Estimation of inspiratory muscle effort using three common indices in various respiratory models, a bench study

Abstract

Background Liberation from mechanical ventilation is a complex therapeutic challenge in the intensive care unit. Estimating inspiratory effort during mechanical ventilation can mitigate lung and diaphragmatic injury, along with weakness and atrophy. During a spontaneous breathing trial, it can be critical to predict over or under assistance to guide safe liberation. While estimation of the inspiratory effort requires special equipment, many other indices have been developed to estimate patient effort, work, and actual muscle pressure. In this bench study, we compare three commonly used maneuvers: airway occlusion at 100 msec (P0.1), airway pressure drop during full occlusion (Pocc), and pressure muscle index (PMI) for their accuracy in predicting the actual muscle effort. Methods A single active lung compartment using ASL5000 was modeled to simulate three common patient care scenarios, including “normal” (airway resistance 5 cm/l/s; compliance 60 ml/cm/H2O), “restrictive” (airway resistance 10 cm/l/s; compliance 30 ml/cm/H2O); and “obstructive” (airway resistance of 20 cm/l/s; compliance of 80 ml/cm/H2O) with respiratory rate of 15/minute, inspiratory time of 1 second (10 % rise, 0% hold, and 10% release while exhalation is passive). A Bellavista 1000e ventilator was used for pressure support of 5 cmH2O and positive end-expiratory pressure (PEEP) of 5 cmH2O. Each index was measured to the inputted Pmus, which ranged from 1 to 30 cmH2O and increased by increments of 1. Results were analyzed using Pearson correlation and regression analysis to predict an associated formula. These were compared to the inputted Pmus using single factor ANOVA followed by post Hoc Tukey test. Formulas from the P0.1 and the Pocc were then compared against previously published equations using single factor ANOVA. Statistics were performed using SPSS 20. P < 0.05 was considered statistically significant. Results All three indices had strong correlations to Pmus, P0.1 [R 0.978, 95% CI 0.97, 0.99, P < 0.001], Pocc [R 0.999, 95% CI 1.1, 1.12, P < 0.001], and PMI [R 0.722, 95% CI 0.61, 0.81, P < 0.001]. The equations to estimate Pmus were: P0.1: 3.95 (P0.1) – 2.05; Pocc: 1.11 (Pocc) + 0.82; and PMI: 1.03 (PMI) + 8.26. A significant difference (P < 0.001) was observed when comparing the inputted Pmus with Pmus estimated from P0.1, Pocc, or PMI. Post hoc analysis showed no difference between Pmus to Pmus estimated from P0.1, Pmus to Pmus estimated from Pocc, and Pmus estimated from P0.1 and Pocc; while comparisons of Pmus estimated from PMI to those from the P0.1 and Pocc revealed significant differences (P < 0.001 and P < 0.001, respectively). When comparing our formula for P0.1 to the previously published formula and the actual Pmus, no significant difference was observed (P 0.261), with post hoc tests revealing no significant differences between any pair. In contrast, a significant difference was found when comparing the formula for Pocc to the previously published formula and the actual Pmus (P < 0.001). Post hoc tests showed no difference between the new formula and Pmus (P 0.99), but a significant difference between Pmus and previous formula (P < 0.001). Conclusions While overall all three methods tested showed good correlation with the actual set Pmus, only P0.1 and the Pocc had strong correlation with the set Pmus in all three settings, suggesting that derived formulas can be useful to estimate muscle effort. PMI did not prove accurate, especially in obstructive scenarios, and may not be relied upon in practice.

Keywords

• pmus
• p0.1
• p occlusion
• pmi