Laryngeal injury following endotracheal intubation: Have you considered reflux?

Abstract

Laryngotracheal injury is an increasingly common complication of intubation and mechanical ventilation, with an estimated 87% of intubated and ventilated patients developing a laryngotracheal injury often preventing their rehabilitation from acute illness. Laryngotracheal injuries encompass a diverse set of pathologies including inflammation and oedema in addition to vocal cord ulceration and paralysis, granuloma, stenosis, and scarring. The existing literature has identified several factors including intubation duration, endotracheal tube size, type and cuff pressures, and technical factors including the skill and experience of the endoscopist. Despite these associations, a key aspect in the sequelae of laryngotracheal injuries is due to reflux and is not clearly related to iatrogenic and mechanical factors.

Laryngopharyngeal reflux is a type of reflux that contaminates the upper aerodigestive tract. The combination of patient positioning and continuous nasogastric tube feeding act to affect the upper aerodigestive tract with acidic and non-acidic refluxate that causes direct and indirect mucosal injury impeding healing.

Despite laryngopharyngeal reflux being an established and recognised causative factor of upper aerodigestive tract inflammatory pathology and laryngotracheal injury, it is very understudied in critical care. Further, there is yet to be an agreed pathway to assess, manage and prevent laryngotracheal injury in intubated and ventilated patients. The incidence of laryngopharyngeal reflux in the intubated and mechanically ventilated patient in the intensive care unit is currently unknown. Prospective studies may allow us to understand further potential mechanisms of upper aerodigestive tract injury due to laryngopharyngeal reflux and herald the development of preventative and management strategies of laryngopharyngeal reflux–mediated upper aerodigestive tract injury in critically ill patients.

Background

Laryngeal injury encompasses several disorders associated with endotracheal tube (ETT) placement including inflammation, oedema, vocal cord ulceration, granuloma, paralysis, and stenosis. It is a highly prevalent iatrogenic consequence of intubation and one that is potentially preventable. Here we examine its aetiology and pathophysiology and postulate techniques to reduce iatrogenic laryngeal morbidity.

Endotracheal intubation (ETI) is a safe and often life-saving procedure that allows physicians to provide a short-term definitive airway. ¹ However, injuries following ETI are common. Laryngotracheal injury (LTI) following prolonged intubation and mechanical ventilation (over seven days) is a well-established concept, with studies as early as the 1960s recognising that tube and cuff pressure have destructive effects on the aerodigestive tract. ² Iatrogenic injury from prolonged intubation can be significant, with sequelae requiring acute and chronic medical care often extending long beyond the initial inpatient stay. ³

From the extant literature it is estimated that up to 87% of adults who require prolonged intubation in Australia suffer a laryngeal injury, with a significant proportion of these patients developing persistent functional comorbidities that prevent their rehabilitation from acute illness. ⁴ Despite this high prevalence and the significant body of evidence linking ETI and iatrogenic LTI, it remains a significantly under-investigated phenomenon.

Symptoms of LTI are both common and troublesome, they include: hoarseness, temporary voice loss, throat clearing, pain and difficulty swallowing post-extubation. ³ Postoperatively, these symptoms (although transient) are encountered almost routinely, but when injuries are encountered from typically lengthier intensive care unit (ICU) stays, they are found to be more prevalent and more serious resulting in dysphonia and dysphagia that frequently last as long as three months post-discharge.^3,5,6 Complications from such prolonged periods of intubation include, but are not limited to, vocal fold oedema, subglottic stenosis, erythema, ulceration, paralysis/paresis/bowing of the vocal cords, laryngeal scar, fibrosis, granuloma formation of glottic webs, and superior laryngeal nerve injury. ⁷ Many factors have been identified as being significant contributors to laryngeal injury; these include duration of intubation, the need for repeat intubation, ETT size, type and cuff pressures as well as technical factors including the skill and experience of the intubating physician. ⁸ Despite these associations, we postulate that a key aspect in the sequelae of laryngeal injury may be attributed to patient factors that are variable in nature and are not related to clearly identifiable iatrogenic and mechanical factors.

ETI and laryngeal injury: an all too familiar occurrence

While there is no accepted consensus, prospective studies suggest that the correlation between laryngeal injury and ETI occurs even just after a few hours. ¹ Laryngeal injury begins within the first few hours and increases with each day of intubation. ¹ It occurs most frequently at two primary sites: glottic injury from direct trauma exerted by the cuff or tube or tracheal injury due to the cuff;¹ these changes are inevitable, and occur as a result of ETT pressure exceeding mucosal capillary pressure leading to a sequelae of ischaemia, oedema and ulceration. ¹

Iatrogenic effects of intubation have been classified in several different ways. Some authors such as Lindholm ⁹ described the magnitude of injury caused to the larynx and trachea, while Santos et al. ¹⁰ described and categorised types of injury and correlated these findings with the duration of intubation. Both Colten House et al. ¹¹ and Benjamin ⁷ stratified their findings by the severity of injury, with the latter further analysing the specific anatomical site of LTI with prolonged intubation. By describing LTI in terms of both chronicity and severity, they stratified injuries into: Grade 1, self-limiting soft tissue injury; Grade 2, haematoma, ulceration and granulation; and Grade 3, stenosis, glottic narrowing and vocal cord fixation.^3,11 A systematic review of 775 patients by Brodsky et al. ³ revealed a high prevalence of laryngeal injury (83%), with dysphonia (76%), pain (76%), hoarseness (63%), and dysphagia (49%) being the most frequently reported symptoms. Further gross analysis of laryngeal injury in these patients suggested that self-limiting soft tissue injury (Grade 1) was the most prevalent (74%) with Grade 2 and Grade 3 injuries arising in 31% and 13% of patients, respectively. ³ Most significantly, longer durations of intubation may result in a greater prevalence and greater severity of laryngeal injury, with Grade 3 injuries having a 125% increase in prevalence between five and ten days of intubation compared to less than five days. ³ A prospective study examining 100 patients with prolonged intubation by Shinn et al. ⁴ revealed that over 57% of patients had acute LTI which was further associated with dyspnoea and dysphonia even at ten weeks post-extubation. The researchers found that higher body mass index, type 2 diabetes and larger ETT size correlated with greater risk. ⁴ Despite this, it is important to note that on evaluation of laryngotracheal structures in 61 patients of a wide age range and intubation period of up to 28 days, Colton House et al. ¹¹ found no statistically significant relationship between LTI, intubation type or ETT size. Despite associations with LTI, the literature currently does not support routine post-extubation visualisation of the larynx and there is no consensus on any appropriate treatment measures.

Early retrospective studies have reported LTIs as common events that resolve by three months. ⁹ These lesions are mostly vocal process granulations and laryngotracheal ulceration. ⁹ Consequently, to prevent pressure-induced lesions, ETT shape and orientation was modified to reduce pressure on structures in the posterior larynx (medial arytenoid cartilages, cricoarytenoid joints and subglottic area).^7,9 Subsequently, a low pressure high volume cuff was also developed.^7,9,12 Nevertheless, following extubation, patients experience varied sequelae of symptoms including dysphonia, salivary retention, cough and globus as well as issues with swallow, aspiration and stridor. ¹³ Despite changes in the operational aspects and design ergonomics of ETI, LTI is still a prevalent pathology, with acute and chronic complications that tend to present at the otolaryngologist’s office in the weeks and months to come.

The placement of a tracheostomy allows patients to spend less time in the ICU, less time under mechanical ventilation, and reduces the frequency of injury to the aerodigestive tract.^14,15 Yet, despite advantages in easier airway care, placement of a tracheostomy comes with laryngotracheal complications including stomal bleeding, infection and granulation, tracheal stenosis, scarring, and fistulating disease to surrounding structures.^16–20 Therefore, another question emerges; when and to whom to give a tracheostomy? The answer to this perplexing question remains elusive and despite over 20 randomised controlled trials and several robust systematic reviews, we are no closer to reaching an answer. Critical care populations have large, complex, and increasingly heterogeneous populations with several different multivariate patient and non-patient factors affecting the decision-making process and subsequent outcome. Perhaps it may be prudent to delve further into disease treatment and patient factor subtypes to classify which population woulds benefit from a tracheostomy and when. ¹⁷ Prolonged intubation and ventilation—especially past day ten—should be reviewed in a multidisciplinary fashion with attention to each patient’s critical illness and wishes going forward. ¹⁷

Ultimately, the literature is clear. There is:

no statistically significant mortality benefit with early tracheostomy;

little to no reduction in ICU stay or reduction in the duration of mechanical ventilation;

minimal benefit from sedation use.^21–28

Going forward, we need to be proactive in the identification of LTI early to reduce acute and chronic patient morbidity and potentially prevent irreversible iatrogenic damage from intubation.

Reflux and laryngeal injury

Reflux has long been thought of as a contributing factor in the pathogenesis of laryngeal injury in intubated patients despite the majority of post-intubation injuries resolving spontaneously.^29,30 Being positioned supine and fed via a nasogastric tube during ETI are contributing factors in allowing gastric and intestinal refluxate to affect the aerodigestive tract. ETI acts as an unavoidable physical insult to laryngotracheal mucosa, with reflux acting as an additional destructive stimulus. ¹² Reflux has also been identified as a key barrier to wound healing and allows infection to precipitate.^7,11

Two distinct yet related disease groups responsible for refluxate-mediated damage are gastro-oesophageal reflux disease (GORD) and more newly identified laryngopharyngeal reflux (LPR). LPR is a distinct inflammatory reflux pathology different to GORD. ³¹ It is defined by the retrograde passage of gastric or duodenal contents beyond the upper oesophageal sphincter, with contamination of the larynx, pharynx and lungs.^31,32 The most common symptoms of chronic LPR in ambulant patients are as follows: hoarseness, globus, throat clearing, cough, dysphagia and excessive phlegm. ³³ Whereas it is normal to experience up to 45 daily episodes of gastric refluxate in the lower and mid oesophagus, even a single exposure of acid and pepsin refluxate to the upper aerodigestive tract (UADT) can result in mucosal injury. ³⁴ The UADT is especially sensitive to gastric refluxate due to the absence of key structural and chemical barriers that ordinarily protect the oesophageal tract. ³³

In the early 1990s, post mortem analysis of patients who had prolonged ETI revealed a high concentration of low pH in their aerodigestive tracts, suggesting that it is a significant factor in the aetiology of LTI. ²⁹ Earlier animal and human models have attributed the effects to GORD, and it has been credited with several different pathologies including (but not limited to) hoarseness, laryngitis, granulomas, laryngospasm, subglottic stenosis, pneumonia and asthma.^29,35,36 Perhaps the initial insult to the mucosa and submucosa of the posterior larynx, vocal process and cricoid cartilage could be attributed to direct pressure from the ETT. ²⁹ The subsequent refluxate prevents healing, adds to ulceration to the deeper structures and contributes to reflex mechanisms such as bronchospasm and laryngospasm that cause further damage.^29,37

Theories supporting the role of acid in LTI gained credence with the elimination of signs and symptoms with the introduction of acid-suppressive agents such as omeprazole. ³⁸ Kamel et al. ³⁹ reported that laryngeal symptoms documented daily by patients with chronic laryngitis changed significantly after six to nine weeks of treatment with omeprazole.^38,39 However, the extensive study using combined multichannel intraluminal impedance and pH probe technology suggested that up to 20% of acid-suppressed patients still experienced signs and symptoms of reflux as well as gross findings on nasendoscopy. ⁴⁰ Investigation into this phenomenon has indicated that several non-acidic components of gastric refluxate, namely pro-enzymes such as pepsin can cause significant mucosal damage through endocytosis into laryngeal and hypopharyngeal cells. ⁴¹ Receptor-mediated uptake of pepsin even in a near neutral environment promotes uptake and activation of the proteolytic enzyme, thereby reproducing symptoms of laryngeal injury despite acid suppression. ⁴¹ Further it is important to note that the review of Rimoli et al. ⁴² on the management of laryngeal granulomas did not find any improvement despite a full course of anti-reflux treatment (b.i.d. proton pump inhibitor (PPI) daily for two weeks). High dose PPI is often used in the intensive care setting; however, other elusive and untreated factors such as pepsin and other gastric and intestinal endopeptidases are not affected by the use of PPIs.

Next steps

So, what does this all mean? Despite several robust prospective studies and literature reviews that examine LTI, authors have yet to agree on comprehensive assessment and management guidelines to prevent and treat these injuries. There are several modifiable and non-modifiable risk factors that exist in the development of LTI, but we do not have an agreed pathway to assess or clinically follow up these patients. LPR is one of the most established and recognised causative factors of LTI and influences several inflammatory disorders and pathologies of the UADT. Could LPR be the missing link in the development of LTI and a precipitating factor that prevents the healing of these injuries post-extubation? This may necessitate that those patients with pre-existing LPR be identified and assessed with post-extubation laryngoscopy as they already have an increased risk of developing LTI.

It is important that we take account of these findings and act to review and assess patients who undergo prolonged intubation on a more routine basis. We believe that the first strategy to preventing and reducing iatrogenic LTI is to identify and screen for those who have an increased risk of developing laryngeal injuries—especially those with prolonged intubation periods, a history of type 2 diabetes mellitus and/or reflux disease. In addition, we suggest increased collaboration with colleagues in speech and language therapy and ear, nose and throat for the development of guidelines for the recognition and management of laryngeal injuries, with formal assessment supported using fibreoptic evaluation of swallowing by speech and language therapists and nasendoscopy of the larynx for ongoing voice and swallowing complaints.

In conclusion, the incidence of LPR in the intubated and mechanically ventilated patient in the ICU is currently unknown. Prospective studies that directly measure the levels of acid, number of reflux events and macroscopic degree of LTI may allow us to understand further potential mechanisms of UADT injury due to LPR and herald the development of preventative and management strategies of LPR-mediated UADT injury in critically ill patients.