Open access Protocol

Can inspiratory muscle training

improve weaning outcomes in difficult
to wean patients? A protocol for a
randomised controlled trial
(IMweanT study)
Mariana Hoffman,1,2,3 Marine Van Hollebeke,2,3 Beatrix Clerckx,2,3 Johannes Muller,3
Zafeiris Louvaris,2,3 Rik Gosselink,2,3 Greet Hermans,4,5 Daniel Langer2,3

To cite: Hoffman M, Van Abstract

Hollebeke M, Clerckx B, Strengths and limitations of this study
Introduction Respiratory muscle dysfunction has
et al. Can inspiratory muscle been associated with failure to wean from mechanical
training improve weaning ►► This study will investigate the effects of inspirato-
ventilation. It has therefore been hypothesised that these ry muscle training in a specifically selected group
outcomes in difficult to wean
patients? A protocol for a patients might benefit from inspiratory muscle training of difficult-to-wean patients and is designed as a
randomised controlled trial (IMT). Evidence, however, is thus far limited to data from randomised controlled trial with a sham treatment
(IMweanT study). BMJ Open small, single-centre studies with heterogeneity in inclusion group with both patients and outcome assessors
2018;8:e021091. doi:10.1136/ criteria, training modalities and outcomes. The aim of this blinded to group allocation.
bmjopen-2017-021091 study is to evaluate the effects of a novel IMT method ►► Patients with weaning difficulties will be uniformly
►► Prepublication history and
on weaning outcomes in selected patients with weaning identified in accordance with a recently introduced
additional material for this difficulties. and highly standardised classification system for
paper are available online. To Methods This study is designed as a double-blind, weaning difficulties.
view these files, please visit parallel-group, randomised controlled superiority ►► Recruitment and monitoring will be facilitated by the
the journal online (http://​dx.​doi.​ trial with 1:1 allocation ratio. Patients with weaning use of an electronic record of ventilator manage-
org/​10.​1136/​bmjopen-​2017-​ difficulties will be randomly allocated into either an ment and weaning progression.
021091).
IMT group (intervention) or a sham-IMT group (control). ►► A novel method for respiratory muscle training (ta-
MH and MVH contributed Ninetypatients (45 in each group) will be needed to detect pered-flow resistive loading) will be applied.
equally. a 28% difference in the proportion of weaning success ►► A limitation of the study is its single-centred design,
between groups (estimated difference in primary outcome which will limit external validity of the collected data.
Received 11 December 2017 based on previous studies) with a risk for type I error
Revised 26 March 2018 (α) of 5% and statistical power (1-β) of 80%. Patients
Accepted 18 April 2018 will perform four sets of 6–10 breaths daily against an Trial status Enrolment into the study have started
external load using a tapered flow resistive loading device in August 2017. Data collection and data analysis are
(POWERbreathe KH2, HaB International, UK). Training expected to be completed in September 2021.
intensity in the intervention group will be adjusted to the Trial registration number NCT03240263.
highest tolerable load. The control group will train against
a low resistance that will not be modified during the
training period. Training will becontinued until patients Introduction
are successfully weaned or for a maximum duration of Background and rationale
28 days. Pulmonary and respiratory muscle function, Failure to wean is a complex problem with
weaning duration, duration of mechanical ventilation, significant clinical and economic impact.1
ventilator-free days and length of stay in the intensive The aetiology of weaning failure is usually
care unit will be evaluated as secondary outcomes. Χ2
multifactorial. Underlying severe respiratory
tests and analysis of covariance with adjustments for
disease with related pulmonary mechanical
baseline values of respective outcomesas covariates will
be used to compare results after the intervention period
derangements, respiratory muscle dysfunc-
between groups. tion, heart failure, metabolic and endo-
For numbered affiliations see Ethics and dissemination Ethics approval was obtained crine disorders and cognitive dysfunction
end of article. from the local ethical committee (Ethische Commissie can all contribute to weaning failure.1–3
Onderzoek UZ/KU Leuven protocol ID: S60516). Results Weaning failure can partly be attributed
Correspondence to
Dr Mariana Hoffman; from this randomised controlled trial will be presented to an imbalance between the ventilatory
​mariana_​hoffmanb@​yahoo.​ at scientific meetings as abstracts for poster or oral demand imposed on respiratory muscles
com.​br presentations and published in peerreviewed journals. and the capacity of the respiratory system

Hoffman M, et al. BMJ Open 2018;8:e021091. doi:10.1136/bmjopen-2017-021091 1

Open access

to generate sufficient driving pressure to maintain tidal muscles, is better adapted to the length–tension char-
breathing.4 5 Respiratory muscle dysfunction in mechan- acteristics of the inspiratory muscles than the isotonic
ically ventilated patients is highly prevalent and can muscle loading applied during MTL. This characteristic
result from inactivity that leads to muscle atrophy and will allow larger tidal volumes to be achieved during IMT
decreased protein synthesis, neuromyopathy, hyperinfla- for a giving external resistance.18 We hypothesise that, in
tion, metabolic disorders, reduced oxygen delivery and analogy with data presented previously in patients with
possibly hypercapnia.6 In addition to ventilator-induced chronic obstructive pulmonary disease (COPD), this
inactivity of the respiratory muscles, general impairments will result in better tolerance of higher training intensi-
in neuromuscular function also frequently develop in ties with subsequent larger improvements in respiratory
patients admitted to the intensive care unit (ICU). This muscle function in comparison to MTL.18
so-called ICU-acquired weakness contributes further to Our primary hypothesis is that high-intensity IMT will
both peripheral and respiratory muscle dysfunction.7 8 improve weaning outcomes in difficult-to-wean patients.
Weaning from mechanical ventilation should be consid-
ered as early as possible to avoid further muscle atrophy, Objective
deconditioning and complications caused by the absence This study will evaluate the effects of high-intensity inspi-
of spontaneous breathing.9 However, this is a challenging ratory muscle strength training with TFRL in comparison
process and requires considerable effort.10 with a sham, low-intensity inspiratory muscle endur-
Intermittent spontaneous breathing periods (eg, ance training on weaning outcomes in difficult-to-wean
using partially assisted modes of ventilation or sponta- patients in the ICU included immediately after weaning
neous modes) are frequently applied during mechan- difficulties have been identified.
ical ventilation and may protect the respiratory muscles
Trial design
from the deleterious effects of prolonged inactivity.11 12
This trial is designed as a parallel-group, randomised
Limited data are also available showing that specific resis-
controlled superiority trial with 1:1 allocation ratio.
tance training of the inspiratory muscles (IMT) can
Participants and outcome assessors will be blinded to
improve both respiratory muscle function and weaning
group allocation. The study was registered in a publicly
outcomes.13–15 Available data indicate that the interven-
accessible clinical trial database (​clinicaltrials.​gov identi-
tion is feasible and safe to apply in mechanically venti-
fier: NCT03240263). All procedures in this study will be
lated patients. From a recent meta-analysis summarising
performed in accordance with the ethical standards of
results from several small, single-centre studies, it
the institutional review board of the UZ/KU Leuven and
appeared that benefits were more evident in patients that
with the 1964 Helsinki declaration and its later amend-
had already failed an attempt to wean.16
ments. Ethics approval was obtained from the responsible
The studies performed so far on IMT in mechanically
local ethical committee (Ethische Commissie Onder-
ventilated patients were however heterogeneous with
zoek UZ/KU Leuven protocol ID: S60516). The protocol
regard to specific inclusion criteria, training modali-
is reported according to the Standard Protocol Items:
ties and outcomes evaluated. Not all studies specifically
Recommendations for Interventional Trials (SPIRIT)
focused on patients with known weaning difficulties and
(see checklist online supplemental material).19
not all studies evaluated weaning related outcomes. Timing
of inclusion of patients was also not consistent between
studies. Finally, most of those randomised controlled
Methods
trials (RCTs) used a mechanical threshold loading (MTL)
Patient and public involvement
device for IMT which might not offer the ideal loading
Patients were not involved in the study design or conduct
characteristics in this specific setting.16
of the study. At this stage, we have not yet decided as to
In the current project, several key elements will be
how to disseminate the results to study participants.
implemented that are hypothesised to improve the
quality of the collected data. First, the focus will specif- Study setting
ically be on patients with known weaning difficulties. The trial will be conducted on both the surgical and
These are the patients that are likely to benefit the most medical ICU in a single centre (University Hospital
from an IMT intervention during mechanical ventila- Leuven, Belgium).
tion.16 Second, IMT will be initiated as soon as weaning
difficulties have been acknowledged. The classification Eligibility criteria
of weaning difficulties will be performed in accordance Patients will be eligible if they have undergone a separa-
with the most recent clinical definitions.17 Third, our tion attempt and they have not been successfully weaned
main outcomes will be weaning success and other wean- within 24 hours of this first separation attempt. The deci-
ing-related parameters such as weaning duration. Finally, sion to start weaning and perform a separation attempt
an alternative, potentially more optimal way of loading will be made by the clinical team caring for the patient.
the respiratory muscles during the IMT sessions will be They will evaluate patients’ ‘readiness to wean’ on a daily
applied. This tapered flow resistive loading (TFRL) IMT, basis. In general, the following guidelines will be used
which is similar to isokinetic loading applied to limb when determining weaning readiness: (1) resolution

2 Hoffman M, et al. BMJ Open 2018;8:e021091. doi:10.1136/bmjopen-2017-021091

 Open access

of the acute phase of the disease for which the patient unsuccessful separation attempt (and in whom no other
was intubated, (2) adequate oxygenation (PaO2 arte- rapidly reversible cause such as fluid overload is identi-
rial oxygen—PaO2/FiO2 of 150–200 requiring positive fied by the treating physician) and who are able to follow
end-expiratory pressure (PEEP) ≤5 to 8 cmH2O and frac- simple verbal commands necessary for the training, will
tion of inspired oxygen—FiO2 ≤0.4 to 0.5), (3) absence of be considered eligible for participation in the study.
fever (temperature <38°C), (4) haemodynamic stability The following exclusion criteria have been defined:
(eg, heart frequency ≤140 bpm), (5) stable blood pressure pre-existing neuromuscular disease, haemodynamic
(BP), no or minimal vasopressors (dobutamine ≤5 μg/ instability (arrhythmia, decompensated heart failure,
kg/min, norepinephrine ≤0.1 µg/kg/min), (6) absence coronary insufficiency), haemoptysis, spinal cord injury
of myocardial ischaemia, (7) adequate haemoglobin above T8, use of any type of home mechanical ventilatory
(eg, haemoglobin >7–10 g/dL), (8) adequate mentation support prior to hospitalisation, any skeletal pathology
and (9) adequate cough.9 20 These criteria can be indi- that impairs chest wall movements such as severe kypho-
vidualised at the treating physicians’ discretion. If the scoliosis, congenital deformities or contractures, poor
overall assessment is positive, a separation attempt will be general prognosis or anticipated fatal outcome.20
performed. Separation attempts and weaning success will
be defined in accordance with a recently developed classi- Recruitment
fication system for weaning outcomes Weaning according Recruitment will be facilitated by the use of an electronic
to a New Definition (WIND).17 record (MetaVision, iMD-Soft, Needham, Massachu-
A separation attempt is defined differently for intu- setts, USA) of the patient’s SBTs and ventilation status
bated patients and tracheotomised patients. For trache- completed by nurses and the treating physician on a daily
otomised patients, a separation attempt is defined as basis. This electronic record will permit identification of
24 hours or more with spontaneous ventilation through difficult to wean patients at the earliest possible moment
a tracheostomy without any mechanical ventilation. For for inclusion according to the prespecified criteria.
intubated patients, a separation attempt is a sponta- Researchers and physiotherapists will have access to this
neous breathing trial (SBT) with or without extubation, electronic record daily to ensure that eligible patients can
or an extubation directly performed without identified be identified and informed about the possibility to partic-
SBT (planned or unplanned extubation). An SBT can ipate in the study (figure 1).
be performed either with the use of a T-tube, low-level
pressure support ventilation (≤8 cmH2O) or continuous Sequence generation, randomisation, allocation concealment
positive airway pressure (≤5 cmH2O) and has a duration mechanism and blinding
of 30–120 min.17 17 21 The SBT will be immediately inter- After providing written informed consent, patients will
rupted in case of poor tolerance. be randomly assigned to one of the two study groups:
Criteria to evaluate the outcome of the SBT are strength+endurance training group (intervention) or
(1) adequate gas exchange (SpO2 ≥85%–90%, PaO2 sham-endurance training group (control). Forty five
≥55–60 mm Hg, pH ≥7.32 and increase PaO2 ≤10 mm Hg), patients will be included in each group (see figure 2).22
(2) adequate ventilatory pattern (respiratory rate ≤30–35/ Opaque sealed envelopes will be prepared and sequen-
min, ∆ respiratory rate <50% and Shallow Breathing tially numbered by a researcher not involved in the study
Index between 60 breaths/min/L and 106 breaths/ (80 for the intervention group and 80 for the control
min/L), (3) haemodynamically stable (heart rate <120– group). Study groups will be also stratified for two factors
140/min, ∆ heart rate <20%, systolic blood pressure (BP) that are known to influence weaning success rates: score
<180–200 mm Hg and >90 mm Hg and ∆ BP <20%) and on the Acute Physiology and Chronic Health Evaluation
(4) subjective clinical signs (no changes in mental well- II scale (APACHE II) on admission and presence of COPD.
being and comfortable, no sweating, no use of accessory The four piles of 40 sequentially numbered envelopes
respiratory muscles, no paradoxical breathing). These (containing 20 intervention and 20 control each) should
criteria can be individualised by the treating physician contain a sufficient number of envelopes in each stratum.
who will decide whether or not to extubate the patient. Patients will be grouped and randomised according
If deemed appropriate in specific situations however, the to the following characteristics: Group 1, APACHE
treating physician may decide to prefer gradual reduc- score <18 and COPD; Group 2, APACHE score >18 and
tion of Pressure Support (PS) or extubation without a COPD; Group 3, APACHE score <18 and non-COPD; and
prior SBT. In case a patient had a successful SBT without Group 4, APACHE score >18 and non-COPD.22 Block
extubation, the reason why the patient was not extubated randomisation with random block sizes of 4 and 6 will be
will be recorded and the training will be continued for performed. After identifying and characterising candi-
maximum of 28 days or until successful weaning has been dates for the study, patients will be allocated into either
achieved. In general, SBTs are performed daily unless intervention or control group by opening randomisa-
contraindicated. Reasons for not performing an SBT will tion envelops from one of the four corresponding strata
be recorded in the patient’s data record. sequentially (from lowest to next highest number).
Difficult-to-wean patients, in whom no termina- Physiotherapists performing the training cannot be
tion of weaning is achieved within 24 hours after a first blinded for the treatment; however, patients, nurses,

Hoffman M, et al. BMJ Open 2018;8:e021091. doi:10.1136/bmjopen-2017-021091 3

Open access

Figure 1 Template of the electronic record (MetaVision, iMD-Soft) where information regarding eligibility criteria can be
recorded. Data regarding ventilatory mode, spontaneous breathing trial (SBT) performance, SBT failure or success, tidal
volume, fraction of inspired oxygen (FiO2), positive end-expiratory pressure (PEEP), tidal volume and pressure support value are
recorded in the system every 30 min. Continuous positive airway pressure (CPAP), pressure support (PS), assisted spontaneous
breathing (ASB), automatic tube compensation (ATC).

treating physicians and outcome assessor will be blinded on their performance provided by the training software
to the group allocation. Training will be scheduled in (BreatheLink Software, HaB International , UK)during
consultation with the treating team and nurses and physi- the training (volume and flow generated during each
cians will be asked not to enter the room during the breath and total number of inspiratory efforts) . The
training unless in case of emergency. Participants will be training sessions will be interrupted if patients report
unblinded and informed about their group allocation intolerable symptoms of dyspnoea or breathing discom-
after completion of the trial. Patients who were allocated fort, when the transcutaneous oxygen desaturation falls
to the control group will be offered strength training with below 85% or when patients starts to cough. Sessions will
the TFRL device on completion of the trial. A blinded consist of four sets of 6 to 10 breaths per set with resting
analyst will perform the statistical analyses. periods of at least 2 min between sets. Patients in the inter-
vention group will perform the training using the TFRL
Interventions
device (POWERbreathe KH2, HaB International, UK). In
Inspiratory muscle training
order to maximise work of breathing performed during
Training will be performed under full supervision of phys-
iotherapists on the ICU. The physiotherapists will monitor the training sessions, the resistance in the intervention
dose of training, amount of work performed and patient’s group will be adjusted on a daily basis to the highest
symptoms during the sessions. Patients will be placed in tolerable load. The external load will be chosen to corre-
a semiupright sitting position in bed (head board at 45 spond with the highest possible percentage of maximal
degrees), aspirated if necessary, and the tracheostomy or inspiratory mouth pressure (PImax) (typically between
tube cuff will be sufficiently inflated in order to prevent 30% and 50% PImax that will still allow inspiratory tidal
potential air leakage during IMT sessions. Before, during volume expansion to at least 70% of the patient’s inspi-
and after every session, vital parameters (respiratory rate, ratory vital capacity. In addition to the ability to perform
heart rate, oxygen haemoglobin saturation and BP) will full vital capacity inspirations, the choice of the highest
be recorded. Patients will be instructed to perform fast tolerable load will also depend on the symptom scores of
and forceful inspirations against the inspiratory resistance patients. The respiratory effort scores reported by patients
and to achieve a full inspiration and expiration at every after the training sessions should ideally range between
breath. Instructions and encouragements during the 4 and 6 on a modified Borg scale. Patients in the control
sessions will be standardised. Visual feedback will be avail- group will perform the training against an external load
able on a laptop screen to provide patients information that will be set to represent a maximum of 10% of their

4 Hoffman M, et al. BMJ Open 2018;8:e021091. doi:10.1136/bmjopen-2017-021091

 Open access

Figure 2 Flowchart of the study design. Blood pressure (BP); Heart rate (HR); Intensive care unit (ICU); Maximal inspiratory
pressure (PImax) ; Peak inspiratory flow (PIF), ; Respiratory rate (RR); Rapid Shallow Breathing index (RSBI); Haemoglobin
saturation (SpO2); Vital capacity (VC).

Hoffman M, et al. BMJ Open 2018;8:e021091. doi:10.1136/bmjopen-2017-021091 5

Open access

PImax) with no adjustments to this load during the entire mechanical ventilation (days since patient was connected
training period. Training in the control group will also to mechanical ventilator), Rapid Shallow Breathing Index
be performed with the TFRL device if the training load (breaths/min/L), VC (L), PImax (cmH2O) and ventila-
(10% PImax) is equal to or higher than 3 cmH2O. Other- tor-free days (days free from the mechanical ventilation
wise, if the training load needs to be adjusted to less than within the treatment period (from start of IMT until
3 cmH2O (ie, in case this minimal resistance corresponds maximum of 28 days after enrolment) (table 1).
to more than 10% PImax), an adapted device that is able
to provide a lower load (POWERbreathe Classic, Hab
Assessment methods
International, UK) will be used.23
PImax will be measured after endotracheal aspiration,
After each session, scores on perceived breathing
with the patient positioned in a semiupright sitting
effort and dyspnoea will be recorded with a modified
position in bed, using a unidirectional valve that will
Borg CR-10 scale.18 20 Training will be continued for a
be attached to the patient’s tracheostomy tube or endo-
maximum duration of 28 days or until the patient is success-
tracheal tube. The valve allows exhalation but occludes
fully weaned from mechanical ventilation.20 Patients still
inspiration. Patients will breathe through this valve and
on mechanical ventilation after 28 days of IMT or who die
will be encouraged to inhale as forcefully as possible
during this period will be considered ‘failure to wean’.
during each inspiratory attempt for an uninterrupted
In patients who will be transferred to other hospitals
during this period registration of the weaning process will period of 25 s.20 27 The valve is connected to a manom-
be continued. We will also make an attempt to continue eter (PFT Systems Pocket-Spiro) and to a computer
the assigned intervention if feasible. that displays pressure/time plots generated during the
Weekly measurements of PImax, forced vital capacity manoeuvres. PImax measurements will be performed
(FVC) and peak inspiratory flow (PIF) will also be used to three times with 2 min rest in between manoeuvres. The
optimise training resistance.20 highest value of the 1 second plateau pressure gener-
ated on one of the manoeuvres will be considered for
Common therapeutic interventions applied in both study arms analysis.
All patients will receive strict glycaemic control, enteral Assessments of FVC will be performed using a spirom-
feeding as soon as possible, no supplemental parenteral eter (PFT Systems Pocket-Spiro) connected to the trache-
nutrition in the first week of ICU stay24 and a standardised ostomy tube or endotracheal tube. After full inspiration
early mobilisation protocol.25 26 Sedation is titrated on the patient will be asked to perform a maximal forced
a daily basis to the minimum dose needed for patient expiration all the way to Residual Volume (RV) followed
comfort unless presence of medical conditions requires by a maximal inspiration until achievement of total lung
heavy sedation (eg, severe acute respiratory distress capacity. The manoeuvres will also be displayed in a
syndrome (ARDS), intracranial hypertension, etc.). All graphic format on the computer screen and values will
patients will follow the institutional protocol for early be recorded. Duringthe FVC manoeuvre, other variables
mobilisation entitled ‘Start to Move ASAP’.25 This protocol will be assessed and recorded: forced expiratory volume
involves passive and active mobilisation according to the in one second (FEV1), peak expiratory flow, FEV1/FVC
individual patient’s clinical status including different and inspiratory capacity. Measurements of FVC will be
postures or activities such as sitting in bed, sitting in the repeated at least three times with a 2 min rest between
chair, cycling or walking.25 the manoeuvres and the best manoeuvre (considering
the reproducibility criteria of differences between
Outcomes manoeuvres not higher than 10%) will be considered for
Primary outcome analysis.20
The primary outcome of this study is weaning success. PIF will also be assessed during the FVC maneuver and
Successful weaning is defined in intubated patients as recorded using the same spirometer. Measurements of
extubation without death or reintubation within the PImax, FVC and PIF will be performed weekly (table 1).
next 7 days whether postextubation non-invasive venti- The RSBI will be measured daily, right before the IMT
lation is used or not, or ICU discharge without invasive session during one minute of spontaneous breathing.
mechanical ventilation within 7 days. For tracheotomised The average value of RSBI at time of inclusion and before
patients, weaning is considered successful when a patient a successful separation attempt (after 24 hours for trache-
is on spontaneous ventilation through tracheostomy otomised patients and at the end of a successful SBT for
without any mechanical ventilation during 7 consecutive intubated patients) will be considered for evaluation and
days or discharged with spontaneous breathing, which- is calculated from the ratio of respiratory frequency to
ever comes first (table 1).17 tidal volume. (table 1)
Weaning duration, ventilator-free days, length of stay
Secondary outcomes and duration of mechanical ventilation will be collected
Secondary outcomes are: duration of the weaning process after weaning success by retrieving data from the patients’
(days since first separation attempt until successful electronic database record from the hospital (MetaVi-
weaning), length of stay in the ICU (days), duration of sion, iMD-Soft) (table 1).

6 Hoffman M, et al. BMJ Open 2018;8:e021091. doi:10.1136/bmjopen-2017-021091

 Open access

Table 1 Trial time schedule—overview of interventions and assessments performed during the study period.
Study period
Enrolment Allocation Postallocation Closeout
Time point Day 0 Day 0 Day 1 Day 7 Day 14 Day 21 Day 28 7 days Day x
or after SSA after SSA
Eligibility screen •
Informed consent •
Allocation •
Interventions
IMT group • • • • • •
(Daily IMT: 4×6–10 breaths)

Sham endurance • • • • • •
(Daily IMT: 4×6–10 breaths

Assessments
MIP • • • • • • •
RSBI • • • • • •

VC • • • • • • •
Weaning success •
Weaning duration •
Length of stay in ICU •
Length of stay in hospital •
• indicates performance of a measure on the day indicated. Arrows represent assessments or interventions that will be performed daily
throughout the study period.
ICU, intensive care unit; IMT, inspiratory muscle training; MIP, maximal inspiratory pressure; RSBI, Rapid Shallow Breathing Index; SSA,
successful separation attempt; VC, vital capacity.

Participant timeline file. The following information regarding patient char-

The time schedule for enrolment, interventions and acteristics will be collected: age, weight, height, gender,
assessments is presented in table 1. diagnosis, classification as difficult or prolonged weaning,
reasons for weaning failure, date of the start of the weaning
Sample size calculation process, date of the first successful separation attempt,
Based on previously observed proportions of weaning date of disconnecting mechanical ventilation, date of
success in a subgroup of patients with weaning difficulties successful weaning and date of ICU discharge. Data on
of 78% in an IMT group and 50% in a control group,16 20 training parameters will furthermore be collected before,
assuming α and β risks of 5% and 20%, a sample of 45 during and after individual IMT sessions on a daily basis.
patients in the IMT group and 45 in the control group
These include data on training volume and training
will be necessary to detect a significant difference between
intensity and tidal volumes achieved during training
groups in our primary outcome: weaning success.16 20
sessions. Borg scores on respiratory effort and dyspnoea
Data management and monitoring after training as well as cardiac frequency, BP, respiratory
The principal investigator will have access to all data of rate and oxygen haemoglobin saturation before and after
patients and training sessions and will be responsible the training sessions will also be recorded. Information
for recollection, adjustments and supervision of training on ventilator settings, and data concerning SBT including
sessions with the physiotherapists of the ICU. The prin- reasons for not performing an SBT, as well as duration
cipal investigator will also try to assess outcome measures and breathing pattern during spontaneous breathing
of included and allocated patients that refuse to continue periods will also be recorded on a daily basis. The ‘start
the training before the end of training duration. to move level’ of each patient will also be registered. This
Data regarding ventilator management will be collected score will reflect the level of general reconditioning exer-
from the individual electronic patient records (Meta- cises that each patient will have received on separate days
Vision, iMD-Soft) as illustrated in figure 1. Data will be during the intervention period.
recorded for each patient in an excel data collection sheet Details of all training sessions (pressure, volume, power
and an example is included as a online supplementary and external work) will also be recorded and saved in the

Hoffman M, et al. BMJ Open 2018;8:e021091. doi:10.1136/bmjopen-2017-021091 7

Open access

training software (BreatheLink Software, HaB Interna- will be submitted to same local Ethic Commission and
tional, UK). communicated to the trial registry. Written informed
A person responsible for data management and moni- consent will be obtained from all patients if awake and
toring will be chosen to supervise data entry during the adequate or from a family member if unconscious. The
study period prior to conduction of statistical analyses. results arising from this RCT will be presented at scien-
Electronic records of patients and training sessions will be tific meetings as abstracts for poster or oral presenta-
stored as a database in password-protected files and will tions and published in peer-reviewed journals. There is
be updated, saved and backed up regularly. Paper docu- no intention of using a professional writer, and author-
ments will be anonymised (every patient will be assigned ship will be based on the collaboration of each member
with a different code), collected and stored in individ- of the research group.
ually coded patient files. Weekly backup of data will be
performed. All researchers will have access to the results Author affiliations
1
Federal University of Minas Gerais, Rehabilitation Sciences Program, Belo
of the analyses. Data will be coded to ensure confidenti-
Horizonte, Brazil
ality and digital information will be password protected. 2
KU Leuven - University of Leuven, Department of Rehabilitation Sciences, Leuven,
Belgium
Monitoring: adverse events and withdrawal 3
Department of Intensive Care Medicine, University Hospitals Leuven, Leuven,
No serious adverse events have been reported to be Belgium
4
caused by IMT.16 To ensure safety, patients will be moni- Medical Intensive Care Unit, Department of General Internal Medicine, University
Hospitals Leuven, Leuven, Belgium
tored before and during the training sessions. Any 5
Laboratory of Intensive Care Medicine, Division of Cellular and Molecular Medicine,
adverse event occurring during the training interventions KU Leuven, Leuven, Belgium
(whether related or unrelated to the intervention) will
be registered and reported to the ICU medical staff and Contributors MH, BC, DL and RG are responsible for the overall development of an
multidisciplinary team. Decisions to withdraw patients ethically sound protocol. MH, BC, DL, RG, MVH, ZL, GH and JM are involved in the
conception and production of the study and the development of the initial protocol.
from the study will be made by the researcher blinded to All authors contributed to the drafting, critical revision and final approval of the
allocation in consultation with the multidisciplinary team document.
or, of course, by the patient himself. Reasons for with- Funding MH was sponsored by Fundação de Amparo a Pesquisa de Minas Gerais-
drawal will be recorded. FAPEMIG/Brazil grant numbers [309494/2013-3 and 442973/2014-4]. ZL is a
postdoctoral fellow of the FWO-Flanders (Fellowship number 12U5618N).
Competing interests None declared.
Statistical analysis
For data analysis, a true intention-to-treat analysis will be Patient consent Not required.
performed on the primary outcome to compare weaning Ethics approval Ethische Commissie Onderzoek UZ/KU Leuven - protocol ID:
S60516.
success between groups. Modified intention-to-treat anal-
ysis will be performed on secondary outcomes that are Provenance and peer review Not commissioned; externally peer reviewed.
not lost to follow-up. Analysis of covariance will be used Open access This is an open access article distributed in accordance with the
for secondary variables for comparison between groups Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which
permits others to distribute, remix, adapt, build upon this work non-commercially,
before and after IMT since it permits comparison of and license their derivative works on different terms, provided the original work is
outcomes between groups correcting for eventual base- properly cited and the use is non-commercial. See: http://​creativecommons.​org/​
line differences as a covariate.28 Statistical significance will licenses/​by-​nc/​4.​0/
be set at p<0.05. Kaplan-Meier curves with log rank tests © Article author(s) (or their employer(s) unless otherwise stated in the text of the
will be used to compare the time to successful extubation article) 2018. All rights reserved. No commercial use is permitted unless otherwise
expressly granted.
and liberation from mechanical ventilation. Patients who
will die or who will be transferred to another unit during
the intervention period will be censored for this analysis;
rates of deaths in both study groups will be compared as
a secondary outcome. Cox proportional hazards regres- References
1. Martin DA, Smith BK, Gabrielli A, et al. Diaphragm weakness and
sion will furthermore be used to adjust this analysis for weaning: a rehabilitation perspective. Respir Physiol Neurobiol
confounding factors that can affect MV duration. 2013;189:377–83.
2. Heunks LM, van der Hoeven JG. Clinical review: The ABC of weaning
Additional and adjusted analysis will be performed by failure-a structured approach. Crit Care 2010;14:245.
intention-to-treat analysis and per protocol analysis.29 30 3. Berger D, Bloechlinger S, von Haehling S, et al. Dysfunction of
Outcomes not assessed in postintervention will be respiratory muscles in critically ill patients on the intensive care unit.
J Cachexia Sarcopenia Muscle 2016;7:403–12.
considered as lost for follow-up. In these cases, data will 4. De Jonghe B, Bastuji-Garin S, Durand MC, et al. Respiratory
not be imputed and values will be labelled as missing data. weakness is associated with limb weakness and delayed weaning in
critical illness. Crit Care Med 2007:35:2007–15.
5. Vassilakopoulos T, Zakynthinos S, Roussos C. Respiratory muscles
and weaning failure. Eur Respir J 1996;9:2383–400.
Ethics and dissemination 6. Dres M, Goligher EC, Heunks LMA, et al. Critical illness-associated
The local Ethic Commission approved this study diaphragm weakness. Intensive Care Med 2017;43:1441–52.
7. Stevens RD, Dowdy DW, Michaels RK, et al. Neuromuscular
protocol (S60516) and the results from the RCT will be dysfunction acquired in critical illness: a systematic review. Intensive
submitted for publication. Any protocol amendments Care Med 2007;33:1876–91.

8 Hoffman M, et al. BMJ Open 2018;8:e021091. doi:10.1136/bmjopen-2017-021091

 Open access

8. De Jonghe B, Bastuji-Garin S, Sharshar T, et al. Does ICU-acquired 20. Martin AD, Smith BK, Davenport PD, et al. Inspiratory muscle
paresis lengthen weaning from mechanical ventilation? Intensive strength training improves weaning outcome in failure to wean
Care Med 2004;30:1117–21. patients: a randomized trial. Crit Care 2011;15:R84.
9. Boles JM, Bion J, Connors A, et al. Weaning from mechanical 21. Ladeira MT, Vital FM, Andriolo RB, et al. Pressure support versus
ventilation. Eur Respir J 2007;29:1033–56. T-tube for weaning from mechanical ventilation in adults. Cochrane
10. Tobin MJ. Advances in mechanical ventilation. N Engl J Med Database Syst Rev 2014:CD006056.
2001;344:1986–96. 22. Doig GS, Simpson F. Randomization and allocation concealment: a
11. Gosselink R, Langer D. Recovery from ICU-acquired weakness; do practical guide for researchers. J Crit Care 2005;20:187–91.
23. Fregni F, Imamura M, Chien HF, et al. Challenges and
not forget the respiratory muscles!. Thorax 2016;71:779–80.
recommendations for placebo controls in randomized trials in
12. Gayan-Ramirez G, Testelmans D, Maes K, et al. Intermittent
physical and rehabilitation medicine: a report of the international
spontaneous breathing protects the rat diaphragm from mechanical placebo symposium working group. Am J Phys Med Rehabil
ventilation effects. Crit Care Med 2005;33:2804–9. 2010;89:160–72.
13. Schellekens WJ, van Hees HW, Doorduin J, et al. Strategies to 24. Casaer MP, Mesotten D, Hermans G, et al. Early versus
optimize respiratory muscle function in ICU patients. Crit Care late parenteral nutrition in critically ill adults. N Engl J Med
2016;20:103. 2011;365:506–17.
14. Mohamed R, Basiouny H EL, Salem M. Response of Mechanically 25. Gosselink R, Clerckx B, Robbeets C, et al. Physiotherapy in
Ventilated Respiratory Failure Patients to Respiratory Muscles the intensive care unit. Netherlands Journal of Critical Care
Training. Med. J. Cairo Univ 2014;82:19–24. 2011;15:66–75.
15. Pascotini S, Denard C, Nunes O, et al. Respiratory muscle training 26. Gosselink R, Bott J, Johnson M, et al. Physiotherapy for adult
in patients weaning from mechanical ventilation. ABCS Health Sci patients with critical illness: recommendations of the European
2013;39:12–16. Respiratory Society and European Society of Intensive Care
16. Elkins M, Dentice R. Inspiratory muscle training facilitates weaning Medicine Task Force on Physiotherapy for Critically Ill Patients.
from mechanical ventilation among patients in the intensive care unit: Intensive Care Med 2008;34:1188–99.
a systematic review. J Physiother 2015;61:125–34. 27. Marini JJ, Smith TC, Lamb V. Estimation of inspiratory muscle
strength in mechanically ventilated patients: The measurement of
17. Béduneau G, Pham T, Schortgen F, et al. Epidemiology of Weaning
maximal inspiratory pressure. J Crit Care 1986;1:32–8.
Outcome according to a New Definition. The WIND Study. Am J
28. Vickers AJ, Altman DG. Statistics notes: Analysing controlled trials
Respir Crit Care Med 2017;195:772–83. with baseline and follow up measurements. BMJ
18. Langer D, Charususin N, Jácome C, et al. Efficacy of a Novel Method 2001;323:1123–4.
for Inspiratory Muscle Training in People With Chronic Obstructive 29. Sedgwick P. Intention to treat analysis versus per protocol analysis of
Pulmonary Disease. Phys Ther 2015;95:1264–73. trial data. BMJ 2015;350:h681.
19. Chan AW, Tetzlaff JM, Gøtzsche PC, et al. SPIRIT 2013 explanation 30. Abraha I, Cherubini A, Cozzolino F, et al. Deviation from intention to
and elaboration: guidance for protocols of clinical trials. BMJ treat analysis in randomised trials and treatment effect estimates:
2013;346:e7586. meta-epidemiological study. BMJ 2015;350:h2445.

Hoffman M, et al. BMJ Open 2018;8:e021091. doi:10.1136/bmjopen-2017-021091 9