Detection of pediatric breathing by CPAP/NIV devices: Clinical experience

Home continuous positive airway pressure (CPAP) and noninvasive ventilation (NIV) are increasingly used in children worldwide.1, 2 The manufacturer's recommend using their device(s) based on a minimal patient's weight. Theoretically, CPAP/NIV devices should thus be able to adequately detect the breathing of patients whose weight is above the minimal weight, and conversely, not in those with a weight below this limit. Consequently, in children, the appropriate choice of a CPAP/NIV device according to the manufacturer's recommendations, should guarantee a correct functioning of the device together with an accurate automatic recording of built-in software data. In a recent bench study, we hypothesized that the detection of the patient's breathing may be expressed as a minimal tidal volume (Vtmin) rather than a minimal weight.3 We observed various Vtmin, independently of the device level category (level I CPAP/NIV devices, with recommended weight >30 kg for CPAP and >13−18 kg for NIV, with no internal battery; level II devices, with recommended weight >10−13 kg, with internal battery and alarms; level III devices, with recommended weight >2.5−5 kg, which are life support ventilators). Importantly, we observed that some home CPAP/NIV devices may be used for CPAP therapy in children with a weight below the manufacturer's recommended threshold, regardless of the device level category. However, these findings need to be confirmed in the clinical setting, with a measurement of Vt. Based on our clinical experience, we have previously observed that the breathing of some patients with a weight below the recommended weight may be detected (Figure 1A,B and Supporting Information S1: Online Figure 1). Surprisingly, on the contrary, we have also observed that some patients with a weight above the recommended minimal weight for CPAP/NIV devices are not correctly detected by the device (Figures 1C,D and 2, Supporting Information S1: Online Figures 2 and 3). This highlights the fact that the detection of the patient's breathing is probably not only related to the patient's weight but also to other parameters, such as Vtmin. Figure 1A,B show an example of the built-in software data of a 5-month-old infant with Pierre Robin sequence, weighing 5.2 kg, who required CPAP because of severe obstructive sleep apneas (OSA). CPAP was started with a device recommended for patients above 13 kg (Stellar 150; ResMed), for practical reasons. The different tracings were present during almost the entire nights, with the exceptions of a few periods where the airflow was reduced with the concomitant disappearance of the respiratory parameters (tidal volume, respiratory rate…) values, indicating that the patient's breathing was probably not detected despite the display of an airflow tracing. Supporting Information S1: Online Figure 1 illustrates the built-in software data of two other infants with Pierre Robin sequence, weighing 3.4 and 5.1 kg, respectively at the time of CPAP initiation. The breathing of the two infants were perfectly detected with a correct display of tracings and persistent values of tidal volume and respiratory rate. However, the accuracy of some built-in software data remains uncertain, as the respiratory rate and/or tidal volume values appeared to be erroneous in some patients (Figure 1A,B and Supporting Information S1: Online Figure 1B). Figure 1C,D show an example of the built-in software data of a 7-year-old girl with facio-craniosynostosis, weighing 14 kg, who required CPAP because of severe OSA. CPAP was first started with a device recommended for patients weighing above 13 kg (Aircurve 10 Vauto; ResMed), and was then switched for a device recommended for >10 kg (BiPAP A40; Philips Respironics). Indeed, Figure 1C and Supporting Information S1: Online Figure 2A and B display the discontinuous tracings of airflow, pressure, and other respiratory parameters due to the intermittent detection of the patient's breathing by the first device. The airflow tracing was present even in absence of tidal volume and respiratory rate values, probably indicating that the patient's breathing was not correctly detected by the device (Figure 1C). As a consequence, data such as the CPAP usage may be inaccurate, as illustrated by the built-in software reporting a ∼4−5 h/night usage while the patient used the device during about 8−10 h/night (Supporting Information S1: Online Figure 2A and B). The second device did not perform better, even though the recommended weight was lower (Figure 1D and Supporting Information S1: Online Figure 2C). The repeated 2 cmH2O-pressure impulses are generated by the device algorithm for the detection and characterization of residual apneas, highlighting the fact that the device falsely detected apneas in the presence of an airflow tracing (Figure 1D). Supporting Information S1: Online Figure 3 shows the built-in software tracings of a 4-year-old child with CHARGE syndrome, weighing 16 kg, who required CPAP for severe OSA. The presence of intermittent tracings also indicates an inappropriate detection of the patient's breathing, despite a weight above the manufacturer's recommendations. In practice, in children, any CPAP/NIV device may be used to deliver constant CPAP. However, the reliability of alarms and built-in software data is not guaranteed if the patient's breathing is not correctly detected. Moreover, it has to be noted that despite the display of airflow tracing with a sensitivity that allow to show residual respiratory events (Supporting Information S1: Online Figure 4), the patient's breathing may not be detected by the device impeding the display of values of some respiratory parameters, and therefore leading to inaccurate built-in software data. The correct detection of patient's breathing is crucial during NIV (Figure 2), to prevent patient-ventilator asynchronies, such as trigger and cycling asynchronies.4 Indeed, delayed triggering or ineffective efforts may be observed in some patients despite a weight above the manufacturer's recommendation (Figure 2A). And even if the clinical consequences of patient-ventilator asynchronies are not clearly determined,5 their occurrence needs to be addressed with an optimization of the device settings or by switching for a more sensitive device whenever it is possible (Figure 2B). An adaptation of the settings (increase of back-up rate, pressure support…) may be attempted if the device is still not sensitive enough to the patient's airflow. The same concern applies for patients treated with autoset CPAP, for which the CPAP variation is based on the detection of residual respiratory effort on the airflow tracing. In the future, it seems important to confirm the results of our bench study on the Vtmin detection by the CPAP devices in the clinical setting. Moreover, the functioning of NIV should also be investigated in children to determine the mechanisms of the patient's breathing detection (breathing effort, inspiratory flow, Vtmin?), and the effect on NIV devices response and accuracy of built-in software data. A guidance for the choice of CPAP/NIV device, in addition to the manufacturer's recommended minimal weight would be helpful in clinical practice. In conclusion, we recommend that a careful testing of the device should be done at CPAP/NIV initiation, with a checking of the built-in software data to ascertain the accurate detection of patient's breathing. Concerning CPAP, independently of the patient's breathing detection, the prescriber should be aware of the possibility of an incorrect patient's breathing detection, false alarms, the inaccuracy of device usage, and the potential lack of pressure, airflow and/or other respiratory parameters tracings display. However, the correct detection of patient's breathing is mandatory when home devices are used to deliver autoset CPAP or NIV. Sonia Khirani: Writing—original draft; conceptualization; methodology; validation; data curation; formal analysis; writing—review and editing; supervision; investigation. Marine Dosso: Data curation; validation; formal analysis; writing—review and editing. Emeline Fresnel: Data curation; writing—review and editing; validation. Charlotte Collignon: Data curation; validation; writing—review and editing. Meryl Vedrenne-Cloquet: Data curation; validation; writing—review and editing. Lucie Griffon: Data curation; validation; writing—review and editing. Brigitte Fauroux: Data curation; validation; writing—review and editing. The authors have no funding to report. The authors declare no conflict of interest. Ethical approval obtained (CPP Sud Ouest et Outre-mer IV; CPP2021-01-013a/2020-A003083-36) and written consent obtained from parents. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.