Lower non-invasive estimated P_mus predicts extubation failure in mechanically ventilated ICU patients

Abstract

We evaluated whether non-invasive estimated inspiratory muscle pressure (P_mus) predicts extubation outcomes in ICU patients. Estimated P_mus, reflecting the pressure generated by respiratory muscles, was measured before and after the spontaneous breathing trial (SBT). Lower pre-SBT estimated P_mus (<4.1 cmH₂O) and post-SBT (<4.4 cmH₂O) were associated with extubation failure (AUC ≈ 0.73). P0.1 and dynamic transpulmonary pressure (P_L,dyn) showed no significant association. Estimated P_mus offers a simple bedside method to assess inspiratory muscle strength and may help identify patients at risk of extubation failure. Further multicenter studies are needed to validate these findings.

Introduction

Extubation timing directly affects critically ill patients: premature extubation increases the risk of reintubation and related complications, whereas delayed extubation prolongs mechanical ventilation, increasing morbidity and ICU stay. ¹ Respiratory muscle strength, reflected by the pressure generated during inspiration (P_mus), is a key determinant for weaning success. ² Historically, measuring P_mus required invasive techniques, such as transdiaphragmatic pressure, which are complex and impractical for routine bedside use. ³ Bertoni et al. ⁴ proposed a non-invasive estimation method using airway occlusion maneuvers, providing a feasible tool to evaluate inspiratory muscle effort during ventilation.

We hypothesized that lower estimated P_mus values measured before and after the spontaneous breathing trial (SBT) could predict extubation failure, even in patients who initially tolerate the SBT.

Method

This prospective observational study was conducted from July 2022 to October 2023 in two mixed medical-surgical ICUs in Brazil. We included adult patients (>18 years) ventilated >24 h who passed a 30-min T-piece SBT and met clinical criteria for extubation. ⁵ Exclusion criteria included tracheostomy, accidental/self-extubation, and upper airway failure, as these do not reflect inspiratory muscle capacity. ⁶ Ventilatory data were collected immediately before and after SBT, including P0.1, estimated P_mus, and dynamic transpulmonary driving pressure (ΔP_L,dyn). Estimated P_mus and ΔP_L,dyn were calculated according to the method described by Bertoni et al., ⁴ which is based on the patient’s inspiratory effort against an occluded airway (ΔP_occ). ΔP_occ is the difference between PEEP and the minimum pressure reached during the inspiratory effort. Three expiratory airway occlusions were performed, each separated by at least three respiratory cycles, and maintained for the duration of a single breath. The mean value of the three ΔP_occ measurements was used to estimate P_mus (P_mus = ¾ × ΔP_occ) and to calculate ΔP_L,dyn [ΔP_L,dyn = (Peak airway pressure – PEEP) – (2/3 × ΔP_occ)].

The primary outcome was extubation failure, defined as reintubation or use of non-invasive ventilation within 48 h post-extubation. ⁷ Statistical analysis included t-tests for continuous variables, Fisher’s exact test for categorical data, univariate logistic regression for significant predictors, and ROC curves to assess the predictive ability of estimated P_mus and identify optimal cutoffs.

Results

Out of 413 patients screened, 169 were included; 159 (94.1%) had successful extubation, while 10 (5.9%) failed (Supplemental Figure 1). Baseline demographics were similar between groups, with no significant differences in age, sex, APACHE II scores, or ICU length of stay (Supplemental Table 1). The success group had higher estimated P_mus before SBT (5.6 ± 2.3 cmH₂O) compared to the failure group (4.0 ± 2.3 cmH₂O; p = 0.032) and a greater increase after SBT (ΔP_mus: 0.46 ± 1.51 vs 0.09 ± 0.19; p = 0.006; Table 1).

Table 1.

Comparison of parameters between the success and failure group.

Parameter	Success (159)	Failure (10)	p Value
P0.1 pre-SBT; (cmH2O)	1.14 ± 0.75	1.1 ± 0.89	0.880
P0.1 post-SBT; (cmH2O)	1.38 ± 1.21	1.28 ± 1.05	0.795
Pmus pre-SBT; (cmH2O)	5.60 ± 2.34	3.95 ± 2.30	0.032
Pmus post-SBT; (cmH2O)	6.04 ± 2.59	4.04 ± 2.31	0.180
PL,dyn pre-SBT; (cmH2O)	3.12 ± 3.08	3.71 ± 4.05	0.564
PL,dyn post-SBT; (cmH2O)	2.62 ± 3.29	4.07 ± 4.11	0.183
∆ P0.1 ;(cmH2O)	0.24 ± 0.71	0.25 ± 0.35	0.958
∆ Pmus ;(cmH2O)	0.46 ± 1.51	0.09 ± 0.19	0.006
ΔPL,dyn ;(cmH2O)	−0.47 ± 2.38	0.36 ± 0.44	0.270

ROC analysis showed an AUC of 0.727 for pre-SBT estimated P_mus and 0.737 for post-SBT values (Figure 1). Cutoffs identified were <4.1 cmH₂O pre-SBT and <4.4 cmH₂O post-SBT, predicting failure with reasonable sensitivity and specificity. P0.1 and ΔP_L,dyn were not associated with extubation outcomes.

Figure 1.

ROC curves of estimated P_mus pre-SBT (blue line), AUC = 0.727 (95% CI, 0.534–0.919); estimated P_mus post-SBT (red line), AUC = 0.737 (95% CI, 0.557–0.917); estimated Δp_mus (green line), AUC = 0.584 (95% CI, 0.490–0.677).

ROC curve: receiver operating characteristic curve; AUC: area under the curve; CI: confidence interval.

Discussion

This study demonstrates that lower estimated P_mus values, even among patients who tolerate SBT, are associated with higher risk of extubation failure, likely due to underlying diaphragmatic weakness. ³ Estimated P_mus offers a direct, non-invasive measure of inspiratory muscle effort, contrasting with P0.1, which reflects respiratory drive and may not correlate with muscle strength. ⁸ While AUC values indicate moderate predictive power, our findings suggest estimated P_mus could complement existing weaning indices, like RSBI, to enhance clinical decision-making. ⁹

The method’s practicality makes it attractive: P_mus estimation requires only standard ventilator functions and simple maneuvers, potentially allowing widespread implementation. Six patients in our cohort had high estimated P_mus (>10 cmH₂O), indicative of excessive effort, yet were successfully extubated, underscoring the complexity of balancing adequate effort without overexertion. Conversely, patients with lower P_mus may fail due to insufficient strength to sustain spontaneous breathing.

Limitations include the observational design, single-center setting, and small number of failure events, which limit statistical power. Moreover, while upper airway failure cases were excluded, other non-muscular causes of failure could confound results. Larger multicenter studies are needed to validate cutoff thresholds and explore integration with multimodal weaning assessments.

Estimated P_mus measured non-invasively before and after SBT shows promise as a predictor of extubation failure in ICU patients, with cutoffs <4.1–4.4 cmH₂O indicating higher risk. Integrating this simple bedside parameter into routine practice could aid in identifying patients at risk, but further multicenter validation is essential.