Monitoring patient-ventilator breath contribution in the critically ill during neurally adjusted ventilatory assist: reliability and improved algorithms for bedside use.

The patient-ventilator breath contribution (PVBC) index estimates the relative contribution of the patient to total tidal volume (Vtinsp) during mechanical ventilation in neurally adjusted ventilator assist mode and has been used to titrate ventilator support. The reliability of this index in ventilated patients is unknown and was investigated in this study. PVBC was calculated by comparing tidal volume (Vtinsp) and diaphragm electrical activity (EAdi) during assisted breaths (Vtinsp/EAdi)assist and during unassisted breaths (Vtinsp/EAdi)no-assist. Vtinsp was normalized to peak EAdi (EAdipeak) using 1) one assisted breath, 2) five consecutive assisted breaths, or 3) five assisted breaths with matching EAdi preceding the unassisted breath (N1PVBC2, X5PVBC2, and PX5VBCEAdi-matching2 , respectively). In addition, PVBC was calculated by comparing only Vtinsp for breaths with matching EAdi (PVBCß2). Test-retest reliability of the different PVBC calculation methods was evaluated with the intraclass correlation coefficient (ICC) using five repeated PVBC maneuvers performed with a 1-min interval. In total, 125 PVBC maneuvers were analyzed in 25 patients. ICC [95% confidence interval] values were 0.46 [0.23-0.66], 0.51 [0.33-0.70], and 0.42 [0.14-0.69] for N1PVBC2, X5PVBC2, PX5VBCEAdi-matching2 , respectively. Complex waveform analyses showed that insufficient EAdi filtering by the ventilator software affects reliability of PVBC calculation. With our new EAdi-matching techniques reliability improved (PVBCß2 ICC: 0.78 [0.60-0.90]). We conclude that current techniques to calculate PVBC exhibit low reliability and that our newly developed criteria and estimation of PVBC-using Vtinsp of assisted breaths and unassisted breaths with matching EAdi-improves reliability. This may help implementation of PVBC in clinical practice. NEW & NOTEWORTHY The patient-ventilator breath contribution (PVBC) index estimates the relative contribution of the patient to tidal volume generated by the patient and the mechanical ventilator during mechanical ventilation in neurally adjusted ventilator assist mode. It could be used to titrate ventilator support and thus to limit development of diaphragm dysfunction in intensive care unit patients. Currently available methods for bedside assessment of PVBC are unreliable. Our newly developed criteria and estimation of PVBC largely improve reliability and help to quantify patient contribution to total inspiratory effort.

Estimation of the diaphragm neuromuscular efficiency index in mechanically ventilated critically ill patients.

BACKGROUND: Diaphragm dysfunction develops frequently in ventilated intensive care unit (ICU) patients. Both disuse atrophy (ventilator over-assist) and high respiratory muscle effort (ventilator under-assist) seem to be involved. A strong rationale exists to monitor diaphragm effort and titrate support to maintain respiratory muscle activity within physiological limits. Diaphragm electromyography is used to quantify breathing effort and has been correlated with transdiaphragmatic pressure and esophageal pressure. The neuromuscular efficiency index (NME) can be used to estimate inspiratory effort, however its repeatability has not been investigated yet. Our goal is to evaluate NME repeatability during an end-expiratory occlusion (NMEoccl) and its use to estimate the pressure generated by the inspiratory muscles (Pmus). METHODS: This is a prospective cohort study, performed in a medical-surgical ICU. A total of 31 adult patients were included, all ventilated in neurally adjusted ventilator assist (NAVA) mode with an electrical activity of the diaphragm (EAdi) catheter in situ. At four time points within 72 h five repeated end-expiratory occlusion maneuvers were performed. NMEoccl was calculated by delta airway pressure (ΔPaw)/ΔEAdi and was used to estimate Pmus. The repeatability coefficient (RC) was calculated to investigate the NMEoccl variability. RESULTS: A total number of 459 maneuvers were obtained. At time T = 0 mean NMEoccl was 1.22 ± 0.86 cmH2O/µV with a RC of 82.6%. This implies that when NMEoccl is 1.22 cmH2O/µV, it is expected with a probability of 95% that the subsequent measured NMEoccl will be between 2.22 and 0.22 cmH2O/µV. Additional EAdi waveform analysis to correct for non-physiological appearing waveforms, did not improve NMEoccl variability. Selecting three out of five occlusions with the lowest variability reduced the RC to 29.8%. CONCLUSIONS: Repeated measurements of NMEoccl exhibit high variability, limiting the ability of a single NMEoccl maneuver to estimate neuromuscular efficiency and therefore the pressure generated by the inspiratory muscles based on EAdi.