Association between abdominal muscle weakness and weaning failure in mechanically ventilated critically ill patients: an ultrasound assessment

Abstract

Background

Ultrasound evaluation of abdominal muscle activity during cough may provide a practical tool to identify patients at risk of failure of liberation from mechanical ventilation (MV). We aimed to investigate the association between abdominal muscle weakness (AMW), assessed by ultrasound during coughing, and failure of liberation from MV within 48 h after the first liberation attempt, with liberation failure within 7 days as a secondary outcome. We also explored its relationship with diaphragm dysfunction (DD) and ICU-acquired weakness (ICUAW).

Methods

We conducted a prospective cohort study including intubated or tracheostomized critically ill patients at high risk of failure of liberation from MV, who had been ventilated for at least 48 h, and had successfully completed a spontaneous breathing trial (SBT). Before the liberation attempt, all patients were assessed for AMW, defined as the overall thickening fraction of the abdominal muscles during cough (TF_abs) measured by ultrasound. DD was assessed using ultrasound and defined by a diaphragm thickening fraction (TFdi) below 29%, while ICUAW was evaluated using the Medical Research Council scale or the simplified peripheral nerve test in uncooperative patients.

Results

A total of 46 patients were included in the study. Nine (20.5%) patients had AMW, ICUAW, and DD simultaneously, whereas 10 (22.9%) had isolated AMW. AMW was not associated with ICUAW (p = 0.28) or DD (p = 0.97). Liberation from MV within 48 h was successful in 30 patients (65.2%) and failed in 16 (34.7%). During the 7-day follow-up, liberation failure occurred in 3 of the 30 patients (10.0%) who remained free from MV within 48 h. Multivariable logistic regression analysis showed a higher probability of liberation failure within 48 h (OR 9.4, CI 1.1–75.2, p = 0.03) and 7 days (OR 14.8, CI 1.6–136.1, p = 0.02) in patients with AMW. ICUAW was associated with liberation failure within 48 h (OR 8.5, CI 1.5–46.8, p = 0.01) and 7 days (OR 20.4, CI 2.7–155.0, p < 0.01), whereas the presence of DD was not related to either outcome.

Conclusions

Ultrasound evaluation of abdominal muscle function during cough, together with an assessment of ICUAW, may help identify critically ill patients at risk of failure of liberation from MV and may support clinical assessment of readiness for liberation.

Supplementary Information

The online version contains supplementary material available at 10.1186/s44158-026-00398-4.

Background

Weaning patients from mechanical ventilation (MV) is a complex and time-consuming process, accounting for nearly 40% of total time spent on MV [1]. The spontaneous breathing trial (SBT) is the standard test used to assess the patient’s ability to sustain spontaneous breathing [1]. However, SBT success does not fully reflect readiness for liberation from MV, which also depends on other factors beyond lung parenchyma, such as respiratory muscle weakness, cough effectiveness, secretion clearance, and neurological status [2, 3]. Consequently, some patients may successfully pass the SBT but still experience delayed failure following the first liberation attempt. As a result, the concept of “readiness for liberation” remains difficult to define and is still influenced by subjective clinical judgment [4–7].

ICU-acquired weakness (ICUAW) and diaphragmatic dysfunction (DD) are well-established factors associated with failure of liberation from MV [1, 8–10]. However, abdominal muscle weakness (AMW) may also impair the weaning process but is often overlooked [4, 11, 12].

Abdominal muscles play a key role in cough effectiveness and secretion clearance. Their dysfunction may impair cough strength, increasing the risk of atelectasis and secretion retention, and potentially contributing to poor outcomes [13–17].

Beyond airway clearance, abdominal muscles contribute to respiratory mechanics [15, 18–21]. Their activity helps optimize diaphragm positioning and enhances its pressure-generating capacity. Furthermore, during expiration, their contraction enhances expiratory efficiency, reducing end-expiratory volume and increasing elastic recoil for the subsequent inspiration [22, 23]. Doorduin et al. showed that abdominal muscle recruitment during weaning may reflect increased respiratory load and compensatory effort in patients at higher risk of failure [19].

Despite these observations, the role of AMW in weaning outcomes remains poorly understood. Assessing abdominal muscle function is challenging, especially in uncooperative patients, as most available methods require patient participation [5, 22–25]. Ultrasound (US) has recently emerged as a simple, noninvasive bedside tool to assess abdominal muscle activity during cough [26].

The present study investigated the associations between AMW, assessed by US during cough, and failure of liberation from MV within 48 h after the first liberation attempt. Additionally, the relationships between AMW, ICUAW, and DD and their overall impact on ventilator liberation outcome were explored.

Study design and methods

This prospective cohort study was conducted between October 2021 and June 2023 at the intensive care unit of Brescia. The study was approved by the local ethics committee (no. 3369). Informed consent was obtained from the patient or, when applicable, from a legal representative; when not feasible, a deferred consent procedure approved by the Ethics Committee was applied, with consent obtained once decision-making capacity was regained or managed in accordance with the approved protocol.

Study population

All patients on MV for at least 48 h were screened, including both intubated and tracheostomized. Patients were recruited after successfully completing SBT and being scheduled for their first liberation attempt from MV (extubation in intubated patients or disconnection in tracheostomized patients). Only patients at high risk of weaning failure, as defined by established criteria from previous studies, were included [27]. The presence of at least one of the following factors was considered to define high risk [27]: age older than 65 years; heart failure as the primary indication for MV; moderate to severe chronic obstructive pulmonary disease (COPD); acute physiology and chronic health evaluation II (APACHE II) score greater than 12 on liberation attempt; body mass index (BMI) of more than 30; clinically assessed inadequate cough reflex or suctioning > 2 times within 8 h before liberation attempt; prolonged use of invasive ventilation (≥ 7 days from the first SBT); and two comorbidities categorized based on the Charlson Comorbidity Index.

The exclusion criteria were as follows: age < 18 years; preexisting neuromuscular disorders that could affect the assessment of abdominal, limb and diaphragm muscle weakness; inability to perform ultrasound of the diaphragm or abdominal muscles; inability to assess ICUAW during the study protocol, using either the Medical Research Council scale sum score (MRCss) [28] or simplified peroneal nerve test (PENT) [29–31].

Study protocol

According to the center’s weaning protocol [32], all patients were assessed for readiness for liberation based on criteria defined by Boles et al. [7]. Readiness for liberation was defined by fulfillment of all following criteria: recovery from the precipitating illness; no sedation, with a Richmond Agitation-Sedation Scale (RASS) score ≥ − 1; only clinically demonstrated adequate cough with minimal or no tracheobronchial secretions and stable cardiovascular status (heart rate ≤ 140 beats per minute, systolic blood pressure between 90 and 160 mmHg); no or low dose of vasopressors. Additionally, the patient needed to have the absence of significant acid–base disturbances, adequate oxygenation (with a PaO₂/FIO₂ ratio > 150 or SpO₂ > 90% with FIO2 ≤ 0.4, PEEP < 8 cmH₂O, and arterial pH > 7.35), and adequate pulmonary function, as evidenced by a tidal volume (TV) > 5 mL·kg⁻¹ and a rapid-shallow-breathing index (RSBI; calculated as the ratio of respiratory rate to tidal volume) < 105 breaths·min⁻¹·L⁻¹ measured under the baseline pressure-support setting.

Once these clinical criteria were met, an SBT was performed using pressure support ventilation (PSV) with an inspiratory pressure of 6 cmH₂O and a PEEP level of 6 cmH₂O for 30 min [33].

In line with current guidelines and previous studies, the same criteria for readiness, SBT performance, and definition of liberation success were applied to both intubated and tracheostomized patients [7, 32, 34].

Patients who passed the SBT were temporarily reconnected to their previous ventilator settings for a resting period [7]. All patients who passed the SBT and then underwent their first liberation attempt were included in the study.

Assessment of AMW, ICUAW, and diaphragm function was performed on the day of the first liberation attempt. Diaphragm ultrasound was conducted immediately after SBT completion. Abdominal muscle ultrasound and ICUAW assessments were performed prior to the liberation attempt, during the resting period under MV at baseline ventilator settings. All ultrasound measurements and ICUAW assessments were conducted by researchers with certified expertise in these evaluations. Readiness for liberation and the decision to proceed with liberation from MV were determined by the treating clinicians in accordance with standard clinical practice [32]. Clinicians were fully blinded to the results of the study-specific muscle assessments.

Following the liberation attempt, patients received either high-flow oxygen or noninvasive ventilation based on clinical judgment, while all tracheostomized patients received high-flow oxygen. Post-liberation respiratory failure was defined according to current recommendations [32], and reconnection to MV was guided by the treating clinical team.

The study protocol is summarized in Fig. 1.

Fig. 1.

Study protocol. Critically ill patients tracheostomized or intubated and mechanically ventilated for at least 48 h were screened daily for liberation from mechanical ventilation. On the day of the first liberation attempt, muscle weakness was assessed and defined as follows: DD with diaphragm thickening fraction (TFdi) < 29%, ICUAW with Medical Research Council scale (MRC) < 48 in cooperative patients or simplified peroneal nerve test (PENT) < 5.26 mV in uncooperative ones, AMW with the sum of thickening fraction during cough of rectus abdominis, transversus abdominis and internal oblique (TFabs) < 127%. Criteria for readiness to wean are modified from Boles et al. [23]. Abbreviations: MV: mechanical ventilation; SBT: spontaneous breathing trial; DD: diaphragm dysfunction; AMW: abdominal muscle weakness; ICUAW: Intensive Care Unit Acquired Weakness; HRC: cardiac frequency; BP: blood pressure; SpO2: arterial oxygen saturation; FiO2: inspiratory oxygen fraction; PaO2: arterial oxygen tension; PEEP: positive end-expiratory pressure; RR: respiratory rate; VT: tidal volume; RSBI: Rapid Shallow Breathing Index

Data collection

According to the WIND (Weaning according to a New Definition) classification [35], patients were categorized into groups according to the time of the liberation process as follows: group 1 (liberation within 1 day from the first attempt; short weaning); group 2 (liberation between 1 and 7 days; difficult weaning); group 3a (liberation after more than 7 days; prolonged weaning); and group 3b (failure to achieve liberation after more than 7 days; weaning failure). This classification was selected as it allows consistent categorization of both intubated and tracheostomized patients [35].

At ICU admission, the following variables were recorded: demographic characteristics, ICU admission severity (Sequential Organ Failure Assessment, SOFA) score, APACHE II, reason for ICU admission, and comorbidities. ICU and hospital length of stay (LOS) and mortality outcomes were also recorded.

Abdominal muscle function was assessed using US by measuring the thickening fraction during cough (TF_cough) of the rectus abdominis (TF_{cough, RA}), transversus abdominis (TF_{cough, TrA}), and internal oblique (TF_{cough, IO}) musclesUS was performed on the patient's right side in the supine position during voluntary or, in the case of uncooperative patients, during cough induced by tracheal suctioning. In cooperative patients, a maximal voluntary cough was requested after standardized verbal instructions, whereas in uncooperative patients, cough was induced to provide a reproducible stimulus [26]. A high-frequency linear-array transducer was positioned at standardized anatomic landmarks: midway between the costal margin and the iliac crest along the right anterior axillary line to measure TF_{cough, TrA} and TF_{cough, IO} and 2–3 cm above the umbilicus and 2–3 cm to the right of the midline to measure TF_{cough, RA}. Using M mode, muscle thickness of each abdominal muscle was evaluated at end inspiration of a tidal volume immediately before the cough maneuver (Tpi) and peak expiration (Tpe) during coughing. The distance between the two internal layers (excluding the aponeurosis) of each muscle was measured offline. An example of the measurement has been included in the supplementary material to provide further clarification (Supplementary Fig. S1). The thickening fraction was calculated using the following equation: TF = (Tpe-Tpi)/Tpi × 100.

Therefore, the overall thickening fraction of the transversus abdominis, internal oblique, and rectus abdominis during cough (TF_abs) was assessed and AMW was defined as TF_abs less than 127% [15, 26]

TF_abs was computed as follows:

$${\mathit{TF}}_{\mathit{abs}}={\mathit{TF}}_{cough,RA}+{\mathit{TF}}_{cough,TrA}+{\mathit{TF}}_{cough,IO}$$

The values used to analyze each muscle were the average of three consecutive measurements.

ICUAW was defined as a Medical Research Council sum score (MRCss) < 48 [36]. In uncooperative patients, ICUAW was evaluated using the bilateral simplified peroneal nerve test (PENT), a validated screening tool with a sensitivity of 100% and a specificity of 85% [29–31]^. Uncooperative patients were defined as those unable to follow simple verbal commands or perform voluntary maneuvers due to impaired consciousness or neurological status [10, 30, 31].

Diaphragm function was assessed in the right hemidiaphragm. Diaphragm thickening fraction (TFdi) was computed as the percentage change in thickness between end expiration at functional residual capacity (i.e., minimum muscle thickness, Tee) and peak inspiration of a tidal inspiration (i.e., maximal muscle thickness, Tpi) as measured in M-mode (TFdi: Tpi-Tee/Tee) [37]. Diaphragm dysfunction (DD) was defined as a TFdi < 29%, calculated as the average of three measurements obtained under SBT conditions immediately after SBT completion [8, 37, 38].

Two experienced operators (LC, SG) independently performed all measurements using a standardized methodology. All ultrasound measurements were independently reviewed by two operators, with discrepancies resolved by a third operator (MB). Post-analyses were conducted using a DICOM viewer (RadiAnt DICOM Viewer, version 2021).

Outcomes

The primary outcome was failure of liberation from MV within 48 h, defined as reintubation (or reconnection to the ventilator in the case of a patient with a tracheostomy) within 48 h after the first MV liberation attempt.

The secondary outcomes were failure of liberation from MV within 7 days; 28-ventilator-free days (VFDs), defined as 28 days on mechanical ventilation or as zero if the subject either died within 28 days of MV or was on MV for more than 28 days. The association between AMW and weaning class (WIND class) was also evaluated.

Finally, the relationship between AMW, ICUAW, and DD and their overall impact on failure of liberation from MV within 48 h and within 7 days, VFDs, and the WIND class was explored.

Statistical analysis

Continuous variables are presented as the mean and standard deviation (SD) or the median and interquartile range (IQR), as appropriate. Categorical variables are presented as counts and percentages. The distribution of the data was assessed graphically.

All variables of interest (AMW, DD, and ICUAW) were examined in terms of failure of liberation from MV within 48 h and 7 days. A sample size of 40 patients was calculated to achieve 80% power (error rate of 0.05) in detecting the association between AMW and failure of liberation from MV within 48 h. The sample size was calculated by considering a proportion of liberation failure among those with 50% AMW and liberation failure among those with no AMW of 10% on the basis of previous studies [26].

ICU admission baseline characteristics were compared between patients with failure of liberation from MV and those without failure via t-test or the Wilcoxon test for continuous variables and Pearson’s chi-square test for categorical variables. The difference in the prevalence of AMW between the ICUAW and DD groups was evaluated using Pearson’s chi-square test.

A multivariable logistic regression was performed to determine the associations between failure of liberation from MV within 48 h and 7 days (dependent variable) and AMW, DD and ICUAW (independent variables). A competing risk regression was performed to explore the associations between AMW and VFDs. This analysis provides a subdistribution hazard ratio (SHR) that links an intervention to failure of liberation from MV while accounting for the competing risk of death [39]. Finally, multinomial logistic regression was used to examine the association with WIND. All regressions were adjusted for age, body mass index (BMI), and SOFA score. Given the sample size, covariates were limited to these variables, considering them a clinically reasonable choice to prevent underpowering.

All statistical tests were two-sided, and the significance level was set to 5%. Data were analyzed using R (version 4.4.2). The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Results

Forty-six patients were included in the study; details of inclusion and exclusion, as well as the reasons, are outlined in Supplementary Fig. S2. Table 1 presents patient characteristics according to liberation status within 48 h. Failure of liberation from MV within 48 h after the first liberation attempt occurred in 16 (34.7%) patients, whereas 30 (65.2%) patients remained free from MV at 48 h. Failure of liberation from MV within 7 days after the first attempt occurred in only 3 of 30 (10.0%) patients who had remained free from MV at 48 h. Only one patient who failed within 48 h was successfully liberated by day 7. No other causes besides respiratory failure were identified as the reason for reconnection to MV. The severity of illness (SOFA) score was greater in patients who failed liberation. No significant difference in MV duration before the first liberation attempt was observed between patients with and without failure within 48 h. Compared with intubated patients, tracheostomized patients had a greater probability of liberation failure within 48 h (9, 56.2% vs. 7, 21.2%). For a better understanding of the study population, Supplementary Table S1 highlights and cross-tabulates the characteristics of patients across the two groups: intubated and tracheostomized.

Table 1.

Patient characteristics according to failure of liberation from mechanical ventilation within 48 h

	Liberation success(48 h) (N = 30, 65.2%)	Liberation failure(48 h) (N = 16, 34.7%)	p
Age (years), mean (DS)	61.8 (10.6)	64.2 (13.0)	0.491
BMI (kg/m2), median (IQR)	27.6 (24.0–32.7)	28.2 (24.1–32.1)	0.892
Male (N, %), mean (DS)	22 (73.3)	8 (50.0)	0.113
ICU admission SOFA score, median (IQR)	4 (4–8)	6 (4–9)	0.052
APACHE II (at LA)	12 (10–14)	14 (12–20)	0.022
Tracheostomy (N, %)	4 (13.3)	9 (56.2)	< 0.01
Comorbidities (N, %)
No comorbidity	8 (26.7)	2 (12.5)	0.511
One comorbidity	9 (30.0)	5 (31.2)
Two comorbidities	7 (23.3)	3 (18.8)
Three comorbidities	6 (20.0)	6 (37.5)
TFabs (%), median (IQR)	0.9 (0.7–1.3)	0.9 (0.5–1.0)	0.212
Reason for admission (N, %)
Trauma	2 (6.7)	1 (6.2)	0.701
Respiratory failure	5 (16.7)	2 (12.5)
Sepsis	2 (6.7)	2 (6.7)
Brain injury	1 (3.3)	2 (12.5)
COVID-19 ARDS	18 (60.0)	9 (56.2)
Other	2 (6.7)	0 (0.0)
Presence of ICUAW (N, %)	13 (43.3)	13 (81.2)	0.011
Presence of DD (N, %)	13 (43.3)	8 (50.0)	0.671
Presence of AMW (N, %)	19 (63.3)	14 (87.5)	0.081
WIND class (N, %)			< 0.011
Short weaning	20 (66.7)	3 (18.8)
Difficult weaning	7 (23.3)	3 (18.8)
Prolonged weaning	2 (6.7)	2 (12.5)
Weaning failure	1 (3.3)	8 (50.6)
Liberation from MV > 7 days	2 (6.6)	6 (38.0)	0.011
Duration of MV	5.0 (3.0–8.1)	7.0 (4.0–15.2)	0.132

Abbreviations: MV Mechanical ventilation, Liberation failure indicates failure of liberation from mechanical ventilation requiring reintubation or reconnection to mechanical ventilation, BMI Body mass index, ICU Intensive care unit, SOFA Sequential organ failure assessment, APACHE II Acute physiology and chronic health evaluation II, TF_abs the sum of the thickening fraction during cough of the rectus abdominis, transversus abdominis and internal oblique, ARDS Acute respiratory distress syndrome, ICUAW Intensive care unit acquired weakness, DD Diaphragm dysfunction, AMW Abdominal muscle weakness, WIND Definition weaning according to a new definition, LA Liberation attempt, Duration of MV Duration of ventilation before the first liberation attempt, N Number of non-missing values. Based on their distribution and types of variables, the data are presented as the median (interquartile range, IQR), mean (standard deviation, DS) or number (percentage)

¹Pearson’s chi-square test

²Wilcoxon rank-sum test

³t-test

Figure 2 shows the distributions of AMW, ICUAW, and DD: only 9 (20.5%) patients had all three conditions simultaneously; 10 (22.9%) patients were affected by isolated AMW, while 8 patients (18.2%) had AMW combined with ICUAW, and 6 (13.6%) patients had DD. Pearson’s chi-square test did not show a significant association between AMW and ICUAW (p = 0.28) or DD (p = 0.97) (Supplementary Table S2).

Fig. 2.

Venn diagram illustrating the prevalence of AMW based on the presence of ICUAW, DD, or both. Abbreviations: AMW, abdominal muscles acquired weakness; ICUAW, intensive care unit acquired weakness; DD, diaphragm dysfunction

Abdominal muscle weakness vs. no abdominal muscle weakness

TF_abs did not significantly differ by liberation failure within 48 h after the first liberation attempt, as they were 0.9 [0.8–1.3] for liberated patients and 0.8 [0.5–1.0] for patients who failed. This trend continued within 7 days, with 0.9 [0.8–1.3] for liberated patients and 0.8 [0.5–1.0] for patients who failed.

The multivariable logistic regression analysis revealed an independent greater probability of liberation failure within 48 h (OR 9.4, CI 1.1–75.2, p = 0.03) and 7 days (OR 14.8, CI 1.6–136.1, p = 0.02) for patients with AMW compared to those without AMW (Table 2, Fig. 3). Patients with AMW had a lower but not significant risk of being liberated within 28 days (an estimated 50% lower risk, p = 0.18) (Supplementary Table S3). The presence of AMW was not associated with the WIND class, and the short-weaning group was used as the reference group (Supplementary Table S4).

Table 2.

Multivariable logistic regression analysis of failure of liberation from MV within 48 h and 7 days

	Liberation failure within 48 h			Liberation within 7 days
Predictor	Odds ratio	CI	p	Odds ratio	CI	p
Presence of AMW	9.4	1.1–75.2	0.03	14.8	1.6–136.1	0.02
Presence of DD	1.2	0.2–5.6	0.74	1.8	0.3–9.7	0.47
Presence of ICUAW	8.5	1.5–46.8	0.01	20.4	2.7–155.0	< 0.01
Age (years)	1.0	0.9–1.1	0.27	0.9	0.7–1.0	0.03
BMI (kg/m2)	0.9	0.8–1.1	0.98	1.0	0.8–1.3	0.34
Admission SOFA score	1.1	0.8–1.4	0.40	1.0	1.0–1.1	0.67

Abbreviations: MV Mechanical ventilation, Liberation failure indicates failure of liberation from mechanical ventilation requiring reintubation or reconnection to mechanical ventilation, AMW Abdominal muscle weakness, ICUAW Intensive care unit acquired weakness, DD, Diaphragm dysfunction, BMI Body mass index, SOFA Sequential organ failure assessment, CI 95% Confidence interval

Fig. 3.

The forest plot displays the results of the multivariable logistic regression analysis, showing that AMW and ICUAW are independently associated with failure of liberation from MV at 48 h (A) and 7 days (B). Abbreviations: AMW: abdominal muscle weakness; ICUAW: intensive care unit-acquired weakness; DD: diaphragm dysfunction; BMI: body mass index; SOFA: sequential organ failure assessment; CI: 95% confidence interval

ICU acquired weakness vs. no ICU acquired weakness

The presence of ICUAW was independently associated with failure of liberation from MV within 48 h (OR 8.5, CI 1.5–46.8, p = 0.01) and 7 days (OR 20.4, CI 2.7–155.0, p < 0.01) (Table 2, Fig. 3) and with a lower risk of being liberated within 28 days (estimated 77% lower risk, p < 0.01) (Supplementary Table S3). The presence of ICUAW was not related to the WIND class (Supplementary Table S4).

Diaphragm dysfunction vs. no diaphragm dysfunction

The presence of DD was not related to either 48-h or 7-day failure of liberation from MV (Table 2, Fig. 3). No significant difference in the probability of liberation (VFDs) or WIND class was found (Supplementary Tables S3 and S4).

Discussion

In this study, US measurements of abdominal muscle and ICUAW, but not TFdi, were independently associated with failure of liberation from MV within 48 h and 7 days in a selected population of patients at high risk of liberation failure who had successfully passed the SBT.

Accurately assessing readiness for liberation from MV remains a major clinical challenge. While SBT is the cornerstone of this evaluation, it does not fully capture the complexity of the liberation process, particularly regarding respiratory muscle function and cough effectiveness.

In this context, these findings suggest that abdominal-muscle US may represent a useful adjunct for assessing abdominal-muscle function in high-risk-of-liberation-failure critically ill patients. Specifically, US-defined AMW during coughing, i.e., when TF_abs is less than 127%, was associated with higher likelihood of liberation failure within 48 h and 7 days, as well as fewer VFDs at 28 days. These findings may have greater clinical implications beyond the commonly referenced failure within 48 h time point [35].

From a physiological perspective, these results highlight the role of expiratory muscles in generating effective cough and facilitating airway clearance. Adequate abdominal muscle function contributes to secretion clearance, prevention of atelectasis, and optimization of diaphragmatic function [5, 40]. As reviewed by Shi et al. [15], severe expiratory muscle weakness commonly develops in critically ill patients and is associated with worse outcomes. However, the underlying pathophysiological relationship between AMW and liberation failure remains incompletely understood, and its clinical relevance has been only partially explored [19]. US may offer a noninvasive and repeatable bedside tool to assess this component [26].

In addition, the overlap among AMW, ICUAW, and DD, and their independent associations with outcomes, were observed. Despite the high prevalence of these conditions, only nine patients presented all three simultaneously. These findings are consistent with previous literature and suggest that different muscle groups may be variably affected during critical illness [41–43].

From a clinical perspective, these data confirm the established impact of ICUAW on liberation outcomes, as it was associated with a higher probability of failure and fewer VFDs [44, 45]. Conversely, lower TFdi did not show a significant association with liberation failure. The role of diaphragm US in predicting outcomes remains debated, as some studies have reported its relationship with liberation success, whereas others have reported conflicting results [26, 38, 42, 43, 45, 46]. Importantly, the present study focused on patients at high risk of liberation failure who had already passed the SBT. In this context, diaphragm function may be less relevant in determining delayed failure. We hypothesize that failure occurring within 48 h, and particularly within 7 days, may be more closely related to abdominal and peripheral muscle function, as well as to cough effectiveness, rather than to diaphragm performance alone.

This study has several limitations. First, the relatively small sample size and the inclusion of a selected high-risk population may limit generalizability. In addition, the inclusion of tracheostomized patients, who are often more impaired and prone to liberation failure, may have contributed to the higher observed failure rates. However, when considering only intubated patients, the extubation failure rate was 20%, which is consistent with reported rates in a high-risk population.

Second, a population at high risk of muscle weakness was selected; it may have contributed to the high prevalence of reduced TFabs values (less than 100%). In the study of Schreiber et al. [26], the threshold used to define AMW was (Tfabs < 127%), which was estimated from the 2.5th percentile of TFabs in healthy subjects. In the mechanically ventilated population, a lower cough TFabs was strongly associated with a greater risk of weaning failure for every 10% decrease in cough TFabs (OR 2.1 [95% CI 1.1–4.4]). Additionally, TFabs demonstrated stronger predictive ability for liberation failure, with an area under the ROC curve of 82% [95% CI, 59–100%]. However, the prevalence of ICUAW and DD in our population is consistent with the current evidence; nevertheless, our findings suggest that the threshold of 127% might not be optimal for critically ill ventilated patients. Unfortunately, the limited sample size and the number of events did not allow for predictive analysis and the identification of a new threshold.

Third, diaphragm excursion was not assessed. Although this technique has been validated for predicting failure of liberation, it does not directly reflect diaphragm contractile function and may be overestimated under respiratory support.

Finally, ultrasound is an operator-dependent technique. However, previous studies [15, 26] have shown moderate reproducibility in assessing and monitoring the abdominal muscle activity in mechanically ventilated patients. Although a standardized acquisition and analysis protocol was used, inter-operator reliability was not formally evaluated in this study. Therefore, future studies involving larger populations are needed to confirm our results and further support the use of this easily repeatable method to assist clinicians in optimizing readiness for liberation from MV.

Conclusion

Ultrasound assessment of abdominal muscle activity during cough may help identify critically ill patients at risk of failure of liberation from MV. These findings support the clinical relevance of assessing abdominal and limb muscle function in evaluating readiness for liberation from mechanical ventilation.

Abbreviations

DD

Diaphragm dysfunction

AMW

Abdominal muscle weakness

ICUAW

Intensive Care Unit Acquired Weakness

TF_abs

The sum of the thickening fraction during cough of rectus abdominis, transversus abdominis and internal oblique

TF_di

Thickening fraction of the diaphragm

MV

Mechanical ventilation

ARDS

Acute respiratory distress syndrome

Weaning > 7 days

More than 7 days of mechanical ventilation from the first SBT to the disconnection attempt

LOS

ICU length of stay

ICU

Intensive care unit

HR

Heart rate

BP

Blood pressure

SpO2

Arterial oxygen saturation

FiO2

Inspiratory oxygen fraction

PaO2

Arterial oxygen tension

PEEP

Positive end-expiratory pressure

RR

Respiratory rate

VT

Tidal volume

RSBI

Rapid Shallow Breathing Index

BMI

Body mass index

SOFA

Sequential organ failure assessment

US

Ultrasound

Authors’ contributions

MB, SP, and NL conceptualized the study. MB, SG, FM, and LC collected the data. SR and MB conducted statistical analysis. MB and SP drafted the manuscript. All the authors critically revised the manuscript. All authors read and approved the final manuscript.

Funding

The authors declare that the study received no funding, and no external funding or sponsorship was obtained. No financial or non-financial support is claimed.

Data availability

I, Michele Bertoni, am the guarantor for this manuscript. I accept full responsibility for the integrity of the work, including the accuracy of the data and analysis. The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The local ethics committee (Fondazione IRCCS Policlinico San Matteo; https://www.sanmatteo.org) approved the study (no. 3369), and informed consent was obtained from all participants.

Consent for publication

I had full access to all the study data and have approved the final version of the manuscript for submission.

Competing interests

The authors declare no competing interests.