Postacute Laryngeal Injuries and Dysfunctions in COVID-19 Patients: A Scoping Review

Objective: To investigate post-acute laryngeal injuries and dysfunctions (PLID) in coronavirus disease 2019 (COVID-19) patients. Methods: Three independent investigators performed a systematic review of the literature studying PLID in patients with a history of COVID-19. The review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Statement. Epidemiological, clinical, hospitalization features, laryngeal diseases, and voice outcomes were extracted from the included papers. Results: Eight papers met our inclusion criteria (393 patients) corresponding to five uncontrolled prospective and three retrospective studies. The most prevalent PLID were vocal fold dysmotility (65%), vocal fold edema (35%), muscle tension dysphonia (21%), and laryngopharyngeal reflux (24%). Posterior glottic stenosis (12%), granuloma (14%), and posterior glottic diastasis (12%) were the most common injuries. Most patients with PLID were obese and had a history of intensive care unit hospitalization, and orotracheal intubation. The delay between the discharge and the laryngology office consultation ranged from 51 to 122 days. The mean duration of intubation ranged from 10 to 34 days. Seventy-eight (49%) intubated patients were in the prone position. The proportion of patients requiring surgical treatment ranged from 39% to 70% (mean = 48%). There was an important heterogeneity between studies about inclusion, exclusion criteria, and outcomes. Conclusions: COVID-19 appeared to be associated with PLID, especially in patients with a history of intubation. However, future controlled studies are needed to evaluate if intubated COVID-19 patients reported more frequently PLID than patients who were intubated for other conditions.

The aim of this scoping review was to investigate post-discharge PLID in COVID-19 patients.

Materials and Methods
The criteria for consideration of study inclusion were based on the population, intervention, comparison, outcome, timing, and setting (PICOTS) framework [13]. Data of the study were independently reviewed by three investigators (JRL, SH, and MC) who extracted findings according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist for systematic reviews [14].

Patient Population
Prospective and retrospective, controlled, uncontrolled, or randomized studies published between December 2019 and May 2022 were included if authors investigated PLID in COVID-19 patients. Studies only reporting symptom outcomes without laryngeal examination were excluded. Patients should have a positive diagnostic for COVID-19 by reverse transcriptase polymerase chain reaction (RT-PCR) testing or serology. The studies were published in English, Spanish, or French peer-reviewed journals. Case reports were not considered in the analysis. The type of study was classified according to the levels of evidence for prognostic studies (I-V) [15].

Intervention, Comparison, and Outcomes
The following outcomes were reviewed for each study: number of patients; mean age; gender; COVID-19 severity; intubation or tracheostomy history and characteristics; duration of hospitalization; post-discharge laryngeal/voice findings and methods of laryngeal outcome assessment. The Tool to Assess Risk of Bias in Cohort Studies developed by the Clarity Group and Evidence Partners was used by two authors (JRL & SH) for the bias/heterogeneity analyses of the included studies [16]. The bias analysis consisted of an evaluation of cofactors that may impact the conclusion of studies.

Timing and Setting
The patients had been cured of COVID-19 and were discharged from the hospital.

Search Strategy
The paper search was conducted with PubMed, Cochrane Library, and Scopus databases by three independent laryngologists (JRL, SH, and MC). Databases were screened for abstracts and titles referring to the inclusion criteria of the present study. Authors analyzed the full texts of the selected publications. The following keywords were considered for the literature search: 'COVID-19'; 'SARS-CoV-2'; 'Larynx'; 'Voice'; 'Laryngeal'; 'Complications'; 'Intubation'; and 'Otolaryngology'.
The features of patients who were in ICU are presented in Table 1. Focusing on studies where authors reported findings of ICU patients, the mean duration of intubation ranged from 10 to 34 days, while the size of the tube ranged from 7 to 8. Seventy-eight (49%) intubated patients were in prone position during the ICU stay. There were 88/237 patients (37%) with a tracheostomy, which was removed after a mean duration ranging from 16 to 70 days.

Laryngeal Abnormalities
Irrespective of the definition and voice quality tools used, dysphonia was found in 70% of patients. Most patients had multiple chiefs of complaints (Table 1). PLID reported in studies are summarized in Table 3. Note that the prevalence of PLID was assessed in studies in which specific data of intubated patients were reported. Thus, the study of Neunheim et al. was excluded from this analysis according to the pooled information from non-intubated and intubated patients. Among COVID-19 patients with dysphonia in the post-discharge follow-up, the most prevalent PLID were vocal fold dysmotility, edema, laryngopharyngeal reflux, and muscle tension dysphonia ( Table 3). The most prevalent laryngeal injuries included posterior glottic stenosis, granuloma, posterior glottic diastasis, and VF immobility. Sandblom et al. reported a positive association between the duration of ICU stay and the severity of swallowing disorders [10]. Among contributing factors, Hans et al. reported that prolonged intubation was associated with an increase in laryngeal injuries, e.g., posterior glottic stenosis [19], while Felix et al. observed that intubation tube size and prone position were both factors associated with laryngeal injuries [12]. Abbreviations: MTD = muscle tension dysphonia; VF = vocal fold.
Some authors reported the need for medical (botox or corticosteroid injections) or surgical (balloon dilatation or laser microsurgery procedures) treatments for the management of laryngeal injuries [11,19]. Overall, the proportion of patients requiring surgical approach ranged from 39% to 70% (mean = 48%) [11,19].

Bias Analysis
Bias analysis is reported in Appendix A. There was no study that compared the prevalence of PLID between COVID-19 discharged patients and those discharged from the hospital for another reason. Three studies were retrospective case-series (EL: IV) and five were prospective uncontrolled studies (EL: III). The patient populations substantially varied from one study to another according to the proportion of intubated versus not intubated patients, and therefore, the severity of the disease. Moreover, the delay between the discharge and the time of examination was not provided in three studies [8][9][10] and may vary from one study to another, which may lead to a comparison bias. Other factors may limit the comparison between studies, e.g., the variability in comorbidity prevalence, the proportion of tracheostomy, the duration of intubation or tracheostomy, and the methods used to evaluate the laryngeal function and PLID (Appendix A). Moreover, many important outcomes that may influence the development of PLID were not investigated in patients, including tobacco consumption, laryngopharyngeal reflux, or a history of previous cancer or radiation. No author reported medical post-discharge care (drugs), which may impact the development of some laryngeal injuries.

Discussion
Coronavirus disease 2019 was found to be associated with many otolaryngological disorders, including smell and taste dysfunctions [20], vestibular neuritis [21], parotitis [22], and facial paralysis [23] or paradoxical vocal fold movement [24]. The association between COVID-19 and laryngeal disorders was initially observed in a first-wave epidemiological study in which 26% of patients with mild-to-moderate COVID-19 reported dysphonia throughout the clinical course of the disease [25]. Since then, the follow-up of patients who were intubated or who had a tracheostomy in ICU suggested the occurrence of mid-term PLID [8][9][10][11][12]. The present review summarized the PLID found in COVID-19 patients. Many factors may limit the draw of a reliable conclusion.
First, most authors suggested that the prevalence of post-intubation or post-tracheostomy PLID was substantially high in COVID-19 patients who were discharged from the ICU. Our data suggested that the prevalence of bilateral or unilateral vocal fold motion disorders, vocal fold edema, or posterior glottic stenosis may reach 16% to 65% of cases. A recent systematic review reported that 13% to 31% of non-COVID-19 patients who were intubated in ICU had moderate-to-severe laryngotracheal injuries [26]. Precisely, grade 3 injuries, including stenosis, hypo/immobility of vocal folds, and/or arytenoid complex, were found in 13% of cases, whereas grade 2 injuries (e.g., hematoma, ulceration, edema, or granulation) concerned 31% of cases [26]. Authors reported that the prevalence of grade 3 and 2 injuries increased to 33% and 18% in patients with an average intubation duration >5 days, respectively [26]. Comparing these data with the findings of the present review, it could appear that COVID-19 patients may present higher rates of PLID but, according to the lack of controlled study, the draw of reliable conclusion remains difficult.
Second, many other biases may limit the interpretation of the study results. The primary bias is an inclusion bias because patients included in the studies were all recruited from otolaryngology or laryngology offices and, therefore, consisted of a dysphonic population. There was no epidemiological study that systematically evaluated the occurrence of voice disorders and PLID in all patients who were discharged from the hospital/ICU. Third, the authors investigated some laryngeal disorders without consideration of other contributing factors that may be associated with the development of PLID. This is particularly the case for muscle tension dysphonia, which was reported as a prevalent condition in COVID-19 patients in two studies [8,18]. Muscle tension dysphonia may develop from gastric or environmental irritants, laryngitis, or even stress, among other conditions [27], which were not investigated in both studies [8,18]. A similar observation may be made for granuloma. Although intubation is an important cause of granuloma, other prevalent etiologies may play a key role in the development of granuloma such as reflux [28]. According to studies, the populations of studies considered in the present review reported high but different rates of comorbidities, which may be an additional limiting factor to precisely study the prevalence of PLID and its association with COVID-19. Indeed, some comorbidities may be associated more frequently with some PLID, such as reflux and posterior laryngeal edema; or diabetes and laryngotracheal stenosis [28,29]. Another factor that may impact the results of studies is the definition of PLID. We observed that the definition of some PLID may vary from one study to another. Precisely, the authors did not define similarly vocal fold hypomobility, which was the most prevalent PLID. The observation of hypomobility of vocal folds is still subjective and many authors did not provide information about the etiology (laryngeal nerve impairment versus arytenoid join ankylosis) [10,11,18]. These biases have to be considered in future studies that aim to investigate PLID in COVID-19.
The present scoping review globally included eight studies (393 patients), which considerably limits the drawing of reliable conclusions. The low number and the low evidence level of studies are, therefore, the primary limitation of this review. The lack of controlled study comparing both prevalence and features of post-intubation/post-tracheostomy PLID between discharged COVID-19 and individuals with a history of intubation or tracheostomy for another condition is another important limitation.

Conclusions
COVID-19 appeared to be associated with PLID in patients with a history of intubation or tracheostomy. However, it remains difficult to determine if the development of post-intubation or post-tracheostomy PLID is more frequent in COVID-19 patients compared with those with a history of intubation or tracheostomy for another condition.

Conflicts of Interest:
The authors declare no conflict of interest. Hans [19] Probably no Probably yes Probably yes Probably yes No

Appendix A
The criteria used and the definition of rating (yes, probably yes, probably no, and no) were explained below. For confounding factors, authors had to exclude some conditions: Yes = exclusion of patients with pre-COVID-19 laryngeal disorders prior the COVID-19 and assessment of other conditions associated with PLID (i.e., reflux, tobacco exposition, trauma or radiation histories); probably yes = exclusion of patients with pre-COVID-19 laryngeal disorders or assessment of other conditions associated with PLID; Probably no = exclusion of only some confounding factors; No = no information provided about exclusion criteria. Population analysis: Yes = different analyses performed according to the intubation/tracheostomy status of patients. No = no provided information. Delay: Yes = information about the delay between the discharge of hospital and the occurrence of PLID in the different populations of patients were available (if applicable; intubated vs non-intubated; tracheotomized vs non-tracheotomized). Probably Yes= details for the entire cohort were available; Probably no = Only delay between COVID-19 diagnosis and PLID occurrence were available; No = no information provided. Intubation/tracheostomy details: Yes = full information for concerned patients were provided (duration, tube size, proning); probably yes ≥50% of information were provided; probably no ≤50% of information were provided; No = no information were provided. Postdischarge care: Yes = information about postdischarge medication/care (e.g., speech therapy) that may influence the development of PLID were provided. No = no information provided.