Complications of percutaneous and surgical tracheostomy

Question 7 from the first paper of 2004 asked the candidates to "outline the potential advantages and disadvantages of a tracheostomy in the weaning of patients from mechanical ventilation." Even though the question was apparently about the specific advantages in weaning from ventilation, the college went on to talk about the complications of the actual procedure. In fact the "disadvantages" section of the model answer is mainly composed of procedural complications, and relates minimally to the effect of tracheostomy on respiratory weaning or mechanics. Other complication-related questions included:

• Question 17 from the first paper of 2023, where the college asked a fairly pragmatic question about the assessment and management of bleeding from a tracheostomy site
• Question 14 from the first paper of 2020, where the college asked very specifically about tracheo-innominate fistula, a rare and hideous complication of tracheostomy. 

The most valuable references for this topic have been the following freely available articles:

• Epstein (2005) "Late complications of tracheostomy."
• Durbin (2005) "Early complications of tracheostomy."
• Cipriano et al (2015) "An overview of complications"...

Complications of the tracheostomy procedure

Haemorrhage

Bleeding of some degree is reported in around 5% of tracheostomies, though major haemorrhage is generally quite rare. For Question 17 from the first paper of 2023, the trainees were expected to separate the causes of post-tracheostomy bleeding into early and late (>72 hours). This seems to be a fairly arbitrary timeframe; for example Bradley (2009) chose 48 hours as the cutoff, whereas LITFL chose 72 hours (and from this, we can guess what the examiners read to get ideas for their SAQs). But that does not matter, as there is nothing that happens to a tracheostomy patient at 48 hours that is uniquely different to what happens at 72 or 96 hours,

• Early tracheostomy bleeding complications (<72 hours)
  • Bleeding from the surgical wound - which can be profuse if it is exacerbated by poor skin tension (eg. an incision which is too large for the tracheostomy tube), or by anticoagulation, or because the patient keeps moving randomly.
  • Inadvertent puncture of an anterior jugular vein. This risk can be modified somewhat by performing unskilled bedside ultrasonography of the pretracheal tissue before you go ahead with your percutaneous tracheostomy. Hatfield et al (1999) reported on a series of such ultrasound-secured tracheostomy procedures during which various little veins were identified and either avoided or ligated prior to tracheal puncture.
  • Injury to the thyroid. A fair few tracheostomies seem to puncture it (Dexter et al, 1995).
  • Injury to the inferior thyroid artery or vein would be very unlucky, but is listed by Bradley (2009) as one of the possibilites.
  • Injury to the trachea. A tracheal tear is less likely in this day of brochoscope-guided tracheostomy, but there is definitely a possibility that the diator, guidewire or subsequent suctioning has resulted in tracheal injury.
  • Haemorrhage from an unrelated pulmonary source that has absolutely nothing to do with the tracheostomy is definitely a possibility
• Late tracheostomy bleeding complications (>72 hours)
  • Perioperative site infection is apparently the most common cause
  • Tracheitis can occur as the result of frequent suctioning or as an unrelated infection
  • Granulation tissue being injured by movement and suctioning is mentioned by all resources, but has no references associated with it, suggesting that
  • Erosion of the tracheal wall which may be more common with an eccentrically placed (i.e. off-centre) tracheostomy
  • Erosion of major paratracheal vessels which also tends to happen if the tracheostomy is eccentric
  • Tracheoinnominate fistula which is the cause of bleeding in less than 1% of bleeding tracheostomy cases, but which is almost uniformly fatal

Subcutaneous emphysema

One needs to rule out a pneumothorax or pneumomediastinum as the cause. One potential reason for this is a "through and through" puncture of the trachea, with gas tracking either into paratracheal tissue or mediastinum under the effect of positive pressure ventilation. An X-ray will identify this complication, and it will require bronchoscopic exploration.
The alternative (much nicer) explanation is a brief leak of airway gas which occurs during the procedure itself, or immediately afterwards due to either  positive pressure ventilation or coughing against a tightly sutured or packed wound. The key is not to suture the wound around the tube, and to minimise the positive pressure support. Remember, you're meant to be weaning them now.  If subcutaneous emphysema happens, it is not the end of the world: just wait a few days and it will go away.

Complications associated with loss of airway and loss of respiratory pressure

During the procedure, there is the potential to either have an unsecured airway or absolutely no airway. Furthermore, there is the period during which the airway circuit is continuous with the atmosphere, and therefore depressurised. This PEEPless time will be poorly tolerated by PEEP-dependent patients, who will desaturate and possibly develop life threatening complications of hypoxia. In summary, here is a point-form list of such complications from Durbin (2005):

Complications of airway failure

• Inadvertant extubation
• Indaventant decannulation
• Loss of airway

Complications of depressurisation

• Derecruitment
• Pulmonary oedema
• Hypoxia
• Cardiac arrhythmia
• Bradycardia
• Haemodynamic instability
• Cardiac arrest and death

Complications of foreign material in the airway

• Bronchospasm
• Aspiration pneumonitis

Injury to surrounding neurovascular structures

Apart from puncturing the jugular, there are other structures around which one might prang on the way into the trachea. Among the possible victims are the vagus nerves, recurrent laryngeal nerves and carotid arteries.

Thyroid injury

One might think that surely the competent surgeon can see where the thyroid is, and concentrate their use of sharpened tools on a thyroid-free area of the operative field. Surprisingly, this is not the case. Thyroid injury seems to complicate both percutaneous and surgical tracheostomy. One may be encouraged by the knowledge that in an autopsy case series many of the percutaneous attempts did in fact skewer the thyroid safely,  without incurring any new haemorrhagic complications (Toye et al, 1986). In other cadaver studies, one in every three attempts at tracheostomy ended up passing through the isthmus of the thyroid (Dexter et al, 1995). Case reports have subsequently demonstrated that it is possible to sustain this complication and not be a cadaver (Duann et al reported on this in 2014; they safely put the tracheostomy though the thyroid isthmus and confirmed its glandular position with CT).

Complications unique to percutaneous tracheostomy

Loss, kinking, fracture or knotting of the guide wire

The use of the guide wire is unique to the percutaneous technique. The most common wire-related complication is wire loss (i.e. it comes out of the trachea accidentally). This can happen ater you have dilated, which is bad. The patient now has a huge hole in their trachea, and no convenient way to guide the tracheostomy tube into it. There are several options one can choose from in this case:

• Push the ETT back down and convert to a surgical tracheostomy
• Explore the defect with your finger and reintroduce the dilator into it, using bronchoscopy to confirm that the dilator is in the correct position (then, pass the guide wire back down into it)
• If the trachea has not been dilated yet, one may consider puncturing it again (ideally, the next ring space down) - however the chances of subcutaneous emphysema are greater, there will be more airway bleeding, and one risks damaging the thyroid.

The wire can become fractured or severed(especially if the needle is advanced back over it for some reason: so, never do that). It can be kinked by dilation, making subsequent dilation difficult (one may be able to exchange the guide wire for a better one, or work around the kink). Case reports of knotted wires do exist, but this must be freakishly rare and more related to dodgy wire construction (" Since submitting the manuscript for publication, it has been brought to the authors’ attention that certain batches of the equipment in question have been withdrawn by the manufacturer.")

Lateral stoma placement

Generally, when one aims to place the stoma one tried to put it dead-center. An exploration of the risks of placing it too laterally are difficult to find in the  published literature, but anecdotally the consequences seem to be related to uneven pressure on the tracheal wall and surrounding structures, leading to tracheal ulceration and late complications such as tracheomalacia and tracheal stenosis. Ultimately, midline is best.

Tracheal ring fracture

This might be obvious (the authors of one case report mention hearing a "crack") or non-obvious (most people never realise it has happened). The consequence is tracheomalacia in the long-term, with inspiratory collapse of the anterior tracheal wall, and stridor.

Fracture of the cricoid

Occasionally thre tracheostomy is sited in a higher than normal position(i.e. the neck was too short). A cricoid ring fracture commonly occurs in this setting (Van Heurn et al, 1996). Necrosis and degeneration of the cricoid tends to follow, as it is either devascularised by the fracture or by the direct pressure from the tube. 

Posterior tracheal wall injury

The posterior tracheal wall is vulnerable in the elderly, short-statured individuals, or patients with COPD who have a thin ectatic trachea.  A good proper posterior wall laceration can give rise to a bilateral tension pneumothorax, which is rapidly fatal. An examination of intentionally lacerated cadaveric and swine airways by Trottier et al (1999) has revealed that this complication usually occurs if you fail to stabilise the guide wire when dilating, and the guide wire ends up withdrawn back into the dilator. The result is an unguided dilator tip, riding up and down the posterior tracheal mucosa. Another popular way to lacerate the trachea is to overinflate the cuff after tracheostomy insertion.

Oesophageal injury

This is a rare complication, usually associated with percutaneous tracheostomy of morbidly obese patients for whom a large pretracheal tissue mass invites the use of excessive force and pressure when dilating. In a case series of 73 obese patients, Byhahn et al (2005) report the case of one female patient with a BMI of 50.8 who ended up sustaining a 4cm anterior oesophageal wall tear in the course of a "Blue Rhino" tracheostomy. Pneumomediastinum usually alerts the intensivist to this complication.

False passage; nontracheal tube position

This complication is virtually unknown with the surgical tracheostomy technique, because you can see the trachea and ensure that nothing but the trachea is being cannulated. The blind percutaneous technique does not offer such a safeguard. One way of preventing this complication is the use of periprocedural bronchoscopy. With the visual confirmation of dilator placement the operator can be assured that the tube is being put in the right place.

Conversion to open tracheostomy

In essence, your failure to safely complete a "minimally invasive" procedure will lead to a maximally invasive surgical repair. The patient being consented for percutaneous tracheostomy must be aware of the possibility that a difficult percutaneous procedure will be abandoned, and a surgical tracheostomy will take place.

Complications unique to surgical tracheostomy

Airway fire

We have no electrocautery available in the ICU, and so this complication is limited to the operating theater. It is also not unique to tracheostomy. Rogers et al reviewed this extensively (2001). Strategies suggested to reduce this risk traditionally include using less oxygen, but these patients are usually critically ill with hypoxia, and in any case it is apparently possible to ignite a PVC endotracheal tube at only 25% FiO₂, So, that's not much of a suggestion. One may instead want to limit the use of diathermy, as it is useless in cutting though the trachea anyway, and the soft tissue bleeding should ideally already be controlled by the time you get to opening the airway.

Early post-tracheostomy complications

Poor secretion management

The tube can be obstructed by mucus or blood clots. This is not strictly speaking a complication of early tracheostomy, but it is usually life threatening in the early stages, when the patient is still slightly delirious and unable to advocate for themselves (or self-suction). Fortunately, inner cannulae can be easily swapped out.  Good humidification is the key.

Dislodgement of a previously well-placed tube

There is little literature about this beyond case reports, and it seems obesity is a major risk factor. Morbidly obese patients are particularly at risk, because frequently the pretracheal tissues are so thick that the tracheostomy tube does not reach a desired position, and has a higher chance of getting dislodged.  El Solh et al (2007) reported a high rate of accidental decannulation followed by extratracheal tube placement in their cohort of morbidly obese patients. Ahuja et al (2013) have reported a scenario where a cuffed endotracheal tube had to be advanced through the defect when surgical emphysema prevented the re-insertion of a dislodged tracheostomy tube. Apart from morbid obesity, agitation and poor tube selection (i.e. too short) are the other most important risk  factors.

Infection, tracheitis, mediastinitis, sternal wound infection

A tracheostomy is considered a "clean-contaminated" wound, given that it is closely associated with the grossness of the phlegmy mucusy airway. In that context, it is surprising that it does not get infected more often. Only rarely does one encounter a necrotising tracheitis, which usually required reintubation and wide tracheal debridement.

A particular concern is tracheostomy with a recent sternal wound, such as following cardiac surgery. A midline sternotomy is a fragile osteomyelitis-prone wound which is positioned directly adjacent to the constant spurting phlegm geyser which is the tracheostomy. Surgeons become understandably concerned by this. Most studies of this phenomenon have confirmed the association between tracheostomy and sternal wound infection( arbitrarily selected, one such study is by Sun et al, 2013).

Swallowing difficulty

The cuff which is inflated in the trachea pushes posteriorly on the oesophagus, which prevents the passage of food boluses. On top of that, the tracheostomy tethers the trachea into a specific position in te neck, which prevents the normal upward movement of the larynx during swallowing. One might deflate the cuff, but this somewhat defeats the purpose of a defended airway. Overall, nasogastric or PEG feeding is often the most convenient and reliable route of nutrition. Remember, you are trying to wean these people- that's not going to happen if they are malnourished from your obstinate persistence with oral diet.

Ulceration of the tracheal mucosa

This is the complication which probably goes on to promote tracheal stenosis (Gibson, 1967), and it is something which can happen in the early post-tracheostomy period. Apart from an awkwardly place tracheostomy (eg. lateral stoma) the next most likely reason for this is an overinflated tube cuff; hence the cuff pressure should be monitored carefully and kept under 30 mmHg.  That said, tracheal ulceration seems to be fairly common among patients with tracheostomy ( (Van Heurn's autopsy series in 1996 revealed ulceration of some sort or another in eleven of the twelve deceased patients) and tracheal stenosis is much less common, suggesting that the association must be weak.

Complications of long-term tracheostomy

Tracheal stenosis

This usually occurs at the level of the stoma. Nearly all patients end up with some degree of narrowing at that level, but for most  it is of no respiratory significance. Risk factors for a significant stenosis include sepsis, stomal infection, hypotension, advanced age, male sex, steroids, tight-fitting or oversized cannula, excessive tube motion (ie. mechanical irritation) and prolonged placement (Epstein et al, 2005).

Stenosis below the stoma is usually due to tracheal ulceration, and may be associated with the cuff being inflated to a stupidly high pressure. But even a normally pressurised cuff can cause ulceration over a prolonged stay in hospital. Stenosis above the stoma is usually due to posterior tracheal lacerations. Additionally, the brutal use of dilators can cause fractured tracheal cartilages to invaginate into the lumen, causing obstruction.

Tracheomalacia

This is a weakening of the tracheal wall, which results from ischaemic injury to the trachea followed by
chondritis and necrosis of tracheal cartilage. In the absence of cartilage, the trachea is a fibrous tube, limp and useless. Under negative pressure from the respiratory muscles it will collapse, resulting in stridor.

Tracheo-innominate artery fistula

This is a topic which, up until the first paper of 2020, the exam candidates could have safely put into the "too rare to be asked about" basked. And then, Question 14 from the first paper of 2020 asked a whoel ten points worth of questions about it. Contributing factors, clinical features and management were the three parts of this question. Thus, much more attention is devoted to it here then would be normally expected for a complication which has an incidence of 1% (Grant et al, 2006). If one needs even more detail than this, one may be able to read about it in Allan et al (2003), if one can somehow get a hold of the full-text article.

What is it? Well. As the name suggests, this is an abnormal passage which forms between the trachea and the innominate artery. As one might imagine, as the consequence,

• blood from the greater vessels is able to make its way into the trachea, or (less likely)
• air from the trachea is able to make its way into the innominate artery

Neither of those possibilities is associated with a long and healthy life. How does this happen? The innominate artery lies adjacent to the trachea and crosses that structure at approximately the 9th tracheal ring. Ergo, a very low tracheostomy tube stands the greatest risk of causing this complication. "The pulsating innominate pounds against the unyielding tracheostomy tube with resultant ischemic necrosis of the tracheal and arterial walls and subsequent rupture", Chew & Cantrell intone poetically (1972).  The best illustration of the ensuing chaos can be found in an old textbook (Wolfe, WG. Complications in thoracic surgery. 1992).  In keeping with the spirit of Deranged Physiology, the image was not only reproduced without permission, but also childishly recoloured. 

Obviosuly, tracheostomy would not be as popular as it currently is, if this were a commonplace occurrence. As meantioned above, it is freakishly rare: 1% is a generous estimate. Goldenberg et al (2000) reported nine case of such fistulas developing in their 1130-patient series, observed over ten years in a busy ENT service. Chew & Cantrell (1972) reported approximately 0.2-0.3%.  If it is going to happen, it will happen within the first 3-4 weeks (peak incidence 7-14 days).

Unsurprisingly, mortality approaches 100%. Karl Shlaepfer, reporting on this in a primitive form of JAMA in 1924,  describes the events as sudden, without any preceding symptoms or warning.

"On the seventeenth day, six days after the permanent removal of the cannula, while the patient was sleeping soundly, a profuse hemorrhage from the mouth and nose occurred, causing sudden death... The wound in the trachea communicated with an opening the size of a pinhead in the innominate artery at the place where the artery crosses the trachea obliquely and where there is a relatively intimate connection between the two structures"

So, under the "clinical features" section below, one could easily put "sudden horrible death".

Contributing factors to the formation of tracho-innominate fistula: 

• A particularly low tracheostomy
• Surgical tracheostomy where excessive dissection interupts the blood supply
• Mediastinal tracheostomy
• Excessive tension on the tracheostomy (eg. where the tube selected is too short for the pretrachea lissues)
• Unusual patient anatomy or behaviour, eg. an abnormal ectatic or tortous trachea, or unusual patient posturing with frequenty extreme neck extension (e
• Hyperextension of the neck during tracheostomy:  marked retraction of the head brings the innominate vessels closer to the level of the suprasternal notch
• A propensity towards poor wound healing: 
  • Prolonged episodes of hypotension
  • Extensive use of vasopressive agents
  • Radiation therapy to the neck
  • Steroid therapy
  • Protein malnutrition 
  • Localised infection at the insertion site
• Problems with the tracheostomy device:
  • Overinflated high pressure cuff
  • A tracheostomy tube which is excessively curved anteriorly (eg. a fixed 90 degree bend), or one which is tied or sutured in a way which angles it anteriorly.

Clinical features of tracheo-innominate fistula

• Airway bleeding is clearly the main feature. 
• A pulsatile tracheostomy tube is occasionally reported
• Infection around the site is an associated feature

Management of tracheo-innominate fistula

• First, overinflate the tracheostomy cuff. In their 2006 review, Grant et al  recommended overinflating the tracheostomy cuff as the optimal first line management, while waiting for the cardiac surgical team to prepare for theater. 
• Next, as the bleeding should be controlled well enough for you to set up for it, intubate the patient orally.  It may be necessary to do this in an emergency anyway, particularly if the blood is coming from the tracheostomy itself, rather than around it. To intubate the patient from above prevents soiling of the lower respiratory tract. Somebody should remove the tracheostomy just as you're advancing the ETT, i.e. at the very last minute. 
• Next, gain control of the haemorrhage by digital compression. This "Utley manoeuvre" (Utley et al, 1972) requires you to stick your index finger into the tracheal stoma, blunt-dissect down along the tracheal wall to separate the artery and the trachea, and put pressure on the innominate artery, compressing it up against the posterior wall of the manubrium. Just like this:
  
• Obviously, this is not going to be a particularly long-term solution. The person with their finger in the patient's mediastinum should get themselves into a comfortable position. They will remain in place until theatres are ready, and then they will escort the patient there. A full median sternotomy will usually follow. Most often, repair is not attempted, and the innominate artery is ligated and resected where it overlies the trachea, so that the arm is now perfused by retrograde flow from the right carotid.

Tracheo-oesophageal fistula

Like the other fistula, this one happens in less than 1% and results from injury to the posterior tracheal wall, or from persistently high cuff pressure more associated with the bygone era of low-volume high-pressure cuffs (Flege, 1967). Eventually, ulceration and erosion leads to the formation of a passage between the trachea and the oesophagus. The presence of a nasogastric tube probably helps this along; it acts as another hard object for the posterior tracheal mucosa to be sandwiched against. One may be alerted to this complication by the presence of undigested food in the trachea.

Persistent stoma

After decannulation, a stoma is supposed to close in a manner analogous to the tract of a pierced ear from whence the ring has been removed. In a minoroty of patients, this does not happen within the normal timeframe. It is not merely annoying - speech and respiration will also be affected. Surgical repair is usually required.

Sternoclavicular osteomyelitis

This is a rare complication, usually associated with an infected tracheostomy site or a displaced tracheostomy tube. In one case, Hashimi et al (2011) report on a sternoclavicular osteomyelitis which resulted from a a misplaced tracheostomy tube that remained in the pretracheal space for 10 days. So, I suppose as long as you don't do something like that, you should be free from this complication.

Pneumonia

A trachestomy is persistently colonised with bacteria, and the tube offers safe passage into the lower airways for these bacteria, bypassing the mucosal defences of the upper airways. This makes the patient more susceptible to pneumonia. An important consideration in whether or not to offer tracheostomy is the impression of how the patient might fare with a hospital-acquired pneumonia. In essence, one would think twice about offering it to the patient whose severe lung disease would not permit recovery from a nasty pneumonia. How likeluy is this complication?  Georges et al (2000) found a post-tracheostomy pneumonia rate of around 26%, of which about half occurred within 5 days of the tracheostomy (presumably, while the patient is still critically ill and defenseless).

Aspiration

Question 7 from the first paper of 2001 asked about aspiration in a patient with tracheostomy, and how one might go about detecting that. A detailed answer was expected, and this complication is treated here with the respect deserving of an SAQ which has not been repeated for over fifteen years.

Suitably elderly studies are offered as references for this question (Elpern et al, 1994; Bone et al, 1974). In short, the causes of aspiration may numerous. One must identify which of the following problems is present:

• Simple feed intolerance with regurgitation
• Impaired swallowing of oral contents
• failure of the tracheostomy cuff to maintain a sealed airway (i.e. cuff is failing to maintain pressure, or the whole tracheostomy has migrated out of the stoma and there is nothing in the trachea)
• Tracheo-oesophageal fistula.

One may wish to go through the following sequence:

• Ensure the airway equipment is not at fault:
  • Check tracheostomy position on CXR
  • Check the cuff for absence of leak
  • Check the seal of the cuff for presence of air leak 
• Ensure the feeding tube is not at fault:
  • Check NGT position on Xray
  • Ensure that the NGT is not fractured and leaking into the oral cavity
• Ensure that there is no feed intolerance
  • Check gastric residual volumes
  • Consider changing to a post-pyloric tube
• Confirm that aspiration of feeds is taking place
  • This is a step which may be omitted.
  • Historically, the "blue dye test" had been used. Evan's Blue had been mixed with the NG feeds, and the tracheal aspirates were observed for blueness. This test is far from reliable, and has been largely abandoned as it neither identifies nor excludes all aspiration, and may give a false sense of security
• Investigate oesophageal intergrity, motility, and swallow coordination: the gold standard is videofluoroscopy, or the modified barium swallow.