Inspiratory Muscle Training as an Adjunct to the Treatment of Weaning Failure in Critically Ill Patients: A Practical Guide

Summary

Respiratory muscle weakness is highly prevalent in patients weaning from mechanical ventilation. Resistive inspiratory muscle training (IMT) can improve inspiratory strength and potentially weaning outcomes. This protocol provides a structured approach to offer IMT to selected patients based on the best available knowledge, facilitating its wider application in critical care.

Abstract

Inspiratory muscle weakness is common among mechanically ventilated patients and is a well-recognized contributor to weaning difficulties, which are associated with poor outcomes and increased mortality. Inspiratory muscle training is a feasible and promising intervention that can improve global respiratory muscle strength and potentially facilitate successful liberation from mechanical ventilation in patients with weaning difficulties.

This article proposes a practical guide supporting clinicians in integrating inspiratory muscle training into the care of mechanically ventilated patients experiencing weaning difficulties, informed by clinical experience and current evidence.

The proposed approach focuses on the application of inspiratory muscle training in patients with weaning difficulties (mostly tracheostomized patients undergoing prolonged weaning), who fulfill readiness-to-wean criteria and are fully cooperative.

Training sessions are typically conducted daily while the patient is disconnected from the ventilator, following an initial assessment of maximal inspiratory pressure and forced vital capacity. These parameters, along with volume displacement observed during each training session and symptom scores evaluated afterward, are used to individualize the initial external load and training progression.

The main components that are taken into account to offer and individualize the training approach include patient selection, respiratory muscle and lung function assessment, load setting modalities, breathing instructions during training, and training progression guided by symptom scores and volume displacement.

Introduction

Respiratory muscle weakness is highly prevalent in mechanically ventilated patients, occurring in up to 80% of patients at the time of weaning¹. This dysfunction is strongly associated with difficulty in liberation from mechanical ventilation, prolonged intensive care unit (ICU) stays, and increased mortality².

Resistive inspiratory muscle training (IMT) is a feasible and promising strategy to improve respiratory muscle function. Meta-analyses report average improvements of 6-7 cmH₂O in maximal inspiratory pressure (PImax) among mechanically ventilated patients^3,4, which may facilitate weaning success, particularly in patients with weaning difficulties³. However, studies focusing specifically on patients with weaning difficulties have reported heterogeneous results, likely reflecting differences in protocol design (timing, intensity, type of loading), breathing instructions, and the level of patient cooperation^5,6,7,8,9,10.

Across studies, resistive IMT has generally been performed once or twice daily, in 2-5 sets of 6-30 breaths, using external loads ranging from 10% to 50% PImax and employing pressure-threshold or (tapered) flow-resistive loading^5,6,7,8,9,10. Despite these variations, most studies demonstrate greater improvements in PImax with resistive IMT than with control conditions. Still, the magnitude of effect is inconsistent: improvements in PImax have ranged from 5 cmH₂O to 33 cmH₂O in intervention groups versus +2 cmH₂O to +15 cmH₂O in controls^5,6,7,8,9,10. Weaning outcomes showed a similar pattern, with average success rates ranging between 55-88% in resistive IMT groups compared to 22-88% in controls^6,7,9,10.

To date, only two studies have investigated the effects of comparable IMT protocols in a similar population of fully cooperative, tracheostomized patients undergoing prolonged weaning^5,10. Their results, complemented by additional physiological findings¹¹, provide the most robust evidence currently available to guide resistive IMT in this patient group. In the randomized controlled trial (RCT) by Martin et al.⁵ patients performed daily fast, deep inspirations against the highest tolerable loads (20-40% PImax) with daily progression, achieving significantly higher weaning success rates and larger gains in PImax than controls who were performing slow, deep inspirations against a sham load (<10% PImax). Van Hollebeke et al.¹⁰ found similar perceived effort, PImax improvements, and weaning success rates between patients training with fast, deep inspirations against the highest tolerable loads (30-50% PImax) and those using a sham load (<10% PImax).

A subsequent exploratory physiological study provided insight into these unexpected differences observed between the sham interventions in Van Hollebeke et al.¹⁰ and Martin et al.⁵, pointing to the different breathing instructions given to patients¹¹. It showed that the external load set on IMT devices does not fully reflect the total load imposed during IMT, as likely internal factors such as high chest wall and lung elastance, airway secretions, and airway obstruction significantly influence the total training load in prolonged weaning patients¹¹. Importantly, performing fast and deep (i.e., powerful) inspirations allowed patients to reach higher end-inspiratory lung volumes even using IMT loads <10% PImax, thereby substantially increasing the total load on the respiratory muscles¹¹. This total load was comparable to that achieved with fast and deep inspirations using IMT loads of 30% PImax and significantly higher than when slow and deep inspirations were performed against IMT loads <10% PImax¹¹. These findings highlight that specific breathing instructions can shape the training stimulus provided to prolonged weaning patients during IMT¹¹. These findings¹¹ also aligned with the similar perceived effort reported by intervention and control groups in the study by Van Hollebeke et al.^10,11, indicating that monitoring perceived effort and dyspnea with Borg scales after each session is a practical approach for clinicians to judge training intensity and progression.

In addition to protocol uncertainties and equipment availability, barriers to implementation include limited skills in PImax assessment and insufficient training in IMT¹². Taken together, these challenges may contribute to the limited clinical adoption of resistive IMT, as confirmed by a recent international survey showing that only about 60% of ICU physiotherapists routinely assess PImax or apply resistive IMT¹².

This article addresses these challenges and proposes a step-by-step protocol informed by clinical experience and the most robust current evidence to guide clinicians in the application of resistive IMT in prolonged weaning and fully cooperative patients. The protocol emphasizes patient selection, respiratory muscle and lung function assessment, load-setting strategies, standardized breathing instructions, and training progression guided by symptom scores and volume displacement.

Protocol

This protocol is clinically applied at University Hospital Leuven, and the previous studies conducted at UZ Leuven, on which it is based, received ethical approval from the Ethics Committee for Research of the University Hospitals Leuven, Belgium.

1. Selection of patients and essential assessments to set the training

1. Identify patients with weaning difficulties and likely respiratory muscle weakness.
   1. Identify invasively mechanically ventilated patients who had a first separation attempt and are not weaned within 24 h after that.
   2. Consider inspiratory muscle weakness as a potential contributing factor to weaning failure in these patients, as its incidence in this population is expected to be high^1,2. Perform an assessment of inspiratory muscle capacity to determine eligibility for inspiratory muscle training.
2. Check for full cooperation and safety criteria.
   1. Screen patients for awakening and cooperation based on the response to the following 5 standardized commands: "Open (close) your eyes", "Look at me", "Open your mouth and put out your tongue", "Nod your head", and "Raise your eyebrows when I have counted up to 5".¹³. Proceed with assessments of lung and respiratory muscle function, as well as the training, if patients respond correctly to at least 4 out of 5 commands.
      ​NOTE: If a patient responds correctly to only 3 of the standardized commands, cooperation should be considered insufficient to reliably perform the measurements. In such cases, cooperation should be reassessed later the same day or on the following day.
   2. Check if patients can follow breathing instructions by asking them to perform a deep inspiration or a series of deep inspirations while connected to the ventilator.
   3. Verify that patients have no contraindications to performing assessments and training (e.g., hemodynamic instability, high intracranial pressure, pneumothorax, positive end-expiratory pressure >10 cmH₂O, fraction of inspired oxygen >50% or respiratory rate > 35/min)¹⁰. If any doubt arises, consulting a physician about the safety of the training is warranted.
      NOTE: Safety criteria must always be checked before performing assessments or training. Cooperation may also vary daily. A summary of the indications and contraindications for performing respiratory muscle assessments and inspiratory muscle training is provided in Table 1.
3. Assessing lung function and respiratory muscle capacity
   1. Ensure availability of a portable digital spirometer and manometer (Figure 1).
   2. Prepare all necessary materials for connection of the portable device to the patient's airway opening (e.g., filters, junction pieces) to ensure proper assessment.
   3. Position the patient in a semi-recumbent position at the time of the assessments and power on the device.
      NOTE: Choose a time when the patient is rested and not immediately before or after demanding activities (e.g., sitting out of bed or spontaneous breathing trials). For safety reasons, the assessments should preferably be performed with the patient in bed. An Ambu bag connected to an oxygen flowmeter (100% fraction of oxygen) should be kept readily available in case of severe desaturation or other rare adverse events during the measurements.
4. Connect a filter to the spirometer to protect both the patient and the instrument from the transmission of pathogens.
5. Confirm that the filter's opening fits securely with the patient's airway opening (endotracheal tube or tracheostomy) without leakage. If needed, use an additional junction piece for a proper fit.
6. Inflate the tracheal cuff to a maximum of 50 cmH₂O to ensure no air leakage occurs during the assessments.
   ​NOTE: The pressure of the tracheal cuff needs to be checked and normalized to its initial value after the measurements are completed
7. Assess Forced Vital Capacity (FVC).
   1. Disconnect the patient from the ventilator and temporarily mute the ventilator alarm. Connect the spirometer to the airway opening.
   2. Instruct the patient to breathe normally first, then perform a forceful deep inspiration followed immediately by a complete and forceful exhalation to residual volume.
   3. Repeat this three times and record the best value.
      ​NOTE: If a portable spirometer is not available and a tapered flow resistive device with a monitoring system is available, consider estimating FVC during the first breaths of a training session (see NOTE at step 3.6.5 and Figure 3).
8. Assess maximal inspiratory pressure (PImax).
   1. Connect the pressure sample line of the manometer to a T-piece with a one-way valve and a junction piece (Figure 1).
   2. Explain to the patient that during the measurements, expiration will be possible, but inspiration will be occluded. Reassure the patient that the procedure is closely monitored.
   3. Disconnect the patient from the ventilator and temporarily mute the ventilator alarm.
   4. Connect the T-piece, and ask the patient to exhale completely. Instruct the patient to perform repeated forceful inspiratory efforts against the occlusion for ~20-25 s¹⁴.
   5. Encourage the patient to sustain each inspiratory effort for ideally at least 1-2 s¹⁵.
   6. Record the plateau PImax (sustained ≥1 s). Repeat at least three times, with a 2-min rest between attempts, and retain the best plateau PImax (Figure 2).
      NOTE: Some patients do not tolerate a prolonged inspiratory occlusion, which may cause psychological distress. In such cases, PImax can be alternatively assessed using a single-breath method. Using the same equipment illustrated in Figure 1, ask the patient to perform one maximal inspiratory effort against the occluded valve after a full exhalation. Repeat this maneuver at least three times and record the highest plateau PImax value. Clinicians should be aware that this approach may underestimate PImax, particularly when patients are unable to perform a full exhalation. If no portable equipment is available to assess PImax, an approximate value can be obtained using the ventilator. Apply an end-expiratory hold (with PEEP set to 0 cmH₂O) and ask the patient to perform repeated maximal inspiratory efforts against the occluded airway for about 20-25 s, or to perform three single maximal inspiratory efforts with a 2-min rest between attempts. Read the most negative airway pressure displayed on the ventilator. Because this maneuver prevents expiration below functional residual capacity, the pressure generated will likely underestimate the PImax measured from residual volume using a one-way valve. Nevertheless, this value can be used to titrate the external load for inspiratory muscle training, with the appropriate adaptations to the breathing instructions during training (see NOTE in step 3.6.4).

2. Choice of the starting IMT modality and training load

1. In patients who cannot tolerate ventilator disconnection
   1. Identify patients who cannot tolerate disconnection from mechanical ventilation due to severe dyspnea. For these patients, initiate IMT using sets of deep and fast inspirations after full exhalations (i.e., from residual volume) performed either against no load or the lowest imposable external load with a device. Provide rest periods on the ventilator in between sets.
      ​NOTE: A recent study showed that in patients with weaning difficulties, at the beginning of a training period, total inspiratory effort is similar when IMT is executed with fast and deep inspirations against either external loads of <10% PImax or 30% PImax¹¹.
2. In patients who tolerate daily unsupported spontaneous breathing
   1. If a tapered flow resistive loading (TFRL) device is available, initially use an intermediate external load of 30-40% PImax.
      NOTE: With TFRL, once the initial threshold is overcome, resistance decreases toward total lung capacity, ensuring larger volume expansion and higher inspiratory flow responses^16,17,18, which are important to target, as training effects are volume- and flow- specific^16,19,20.
   2. If a pressure threshold loading device is available, initially use an external load of 20-30% PImax.
      ​NOTE: With threshold loading, the absolute load remains constant during inspiration and consequently the relative load increases with lung volume^16,17. To allow effective training by targeting higher volume expansions and inspiratory flows^16,19,20, start with slightly lower loads (20-30% PImax), as higher loads (30-40% PImax) may limit these responses¹⁸. Although both tapered flow-resistive and pressure-threshold loading devices can be used for IMT, they have not been directly compared in clinical studies involving patients with prolonged weaning that assessed clinical outcomes such as PImax improvement or weaning success. Therefore, current evidence does not support the superiority of one modality over the other.

3. Performing IMT in patients

1. Schedule the training daily, initially at a time when patients are rested and not engaged in other demanding activities.
2. Re-evaluate safety and cooperation (step 1.2 and Table 1).
3. Explain to patients that brief periods off the ventilator are required for performing the breathing exercises.
4. Explain that they will perform sets of deep and fast inspirations after full exhalations, followed by rest periods on the ventilator between sets.
5. Perform IMT using no external load.
   1. Set the ventilator support to zero or disconnect the patient from the ventilator (if the patient is disconnected from the ventilator, temporarily mute the ventilator alarm). Instruct the patient to exhale completely before performing the first inspiration of the set.
   2. Monitor inspiratory volumes either directly on the ventilator (with the support of the ventilator set at zero) or by connecting the patient to a portable spirometer.
   3. Instruct the patient to perform 4-10 deep and fast inspirations before restoring the support of the ventilator.
      ​NOTE: If the spirometer software provides visual feedback on inspiratory volumes and/or flows (i.e., power), encourage patients to use it to maximize both.
   4. Restore the support of the ventilator to let the patient rest for approximately 2 min.
   5. Repeat steps 3.5.1-3.5.4 to perform 4-8 sets of 4-10 deep and fast inspirations, aiming for a total of 30-32 inspirations per session.
6. Perform IMT using an external loading device (from lowest imposable to higher loads).
   1. Provide each patient with single-use and/or disinfected materials to perform the training.
      ​NOTE: When using a mechanical threshold device, supply the device for personal use, together with an appropriate junction piece for connection to the airway opening if the device does not fit directly. When using an electronic tapered flow resistive device, provide a personal electronic valve and appropriate junction piece for connection to the airway opening.
   2. Explain to the patient that fast and deep inspirations will be performed against resistance based on their inspiratory muscle strength through the external device.
   3. Inflate the tracheal cuff to about 50 cmH₂O to ensure no air leakage occurs during the training.
      ​NOTE: The pressure of the tracheal cuff needs to be checked and normalized to its initial value after training is completed.
   4. Disconnect the patient from the ventilator and temporarily mute the ventilator alarm. Connect the external device to the airway opening and instruct the patient to exhale completely through the device before performing the first inspiration of the set.
      ​NOTE: If PImax was measured using the ventilator (i.e., obtained close to functional residual capacity; see NOTE at step 1.3.8), the load intensity expressed relative to PImax should be applied to inspirations initiated after a relaxed exhalation (close to functional residual capacity), rather than after a full exhalation.
   5. Instruct patients to perform 4-10 deep and fast inspirations before restoring the support of the ventilator.
      ​NOTE: If a tapered flow resistive load is used, be aware that the first breaths are performed for calibration of the device. Patients should be instructed to perform these breaths maximally from residual volume. The inspiratory volumes recorded during these initial breaths may serve as an estimate of FVC if a spirometer is not available. These volumes can then be compared to the inspiratory volumes of subsequent breaths performed against the set external load to determine whether the patient can inhale ~ 70% of the estimated FVC against the training load.
   6. Restore the support of the ventilator to let the patient rest for approximately 2 min.
   7. Repeat steps 3.6.4-3.6.6 to perform 4-8 sets of 4-10 deep and fast inspirations, aiming for a total of 30-32 inspirations per session. Figure 3 shows an example of a recorded training session.
      NOTE: If the external device software provides visual feedback on inspiratory volumes and/or flows (i.e., power), encourage patients to use it to maximize both. If needed, supplemental oxygen can be administered during training to prevent desaturation. A flow of 1-6 L/min (delivering approximately 24-44% fraction of inspired oxygen, FiO₂) may be used²¹, depending on the FiO₂ the patient requires when connected to the ventilator. Oxygen can be administered through a tube connected to the oxygen flowmeter at one end and positioned behind the valve opening at the other end. Continuous monitoring of oxygen saturation throughout the training session will ensure patient safety. For patient safety, the training session should be aborted if any of the following occur: intolerable dyspnea, oxygen saturation <85%, persistent coughing, heart rate <50 or >140 beats per minute or a >20% increase or decrease from resting heart rate, mean arterial pressure < 60 mmHg, systolic blood pressure >180 mmHg or mmHg or diastolic blood pressure >120 mmHg, or a >20% increase or decrease from resting blood pressure, development of a new arrythmia or evidence of coronary ischemia²² (Table 1). Some practical challenges may arise during IMT sessions. When increased airway secretions are present, it is good practice to aspirate them or perform airway clearance techniques before starting the training session. The same applies to interruptions caused by expectoration of secretions during training. Low inspiratory volumes or difficulty achieving consistent valve opening may be related to suboptimal load intensity or to suboptimal starting lung volume (see NOTE to steps 2.2.1 and 2.2.2). Encouraging patients to exhale fully before each inspiration can improve performance. If early fatigue occurs and the patient is not able to complete 8 inspirations consecutively, the total 30-32 inspirations may be distributed using an alternative combination of sets and inspirations (for example, 8 sets of 4 inspirations) with additional rest between sets as needed. If patients experience difficulty synchronizing full expirations with fast, deep inspirations, simple verbal cues during each breath to guide the timing of the full exhalation and the initiation of the inspiration may help. Reminding the patient to perform the inspiration continuously without pausing once initiated can further improve coordination. When available, visual feedback can also assist patients who have difficulty with timing or synchronization.

4. Daily re-evaluation of training load, modality, and progression

1. Assess inspiratory volumes during training (if a monitoring system is available with a tapered flow resistive loading device).
   1. Visually inspect breath-by-breath or average inspiratory volumes recorded during the completed training session.
   2. Determine whether the average inspiratory volumes during the completed training session were below or above 70% FVC (which can be estimated either from the first breaths performed against the lowest possible load or measured with spirometry; see step 1.3.7 and Figure 3).
      NOTE: The target of ~70% FVC was selected to ensure optimal stimulation of the inspiratory muscles across their full shortening range, as training effects are volume-specific^10,16,19.
2. Evaluate symptom scores after training.
   1. Print a modified Borg Category-Ratio-10 (CR-10) scale with anchors for assessing perceived breathing effort and dyspnea sensation²³.
   2. Ask patients, immediately after each session, to rate perceived breathing effort and dyspnea sensation during the completed session.
3. Decide to progress with training load based on daily assessments.
   1. Target Borg CR-10 scores between 4 and 6 and/or average inspiratory volumes around 70% FVC to guide the external load adjustments.
   2. Adjust the load for the next session according to the following criteria: if the Borg scores are < 4/10 and/or average inspiratory volumes were >70% FVC, introduce an external load if no load is used, or increase the load if an external load is already used; if the Borg scores are > 6/10, decrease the external load.
      NOTE: Re-evaluate load progression during the session if inspiratory volumes are larger than expected (when breath-by-breath visual feedback is available) or if the increased load is poorly tolerated, excessively limits volumes, or prevents consistent valve opening throughout the session. A summary of the practical indicators that can help guide decisions on whether to increase or decrease the external load during IMT is provided in Table 2.

5. Weekly assessments

1. Repeat lung function and inspiratory muscle capacity assessments (step 1.3) weekly to monitor progress.
   NOTE: At the beginning of the training period, weekly assessments may be very demanding for these patients, and performing both assessments and training on the same day may be challenging. If assessments and training are scheduled on the same day, perform the assessments before training.

Results

Following this protocol, patients undergoing prolonged weaning generally demonstrate improvement in inspiratory muscle strength, as assessed with weekly PImax measurements. At baseline, these patients often have PImax of 30-40 cmH₂O^5,10, which are very low values, considering the recently identified cut-off to define inspiratory muscle weakness was 62 cmH₂O for female subjects and 83 cmH₂O for male...

Discussion

Resistive IMT is a safe and feasible strategy to address respiratory muscle dysfunction in fully cooperative patients undergoing prolonged weaning. In both RCTs by Van Hollebeke et al.¹⁰ and Martin et al.⁵, no adverse events were reported during IMT sessions, and in the trial by Van Hollebeke, approximately 75% of sessions could be completed, indicating a good adherence to the training protocol¹⁰.

Notably, previous studies...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Air filters	Lemon Medical	1011-01
Connector pressure sample line to t-piece	M.E.C Belgium	MIP/MEP kit (https://mecrd.eu/products/lung-functions/pft-systems-pocket-spiro/)
Electronic training device	POWERbreathe International Ltd, UK	POWERbreathe KH2 (https://www.powerbreathe.com/product/powerbreathe-kh-series/model-kh2/)
Gas sampling line disposable – 2 m	GE Healthcare	2097307-001
Junction piece – lung function assessment	Covidien	609/5252
Junction piece - PImax assessment and execution of training	Dräeger	Elbow, reusable (8412235)
Laptop	N/A	N/A
One-way valve	M.E.C Belgium	MIP/MEP kit (https://mecrd.eu/products/lung-functions/pft-systems-pocket-spiro/)
Portable spirometer and manometer	M.E.C Belgium	Pocket-Spiro USB/BT100 (https://mecrd.eu/products/lung-functions/pft-systems-pocket-spiro/)
Printed Borg CR-10 scales	N/A	N/A
Resistive loading electronic valve	POWERbreathe International Ltd, UK	POWERbreathe K Series Valve Head (https://www.powerbreathe.com/product/powerbreathe-kh-series/model-kh2/)
Spirometry software	M.E.C Belgium	MEC Patient Data Interface (PDI) software (https://mecrd.eu/products/lung-functions/pft-systems-pocket-spiro/)
T-piece	M.E.C Belgium	MIP/MEP kit (https://mecrd.eu/products/lung-functions/pft-systems-pocket-spiro/)
Training device software package	POWERbreathe International Ltd, UK	Breathe-Link software, version 3.3.2a (https://www.powerbreathe.com/product/powerbreathe-kh-series/model-kh2/)