The practice of tracheostomy decannulation—a systematic review

Decannulation is an essential step towards liberating tracheostomized patients from mechanical ventilation. However, despite its perceived importance, there is no universally accepted protocol for this vital transition. Presence of an intact sensorium coordinated swallowing and protective coughing are often the minimum requirements for a successful decannulation. Objective criteria for each of these may help better the clinical judgement of decannulation. In this systematic review on decannulation, we focus attention to this important aspect of tracheostomy care. Electronic supplementary material The online version of this article (doi:10.1186/s40560-017-0234-z) contains supplementary material, which is available to authorized users.


Background
Tracheostomy is a common procedure in patients requiring prolonged mechanical ventilation (MV) and airway protection in intensive care unit (ICU) [1]. The process of weaning from tracheostomy to maintenance of spontaneous respiration and/or airway protection is termed "decannulation". This apparently simple step requires a near perfect coordination of brain, swallowing, coughing, phonation and respiratory muscles [2]. However, multifactorial aberrations in this complex interplay can result in its failure. Moreover, inappropriate assessment of the above factors increases the risk of aspiration during and after the decannulation process. Old age, obesity, poor neurological status, sepsis and tenacious secretions are the predominant reasons of failed decannulation [3].
Inability to speak with tracheostomy tube (TT) in situ results in significant anxiety and depression amongst patients [4]. More often than not, the process of decannulation is slow and prolonged leading to increased ICU stay, nosocomial infections and costs [5]. Provision of optimal tracheostomy care can help discharge these patients with TT in situ to ward, high dependency unit (HDU) and/or home. Repeat assessment and decannulation can then be performed during follow-up visits.
Several studies have emphasized the importance of decannulation within the ICU due to better and focused care compared to HDU or ward [6,7].
Inspite of the relevance and importance of decannulation, there is no universally accepted protocol for its performance. Variability in existing algorithms [8], nonrandomized study design [9] and ambiguity in the screening, technique and monitoring of decannulation limits our understanding in this important area of care. In order to better understand the various practices of tracheostomy decannulation, we performed the present systematic review of the process of decannulation.

Criteria for including studies
Case series, case-control, prospective, retrospective, randomized or non-randomized studies or surveys dealing with the process of decannulation were all included in this systematic review.

Patients
Adult patients aged above 18 years and admitted in ward, operation theater, ICU or HDU were included.

Interventions
Patients with surgical or percutaneous dilatational tracheostomy who were subjected to the process of decannulation during weaning from MV were included.

Outcome measures
Primary outcome measure assessed was success of weaning defined by a period of spontaneous breathing without having to resort to non-invasive ventilation (NIV) support or re-insertion of TT.

Identification of studies
Two independent reviewers searched the electronic database PubMed using mesh words "Tracheostomy", "Decannulation" and "Decannulation process" as title for the intervening period from 1995 to 2016 for identification of studies. The third independent reviewer then screened the two lists, removed the duplicates, and then searched for the abstracts which fulfilled the inclusion criteria. Full texts of the selected abstracts were then retrieved. Studies which further detailed the aspects of the "process of decannulation" were included. References of the included studies were further searched for any additional relevant studies not identified through our former search.

Study selection
Only studies wherein full texts were available were finally included. In case full-text article was not available for a selected study, the institutes e-library using "ERMED consortium" and/or "Clinical key" were used for free access to journals. In the event free access to full text was still not available then the authors were contacted directly for copies. English language and full-text restriction were used for inclusion of relevant studies.

Data extraction
Author (s), year of publication, country, type of study (observational, cohort, case-control, randomized and or survey), characteristics of patients, nature and severity of illness, site of care (ward, OT, HDU or ICU), method by which tracheostomy was performed [surgical or percutaneous dilatational (PCD)], length of MV prior to decannulation, criteria and method of decannulation used, outcomes in terms of success or failure of decannulation, definition of failed decannulation and limitations of study were all assessed. For completeness of data, any missing information was retrieved by directly contacting the respective authors.

Quality assessment
The methodological quality of randomized controlled studies was assessed by Jadad scale while non-randomized studies were assessed using the "Q-Coh" tool for cohort studies in systematic reviews and meta-analyses [10]. Q-Coh is a 9-point tool which incorporates the attributes of design, representativeness, and comparability of groups, exposure measures, and maintenance of comparability, outcome measures, attrition, statistical analysis and the overall assessment of each study.

Results
Our PubMed database search yielded 62 articles published between January 1995 and December 2016. Fourteen articles were excluded as they were not related to the process of decannulation. Another 9 articles from paediatrics were also excluded. The remaining 39 articles included 24 observational studies, 5 case series, 1 case report, 4 editorials and special comments, 2 questionnaires or expert opinions and 1 systematic review. There was no randomized controlled study. Six studies each were further excluded owing to non-availability of data and use of language other than English. The final number of full-text studies thus included in our analysis was 18. The step-wise selection of studies along with reasons for exclusion was as enumerated in Fig. 1. After analyzing the selected studies, we decided to perform a systematic and critical review of the existing studies on decannulation due to lack of statistical requirements for a meta-analysis.
Majority of the studied tracheostomized patients had illnesses chronic in nature [8,11]. The clinical spectrum included patients with stroke, quadriplegia, GBS, head trauma, acute exacerbations of chronic obstructive pulmonary disease, obstructive sleep apnea, restrictive lung disorder, acute respiratory distress syndrome, cardiac failure, cancer and postoperative neurosurgical, cardiothoracic and abdominal patients. The study by Kenneth B et al. specifically included critically ill obese patients with an average body mass index of 41.9 ± 14 [21]. Few studies [8,13,20,23] reported the severity of the illness of included patients.
Ten out of 18 studies did not report whether the tracheostomy was performed by surgical or percutaneous technique. There were 2 studies each with tracheostomies performed by either surgical [8] or percutaneous [20] technique, while 5 studies included patients with both techniques [13,14,19,21,25].
The duration of MV prior to decannulation was quite variable. It was lesser than 3 days in the study by Guerlain J et al. [18] to as long as 2224 days with Bach et al. [12].
Readiness to decannulate was assessed by qualitative and quantitative determinants of coughing and swallowing in different studies. Peak cough flow (PCF) [20] and maximum expiratory pressure (MEP) [8] were used as quantitative measures of coughing. Swallowing was mostly assessed subjectively via gag reflex or dye test [2], except in the study by Wranecke et al. wherein fibreoptic endoscopic evaluation of swallowing (FEES) was used for objective assessment [23].
Specific method of decannulation was mentioned in all studies except two [19,21]. Patients satisfying the criterion for decannulation were initially switched over to a smaller downsized fenestrated or non-fenestrated TT, which was later uncuffed and/or capped for a variable observation period before being finally removed. However, capping without downsizing [4,13] and abrupt TT removal was also reported [9]. While spontaneous respiratory workload post downsizing TT was monitored in most studies, Bach et al. used NIV support to decrease the breathing workload [12].
While the primary outcome in most studies was a successful decannulation, the secondary outcomes were quite variable. These secondary outcomes included survival, length of stay, prediction factors for success, and utility of a particular assessment technique [16] or a screening tool [25]. In most studies, a successful decannulation occurred when there was no need of reinsertion of TT. However, the period of observation during which re-insertion was averted varied widely from a minimum of 24 h [18] to 3-6 months [8] and/or until discharge from the unit or hospital [14,19]. The success rate of decannulation in the studies varied from as low of 23% [25] to as high as 100% [23].
The authors concluded from the studies that identification of patients ready for decannulation via objective assessment of swallowing (FEES) [16], coughing [PCF or peak [12,18] inspiratory flow (PIF)] and use of a scoring (QsQ) system [26] performed by a multidisciplinary decannulation team in ICU may prove to be more successful.
According to the Q-Coh tool [10] majority of the studies were of low quality, except the study by Ceriana et al. [8], Chaote et al. [14] and Wranecke et al. [13]. Details of all attributes of the Q-Coh tool were as depicted in the Additional file 1: Table S1.    After this systematic review, we designed a protocolized bedside decannulation algorithm for use in our ICU (Fig. 2). This protocol is being currently studied in a prospective randomized manner to assess its feasibility in adult mechanically ventilated ICU patients.

Discussion
Decannulation in tracheostomized patient is the final step towards liberation from MV. Despite its relevance, lack of a universally accepted protocol for decannulation continues to plague this vital transition. In order to focus attention on various practices of the process of tracheostomy decannulation, we decided to do this systematic review. The main finding from this review is that there is no randomized controlled study on this critical issue. Several individualized, non-comparative and non-validated decannulation protocols exist. However, a blinded randomized controlled study, either comparing protocolized and non-protocolized (usual practice) decannulation or comparing two different decannulation protocols, is urgently needed.
After ascertaining intactness of sensorium, further identification of patient's readiness to decannulate is mostly based on the assessment of coughing and swallowing. More often than not these assessments are based on subjective clinical impression of the physician who may or may not be the most experienced one at the time of decannulation. This is an avoidable lacuna in care of tracheostomized patients. Busy units and busy physicians may devote minimal time for this transition. Protocolized decannulation in our opinion may guarantee consistency and objectivity of care.
As is obvious from the studies included in our systematic review, assessments were mostly subjective, although objective FEES [16] and of coughing with PCF [12] or PIF [18] have also been attempted. Endoscopic evaluation of swallowing though technically demanding provides an objective assessment. However, studies in support of this approach are limited. Only two studies Fig. 2 Decannulation algorithm [16,23] out of 18 incorporated fibreoptic endoscopic evaluation of vocal cords and/or swallowing prior to decannulation. Warnecke T et al. in their study performed a mandatory step of FEES in their decannulation process [16]. In a recent retrospective study by Cohen et al., a three-step endoscopic confirmation of vocal cord mobility and normal supraglottis was ascertained prior to immediate decannulation [23]. He considered immediate decannulation as a safer and shorter alternative for weaning in tracheostomized patients as compared to traditional decannulation. When so many decannulations can happen without FEES, then what extra benefit does this technically demanding step offer over clinical swallowing evaluation (CSE) needs to be ascertained. Graves et al. [11] also concluded about good success rate without fibreoptic evaluation prior to decannulation of long-term MV patients. Availability and technical expertise of FEES needs to be ensured before including it in any decannulation protocol.
Similarly, subjective assessment of coughing is the usual norm. Only Bach et al. [12] in 1996, Ceriana et al. [8] in 2003, Chan LYY et al. [15] in 2010 and Guerlain J et al. [18] in 2015 used an objective measure of an effective cough to decide about decannulation. PCF, MEP and PIF are all parameters used by these investigators as measures of an effective cough. However, superiority of one over the other is undecided.
The adopted method of decannulation is also variable. While some authors preferred TT occlusion after downsizing to fenestrated or non-fenestrated tube [8,11], others straight away capped the TT without downsizing [4,13], while some abruptly removed the TT [9,14]. The choice of the method is based on patient's tolerability of the procedure and also on the physician's experience. There exists no universally accepted method. Furthermore, discrepancy also exists in the period of observation before which decannulation is deemed successful. Probably, a combination of factors like the period of MV prior to decannulation, anticipation of neuromuscular fatigue on account of respiratory workload and protection of airway all play a role.
The self-confessed limitations of the included studies were as depicted in Table 2. Specific illness group, small sample size, retrospective design, and non-standardized, non-protocolized and non-validated method of decannulation are the major limitations of the included studies. But above all, absence of a randomized controlled study in this aspect of care is a major hurdle. The previously published systematic review on tracheostomy decannulation was by Santus P et al. [26] in 2014. Our systematic review has included 10 of these studies apart from addition of another 8. While he compared primary and secondary outcomes of included studies, our review is much more exhaustive in that it incorporates the relevant details of 18 studies in a concise tabular form. Our systematic review also incorporates the Q-Coh tool [10] to assess the methodological quality of included cohort studies. As none of the studies included are of desired quality, the need for randomized controlled study on decannulation cannot be over emphasized. However, our systematic review also has several limitations. We have not searched other databases like Google Scholar, Scopus or EMBASE and also not included non-English language articles.
Our protocolized decannulation algorithm (Fig. 2) incorporates easy to use bed-side checklist for evaluation of patients deemed fit for decannulation. The screening checklist includes assessment for intactness of sensorium, characteristics of secretions and need and frequency of suctioning, effectiveness of swallowing and coughing, patency of airway and successfulness of a prolonged spontaneous breathing trial (SBT). The patient should be conscious, oriented and be able to maintain a patent airway. Secretions should be easy to handle by the patient and frequency of suctioning should be less than 4 in the previous 24 hours. The patient must be able to swallow liquids/semisolids without risk of aspiration, have adequate cough with good peak expiratory flow rate (PEFR) (>160 L/min) and be able to maintain a patent airway. Patency of the airway can be assessed bedside by simply deflating the cuff and occluding the TT with a gloved finger for testing phonation of the patient. In patients with prolonged MV of greater than 4 weeks, the duration of successful SBT should preferably be 48 hours or more. After the initial screening checklist, decision about the decannulation technique is based on the duration of MV and presence of neuromuscular weakness. Patients with less than 4 weeks of MV and with no suspicion of neuromuscular weakness are subjected to a corking trial. This trial involves blocking the existing TT after cuff deflation followed by careful instructions to the bedside nurse/physician to re-inflate the cuff in case of respiratory distress. Depending on the tolerability and absence of any distress the TT is decannulated. However, in case of a failed corking trial the TT can be downsized and blocked followed by a period of careful observation for few hours. If the observation period is not associated with any respiratory distress decannulation can then be performed. Patients who failed the corking trail as well as downsizing & blocking and are in respiratory distress need immediate upsizing of the TT to resume ventilation. Further assessment warrants a FOB examination to explore the cause of failure. In patients with MV for more than 4 weeks and with suspicion of neuromuscular weakness the decannulation technique is that of downsizing and blocking. In case of failure and respiratory distress, approach remains same as above. This protocol is currently under evaluation in our unit via a randomized study.