Role of ventilator and ultrasound parameters in predicting extubation success

Abstract

The gold standard method for an accurate and easy way to predict successful weaning has not been established. The study was aimed to determine the predictive values of airway occlusion pressure (P0.1), parasternal intercostal muscle, and diaphragm for successful extubation and their correlation with ICU length of stay and duration of mechanical ventilation. We recruited patients who had been admitted to ICU and mechanically ventilated for at least 48 h. Patients who had successfully passed the spontaneous breathing trial (SBT) and planned for extubation were screened. P 0.1 parameter was obtained from the ventilator. We measured parasternal intercostal muscle thickness (PIMT), parasternal intercostal muscle thickening fraction (PIMTF), diaphragm thickness (DT) and diaphragm thickening fraction (DTF). Primary outcome was to determine the predictive values of P 0.1, PIMTF & DTF for successful extubation. Of the 126 enrolled patients, 101(80.1%) were successfully extubated and 25 (19.8%) had extubation failure. The predictive thresholds for successful extubation for P 0.1 was 1.8 cm of H₂O (88% sensitivity and 81% specificity with area under receiver operating characteristic curve (AUC) of 0.907, for PIMTF was 12.5% (100% sensitivity and 98% specificity with AUC 0.999), and for DTF was 22.8% (84% sensitivity and 94% specificity with AUC 0.907). P0.1, PIMTF, and DTF are independently linked to extubation outcomes and a combination of these three variables could serve as a valuable tool for accurate prediction of successful extubation.

Introduction

Mechanical ventilation is a cornerstone intervention in critical care for managing cardio-respiratory failure. Optimizing the timing of weaning is imperative, as both premature and delayed extubation are associated with increased morbidity, mortality, and healthcare burden¹. Premature extubation predisposes patients to cardiovascular stress, hypercapnia, hypoxemia, and re-intubation, whereas delayed weaning increases the risk of ventilator-associated pneumonia (VAP), diaphragmatic atrophy, and prolonged ICU stay^2–4. The importance of a comprehensive approach to weaning is further supported by the ENIO International Prospective Study, which highlighted the critical role of neurological factors, such as level of consciousness and secretion management, in predicting extubation success in neurocritical care patients⁵.

The airway occlusion pressure at 100 milliseconds (P 0.1) reflects neural respiratory drive and can be non-invasively assessed via modern ventilators^6,7.

Ultrasonography is a safe, radiation-free, non-invasive, easily available bedside tool^7–11. Ultrasound-based evaluation of the parasternal intercostal muscle thickening fraction (PIMTF) and diaphragm thickening fraction (DTF) provides a non-invasive and dynamic assessment of respiratory muscle function¹⁰. While these parameters have been individually studied, their combined predictive value for extubation success remains unexplored¹².

This study aimed to evaluate the predictive accuracy of P 0.1, PIMTF, and DTF in determining extubation outcomes and their correlation with ICU length of stay and duration of mechanical ventilation.

Materials and methods

This study was a single-centre, prospective observational study conducted in the intensive care unit (ICU) of a tertiary care hospital. Approval from Institutional ethics committee of St John’s medical college, hospital was obtained (IEC 419/2021), CTRI − 2022/09/058080 dated 12th November 2021). All the procedures were followed in accordance with the ethical standards of the responsible committee on human experimentation (institutional or regional) and with the Helsinki Declaration of 1975. Strengthening the reporting of observational studies in epidemiology (STROBE) guidelines were followed (Strobe checklist). Written informed consent was taken from legally acceptable representative of the patient. The study was conducted over a period of 18 months.

Patients intubated and mechanically ventilated for a minimum of 48 h, successfully passed the SBT, were on weaning mode [pressure support-positive end expiratory pressure (PS/PEEP)] and planned for extbation as decided by the treating team were included in the study. Patients who were tracheostomised, were on T-piece, failed extubation, having diaphragm paralysis or having poor ultrasound window were excluded from the study.

Each patient was screened just before extubation. Comorbidities, reason for intubation, number of ventilator days, sequential organ failure assessment (SOFA) score, acute physiology and chronic health evaluation (APACHE II) score (after 24 h of admission) were collected. Rapid Shallow Breathing Index (RSBI), (Respiratory rate (RR) / tidal volume (VT) in litres) was also calculated.

P 0.1 was measured from the Maquet Servo-s ventilator during spontaneous breathing. Three values were taken covering 3–5 respiratory cycles, and an average was taken¹³. All Ultrasound-based evaluations were performed by the principal investigator under supervision. Patients were evaluated with a 12 MHz liner array probe for better resolution.

Parasternal intercostal muscle was measured at the level of second right intercostal space, 5 to 6 cm lateral to the sternal edge¹⁴. A linear array probe was placed perpendicular to the second right intercostal space. A three-layered biconcave structure with two linear hyperechoic membranes and a medial portion with muscle echotexture that corresponds to the right parasternal intercostal muscle was visualized. After initial identification of the parasternal intercostal muscle by 2D imaging, the measurements were taken using the M mode. The thickness of the parasternal intercostal muscle was measured at end-expiration and end-inspiration by freezing the images in M mode (Fig. 1A). PIMTF was then calculated as a percentage by the formula:

Fig. 1.

(A) Ultrasonographic visualisation of parasternal intercostal muscle with B-mode. (B) Ultrasonographic visualisation of diaphragm with M-mode showing an increase in diaphragmatic thickness during inspiration and decrease in thickness at expiration.

Three measurements were taken, and an average was noted.

Diaphragm thickness (DT) measurements were taken in a semi-recumbent position with a head-end elevation of 30 to 45°¹⁵. A linear probe was placed perpendicular to the chest wall in the mid-axillary line between 8th and 10th intercostal space. The diaphragm identified as the hypoechoic structure between two hyperechoic lines, which correspond to the pleura and peritoneum (Fig. 1B). After initial identification of the diaphragm by 2D imaging, the measurements taken using the M mode. The thickness of the diaphragm was measured as the perpendicular distance between the pleural and peritoneal lines. The thickness of the diaphragm increases during inspiration and decreases during expiration. The thickness of the diaphragm at inspiration was thus measured at the point of maximum thickness, and the thickness at expiration was measured at the point of least thickness (Fig. 1B). DTF was then calculated as a percentage by the formula. An average of three readings were taken.

Fig. 2.

(A) Receiver operating characteristic (ROC) curve analysis of diaphragm thickening fraction (DTF), parasternal intercostal muscle thickening fraction (PIMTF) and P0.1 in predicting successful extubation. (B) Receiver operating characteristic (roc) curve analysis of P 0.1, (C) PIMTF and (D) DTF.

Patients were categorized as successful extubation when they tolerated extubation for more than 48 h. Patients who got intubated within 48 h of extubation were considered as failed extubation. Primary outcome of our study was to determine the predictive values of P 0.1, PIMTF , and DTF in successful extubation from mechanical ventilation. Secondary outcomes were to study the correlation of P 0.1, PIMTF, and DTF with ICU length of stay and the duration of mechanical ventilation.

Statistical analysis

Data analysis was performed using SPSS version 24. Sample size calculations were based on data from the study by Turton et al.¹⁰. A total of 126 sample size was required to achieve a statistical power of 80% and a significance level of 5% and anticipated 10% loss to follow-up.

Categorical data was presented as frequency and percentages. All continuous data was described using the mean and standard deviation, or median and interquartile range based on the distribution of the data. “Independent t test”or “Mann Whitney U test” was applied to compare the parameters between successful extubation and failed extubation groups. Receiver operating characteristic curve (ROC) analysis was done for each of the extubation parameters and predictive cutoffs for P0.1, PIMTF, and DTF were calculated from the Youden’s index. Pearson’s correlation was also performed for these 3 variables with ICU length of stay and duration of mechanical ventilation.

Results

Of 126 study participants, 25 (19.8%) had failed extubation, and 101 (80.2%) had successful extubation. The mean age was 49.5 (17) years and predominantly were males (68%). The mean APACHE II was 15.2(5.5) in successful extubation group and 18.5(5.2) in failed extubation group with p value of 0.009. The mean SOFA was 4.36(2.39) in successful extubation group and 5.56 (2.1) in failed extubation group with p value of 0.023. The main indication for intubation was extra pulmonary sepsis (38.85%) (Table 1).

Table 1.

Demographics & clinical characteristics.

Parameters	All (n = 126)	Successful extubation ( n = 101)	Failed extubation ( n = 25)	p value
Age*	49.5 (17)	48(17)	53(17)	0.207
Male/female	86(68)/40(32)	66(65.3)/35(34.6)	20(80)/5(20)	0.159
Indication for intubation
Pulmonary Neurological Cardiac Extra-Pulmonary sepsis Others	36(27.9) 24(19.9) 11(8.7) 49(38.85) 6(4.7)	30(29.7) 21(20.7) 5 ( 4.9) 42 (41.5) 3(2.9)	6(24) 3(12) 6 (24) 7(28) 3(12)	0.906
SOFA	4.6(2.38)	4.36(2.39)	5.56(2.10)	0.023
APACHE II	15.9(5.5)	15.2(5.5)	18.5(5.2)	0.009
Duration of mechanical ventilation	5(4–9.25)	5(4–8)	8(5–11)	0.054

*mean (standard deviation SD), ^median (interquartile range), values (%).

Among the ventilator parameters mean RSBI was lower in successful extubation group and higher in failed extubation group [55 vs. 67, (p = 0.001)] (Table 2). Mean pressure support (PS) was significantly lower (5.3 cm of H₂O) in successful extubation group and higher (6.2 cm of H₂O) in failed extubation group (p = 0.016) (Table 2). The mean P 0.1 in the successful extubation group was significantly lower [1.33 (0.43) cm of H₂O] than in the failed extubation group [2.27 (0.53) cm of H₂O, p value < 0.001] (Table 2). The mean PIMTF in the successful extubation group was lower (8.77%) than failed extubation (16.05%) group (p = 0.001). The mean DTF in successful extubation group was higher (34.43%) than in failed extubation group (22.39%) (p < 0.001).

Table 2.

Respiratory and ultrasonographic parameters.

Parameters	All (n = 126)	Successful extubation ( n = 101)	Failed extubation ( n = 25)	p value
RSBI(BREATHS/MIN/L)	57.6(8.4)	55(6)	67(6.2)	0.001
HR at extubation(BEATS/MIN)	79(11)	77(11)	88(11)	< 0.001
MAP at extubation(MM OF HG)	79(8)	78(8)	83(6)	0.009
P/F RATIO	246(28)	251(27)	225(22)	< 0.001
PEEP( CM OF H2O)	6.12(1.07)	6(1.07)	6.4(1.04)	0.156
PS( CM OF H2O)	5.57(1.97)	5.3(2.05)	6.2(1.45)	0.016
P 0.1 (CM H2O)	1.51(0.58)	1.33 (0.43)	2.27 (0.53)	< 0.001
Parasternal Intercostal muscle thickness inspiration (mm)	4.14(0.41)	4.10 (0.41)	4.31(0.41)	0.026
Parasternal Intercostal thickening fraction (PIMTF-%)	10.22(3.40)	8.77(1.69)	16.05(2.08)	< 0.001
Diaphragm thickness inspiration (mm)	2.62(0.36)	2.64 (0.37)	2.52(0.28)	0.123
Diaphragm thickening fraction (DTF-%)	32.04(8.04)	34.43 (6.75)	22.39 (5.12)	< 0.001

(RSBI-Rapid Shallow breathing index, HR-Heart Rate, MAP- Mean Arterial Pressure, P/F ratio- Pao2/Fio2, PS- Pressure Support, PEEP- Positive End Expiratory Pressure).

Based on the point of maximum Youden’s Index, P 0.1 had a cut-off of 1.8 cm H₂O with the demonstrated sensitivity of 88% and a specificity of 81% (AUC – 0.907), PIMTF cut-off of 12.5% with a sensitivity of 100% and a specificity of 98%(AROC- 0.881) and DTF cut-off of 22.8% with a sensitivity of 84% and a specificity of 94%(AUC-0.907) for predicting successful extubation from mechanical ventilation. All the three parameters with different cut offs combined and individual ROC curves presented in Fig. 2A, B,C, D.

The positive predictive value (PPV) for P0.1was 53% and Negative predictive value (NPV) was 97%. The PPV and NPV were 93% and 100% for PIMTF respectively. The PPV and NPV for DTF were 78% and 96% respectively. If all three parameter were positive based on the cut off, the sensitivity was 76% and a specificity was  100%. When we studied combining two parameters together the best combination we observed was for P 0.1 and PIMTF with sensitivity was 88% and specificity was 100% when both the tests results were positive. For P0.1 and DTF sensitivity was 76% and specificity was 97%. For DTF and PIMTF sensitivity was 84% and specificity was 99%.

DTF had a negative correlation with ICU length of stay with Pearson’s correlation coefficient of 0.200 (p < 0.05). P 0.1 and PIMTF had a positive correlation with ICU length of stay with Pearson’s correlation coefficient of 0.393 and 0.470 respectively (p = 0.001). Similarly, P 0.1 and PIMTF had positive correlation with duration of mechanical ventilation with Pearson’s correlation coefficient of 0.229 and 0.213 respectively, with significant p value. DTF did not have a statistically significant correlation with the duration of mechanical ventilation (Table 3).

Table 3.

Correlation of P0.1, Parasternal intercostal muscle thickening fraction (PIMTF) and diaphragm thickening fraction (DTF)  with ICU length of stay & duration of mechanical ventilation.

ICU Length of stay	Pearson’s Correlation Coefficient	p value
P 0.1	0.393	0.001
DTF	-0.200	0.025
PIMTF	0.470	0.001

Duration of MV	Pearson’s Correlation Coefficient	p value
P 0.1	0.229	0.010
DTF	− 0.057	0.526
PIMTF	0.213	0.017

Discussion

Weaning from mechanical ventilation is a critical phase in the management of ICU patients, and identifying reliable predictors of successful extubation remains a key challenge. Emerging indices such as P 0.1, PIMTF, and DTFhave shown promise in predicting weaning outcomes.

This study evaluated these parameters in a cohort of 126 mechanically ventilated patients, of whom 101 were successfully extubated, while 25 had extubation failure. The findings provide valuable insights into the utility of these indices in clinical practice and highlight their relevance in comparison to prior studies. A recent multicenter study also demonstrated the utility of diaphragm ultrasound in assessing weaning outcomes in COVID-19 patients¹⁶.

The demographic profile of our study population, with a mean age of 49.5 years and a male predominance (68%), aligns with previous studies. For instance, Abdehwahed et al. reported a mean age of 43.49 years in successfully weaned patients and 40.85 years in the failed weaning group, while Kumar et al. reported a lower mean age of 35.5 years^17,18. The male predominance observed in our study is consistent with findings from other studies^17,19–21. Thille et al., noted higher reintubation rates among males within 48 h post-extubation²². In our cohort, 80% of the failed extubation group were male, further supporting the association between male gender and poorer weaning outcomes.

The primary indication for intubation in our cohort was extrapulmonary sepsis (38.8%), followed by respiratory, neurological, and cardiac causes. (Table 1). This contrasts with studies by Guojun et al., where pneumonia and congestive cardiac failure were the leading indications, and sepsis accounted for only 7% of cases²³. Similarly, Alam et al. and Funda Gok et al. focused on weaning outcomes in polytrauma patients, highlighting the variability in patient populations across studies^24,25. Our inclusion of a mixed ICU population enhances the generalizability of the findings.

Higher organ failure scores, as measured by SOFA and APACHE II, were significantly associated with extubation failure in our study. This suggests that patients with greater physiological derangement are at increased risk of weaning failure. However, this finding diverges from studies by Alam et al. and Funda Gok et al., which did not observe a similar association. This discrepancy underscores the need for further investigation into the role of organ dysfunction in weaning outcomes and highlights the potential influence of population-specific factors^25,26.

The extubation failure rate in our study was 19.8%, consistent with the range of 10–20% reported in similar studies^3,27,28. However, higher failure rates have been observed in specific subpopulations, such as patients with ICU-acquired weakness. For example, Jung et al. reported an extubation failure rate of 50% in patients with ICU-acquired weakness, a finding supported by the observation that 80% of such patients in his cohort had diaphragm dysfunction²⁹. This highlights the critical role of diaphragm function in weaning success and suggests that targeted assessment of diaphragm activity may improve outcomes in this high-risk group.

Among the weaning parameters assessed, P 0.1, PIMTF, and DTF demonstrated significant differences between the successful and failed extubation groups (Table 3). Our findings are consistent with a meta-analysis by Sato et al., which reported a pooled sensitivity of 86% and specificity of 58% for P 0.1 in predicting weaning outcomes³⁰. However, the cutoff values for P 0.1 varied across studies, with Guojun et al. reporting a cutoff of 3.9 cm H₂O for weaning failure, compared to our cutoff of 1.8 cm H₂O²³. Similarly, PIMTF and DTF were strong predictors of extubation success, with lower PIMTF and higher DTF values correlating with favorable outcomes. Our findings align with studies by Dres et al.²⁶, who identified a PIMTF cutoff of ≥ 9.5% for predicting weaning failure, and Umbrello et al.³¹, who highlighted the inverse relationship between diaphragm and parasternal intercostal muscle activity during weaning. Guojun et al.²³ reported a PIMTF cutoff of ≥ 13.15% for weaning failure, while Kumar et al. identified a cutoff of ≥ 15.38%¹⁸. We found PIMTF cut off of 12.5%. These variations underscore the need for standardized measurement protocols and population-specific cutoff values. A prior study emphasized the importance of key weaning parameters like respiratory drive and diaphragm function, in predicting weaning success³². Our study helped to know interaction between diaphragm and parasternal intercostal muscles.

For DTF, our study identified a cutoff of 22.8% for predicting successful extubation, with sensitivity and specificity comparable to those reported in prior research. Ali and Mohammad et al.²¹ reported a DTF cutoff of 30%, while Blumhof et al.³³ and Di Nino et al.³⁴ identified cutoffs of 20% and 30%, respectively. Abdelwahed et al.¹⁷ also found that a higher DTF was associated with successful weaning. Study by Farghaly et al.²¹ and Ferrari et al.¹⁹ reported similar trends, further reinforcing the role of diaphragm function in predicting extubation outcomes. These variations in cutoff values highlight the influence of patient population differences, measurement techniques, and clinical contexts, underscoring the need for standardized protocols to optimize DTF assessment in weaning decisions.

The multifaceted approach to weaning assessment is necessary, as reliance on a single parameter may be insufficient given the complex nature of weaning failure. Al-Husinat et al. had proposed an algorithm which integrated all these parameters for process of weaning in critically ill patients³⁵.

Furthermore, we observed significant correlations between these indices and clinical outcomes. Higher DTF, lower P 0.1 and PIMTF values correlated with reduced ICU length of stay. P 0.1 and PIMTF had statistically significant positive correlation with duration of mechanical ventilation. These findings suggest that integrating these parameters into weaning protocols could enhance decision-making and improve patient outcomes.

Strengths and Limitations.

This is possibly a first study which measured 3 important parameters (P 0.1, PIMTF, and DTF) for weaning assessment. The impact of each of these parameters on successful extubation was measured individually as well as in combination, which improved the accuracy of the prediction of successful extubation outcomes. To avoid inter-observer variations, which are very common in ultrasound measurements, all the measurements were taken by a single observer, and intra-observer variations were avoided by taking an average of 3 readings of these variables. Our primary outcome was successful extubation rather than a spontaneous breathing trial which is more appropriate and pragmatic outcome.

One of the limitations was a single-centre study. Secondly, although planned for consecutive enrolment of the patients, convenient sampling was done which may have caused selection bias. As ultrasonography-based indices are operator dependent, training to measure these indices is necessary. We performed a single measurement of diaphragm and parasternal intercostal muscles at the time of extubation only. The change in thickness and thickening fraction with respect to the duration of mechanical ventilation were not measured. The effect of different PS on thickness and thickening fraction was not studied. Assessment for cardiac causes of weaning failure was not performed³⁶. We did not include number of failed SBTs before first extubation attempt.

Conclusion

From our study, P 0.1 ≤ 1.8 cm of H₂O, PIMTF ≤ 12.5%, and DTF ≥ 22.8% could predict successful extubation. Further studies are needed to validate the results of our study.

Weaning from mechanical ventilation is an important aspect for an intensivist. P 0.1, PIMTF and DTF are emerging parameters that help to predict successful extubation with good accuracy. They are easy to perform at the bedside, and each parameter has good correlation with extubation outcome. Among the three parameters studied, we suggest to use one ventilator parameter P0.1 and one ultrasound parameter for predicting extubation success.

Abbreviations

APACHE II

Acute physiology and chronic health evaluation

BP

Blood pressure

CPAP

Continuous positive airway pressure

DTF

Diaphragm thickening fraction

HR

Heart rate

ICU

Intensive care unit

MAP

Mean arterial pressure

PIMTF

Parasternal intercostal muscle thickening fraction

P 0.1

Airway occlusion pressure at 100 msec

PR

Pulse rate

PS

Pressure support

PEEP

Positive end expiratory pressure

RR

Respiratory rate

SBT

Spontaneous breathing trial

SOFA

Sequential Organ Failure Assessment Score

TV

Tidal volume

USG

Ultrasound

Author contributions

MR-Concept, design, statistical analysis and writing a manuscript. AAH- Concept, Design, writing and approving final draft. All authors have read and approved the final manuscript.

Data availability

Data is provided within the manuscript.

Declarations

Ethics approval and consent to participate

Approval from institutional ethical committee of St John’s medical college, hospital was obtained. Consent from legally acceptable representative (LAR) was obtained and then patients were enrolled.

Competing interests

The authors declare no competing interests.