.53	t test for proportions
H/O previous hospital admission with respiratory distress	9	6(66.7%)	3(33.3%)	0.16	t test for proportions

• Multiple patients had two or more of the parameters present

The commonest diagnosis of NMD in our study population were that of GBS (13, 30.9%), MND (8, 19.04%) and Myasthenia Gravis (7, 16.7%). Muscular weakness/diaphragmatic dysfunction was the commonest cause of Type I RF seen in 12 (28.6%) patients, followed by community acquired pneumonia seen in 10(23.8%) patients and aspiration pneumonia seen in 6(14.3%) patients. The values of the diaphragmatic parameters measured bed side in our patients are given in Table 3. No statistically significant difference were found between patients with HFNC success or failure based on ultrasound assessment of muscle weakness.

Table 3.

Comparison of ultrasound parameters of diaphragmatic movements between HFNC success and failure

Parameters	HFNC success (n = 29)	HFNC failure (n = 13)	p value	Statistical test used
Shortening fraction %, mean (sd)	18.9 (10.2)	16 (6.9)	0.36	t test for means
Maximal excursion mm, mean (sd)	2.9 (1.1)	2.6(2.3)	0.57	t test for means
Excursion quiet breathing mm, mean (sd)	2(1.4)	1.4(1.1)	0.18	t test for means
Thickness of diaphragm mm, mean (sd)	1.2(0.5)	0.9(0.4)	0.06	t test for means

There were 4 patients with kyphoscoliosis out of whom 3 were polio sequelae and one patient had muscular dystrophy. Out of these 4 patients with kyphoscoliosis, 3 were admitted with exacerbation of their underlying restrictive lung disease while the other had pulmonary embolism following prolonged bedridden state. Out of these 3 patients who had Type I RF due to exacerbation, all three failed HFNC trial and required NIV support (p = 0.007). On univariate analysis, it was found that SAPS score, radiological involvement, serum albumin levels, CRP and exacerbation of underlying restrictive lung disease at the time of admission were the factors that had statistically significant relationship with HFNC success/failure. Out of the 42 patients enrolled in our study, 25(59.5%) patients had Chest X ray changes. Chest x ray changes ranged from non homogenous opacities to homogenous lobar consolidation. Sixteen (94%) of the 17 patients with unilateral involvement in Chest X ray had HFNC success while 3(37.5%) of the 8 patients with bilateral involvement had HFNC failure. Bilateral involvement on Chest Xray had a statistically significant relationship with HFNC failure (p = 0.044, Fishers Exact test used). On multiple logistic regression, exacerbation of restrictive lung disease, serum albumin and SAPS 3 scores were found to be independent risk factors for HFNC failure.

ROX index which is defined as the ratio of pulse oximetry/fraction of inspired oxygen to respiratory rate was calculated in all patients at 2 h and again at 12 h and compared with the clinical outcome in the form of HFNC success or failure. ROC (receiver operating Curve) analysis was done with these two values of ROX index and the AUC (area under curve was calculated). AUC of ROX at 12 h was found to be higher than the AUC of ROX at 2 h post HFNC application(0.81 v/s 0.59). Taking ROX at 12 h > 4.88 as the threshold, sensitivity parameters of ROX as a predictor of HFNC failure was calculated. ROX > 4.88 at 12 h had a sensitivity of 80.3% and specificity of 69.23% and an accuracy of 76.2% of predicting HFNC success. ROC curve analysis of ROX is shown in Fig. 2 and the diagnostic contingency table of ROX and its sensitivity analysis shown in Table 4.

Fig. 2.

ROC curve analysis of ROX at 2 and 12 h with HFNC success

Table 4.

Diagnostic contingency table of ROX with HFNC success/failure and sensitivity analysis

	HFNC success	HFNC failure	Sensitivity analysis with 95% CI
ROX 12 h > 4.88	24	3	Sensitivity	80.31%	60.3% to 92%
			Specificity	69.23%	38.6% to 90.9%
ROX 12 h < 4.88 h	5	10	Positive Likelihood Ratio	2.58	1.1 to 5.95
			Negative Likelihood Ratio	0.3	0.13 to 0.67
			Positive Predictive Value	85.2%	71.4% to 93%
			Negative Predictive Value	60%	40.3% to 76.9%
			Accuracy	76.2%	60.6% to 87.9%

Discussion

In recent times, HFNC has played an important role in management of respiratory failure in critically ill neurological patients. Studies have shown that HFNC helps preventing re intubation in this subgroup of patients. In a recently published study, HFNC was given to patients in a neurointensive unit to reduce work of breathing following extubation [7]. HFNC was successful in preventing re intubation in 76.5% of such cases [7]. The authors also found that HFNC worked well in both acute hypoxaemic and hypercapnic respiratory failure [7]. Use of high flow rates from the outset was done in this study and is believed to have helped in managing respiratory distress and improved ventilation. In order to improve patient comfort, HFNC has several physiological effects, such as removing CO2 from anatomical dead space, overcoming expiratory resistance and delivering positive end-expiratory pressure (PEEP), opening atelectatic alveoli, delivering FiO2 consistently, and supplying warm, fully humidified air to maintain mucociliary function [7]. In another retrospective case series study, in patients with both traumatic and non traumatic brain injury, role of HFNC was compared with COT [8]. It was found that HFNC led to reduced need for intubation, shorter hospital and ICU stay [8].

However in the large RCTs which have looked in the role of HFNC in managing acute hypoxaemic failure, patients with neurological disorders are few in number. In the FLORALI trial, only 7% of the patients enrolled had neurological issues. None of them had neuromuscular disorders [9]. In the recently published RENOVATE trial also, no neurological patient were included [10].There has been no studies to compare head to head the efficacy of NIV and HFNC in managing patients of neuromuscular disorders. The only study we found in literature which looked at the efficacy of managing respiratory failure in NMD patients, is the study by Lionello et al. [4]. The authors in this retrospective case series applied HFNC to patients with NMD who could not tolerate NIV [4]. Three of the nine patients (33.33%) who were applied HFNC, failed HFNC treatment and were subsequently intubated or expired [4]. However the authors in this study did not look into the risk factors associated with HFNC failure.

The patients with neuromuscular disorders present their own unique challenges in form of diaphragmatic dysfunction, muscle and bulbar weakness, impaired cough and gag reflexes. In our study, 5/42(11.9%) patients had bulbar dysfunction and swallowing difficulties and required PEG tube for feeding. However no statistically significant association was found between them and HFNC success/failure (Table 2). The interface of oxygen delivery in this subset of patients should be such that the risk of hypercapnia is not aggravated while the risk of aspiration be low while maintaining ventilation.

NIV/BIPAP is commonly associated with intolerance and discomfort which leads to asynchrony. Inappropriate breathing efforts, raised respiratory rate occurs as a result of asynchrony. These excessive breathing efforts trigger inspiratory support [11]. Episodes of inspiratory support may disrupt normal respiratory-swallowing coordination, leading to swallowing followed by inhalation, which increases the likelihood of aspiration [11]. This discomfort and intolerance associated with NIV/BiPAP is ameliorated to a large extent by use of HFNC although some studies have shown that HFNC too can cause abnormalities in swallow reflex with persistently high flow rates of more than 40Lt/min [12, 13]. Oral suctioning is also easier in patients with HFNC and thus prevents pooling of oral secretions. A study by Wang et al., showed that the humidified air of HFNC resulted in decreased viscosity of sputum in neurological patients admitted in an ICU [8]. All these factors lead us to believe that HFNC might be useful in NMD patients with high risk of aspiration or with aspiration pneumonia. This is reflected in our study in which all the 6 patients with aspiration pneumonia could be successfully managed with HFNC and did not require any NIV or IMV support.

NIV/BiPAP masks with port for bronchoscope is not readily available and manoeuvring the scope through the NIV port requires expertise. On the other hand, bronchoscopy is much easier to perform through the oral route in patients on HFNC with atelectasis due to mucous impaction. Chest physiotherapy followed by coughing out of inspissated secretions can also be readily done in these patients. In the FLORALI study, we find that more than 60% of patients with Type I RF had a diagnosis of pneumonia and HFNC yielded a higher 90 day survival rate than NIV or COT [9]. In our study we found that 8 out 10(80%) patients with CAP had a favourable outcome on HFNC. However, radiological involvement of lung played an important role in these patients as we found that patients with bilateral lung involvement on Chest X ray had significantly higher chance of HFNC failure on univariate analysis.

Four patients in our study had kyphoscoliosis out of which 3 (75%) had HFNC failure and all of these three patients had exacerbation of their underlying restrictive lung disease. This fact segregates a specific subgroup of patients with NMD who are at a high risk of HFNC failure. Kyphoscoliosis patients have decreased chest wall compliance, increased work of breathing and hypoventilation. This coupled with the muscular weakness in patients with NMD make them prone to type II respiratory failure, a subgroup in which the hypercapnic tendency is exacerbated by HFNC administration. HFNC at high flow rates can produce a modest PEEP of 2–5 cm of H2O that too dependent on mouth opening and respiratory rates [14]. Patients with stiff chest wall and restrictive lung disease require pressure support (PS) to generate an adequate tidal volume along with a constant PEEP for alveolar recruitment, for which NIV/BiPAP support is required. Studies have shown that HFNC is also not efficient in increasing tidal volume or alveolar ventilation in patients with hypercapnic failure or restricted chest wall movements [14, 15]. Risk of HFNC failure hence is higher in this subgroup of NMD patients with thoracic skeletal changes.

In our study, serum albumin levels were found to be an independent risk factor of HFNC failure with Odds ratio of 6.9 Low seum albumin levels are associated with muscle weakness, including the weakness of respiratory muscles necessary for ventilation in NMD patients [16]. Furthermore, albumin is crucial in maintaining plasma oncotic pressure. Low albumin levels may lead to pulmonary interstitial oedema which would impair gas exchange and decrease the effectiveness of HFNC [17].Hypoalbuminemia is associated with malnutrition and is an acute phase reactant which mirrors the inflammatory milieu of the body. It is frequently a sign of chronic inflammation which may lead to decreased HFNC tolerance and increased disease burden [18]. Since patients with NMD already have a significant risk of ventilatory failure, the presence of low albumin may indicate increased vulnerability and further supports its use as a prognostic factor in this patient group.

SAPS 3 score was found in our study to be a risk factor of HFNC failure with Odds ratio of 5.1.These scores reflect a greater burden of systemic problems, such as haemodynamic instabilities and metabolic imbalances [19]. People with NMD have weak respiratory muscles, poor cough reflex and limited ventilatory capacities. Even small amounts of physiological stress can lead to serious clinical decline. SAPS 3 scores were calculated in the RENOVATE trial, however, inference was not drawn regarding relationship between HFNC failure and high SAPS 3 scores [9]. In multiple studies consisting of heterogenous patient population with Type I respiratory failure, SAPS 2 scores were found to be risk factors for HFNC failure [20, 21]. More studies are required in the future to further evaluate the predictive power of SAPS 3 in patients with Type I respiratory failure on HFNC.

In our study we also looked at ROX scores and its ability to predict HFNC success or failure. ROX > 4.88 at 12 h emerged as a better predictor of HFNC success than that at 2 h (AUC 0.81 v/s 0.59). ROX > 4.88 at 12 h had a sensitivity of 80.3% and specificity of 69.23% and an accuracy of 76.2% of predicting HFNC success. In the seminal study by Roca et al., in the validation cohort, ROX > 4.88 at 12 h predicted HFNC success with a sensitivity of 86.8% and a specificity of 52.2% [22]. However, no patients with neuromuscular disorder as a comorbidity was included in the sample of patients suffering from acute hypoxaemic failure. There is a a lack of studies in which a ROX cut off value has been looked at to predict HFNC failure specifically in patients with neuromuscular disorders. In a recent meta analysis done in 2023, the authors found that ROX had an overall AUC of 0.771 (95% CI: 0.67–0.84) and the overall optimal cut-off value of ROX to predict HFNC success was 5.23. At this cut-off, sensitivity was 0.732 (95% CI: 0.58–0.85) and specificity was 0.69 (95% CI: 0.54–0.81) [23].

This study had several limitations. It was a single-centre, retrospective analysis which limited generalisability and introduced selection bias. The absence of a control group receiving standard oxygen therapy or exclusive NIV restricted direct comparisons of treatment efficacy. Detailed physiological measurements, including serial assessments of diaphragmatic function and work of breathing, were not uniformly captured. The heterogeneous neuromuscular disorder cohort may have introduced variability in clinical presentation and outcomes. Lastly, ABG in patients were done 12 h after application of the HFNO. ABG done at 6 h could have allowed us to recognise HFNO responders or failures earlier.

Conclusion

In our knowledge, this is the largest study which has looked into role of HFNC in managing Type I respiratory failure patients with NMD. Certain subgroups of NMD like those with skeletal changes and kyphoscoliosis, hypoalbuminaemia and high SAPS 3 scores were associated with HFNC failure. Larger RCTs are required to better evaluate the role of HFNC in NMD patients.

Authors’ contributions

AC, SD, MD, MS were involved in patient enrolment, work up of patients, data entry and manuscript preparation. AC, SD, MS, RW and SK were involved in study formulation, manuscript preparation and statistical analysis. AC, SD and MS involved in data analysis, statistical evaluation and manuscript preparation. All the authors have read and approved the final version of the manuscript and agreed to be accountable for all aspects of the work.

Funding

No funding was received for this study.

Data availability

Data can be shared on reasonable request. Person to be contacted to request for raw data - Dr.Amartya Chakraborti, Mobile number +91-9599700325, email -amartya86@gmail.com176, Rajdanga School Road, House -Ambar, Kolkata -700107, India.

Declarations

Ethical approval

The Institute Ethics Committee of Bhagwan Mahavir hospital waived the requirement for ethical approval because the study was retrospective and not conducted as a formal research project and only de-identified data were used in the analysis(IRB min no:325 dated 11.12.2019). Informed consent to participate was obtained from all of the participants in the study.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable.