Breath detection by transcutaneous electromyography of the diaphragm and the Graseby capsule in preterm infants

Objective: To compare triggering, breath detection and delay time of the Graseby capsule (GC) and transcutaneous electromyography of the diaphragm (dEMG) in spontaneous breathing preterm infants. Methods: In this observational study, a 30 minutes respiration measurement was conducted by respiratory inductance plethysmography (RIP), the GC, and dEMG in stable preterm infants. Triggering was investigated with an in vitro set-up using the Infant Flow^® SiPAP^TM system. The possibility to optimize breath detection was tested by developing new algorithms with the abdominal RIP band (RIP_AB) as gold standard. In a subset of breaths, the delay time was calculated between the inspiratory onset in the RIP_AB signal and in the GC and dEMG signal. Results: Fifteen preterm infants with a mean gestational age of 28 ± 2 weeks and a mean birth weight of 1086 ± 317 g were included. In total, 14 773 breaths were analyzed. Based on the GC and dEMG signal, the Infant Flow^® SiPAP™ system, respectively, triggered 67.8% and 62.6% of the breaths. Breath detection was improved to 99.9% for the GC and 113.4% for dEMG in new algorithms. In 1492 stable breaths, the median delay time of inspiratory onset detection was +154 ms (IQR +118 to +164) in the GC and −50 ms (IQR −90 to −22) in the dEMG signal. Conclusion: Breath detection using the GC can be improved by optimizing the algorithm. Transcutaneous dEMG provides similar breath detection but with the advantage of detecting the onset of inspiration earlier than the GC.