Discussion
The relationship between COVID-19 and laryngeal disorders was initially observed in an epidemiological study in which 26% of COVID-19 patients reported dysphonia throughout the clinical course of the disease.10 The potential laryngeal reach by the virus was supported in basic science studies reporting significant angiotensin converting enzyme-2 (ACE2) expression in vocal fold and laryngeal tissues.11
In the present study, we observed a high prevalence of laryngeal injuries in patients with a history of severe-to-critical COVID-19 and intubation. Depending on the intubation duration, the most common findings included posterior commissure hypertrophy and laryngeal edema, posterior glottic stenosis, and granuloma. Some recent studies reported similar laryngeal injuries in patients with a post-COVID-19 history of intubation.12,13 Naunheim et al . observed vocal fold immobility (40%), posterior glottic stenosis (15%), subglottic stenosis (10%), laryngeal edema (10%), LPR (10%) and posterior glottic diastasis (10%) in a cohort of 20 adults with a history of post-COVID-19 intubation.12 In the same way, Neevelet al . reported substantial prevalence of vocal fold motion impairment (50%), early glottic injury (39%), subglottic/tracheal stenosis (22%), and posterior glottic stenosis (17%) in 24 patients who required endotracheal intubation for a severe COVID-19.13 Rouhani et al. showed that 19% of COVID-19 patients with a history of tracheostomy in intensive care unit had vocal fold immobility and subglottic stenosis at 2-month postdischarge.7 More recently, Felix et al . observed laryngotracheal lesions in 40% of patients with a history of post-COVID-19 intubation, including posterior glottic or subglottic stenosis (17%), granuloma (16%) and hypermia of glottis (6%).14 In the study of Felix et al ., 60% of patients had normal laryngeal examination.14
Whatever the intubation indication, the laryngeal injuries observed in this study are known to arise after endotracheal intubation. The majority of lesions developed in the posterior laryngeal region, which may be due to the greatest pressure and trauma from the endotracheal tube during prolonged intubation in prone position.12The influence of prone position seems to be an important factor according to studies that reported the development of laryngeal injuries only 3 days post-intubation.15 The mean delay of the development of symptoms (dysphonia, dysphagia or dyspnea) in our patients was 3 months post-intubation, which corroborate the findings of previous studies.14
The mechanisms underlying the high prevalence of laryngeal injuries post-COVID-19 intubation remain poorly understood. The COVID-19 infection is associated with endothelial dysfunction, systemic prothrombotic state, microvascular injury, mucosa necrosis, and healing process impairments.16 In that way, it seems conceivable that the ACE2-related infection of laryngeal cells by the virus may lead to a local inflammatory reaction, which may weaken the laryngeal tissues. However, to date, this hypothesis remains not demonstrated. Moreover, it is unclear if COVID-19 is associated with a higher prevalence of post-intubation laryngeal injuries than other diseases.
The two most prevalent lesions in our study were posterior glottic stenosis and posterior commissure hypertrophy/laryngeal edema. Interestingly, we observed that the duration of intubation was a predictor of the development of posterior glottic stenosis. This observation support the findings of Hillel et al . who reported that duration of intubation, ischemia, and diabetes mellitus were significant risk factors for the development of posterior glottic stenosis.17 Similar studies corroborated the relationship between the duration of intubation and the development of posterior laryngeal lesions.18,19 An additional potential factor that may increase the laryngeal inflammation is reflux. COVID-19 patients commonly require moderate to high positive end-expiratory pressure,20 which may increase the stomach pressure and the back flow of gastric content into the laryngopharyngeal cavity. The deposited pepsin into the laryngeal tissue may, therefore, decrease the defense mechanisms of laryngeal mucosa,21 increasing the risk of injuries and lesions. The posterior commissure hypertrophy and laryngeal diffuse edema are furthermore two prevalent findings associated with LPR.21 From a pathophysiological standpoint, the development of posterior glottic stenosis is related to ulceration of mucosa and cartilage, inflammation with granulation, and fibrous contraction,22 which are related to the endotracheal tube pressure. As granulation tissue matures, it may assume the smooth, regular, and round shape of a granuloma,23 which are moreover found in the present study in 13.8% of patients.
Many comorbidities may be associated with the development of laryngotracheal injuries, including type 2 diabetes mellitus, obesity, hypertension, cardiovascular disease, or smoking.24 In the present study, we did not observe such association, but our cohort was not statistically powered to detect differences between subgroups of patients. The high prevalence of hypertension, cardiovascular disease and diabetes were just related to the inclusion of severe-to-critical COVID-19 patients in whom these conditions have a critical impact on the infection severity form. Tracheotomy is commonly considered as a relevant factor in the reduction of the occurrence of laryngeal lesions.25 In our study, the patients who benefited from tracheostomy reported similar proportions of laryngeal injuries than those who had no tracheostomy, which is attributed to the delay between the intubation and the tracheostomy decision (>14 days). Indeed, in our hospital, this delay was due to the greater risk of contaminating health professionals during an early procedure.
The management of laryngeal injuries may involve medical and surgical approaches. Only 16 patients benefited from surgical approaches after medical therapy failure, corresponding to 39% of cases. The high prevalence of posterior commissure hypertrophy, granuloma and diffuse laryngeal edema may explain this high rate of medical therapy success.
The present study has several limitations. The small sample size and the lack of control group evaluating the prevalence of post-intubation laryngeal injuries in patients without COVID-19 history are the most important limitations. Moreover, we did not assess some important ICU outcomes, including the tube size or the lung pressure of mechanical ventilation device, which may have a significant impact on the development of laryngeal injuries. Finally, we did not have sufficient follow-up to determine the mid-to-long term effectiveness of surgical procedures.