Factors associated with difficulties in weaning patients with tracheostomies from IMV include an imbalance between the workload and the force efficacy of the ventilatory pump, further limited by any compromise of the upper airways (for example, bulbar dysfunction, low cough efficiency, vocal cord dysfunction and airway stenosis and/or oedema).

Even brief periods of IMV may result in diaphragm weakness due to a major decline of diaphragmatic contractile force together with atrophy of diaphragm muscle fibres. IMV, sepsis and malnutrition leading to mitochondrial oxidative stress and inflammation, resulting in diaphragmatic muscle atrophy and contractile dysfunction.13 Other factors contributing to respiratory muscle dysfunction are advanced age and use of medications like corticosteroids and neuromuscular blocking agents.14 The loss of efficacy from the ventilatory pump contributes to imbalance between the work of breathing and respiratory muscle reserve.3 Therefore, the first step in weaning from tracheostomy comprises reconditioning the respiratory pump and reducing workload imposed during spontaneous breathing.8

Generally, tracheostomy tubes are placed to facilitate weaning as they do decrease dead space, are generally more comfortable for the patient and permit a gradual decrease in ventilatory support until the patient has recovered enough to maintain independent unassisted ventilation. Often this involves nocturnal ventilatory support with increasing duration of unassisted breathing during waking hours until no longer requiring assisted ventilation. However, there is a subset of patients who are not able to discontinue mechanical ventilation with this approach, in whom NIV may be able to facilitate this transition. This favourable effect is due to several physiological advantages with NIV, as it allows: reduction of inspiratory effort and work of breathing, counters intrinsic positive end-expiratory pressure, recruitment of collapsed alveoli and thereby increase dynamic compliance.15 In addition, NIV may play a decisive role in allowing decannulation, since patients can initiate NIV with the tracheostomy tube still in place and continue to receive non-invasive support during and after decannulation. As NIV applies positive airway pressure, it may help maintain upper airway patency by preventing oro-pharyngeal collapse and potentially minimizing effects of vocal cord dysfunction. However, in some patients after tracheostomy, the continuity and integrity of the trachea may be damaged, which adds extra difficulties to NIV.16

One approach is to confirm upper airway patency using bronchoscopy, after which NIV can be applied using a facial mask, with the cuff deflated and the tracheostomy capped. This will permit a check for tolerance and clinical stability, after which decannulation could be considered.17

In patients with neuromuscular respiratory weakness or failure, combined non-invasive assistance of both the ventilatory pump and forced expiratory flow is needed to wean these patients off ventilatory support. Cough augmentation techniques (CAT) may play a decisive role in secretions management and reducing weaning failure rates, especially in those patients with neuromuscular respiratory weakness and clinical conditions associated with excessive respiratory secretions. The CAT is a useful adjunct and should be applied before each period of NIV application.18

The creation and development of respiratory high-dependence care units is crucial for the management of patients with tracheostomies, especially in view of the weaning process.19 It is also important to highlight that the process of decannulation includes a multidisciplinary approach between doctors, nurses, physiotherapist and speech therapists.20

Udwadia et al.2 performed a study in ICU enrolling 22 patients with difficult weaning from IMV, 9 of them had tracheostomies. Nasal NIV was used while the cuff was deflated and the tube was occluded. Patients who tolerated NIV were transferred from the ICU to a general or high dependency ward within 24−48 h (only two patients did not tolerate NIV, but not explicitly specified whether they had a tracheostomy or endotracheal tube). In this study tracheostomies were closed in all patients who were weaned from IMV (Table 1).

Körber W et al.21 reported that 32 of 37 patients with tracheostomy who had been on long-term ventilation were successfully weaned off IMV using a volume-controlled intermittent ventilation via an individually adapted face mask. Out of 37 weaned patients, 29 were discharged with home NIV.

Schöenhofer et al.22 reported their ten-year experience, in weaning centre, on 306 tracheostomy patients, 59% with COPD, using a protocol directed T-piece weaning strategy. They applied NIV in patients with ongoing hypercapnia (PaCO2 > 50 mmHg (COPD) or >45 mmHg (restrictive disease) after 24 h without ventilatory support (after capping the tracheostomy). They identified 114 hypercapnic patients, 96 of whom were maintained off IMV and discharged home on NIV; the most common condition was COPD and the rest had either thoracic cage disorders or neuromuscular disease. No data was presented on the percent undergoing successful decannulation, but it was clear that NIV was fundamental in maintaining independent ventilation to permit discharge home.

In a similar report, Quinnell et al.23 described a nine-year experience in their weaning centre, focusing on 67 COPD patients with prolonged mechanical ventilation and tracheostomy tubes. They were able to wean 65 (96%) with 62 surviving to discharge, of which NIV was initiated in patients who were unable to tolerate unsupported spontaneous breathing trials without developing respiratory distress or hypercapnia, defined as PaCO2 > 56 mmHg. NIV was used to wean 40 and used as long term ventilatory support in 25. Although not explicitly stated, all except two were discharged without tracheostomies, with NIV provided via a nasal or orofacial mask. A one-way speaking valve was also used to facilitate communication and speech. NIV was associated with a better long-term survival.

Heinemann et al.24 analysed 117 COPD patients admitted to a weaning centre. Weaning was achieved in 82 patients (although it was not discriminated how many had tracheostomies). In patients with tracheostomies, NIV was applied according to clinical and laboratory (arterial blood gas analysis) evidence of respiratory failure. NIV was initiated in 62 patients after extubation or decannulation, and 39 patients were discharged home on NIV. However, the study did not clearly state the number/percent of patients in which NIV was necessary to achieve weaning from tracheostomy after decannulation. The overall success rate in weaning from IMV was 70.1% (no distinction made between endotracheal tube and tracheostomy). The initiation of home NIV after successful weaning was an independent prognostic factor for survival to one year.

In a retrospective observational study on 26 patients with tracheostomies, Ibrahim et al.25 proposed the use of NIV as an alternative technique to facilitate discharge of patients with tracheostomies who leave the ICU and eventually leave the hospital. NIV was connected to the tracheostomy tube, and 20 patients were discharged from the ICU and 16 from the hospital (no data on decannulation or need for home NIV).

Ceriana et al.26 reviewed their experience with patients requiring prolonged mechanical ventilation and tracheostomies. Over a 13-year period, they retrospectively analysed prospectively collected data on their cohort of patients. Out a total of 587 patients, they were able to enrol 51 (9%) into their NIV decannulation protocol, of which 46 (8%) were successfully weaned and discharged on home NIV. They identified patients as candidates for NIV weaning as those who could maintain unassisted breathing and had favourable criteria for decannulation, but with signs of inadequate ventilation defined as a PaCO2 > 50 mmHg, and/or an increase in PaCO2 ≥ 5 mmHg since suspension of IMV, and/or pH < 7.33 within seven days of spontaneous breathing. These patients underwent a protocol which included downsizing by 1 mm the cannula internal diameter without fenestration, capping of the tracheostomy cannula, and increasing periods of nocturnal NIV with the goal of achieving at least 4 consecutive hours of NIV (after two nights with good adherence they were switched to a domiciliary bilevel ventilator). After conducting an upper endoscopy to check the airways, the investigators decannulated patients who were admitted to a domiciliary NIV program. The mean time needed to protocol completion was 7.2 days. Home ventilators were set with a mean IPAP of 17.1 cmH2O and mean EPAP of 4.2 cmH2O. None of the patients required surgical closure of their tracheal stoma. Five patients failed due to claustrophobia associated with the mask and/or poor adherence to the mask. After one year, the survival rate was 82% and only one of the surviving patients was switched back to IMV. It was possible to decannulate all patients in whom NIV was successfully used for weaning from IMV.

In a one-year, prospective, multicentre study of respiratory care units in Spain, Sancho et al.27 identified 231 patients with tracheostomies out of a population of 4609, requiring prolonged mechanical ventilation. Of these 231 patients, 198 (86%) were successfully weaned and of that group, 40 (21%) needed NIV during weaning, which was applied when patients were not able to progress beyond 18 h on their spontaneous breathing trial for 5 consecutive days. The ventilator was set in pressure-support mode to achieve a tidal volume of approximately 8−10 mL/Kg. In this study, the tracheostomy stoma was either capped with a tracheal button or the tracheostomy tube was capped, with the cuff deflated and inner cannula replaced with a fenestrated tube. Bronchoscopy was performed to evaluate upper and lower airways for possible lesions. No differences were found between those with a capped tracheostomy tube or tracheal button. Treatment focused on nocturnal support and ventilatory goals included a PaCO2 < 45 mmHg and less than 5% of time with oxygen saturation under 90%. In subgroup analysis, they found obstructive sleep apnea, congestive heart failure and chronic renal failure as underlying conditions of patients requiring NIV. Prior use of home CPAP, NIV or chronic hypercapnia identified those most likely to need NIV during weaning. Baseline hypercapnia was also significantly increased with an average PaCO2 of 50 ± 10 mmHg. Although not explicitly reported, it appears that all the 40 NIV patients were successfully decannulated.

Bonnici DM et al.28 conducted a prospective study of patients with tracheostomies and neuromuscular and/or chest wall disorders, postsurgical, COPD and obesity-related respiratory failure. They found that 382% of the patients were weaned to self-ventilation. In this cohort, 24% of patients were discharged with nocturnal NIV and 1.9% of patients were discharged with both nocturnal and intermittently daytime NIV. However, numbers for patients on NIV and the decannulation success rate were not explicitly reported.

Bach et al.29 conducted a prospective study on 49 patients, including 37 with tracheostomies, with primarily neuromuscular insufficiency. When patients were clinically stable, the tracheostomy tube was switched to a fenestrated cuffed one, which was then capped, and the patients were started on NIV using mouthpiece and nasal mask while awake (during sleep they would use a nasal interface or lipseal). Mechanical insufflation-exsufflation was applied when needed. After decannulation (n = 32), 26 patients required home NIV (later, it was possible to stop NIV in 9). Decannulation was successful in 32 of the 37 patients.

Duan et al.30 performed a single center feasibility study to compare the NIV vs. a conventional strategy to accelerate weaning of patients with tracheostomies. This study enrolled 15 patients in the NIV group and 17 cases in control group. At the initial stages, all of the patients underwent conventional weaning strategies. If multiple spontaneous breathing tests failed, the patient was enrolled in this study. In the NIV group, the tracheostomy cuff was deflated, and the tube was capped. NIV was initially set to an IPAP 10 cmH2O and EPAP 4 cmH2O (titration to 8 mL/kg tidal volume), with facemask. During the application of NIV, airway management was especially important especially in patients with weak cough. If the sputum production decreased and was thin, and the ability to cough was strong, the tracheostomy tube was capped for more than 48 h. If the patient tolerated this trial, the tracheostomy tube could be removed. NIV allowed weaning from IMV in 14 patients. In this study, NIV reduced nosocomial pneumonia, shortened the days on ventilator weaning and ICU length of stay for these patients with tracheostomies. From the 14 patients weaned from IMV using NIV, decannulation was possible in 6.

In a retrospective study by Pu et al.16 conventional weaning approach in those with a tracheostomy (n = 24) was compared to a sequential invasive-noninvasive ventilation strategy (n = 26). The latter patients were started on NIV with a nasal of full-face mask, while tracheostomy was plugged and the cuff deflated. Weaning was possible in 21 patients from the sequential group (however it was not disclosed whether continued nocturnal NIV was required). PaO2 at 1 and 24 h after withdrawal of IMV was significantly higher in the sequential group. Also, the sequential group presented significantly shorter duration of IMV and incidence of ventilator-associated pneumonia, and a significantly higher rate of successful weaning. Although the decannulation success rate was not explicitly reported, it appears that all the 21 patients weaned from IMV in the NIV group were decannulated.

A retrospective study by Budweiser et al.31 evaluated the usefulness of a tracheostomy retainer during decannulation. From 384 prolonged weaning patients, a tracheostomy retainer was inserted in 166. After decannulation, NIV was applied to 135 of them, with a first decannulation success rate of nearly 72% (119 of 166), and 63 patients adapted to home NIV.