Mechanical ventilation is a life-saving treatment in critically ill patient; however it is also accompanied by many complications [1]. So, it is advisable to liberate patients from mechanical ventilation as soon as the underlying cause that led to the mechanical ventilation has improved, and the patient is able to preserve spontaneous breathing with good gas exchange [2].

Patients who wean successfully have less morbidity, mortality, and supply utilization than patients who require prolonged mechanical ventilation [3]. Therefore, once mechanical ventilation commenced; planning for weaning should starts [4]. Multiple predictors of successful weaning have been studied; mostly displaying good sensitivity but low specificity [5]. Weakness of diaphragmatic muscle considers a major cause of weaning failure from mechanical ventilation [6].

The maximal inspiratory pressure (Pi_max) represents the extreme pressure generated against an occluded airway during inspiration and is used frequently to evaluate the diaphragmatic muscle capacity in intensive care unit [7]. Airway occlusion pressure (P_0.1) represent the pressure generated at the airway opening at the first one hundred ml seconds after inhalation against an occluded airway and it is an adequate measure of the central respiratory drive [8].

The aim of this study was to measure the predictive power of the respiratory muscle determinants including Pi_max, P_0.1, and P_0.1/Pi_max in weaning outcome of critically ill patients undergoing invasive mechanical ventilation.

This prospective observational study was conducted in the Surgical Intensive Care Unit (SICU) in Aswan University Hospital during the period from March 2017 to December 2017. The hospital ethics committee approved the study and a written informed consent was given by surrogate decision maker.

All adult patients who required invasive mechanical ventilation for at least 24 hours were included in the study and considered eligible for weaning when they attained the following criteria: resolution of the acute episode for which the patient was placed on ventilator, low-level pressure support (8 cm H₂O) and PEEP level (≤ 5 cm H₂O), adequate oxygenation (partial pressure of arterial oxygen (PaO₂)/fraction of inspired oxygen (FiO₂) ≥ 150), FiO₂<0.5, alert, and stable cardiovascular status (heart rate ≤ 120/ min, systolic blood pressure higher than 90 mmHg and lower than 160 mmHg) in the absence of any vasoactive support therapy [9].

The patients underwent a SBT for two hours by putting them on spontaneous mode of weaning with low-level pressure support (8 cmH₂O) and PEEP level (≤ 5 cm H₂O) using GE ventilator (Carescape R860, USA).

Patients who had severe ICU acquired neuromyopathy, primary unilateral/bilateral absence of diaphragmatic mobility, previously failed SBT, or with tracheostomy, were excluded from the study.

All the eligible patients for the study were evaluated by Demographic data (Age, Sex, weight and height), diagnosis on ICU admission, vital signs (Heart rate, Blood pressure, and Respiratory rate), oxygen status by pulse oximetry (SPO₂) and arterial blood gas. Ventilatory data including spontaneous exhaled tidal volume, respiratory rate, minute ventilation, and weaning predictors including, Pi_max, P_0.1 (measured five times over a period of 60-90 s and the average of these measurements was taken) and lastly P_0.1/ Pi_max, were measured using GE ventilator (Carescape R860, USA).

Patients who passed SBT without deterioration were extubated and received oxygen through Venturi mask 40%. However, SBT failure was considered if the patient developed a decreased level of consciousness, diaphoresis, RR >35 breaths/min, haemodynamic instability (heart rate >140, systolic blood pressure >180 or <90 mmHg) or signs of increased work of breathing [10].

Statistical analysis

SPSS (Statistical Package for Social Science) software program version 19.0 (SPSS Inc., Chicago, IL) was used for data recording and handling. Data presented as (mean ± SD) for continuous variables. Student’s t tests were used for the comparison of continuous variables and chi-square tests were used to compare numerical variables. Nonparametric tests were used for abnormal distributed data in the current study.

To assess the accuracy of each weaning index, Receiver operator characteristic curves (ROC) were used and the non-parametric method of Delong was used to calculate the area under the ROC curves (AUC) for each weaning index [11]. P-value <0.05 considered significant.

During the study period, we evaluated 55 patients ready for weaning. Five cases were excluded, three of which had hypotension (systolic blood pressure less than 90 mmHg) and two cases had disturbed conscious level. Among the fifty patients underwent SBT, 30 patients successfully passed SBT and weaned from mechanical ventilation (group A) while 20 patients failed SBT and not weaned from mechanical ventilation (group B).

The demographic and clinical data did not differ considerably between the two study groups (Table 1). Mean values for different weaning indices used to guide the success of weaning are shown in (Table 2) and there was a substantial difference between the two study groups as regard Pi_max, P_0.1, and P_0.1/Pi_max.

Table 1: Demographic and clinical data of the studied groups (n=50).

Table 2: Weaning indices among the studied groups.

Analysis of the predictive values of the studied indices regarding weaning success showed that, Pi_max with a cutoff value of ≤ -22 cm H₂O had the greatest sensitivity, positive predictive value, and negative predictive value (91.67, 87.3, and 87.2 respectively) compared with P_0.1 and P_0.1/Pi_max (Table 3). Also, Pi_max had an excellent area under the curve (AUC=0.93) with the highest diagnostic accuracy (0.87) among all the studied indices) (Figures 1-3).

Figure 1: Area under receiving operating characteristic curve for Pimax ≤ -22 to predict successful weaning AUC=(0.93).

Figure 2: Area under receiving operating characteristic curve for central respiratory derive (P_0.1)>2 to predict successful weaning, AUC=(0.72).

Figure 3: Area under receiving operating characteristic curve for P_0.1/Pi_max ≤ 0.13 to predict successful weaning, AUC=0.63).

Table 3: The diagnostic performance tests of each index used to predict weaning success in our study.

The current study focused on evaluating the diagnostic accuracy of some weaning indices for respiratory muscle determinants including (Pi_max, P_0.1 and P_0.1/Pi_max) in predicting the outcome of weaning in critically ill patients undergoing mechanical ventilation [12,13]. The present study concluded that; Pi_max, P_0.1 and P_0.1/Pi_max reading values were significantly different among the weaned and not weaned groups. the diagnostic performance tests,Pi_max diagnosis on admission to the intensive care unit

During correlating weaning outcome with Pi_max values in our study, there was a higher negative value of Pi_max in the successfully weaned group compared with the failed one. Similar results were reported by dos Santos Bien et al. [14]; however those results were in contrast with previous studies [15].

Regarding P_0.1 values, they were significantly different between patients who had succeeded and those who had failed weaning (-2.40 ± 0.55) cm H₂O vs. (-1.95 ± 0.63) cm H₂O. This variation could be as this test mainly affected by impaired neurological drive which is widely variable between patients.

Pi_max.P_0.1 P_0.1Pi_maxPi_maxP_0.1 P_0.1Pi_maxPi_maxP_0.1 Pi_maxP_0.1 P_0.1Pi_max

The best cut off value of P_0.1 that could predict weaning success varied among studies from 0.5 to 1.5 cm H₂O to lesser than 4.2 cm H₂O [16], Conversely, de Souza et al. suggested that a value of P_0.1 higher than 2.33 cm H₂O was associated with weaning failure [17], which is comparable with our result, in which the best cut off point of P_0.1 that predict weaning success was >2 cm H₂O.

Nemer et al. [18] found that, P_0.1/Pi_max ratio <0.14 was highly associated with weaning success, this result was comparable with our study, in which P_0.1/Pi_max ratio with a cutoff value less than 0.13 was associated with successful weaning, but area under ROC curve was inaccurate (0.63).

Our study is a single center study with small sample size, so we need many future studies with a large numbers of subjects to emphasizing our results. Moreover, this study involved patients in the surgical intensive care unit that limited generalization of our results.

Pi_max provides appreciated data with greater accuracy to assess inspiratory muscle strength and predicting weaning success in mechanically ventilated patients than did P_0.1, and P_0.1/Pi_max ratio.

The authors appreciated the surgical intensive care unit residents and nurses for their help in this research. No fund was paid by any institution.