Failure of standard tracheostomy decannulation criteria to detect suprastomal pathology

Summary

Meeting established criteria for tracheostomy decannulation should improve success, although there will be a small proportion of patients with a tracheostomy who fail decannulation. Failure rates depend on patient characteristics and disparity between institutional practices and expert opinion. However, there are no widely accepted published failure rates, or agreement on the time‐point at which failure is assessed. We present a patient who had evidence of readiness for decannulation, but had immediate failure due to extrinsic tracheal compression, which proved difficult to diagnose and required surgery to resolve. Capping the tracheostomy before decannulation may or may not have given rise to suspicion of potential failure and this practice requires further evaluation as it is not without risk. For subglottic, but suprastomal lesions, nasendoscopy is not of value. It is important to decannulate patients in a safe environment, preferably early in the day to allow post decannulation observations and interventions should they become necessary, and with the close involvement of the multi‐professional team. This report illustrates the failure of our standard Tracheostomy decannulation criteria, and direct upper airway view to identify suprastomal tracheal pathology, and we discuss the potential for additional criteria which may have identified the issue before decannulation attempts.

Introduction

Approximately 17,000 tracheostomies are performed annually in the UK [1, 2] for a variety of reasons, including upper airway obstruction; pulmonary hygiene; prolonged tracheal intubation; and to aid weaning from mechanical ventilation [3]. Tracheostomy decannulation is the final step in a process of tracheostomy weaning, which involves the re‐establishment of glottic airflow and subjective assessment of competent oral and pulmonary secretion management. The proportion of patients with a tracheostomy who proceed to decannulation is variable and largely unknown. In our institution, approximately 60% of our tracheostomy patients proceed to decannulation during their inpatient stay. Standardised multi‐professional decannulation criteria allow us to determine decannulation readiness with a relatively low rate of failure. In the following case report one of these failures is analysed.

Report

A 66‐year‐old woman was admitted for craniotomy and debulking of a frontal meningioma. The patient was extubated uneventfully following surgery but then developed recurrent supraventricular tachycardia requiring pharmacological cardioversion, cranial diabetes insipidus and a chest infection. At 22 days following surgery, she required re‐intubation (size 7.0 mm I.D.; grade 2 Cormack and Lehane view with a gum elastic bougie being required) and was transferred to the intensive care unit (ICU) for management of respiratory failure. An embolus of the right pulmonary artery was noted in a subsequent scan.

Four days following re‐intubation, the patient had a percutaneous tracheostomy (size 8.0 TRACOE^® twist, TRACOE medical GmbH, Nieder‐Olm, Germany) and underwent weaning from ventilation over the course of 4 weeks. She experienced progressive periods of in‐line speaking valve and tolerated cuff deflation/speaking valve during ventilator‐free periods. Following liberation from ventilation the patient tolerated cuff deflation/speaking valve without difficulty.

Following discharge from ICU, the patient re‐commenced an oral diet, and met our criteria for decannulation. Fibreoptic endoscopic evaluation of swallow was completed to assess her swallow, given reports of globus. The fibreoptic endoscopic evaluation of swallow showed no dysphagia or upper airway abnormality. It was hypothesised that the globus was due to the presence of either the nasogastric tube or the tracheostomy.

Immediately following decannulation, the patient had an audible wheeze, progressive stridor, increased work of breathing, tachycardia and oxygen desaturation to < 80%, despite delivery of 15 l.min^‐1 oxygen. A size 7.0 TRACOE^® Twist tube was inserted which alleviated the adverse respiratory signs.

Following the failed decannulation a nasendoscopy demonstrated normal glottic opening and upper airway. The tracheostomy could not be viewed through the vocal cords but some crusting/granulation was noted. The assessment could not view the suprastomal/subglottic region therefore microlaryngoscopy under general anaesthesia was arranged. This revealed a normal glottis and immediate subglottic area with minimal granulation, but severe lateral suprastomal tracheal collapse was observed obliterating the trachea distal to the cricoid bone along approximately three tracheal rings. Steroid injection and three balloon dilatation attempts failed to improve the collapse. Supraglottic jet ventilation allowing Hopkins rod inspection of the distal trachea revealed no abnormality distal to tracheostomy stoma.

The diagnostic dilemma required a second microlaryngoscopy which confirmed collapse of the right suprastomal tracheal wall to the left. Attempts to laser the right tracheal wall failed to create an adequate airway and extrinsic compression was queried.

Ultrasound revealed a large multinodular, but not clinically palpable, thyroid goitre. Subsequent Computed Tomography scan of the neck (Fig. 1) confirmed a goitre with multiple foci of calcification. There was asymmetrical enlargement of the right lobe, with tracheal deviation to the left and narrowing of the tracheal lumen. The narrowest luminal area measured approximately 20 mm² with a slit‐like airway above the tracheostomy. The enlarged right thyroid lobe extended to the thoracic inlet but without significant retrosternal extension. The thyroid tissue extended into the right tracheo‐oesophageal groove and distorted the right postero‐lateral tracheal wall.

Figure 1.

(a) Coronal and (b) sagittal plane computed tomography scans demonstrating the suprastomal goitre.

A total thyroidectomy and tracheal resection was performed and this allowed decannulation intra‐operatively and discharge home 7 days later.

Discussion

There is currently no universally accepted method of establishing decannulation readiness [4], but it is determined by expert opinion and individual institutional protocols [5]. The published literature which supports this clinically important decision consists of observational studies, case series and questionnaires, but is not supported by randomised controlled trials [4, 6].

In our institution, the multi‐professional decision to decannulate is determined formally by the following factors:

1. The initial reason for the tracheostomy placement should have resolved

2. The patient should be able to manage oral secretions, as determined by tolerance to cuff deflation over a 24‐h period

3. The patient should be able to generate glottic flow and produce sounds through speaking valve tolerance for a minimum of 4 h

4. The patient should demonstrate an ability to manage pulmonary secretions determined by subjective and objective measures of cough strength.

Downsizing to a smaller unfenestrated tracheostomy tube often allows step 3 to occur, but ‘capping’ or ‘corking’, which involves completely obstructing the tracheostomy tube has been discontinued in our practice for at least a decade. Whilst we have the availability of both fibreoptic endoscopic evaluation of swallow and nasendoscopy, these assessments are only requested when there is a clinical concern. The majority of our decannulations occur without the need for additional assessments.

In the case presented, our institutional decannulation criteria failed to identify the presence of suprastomal tracheal pathology. The patient was fully tolerant to cuff deflation which indicated an ability to manage oral secretions. The patient also demonstrated excellent voice quality when using a one‐way speaking valve, which indicated that the suprastomal tracheal segment had sufficient space to allow airflow through the glottis during expiration. Indeed, the fibreoptic endoscopic evaluation of swallow view of the glottis before decannulation revealed normal vocal cord movement. We presume that the tracheostomy tube itself was splinting the suprastomal tracheal segment against the opposing extrinsic compressive force of the enlarged thyroid gland. The speaking valve may have also contributed to an increase in positive subglottic tracheal pressure [7] which, in combination with the tracheostomy tube splinting effect, elicited sufficient expiratory glottic airflow to mask potential complications.

It is difficult to know whether ‘capping’ or ‘corking’ this particular tracheostomy before the decannulation event may have revealed the potential for failure. The practice of capping is adopted by some hospitals and ear, nose and throat specialists as an essential stage in the journey towards decannulation [8, 9, 10]. It could be argued that capping may specifically indicate the integrity of the suprastomal tracheal segment during inspiration. The cap prevents tracheostomy tube flow during both phases of the respiratory cycle and inspiratory flow through the suprastomal airway must be increased, compared with inspiratory flow, when using the one‐way speaking valve. It is also reasonable to suppose that the sub‐glottic negative pressures generated during inspiration may be sufficient to test the structural integrity of the suprastomal trachea. If the suprastomal airway lacked integrity, inspiratory pressure changes may have decreased the tracheal diameter and generated stridor during a capping trial.

It is also difficult to appreciate whether nasendoscopy should have been performed before the initial decannulation event. There was no clinical indication that the glottis and upper airway were compromised because the patient was producing a clear unimpeded voice. The fibreoptic endoscopic evaluation of swallow found normal appearances of the vocal cords and unremarkable supraglottic anatomy. Although the nasendoscopy performed after failed decannulation reported that the tracheostomy tube could not be viewed through the vocal cords, this is not an uncommon nasendoscopic finding in a procedure which uses a short endoscope. In addition, viewing the tracheostomy tube through the cords will be determined by the tracheostomy position, presence of glottic secretions and ability to deeply insert the nasendoscope without causing too much patient discomfort. This report demonstrates that traditional criteria for determining suitability for tracheostomy decannulation may have pitfalls, highlighting the need for clinicians to be conscious of potential failure.