AUTHOR=Mentens Xante , Vanhees Janne , Paulussen Jolien , Installé Sophie , Van Ostaeyen Anse , Ides Kris , Jouret Nathalie , Van Hoorenbeeck Kim , Verhulst Stijn TITLE=Predicting hypercapnia and hypoxia by the ventilator's built-in software in children on long-term non-invasive ventilation: A pilot study JOURNAL=Frontiers in Pediatrics VOLUME=11 YEAR=2023 URL=https://www.frontiersin.org/journals/pediatrics/articles/10.3389/fped.2023.1158396 DOI=10.3389/fped.2023.1158396 ISSN=2296-2360 ABSTRACT=Introduction

Follow-up of children on long-term non-invasive ventilation (NIV) could be improved by telemonitoring, using the ventilator's built-in software (BIS) parameters as alternative for in-hospital sleep studies to reduce costs, enhance patient independence and contribute to early detection of infections. This pilot study investigated whether analysis of BIS parameters can predict abnormal nocturnal transcutaneous CO2 (TcCO2) and saturation (SpO2) measurements in children on long-term NIV.

Methods

Children on long-term NIV in follow-up at the Antwerp University Hospital were retrospectively included. Nocturnal TcCO2 and SpO2 measurements were collected together with BIS parameters at three different time points: the night of the sleep study (BIS1), mean values from 48 h (BIS2) and 72 h (BIS3) before the sleep study. Predictions were calculated for following outcome measures: % recording time TcCO2 > 46.9 mmHg (%RT TcCO2; abnormal if ≥2%), recording time SpO2 < 93% (RT SpO2; abnormal if >1 h), abnormal TcCO2 or SpO2, mean TcCO2, mean SpO2.

Results

69 patients were included. %RT TcCO2 was separately predicted by reached tidal volume2 [OR 0.97 (0.93; 1.00); p = 0.051; AUC = 30%] and reached IPAP1 [OR 1.05 (1.00; 1.10); p = 0.050; AUC = 66%]. Leak1 predicted RT SpO2 [OR 1.21 (1.02; 1.43); p = 0.025; AUC = 84%]. Mean TcCO2 correlated with reached tidal volume2 (R2 0.10, p = 0.033).

Discussion

Certain BIS parameters can predict nocturnal hypercapnia and desaturation in children on long-term NIV. Future studies with larger sample sizes are warranted to further investigate the predictive value of the identified BIS parameters.