Feasibility of a Hypoxic Challenge Test Under Noninvasive Ventilation Versus Oxygen in Neuromuscular Patients with Chronic Respiratory Insufficiency

Introduction

In 2018, 4 billion passengers were carried by air transport according to the International Civil Aviation Organization (ICAO's Annual Report of the Council, 2018). Among passengers planning air travel, patients with respiratory diseases are at high risk of in-flight complications (Noble and Davidson, 1999; Ergan et al., 2018). Indeed, a commercial aircraft cabin is pressurized to a maximum altitude of 2,400 m and the oxygen pressure is equivalent to a fraction of inspired oxygen (FiO₂) of 15.1%. In this condition, the partial pressure of arterial oxygen (PaO₂) falls to 60–75 mmHg and the oxygen saturation (SpO₂) measured by pulse oximetry is between 89% and 94% in healthy subjects. This results in moderate hyperventilation and moderate tachycardia.

In 2011, the British Thoracic Society (BTS) recommended performing a hypoxic challenge test (HCT) for all patients planning air travel with severe restrictive lung disease (vital capacity <1 l), especially for those with hypoxemia and/or hypercapnia, as well as for those requiring ventilatory support (Ahmedzai et al., 2011). If the SpO₂ at the end of the test remains >85%, oxygen supplementation is not required. Below this value, it is recommended to add 2–4 l/min of oxygen through a nasal cannula. However, the correction of hypoxemia in patients with chronic respiratory failure due to advanced neuromuscular disease can be achieved by ventilatory support, which allows the patient to reach a level of ventilation equivalent to that of a normal subject (Mestry et al., 2009). Few studies have investigated the use of a ventilator instead of oxygen to correct hypoxia during air travel in these particular patients.

This study was designed to evaluate the feasibility of an HCT under noninvasive ventilation (NIV) and to compare the correction of hypoxemia by increasing ventilation with NIV to that by oxygen therapy during an HCT in patients with chronic respiratory failure due to neuromuscular diseases already treated with home ventilatory support.

Methods

We conducted a comparative, observational study in the respiratory medicine department of the Pitié Salpêtrière Hospital (Paris) over a 6-month period. We extracted retrospectively the data from the medical files from patients tested before air travelling. The study was approved by the ethics committee from the Institutional Review Board of the French Learned Society for Respiratory Medicine (CEPRO 2012-039). Written informed consent was obtained from all subjects.

Patients were included if they suffered from chronic respiratory failure due to neuromuscular disorders and required nocturnal mechanical ventilation, and if they were planning a flight. Patients with nontreated cardiovascular diseases, recent myocardial infarction, cardiac arrhythmia, pregnancy, Forced Expiratory Volume in the first second < 30% due to obstructive disease, pulmonary arterial hypertension, infectious tuberculosis, or ongoing pneumothorax with persistent air leak and major hemoptysis were excluded (Ahmedzai et al., 2011). Data collected were age, diagnosis, sitting and supine force vital capacity, SpO₂, expired CO₂ (pressure, end tidal, carbon dioxide [PetCO₂]), room air blood gases, and daily adherence to home mechanical ventilation.

The HCT was performed with a ventilated canopy placed over the patient's head or the patient's home ventilator (Fig. 1A) and connected to a hypoxia generator (HYPOXICO, Inc., Jalhay, Belgium). The extraction of O₂ was adapted to obtain a stable FiO₂ of 15% ± 0.2%. FiO₂ was measured using an FiO₂ sensor (MAXTEC, Inc., UT) placed inside the canopy. The measurement of SpO₂, PetCO₂, heart rate, and respiratory rate was performed using an integrated monitor (Capnocheck^® Plus; Smiths Medical PM, Inc., WI), including pulse oximetry and nasal cannula for measuring CO₂. Oxygen was administered using another nasal cannula previously placed in superposition of the Capnocheck cannula. The HCT was performed with the patients seated at rest in four successive measurements (15 min each): (i) spontaneous ventilation, room air (baseline), (ii) under the canopy, spontaneous ventilation, FiO₂ of 15%, if the test was positive (SpO₂ <90%), (iii) under the canopy, spontaneous ventilation, FiO₂ of 15%, with an oxygen flow of 2 l/min, and finally (iv) under patient's home ventilator, FiO₂ of 15% (Fig. 1B). Dyspnea and headaches were measured at the end of each period by visual analog scales (VAS) rated from 0 to 10.

OPEN IN VIEWER

FIG. 1.

(A) HCT with NIV. The NIV device is placed under a canopy connected to a hypoxic generator. The gas used by the NIV to ventilate the patient is at an FiO₂ of 15% during the test. A PetCO₂ sensor is connected to the system. (B) Diagram of the four stages of the test: at baseline, during the HCT, the HCT with 2 l/min oxygen flow (HCT + O₂), and the HCT with NIV (HCT + NIV). FiO₂, fraction of inspired oxygen; HCT, hypoxic challenge test; NIV, noninvasive ventilation; PetCO₂, pressure, end tidal, carbon dioxide.

We chose a positive value for the HCT of <90% SpO₂, which is different from that of the BTS recommendation (positive <85% of SaO₂). Indeed, patients with neuromuscular disease or extrapulmonary restrictive chest wall deformity are more at risk of hypoxemia during an HCT, even with a resting saturation of >95% (Masa et al., 1997). Our choice of 90% was selected for two reasons. First, the related effect of hypoxemia, such as an increase in respiratory rate, was observed at <90% of the SpO₂, suggesting a lower respiratory tolerance. Second, the BTS threshold is common to all respiratory disease, broadly based on studies that include obstructive lung disease. Our patients had normal lung parenchyma associated with reduced ventilatory adaptation in a hypoxemia condition. Thus, a threshold of 90% appeared to be more clinically relevant for this neuromuscular population. It was a choice of caution; it is difficult to formally prove that a patient with severe diaphragmatic dysfunction can remain with an SpO₂ between 85% and 90% for a long time.

A p-value of <0.05 was considered statistically significant. Quantitative variables were compared using a Friedman's test and a post hoc Dunn test and the results expressed as their standard value (median and interquartile range). Qualitative variables were compared using Fisher's exact test and are expressed in terms of numbers and percentage. All statistical analyses were performed using GraphPad Prism software.

Results

Thirteen patients were included, of whom seven had amyotrophic lateral sclerosis. Their characteristics are presented in Table 1.

At baseline, median blood gas measures were a PaO₂ of 82 mmHg and a partial pressure of carbon dioxide, arterial (PaCO₂) of 41 mmHg. All patients performed the HCT. The median SpO₂ was 95% at baseline. By the end of the test, the SpO₂ of 11 patients had fallen to <90% (Fig. 2). Among these patients, the SpO₂ of two had fallen to <85%. The median fall of SpO₂ was 7 [5.5–8.5]% relative to the baseline SpO₂ (p < 0.001). There was no significant change in the PetCO₂ during spontaneous ventilation at the end of the HCT.

OPEN IN VIEWER

FIG. 2.

SpO₂ at baseline and at the end of the HCT. Each line represents one patient (n = 13). Two patients shared lines from 96% to 89% and 93% to 88%. The dashed line indicates the positivity threshold of the test. SpO₂, oxygen saturation.

An oxygen flow of 2 l/min allowed a correction of SpO₂ to >90% for all positive HCT patients (+9 [6–12]%, p = 0.0029). In comparison, patient's home ventilator significantly increased the SpO₂ by 8 [7–12]% (p = 0.016). Oxygen was not more effective than NIV for SpO₂ correction during the HCT (Fig. 3). There was a significant decrease in PetCO₂ only with NIV (−10 [−16 to −7.5] mmHg, p = 0.04). The HCT was not associated with a significant increase in the respiratory rate and the VAS score for dyspnea. Three tests were prematurely stopped: one because of dyspnea (at the end of HCT), one because of dyspnea and headache (at the end of HCT + O₂ period), and one on request by the patient (at the end of HCT). No other patients reported a change in the VAS for dyspnea or headache.

OPEN IN VIEWER

FIG. 3.

Comparison of SpO₂ (A), PetCO₂ (B), and respiratory rate (C) of patients with a positive HCT who have completed all periods of the protocol: at baseline, during the HCT, the HCT with 2 l/min oxygen flow (HCT + O₂), and the HCT with NIV (HCT + NIV). Median and interquartile range, *p < 0.05, **p < 0.01.

Discussion

Our study shows that a preflight HCT is feasible under NIV in neuromuscular patients. The study also shows that SpO₂ drops below 90% under a 15% FiO₂ in majority of such patients (11 out of 13 in this study), and that NIV corrects hypoxemia in all cases.

HCT is achievable for patients using NIV under a canopy coupled to a hypoxia generator. In contrast to other already available methods using a Douglas bag or a Venturi mask, it is easily achievable directly at the patient's bedside, making it accessible to patients with motor disabilities or respiratory failure. Its speed of acquisition (45 min for a complete test) allows its routine use. Moreover, unlike conventional methods, it allows the evaluation of alternative therapies, such as mechanical ventilation. The order of different periods of the test was planned to avoid the necessity of re-establishing a stable hypoxia multiple times, which would have lengthened the experiment. However, a potential effect of the sequence cannot be excluded and the results could be different with another sequence.

NIV was always as effective as oxygen therapy in our group of patients. NIV appeared to be an acceptable alternative therapy, with physiological advantages like decreasing CO₂ level. Furthermore, Winck et al. (2010) showed that a patient with a neuromuscular disease, who is correctly ventilated, without supplemental oxygen, can maintain SpO₂ throughout a flight. In our study, the PetCO₂ did not significantly decrease during the HCT because of a possible lack of ventilatory adaptation in this particular population. Indeed, hypoxemia normally induces a stimulation of the central respiratory drive leading to an augmentation of ventilation to maintain the oxygen level. In the case of normal respiratory mechanics, it should have led to a significant decrease in CO₂ level. The interpretation of the PetCO₂ is limited by the measurement method and should be considered with caution, given the potential for washout of this signal by NIV flows. However, the evolution of this parameter is in favor of the absence of hypercapnia during an HCT with oxygen or NIV.

We can assume that this favorable effect of NIV would be even more obvious for patients with more advanced respiratory failure. Even if not significant, probably because of the small number of subjects, the increase in respiratory rate at the initiation of hypoxia was well corrected by NIV. Importantly, this respiratory frequency was always higher than the minimum frequency set on the ventilator and therefore corresponded to the patients' “voluntary” frequency. These results suggest that the respiratory effort under NIV is reduced, and therefore the risk of respiratory exhaustion during longer exposure to hypoxemia, such as during air travel prolonged several hours, may decrease with NIV.

Conclusion

We conclude that performing an adapted HCT for home-ventilated patients with a neuromuscular pathology is useful to choose a personalized treatment for air travel. NIV allows sufficient hypoxemia correction for certain patients and can be used instead of oxygen therapy during air travel for this population. NIV can therefore be a new alternative for patients planning to take a flight.