Identify Patients at Risk of Moderate or Severe OSAHS During ENT Examination Authors: Nicolas Lenay, Bruno Toussaint, Phi Linh Nguyen‐Thi, Nicolas Carpentier, Duc Trung Nguyen Published: 2026 ABSTRACT ABSTRACTObjectivesTo identify anatomical factors predictive of moderate/severe obstructive sleep apnea‐hypopnea syndrome (OSAHS) during routine ENT examinations in awake patients.MethodsA total of 110 patients who snore or are suspected to have OSAHS were prospectively included and underwent a standardized ENT examination coupled with recorded nasofibroscopy and sleep recording. Anatomical assessment of the upper airway was performed independently by two examiners using the recorded videos. The shape, size, and position of the epiglottis, the shape of the retro‐pharyngeal space, the modified Cormack–Lehane score, and the tongue base were evaluated. An interexaminer agreement was calculated. A multivariate logistic regression model was used to assess independent predictors of moderate/severe OSA, enabling the construction of a screening model. Thresholds were defined using the ROC curve.ResultsInterobserver agreement was excellent for the assessment of epiglottis shape/aspect (κ = 0.95), contact with the posterior pharyngeal wall (κ = 0.92), vallecula visibility (κ = 0.89), epiglottis size (κ = 0.86), and modified Cormack–Lehane score (κ = 0.86). Age > 45 years, male sex, retrovelar narrowing < 1 cm, presence of mega‐epiglottis and modified Cormack–Lehane score were significantly associated with moderate/severe OSA. Multivariate analyses revealed 3 predictive factors for moderate/severe OSAHS: age > 45 years (aOR = 4, 95% CI [1.44–11.12], p = 0.008), retrovelar narrowing < 1 cm (aOR = 2.8, 95% CI [1.05–7.6], p = 0.04), and the presence of a mega‐epiglottis (aOR = 4.1, 95% CI [1.49–11.23], p = 0.006).ConclusionMeticulous anatomic assessment of the upper airway during routine ENT consultations is a relevant tool for identifying patients at risk for moderate/severe OSAHS.Level of Evidence2 Trial Registration: Clinical Trials: NCT05808868. Introduction Obstructive sleep apnea‐hypopnea syndrome (OSAHS) is defined as an alteration in nocturnal ventilation related to partial or complete collapse of the upper aerodigestive tract (UA). The pharyngeal collapse responsible for nocturnal respiratory abnormalities can be explained by three factors: anatomical narrowing of the upper airway, increased pharyngeal compliance, and loss of efficiency of the pharyngeal dilator muscles [1]. The prevalence of OSAHS has been estimated to be between 9% and 38% [2]. However, the condition remains underdiagnosed [3]. In the absence of treatment, OSAHS leads to excess mortality particularly due to cardiovascular, cardiometabolic, neurological, psychosocial, and road accident complications [4, 5]. Therefore, the identification of patients at risk of OSAHS through early and targeted screening and the implementing effective treatments to reduce these complications has become a public health issue [6]. Many screening methods are now validated to identify patients at risk of OSAHS. The Berlin questionnaire allows the identification of patients at risk of OSAHS, as well as its severity [7]. The STOP‐BANG questionnaire was developed for OSAHS screening in the general population [8]. The NoSAS score, composed of one subjective criterion (snoring) and four objective criteria (neck circumference, overweight or obesity, age > 55 years, and male sex), has a sensitivity and positive predictive value (PPV) of 94.3% and 87.6%, respectively, for the diagnosis of OSAHS in suspected OSAHS patients [9]. However, these screening questionnaires do not account for UA anatomical abnormalities. UA examination coupled with fiberoptic pharyngolaryngoscopy is a routine physical checkup during ENT consultations. This examination can identify patients at risk of OSAHS and perform a pretherapeutic evaluation [10, 11] to define patient profiles and personalize patient care. There are multiple sites of UA collapse. A systematic review of drug‐induced sleep endoscopy (DISE) revealed a distribution of obstructive sites of 58.8% for the soft palate, 43.2% for the base of the tongue, 29.9% for the lateral walls of the pharynx, and 22.4% for the epiglottis [12]. However, the laryngeal level also plays a significant role in the mechanism of OSAHS [13]. Endoscopy under induced or natural sleep is reliable for identifying obstructive sites [14]. However, these techniques are time‐consuming and require substantial resources (access to the operating room or dedicated hospital room), specific equipment, and qualified personnel. In contrast, the standardized and meticulous ENT clinical examination coupled with fiberoptic pharyngolaryngoscopy is inexpensive, rapid, and easily accessible. It allows the identification of patients at risk of OSAHS, the detection of obstructive sites in the UA, and the optimization of therapeutic management by differentiating patients' profiles [15, 16]. This study aimed to identify anatomical factors predictive of moderate/severe OSAHS during a routine ENT examination in awake patients. Methods Study Population From May 2021 to August 2022, 110 consecutive adult patients consulting for snoring, suspected OSAHS, or mandibular advancement device placement in our ENT‐HNS department were prospectively included in this study. Patients with nonrespiratory sleep disorders (e.g., central sleep apnea, narcolepsy, or restless legs syndrome), age < 18 years, a history of UA surgery that potentially modified the UA anatomy (e.g., palatoplasty, surgery for ENT cancers), or a history of ENT tumors were not included. Patients with poor‐quality video recordings of fiberoptic pharyngolaryngoscopy were also excluded. Patient informed consent was obtained before inclusion in the study. This study was approved by the institutional review board of the University Hospital of Nancy, France. At inclusion, all patients underwent an ENT consultation by a junior or senior ENT physician (the senior ENT surgeon had 13 years of experience in sleep medicine and endoscopy, while the junior ENT has 1 year of experience in sleep medicine), including a standardized ENT clinical examination according to an established grid with a systematic fiberoptic pharyngolaryngoscopy that was recorded. Sleep recording was performed within 2 months after the ENT visit. The endobuccal evaluation included scoring the size of the tonsils according to the Friedman classification [17], characteristics of the soft palate and uvula (a uvula is considered long if its length is > 1.5 cm or wide if its width is > 1 cm [18]; the EK sign corresponds to the presence of horizontal linear striations on the uvula [19]), the volume of the tongue (relative macroglossia is defined by the presence of a dental impression on the lateral edges of the tongue or an overhang of the tongue at rest on the dental arches), and the soft palate to tongue ratio using the modified Mallampati score [17]. Fiberoptic pharyngolaryngoscopy was performed under local anesthesia (two sprays of 10% lidocaine [xylocaine pump spray 10%] were applied to each nasal cavity) with the patient in a vertically seated, neutral head position and awake. An Olympus flexible fiber–optic endoscope measuring 3.2 mm (Olympus Corporation of the Americas, Center Valley, PA) was used. Nasal cavities were assessed via rhinoscopy for their anterior part and then via fiberoptic pharyngolaryngoscopy up to the nasopharynx. Nasal obstruction was defined subjectively by the patient, while allergic rhinitis was diagnosed after an allergological assessment (allergist consultation or prick test). The tip of the fiberoptic laryngoscope was then placed at the entrance to the oropharynx to assess retrovelar narrowing (Figure 1). Visual estimation of the 1 cm threshold was standardized using the distal tip of the flexible endoscope (diameter 3.2 mm) as a reference; a space was considered < 1 cm if it was less than approximately three “scope‐widths.” This technique is a validated practice in endoscopy to overcome the “fish‐eye” distortion of wide‐angle lenses, where absolute distances cannot be measured by a simple screen‐ruler. Video recording began when the device was placed at the soft palate level to evaluate the tongue base, the vallecula, the epiglottis, the hypopharynx, the larynx, and the retrobasilingoepiglottic space. The procedure was videotaped using the Storz video system (Atmos servant s 61, Xion camera, with Divas mini software; Karl Storz, Tuttlingen, Germany). Example of retrovelar narrowing: non‐pathological (left image) and retrovelar narrowing < 1 cm (right image). The camera head is at the entrance to the oropharynx. Sleep recording was performed either by home ventilatory polygraphy (NOX T3 Portable Sleep Monitor; ResMed, San Diego, CA) or by polysomnography (ALICE 5 and ALICE 6; Philips Respironics, Amsterdam, the Netherlands) in our hospital sleep laboratory. All tests were scored according to the AASM 2012 criteria. Because moderate/severe OSAHS patients require specific management, we performed subgroup analyses: the “absent/mild OSAHS” group (including no OSAHS or mild OSAHS) with an apnea/hypopnea index (AHI) < 15/h and the “moderate/severe OSAHS” group (including moderate or severe OSAHS) with an AHI ≥ 15/h. Analysis of Recorded Videos Analysis of upper aerodigestive anatomy was performed using a standardized evaluation grid based on videos recorded by two independent raters, a junior and senior ENT physician (different from the senior physician who examined the patients). The analyses were performed blinded to the patients' clinical history and sleep recording results. The first step of the morphological analysis was performed when the fiberoptic pharyngolaryngoscope was placed at the soft palate level. We assessed the visibility of the vallecula, the epiglottis size (normal or mega‐epiglottis defined as an epiglottis completely or almost completely covering the pharyngo‐epiglottic folds or a transverse dimension, estimated ≥ 2/3 of the largest transverse dimension of the hypopharynx) [16], the epiglottis shape (defined as flat or flattened like a tree leaf, semicircular or C‐shaped, tubular or U‐shaped), the epiglottis position (assessed according to the virtual, central axis of the epiglottis passing through its base, visible either on its lingual or laryngeal side, relative to the vertical plane passing through the pharyngo‐epiglottic folds: the epiglottis position was defined as forward, vertical, or backward if this axis was anterior, superposed, or posterior to this plane, respectively), and the presence of the epiglottis contacting the posterior pharyngeal wall. The retropharyngeal‐epiglottic aerospace (RPEA) was defined according to its shape: the width was defined as the greatest virtual distance between two lateral ends of the endopharynx, and the depth as the greatest anteroposterior distance between the posterior face of the epiglottis and the posterior pharyngeal wall (Figure 2). According to these two parameters, an almond shape corresponds to a space that was significantly wider than it is deep; an oval shape corresponds to a depth greater than half the width; and a diamond shape corresponds to a space whose posterior boundary is marked by an angle and associated with a semicircular or tubular epiglottis. The modified Cormack–Lehane score was evaluated according to the visibility of the arytenoids and vocal cords [15]. The V‐sign (the appearance of a V shape) corresponds to a central angle in the posterior pharyngeal wall opposite the epiglottis (corresponding to fat accumulation in parapharyngeal spaces). (A) Example of a mega‐epiglottis. The widest transverse dimension of the epiglottis (a) is > 2/3 the transverse dimension of the hypopharynx (b), retro‐basilinguo‐epiglottic space (c) is defined according to shape: Width (d) is defined as the largest virtual distance between two lateral ends of the pharynx, and depth (e) is defined as the largest anteroposterior distance between the posterior face of the epiglottis and the posterior pharyngeal wall. (B) The V‐sign. (C) The different shapes of the epiglottis, from top to bottom: Flat or flattened like a tree leaf, semicircular or C‐shaped, tubular or U‐shaped. (D) Example of basilingual muscular hypertrophy: Top image, a non‐pathological language base, the lower part shows basilingual muscle hypertrophy. The second stage involved analysis of the lingual tonsils, using the lingual tonsil hypertrophy (LTH) classification system [20], and basilingual muscular hypertrophy, defined as tissue overlap > 1/3 of the lingual surface of the epiglottis with lingual tonsils less than or equal to Grade 1, when the nasofibroscope was placed over the epiglottis. Statistical Analysis Statistical analyses were performed with SAS version 9.1 software (SAS Institute Inc., Cary, NC, USA). Categorical variables are expressed as percentages and as means for quantitative variables. The normality of continuous variables was tested with the Shapiro–Wilk test. Characteristics were compared between groups with the χ 2 test or Fisher's exact test for categorical variables and the Wilcoxon–Mann–Whitney test for continuous variables. Interrater reliability was assessed using kappa (κ) coefficients. The Landis and Koch classification was applied to interpret the κ coefficients: ≤ 0.4 = low to moderate agreement, 0.4–0.6 = moderate agreement, 0.6–0.8 = substantial agreement, and 0.8–1 = high or excellent agreement. To identify independent factors predictive of moderate/severe OSAHS, a multivariate logistic regression model was constructed to adjust for all significant parameters. All variables significant at p levels ≤ 0.10 in the bivariate analyses were included in the multivariate logistic regression model, performed with ascending stepwise selection. A predictive model of moderate/severe OSAHS was subsequently constructed. The predictive thresholds for moderate/severe OSAHS compared with absent/mild OSAHS were identified using the ROC curve. A two‐tailed p value < 0.05 was considered statistically significant. Results A total of 110 patients (38 women and 72 men; mean age 49.1 ± 12.2 years; age range 25–74 years) were included in the study. The “absent/mild OSAHS” group included 28 patients (mean age 43.4 ± 10.3 years), and the “moderate/severe OSAHS” group included 82 patients (mean age 51.1 ± 12.3 years) (p = 0.003, Wilcoxon–Mann–Whitney test). The mean body mass index (BMI) was significantly different between the “absent/mild OSAHS” group (27.56 ± 5.43 kg/m2) and the “moderate/severe OSAHS” group (30.11 ± 6.39 kg/m2) (p = 0.02, Wilcoxon–Mann–Whitney test). There were more men in the “moderate/severe OSAHS” group than in the “absent/mild OSAHS” group (p = 0.046, χ 2 test) (Table 1). TABLE 1Relationship between presence and severity of obstructive sleep apnea–hypopnea syndrome and ENT clinical examination parameters.Absent/mild OSAHSModerate/severe OSAHS p Age, years—no. (%)0.002* ≤ 4518 (16)25 (23)> 4510 (9)57 (52)Gender no. (%)0.046* Men14 (13)58 (53)Women14 (13)24 (22)Retrognathia no. (%)0.11 † Yes3 (3)2 (2)No24 (22)76 (69)Characteristics of the soft palate no. (%)0.26* Normal4 (4)20 (18)Long/thick/peri velar mucous membrane24 (22)62 (56)Characteristics of the uvula no. (%)0.84* Normal7 (6.4)19 (17)Long or presence of EK sign21 (19)63 (57)Ogival palate, no. (%)0.08* Yes13 (12)24 (22)No14 (13)57 (52)Relative macroglossia, no. (%)0.11* Yes17 (15)62 (56)No11 (10)19 (17)Size of the palatine tonsils, no. (%)0.78 † 05 (5)18 (16)117 (15)42 (38)24 (4)17 (15)32 (2)4 (4)Modified Mallampati score, no. (%)0.5* 1 ou 29 (8)21 (19)3 ou 419 (17)61 (55)Nasal obstruction, no. (%)0.052* Yes17 (15)65 (59)No11 (10)17 (15)Nasal valve syndrome, no. (%)0.999 † Yes1 (1)4 (4)No27 (25)78 (71)Hypertrophy of the inferior turbinates, no. (%)0.59* Yes11 (10)37 (34)No17 (15)45 (41)Allergic rhinitis, no. (%)0.27 † Yes4 (4)6 (5)No24 (22)76 (69)Retrovelar narrowing < 1 cm, no. (%)0.001* Yes11 (10)60 (55)No17 (15)22 (20)* χ 2 test. † Fisher's exact test. Relationship between presence and severity of obstructive sleep apnea–hypopnea syndrome and ENT clinical examination parameters. χ 2 test. Fisher's exact test. Table 2 shows the interobserver agreement for rating anatomical parameters based on the recorded videos. Concordance was excellent for the assessment of epiglottis shape/appearance (κ = 0.95), contact with the posterior pharyngeal wall (κ = 0.92), visibility of the vallecula (κ = 0.89), epiglottis size (κ = 0.86), and modified Cormack–Lehane score (κ = 0.86). The agreement remained substantial regarding assessing the RPEA shape, V‐sign, tongue base muscular hypertrophy, lingual tonsil size, and epiglottis position (κ ranging from 0.62 to 0.80). TABLE 2Relationship between severity of obstructive sleep apnea–hypopnea syndrome and nasofibroscopic examination of the pharyngolaryngeal tract.OSAHSJuniorSeniorJuniorSeniorKAbsent/mildModerate/severe p Absent/mildModerate/severe p Visibility of the vallecula, no. (%)0.95* 0.44* 0.89Yes6 (5)18 (16)7 (6)15 (14)24 (22)22 (20)No22 (20)64 (58)21 (19)67 (61)86 (78)88 (80)Epiglottis size, no. (%)0.004* 0.005* 0.86Normal size16 (15)22 (20)16 (15)23 (21)38 (35)39 (35)Mega‐epiglottis12 (11)60 (55)12 (11)59 (54)72 (65)71 (65)Shape/aspect of epiglottis, no. (%)0.34* 0.31* 0.95Flat or flattened12 (11)44 (40)11 (10)42 (38)56 (51)53 (48)Semicircular (C‐shaped)11 (10)31 (28)12 (11)33 (30)42 (38)45 (41)Tubular (U‐shaped)5 (5)7 (6)5 (5)7 (6)12 (11)12 (11)Epiglottis position, no. (%)0.18 † 0.46 † 0.72Vertical18 (16)62 (56)22 (20)61 (55)80 (73)83 (75)Tilted backward9 (8)20 (18)5 (5)20 (18)29 (26)25 (23)Tilted forward1 (1)01 (1)1 (1)1 (1)2 (2)Contact to the posterior pharyngeal wall, no. (%)0.98* 0.84* 0.92Yes11 (10)32 (29)10 (9)31 (28)43 (39)41 (37)No17 (15)50 (45)18 (16)51 (46)67 (61)69 (63)Shape of the retropharyngeal‐epiglottic‐aerospace (RPEA), no. (%)0.22 † 0.19 † 0.80Almond8 (7)18 (16)8 (7)15 (14)26 (24)23 (21)Oval12 (11)45 (41)12 (11)46 (42)57 (52)58 (53)Runded7 (6)10 (9)7 (6)11 (10)17 (15)18 (16)Lozenge1 (1)9 (8)1 (1)10 (9)10 (9)11 (10)Modified Cormack–Lehane score, no. (%)0.01 † 0.04 † 0.8614 (4)2 (2)4 (4)2 (2)6 (5)6 (5)210 (9)28 (25)9 (8)32 (29)38 (35)41 (37)39 (8)47 (43)12 (11)45 (41)56 (51)57 (52)45 (5)5 (5)3 (3)3 (3)10 (9)6 (5)V sign no. (%)0.15 † 0.51 † 0.78Yes2 (2)16 (15)2 (2)12 (11)18 (16)14 (13)No26 (24)66 (60)26 (24)70 (64)92 (84)96 (87)Size of lingual tonsils, no. (%)0.17 † 0.35 † 0.6203 (3)1 (1)1 (1)04 (4)1 (1)111 (10)37 (34)14 (13)48 (44)48 (44)62 (56)210 (9)29 (26)8 (7)19 (17)39 (35)27 (25)33 (3)14 (13)4 (4)14 (13)17 (15)18 (16)41 (1)1 (1)1 (1)1 (1)2 (2)2 (2)Muscular hypertrophy of the tongue base, no. (%)0.77* 0.97* 0.75Yes9 (8)24 (22)10 (9)29 (26)33 (30)39 (35)No19 (17)58 (53)18 (16)53 (48)77 (70)71 (65) Note: Interobserver agreement. K simple Kappa coefficient.* χ 2 test. † Fisher's exact test. Relationship between severity of obstructive sleep apnea–hypopnea syndrome and nasofibroscopic examination of the pharyngolaryngeal tract. Note: Interobserver agreement. K simple Kappa coefficient. χ 2 test. Fisher's exact test. Regarding the clinical parameters of the standardized ENT examination (Tables 1 and 2), significant differences were observed between the subgroups “absent/mild OSAHS” and “moderate/severe OSAHS” for retrovelar narrowing < 1 cm (p = 0.001, Fisher's exact test), epiglottis size (p = 0.004, χ 2 test), and the modified Cormack–Lehane score (p = 0.01, Fisher's exact test). In multivariate analyses, the logistic regression model included the following variables: sex, age (≤ 45 and > 45 years), BMI, nasal obstruction, retrovelar narrowing, epiglottis size, and modified Cormack–Lehane score. Three independent parameters predictive of moderate to severe OSAHS were identified: age > 45 years (aOR = 4; 95% CI [1.44–11.12], p = 0.008), retrovelar narrowing < 1 cm (aOR = 2.8; 95% CI [1.05–7.6], p = 0.04), and the presence of a mega‐epiglottis (aOR = 4.1; 95% CI [1.49–11.23], p = 0.006) (Table 3). This allows establishing a model to identify the presence or absence of moderate/severe OSAHS during routine ENT examination coupled with fiberoptic pharyngolaryngoscopy in awake patients: TABLE 3Multivariate analysis of independent predictors of moderate or severe obstructive sleep apnea‐hypopnea syndrome.Wald's K 2 test p Odds ratio95% CIHosmer‐Lemeshow goodness‐of‐fit test2.270.89PredictorsIntercept = −1.08Age, years (age > 45 years vs. age ≤ 45 years)7.050.00841.44–11.12Retrovelar narrowness (narrow ≤ 1 cm vs. absence narrowness)4.240.042.81.05–7.6Epiglottis appearance (mega‐epiglottis vs. normal epiglottis)7.450.0064.11.49–11.23Abbreviation: CI = confidence interval. Multivariate analysis of independent predictors of moderate or severe obstructive sleep apnea‐hypopnea syndrome. Abbreviation: CI = confidence interval. Score = −1.08 + 4*age (0 if ≤ 45 years, 1 if > 45 years) + 2.8*retrovelar (velar narrowing < 1 cm: 1 if yes, 0 if no) + 4.1*megaepiglottis (1 if yes, 0 if no). The ROC curve was used to determine the score thresholds for predicting moderate/severe OSAHS (Figure 3A). If the threshold is set at 3, the sensitivity is 61%, and the specificity is 80%. If the threshold is set at 5.7, the sensitivity is 71%, and the specificity is 63%. However, the threshold at 3 had the highest Youden index (Figure 3B–C). Determination of predictive thresholds for moderate/severe OSAHS. (A) According to the ROC curve for the calculation of sensitivity and specificity. (B) Intrinsic validity. (C) Variation in sensitivity, specificity, and Youden index according to the threshold value. Discussion Identifying patients at risk of OSAHS must be a priority of public health due to the significant complications and mortality linked with OSAHS and the economic impacts in the absence of diagnosis and early treatment [21]. The disease's nonspecific symptoms (daytime somnolence, nonrestorative sleep, etc.) and the exclusively nocturnal clinical signs (snoring, respiratory pauses) explain the difficulty in identifying patients when no partner is present. This study clarifies the decisive role of clinical examination by ENT specialists in identifying obstructive anatomical factors predictive of OSA. Using an OSAS predictive model based on epidemiological and clinical data offers ENT specialists an interesting tool for suspecting OSAS in some patients. We confirm the value of a detailed examination of the hypopharynx as a nonnegligible anatomical site of OSAHS, specifically in detecting a mega‐epiglottis. We propose new anatomical parameters for analysis, such as the shape of the epiglottis or the presence of the V‐sign of the posterior pharyngeal wall, whose analyses did not reveal any significant differences between the groups. Currently, many screening methods for OSAHS are available. Screening questionnaires such as the NoSAS score or the STOP‐BANG questionnaire have the advantage of being accessible for all physicians and are easy to perform. However, they rely on subjective criteria without considering the anatomical abnormalities of the UA. Sun et al. proposed using artificial intelligence to build predictive models of OSAHS from anthropometric data, interviews, and medical history. They reported promising results that require further studies for use in practice [22]. However, the parameters studied are indirect indicators of a reduced VAS score of the lumen. As OSAHS results from the collapse of the UA during sleep, identifying the involved anatomical elements and directly assessing the UA lumen status (container and content) using a standardized ENT examination coupled with fiberoptic pharyngolaryngoscopy plays essential roles in diagnosing and managing this pathology. Friedman et al. reported an interest in standardized clinical examination of the UA for identifying patients at risk of OSAHS, and several classifications have been proposed [10, 17, 23, 24]. The difficulty of assessing pharyngeal obstructive sites during consultation is explained by the muscular tone of the awake upper airway, the upright position of the patient in the consultation chair (as opposed to the nocturnal dorsal decubitus position), a static assessment of anatomical parameters, and the presence of multiple concomitant obstructive sites (found in 73% of patients) [25]. To ensure reproducibility, we chose to measure these parameters under typical consultation conditions without a dynamic maneuver (Muller's maneuver—forced inspiration with the mouth closed and nose pinched—presents contradictory results and tends to underestimate retrobasilingual obstructive mechanisms [26]) and assess patients in the seated position. Patient position during sleep plays a critical role in obstructive mechanisms, as described by Lee et al. [27]. Evaluating patients in the supine position during consultation is one way of reproducing the obstructive mechanisms observed during sleep. However, this position is uncomfortable for the patient and the ENT doctor and requires additional consultation time. In addition, a tilting chair or stretcher must be set up in a consultation room. In our study, the clinical and anatomical parameters used to predict moderate/severe OSAHS were age > 45 years, retrovelar narrowing < 1 cm, and a mega‐epiglottis. Regarding the evaluation of the oropharynx, an endoscopic study of the soft palate, which involves assessing the lumen of the air passage, seems more relevant than an endobuccal examination, whose study of the uvula and soft palate correlates with retrovelar narrowing. According to the literature, the soft palate plays a significant role in the obstructive phenomenon of OSAHS (in 58.8% of cases) during induced sleep endoscopy [13]. The absence of a supporting bony structure and neuronal degeneration explain the increased narrowing and collapse of the upper airways at this level [28]. Regarding the endoscopic evaluation of laryngeal and hypopharyngeal sites, we used our team's previously published classification systems [16]. Our study revealed high interobserver correlations for all parameters, especially for the modified Cormack–Lehane score and the size and position of the epiglottis. These results can be explained by the easy and unambiguous characteristics of these parameters. Although objective elements are well‐defined, these criteria still contain subjective elements. However, we believe their use in routine practice is easy and reliable. The subjectivity of their evaluation can explain the poorer correlations of the RPEA, V‐sign, and tongue base muscular hypertrophy. Our endoscopic analyses considered more parameters than those in the published studies, allowing us to construct a predictive model for identifying patients at risk for moderate/severe OSAHS. The advantage of this model is the combination of only three criteria: demographic (age), velar (narrowing < 1 cm), and laryngeal (epiglottis size) criteria. The calculation is easy to perform and facilitates extended sleep studies and early identification of patients who are challenging to screen using conventional methods: patients with risk factors for OSAHS but no associated sleep disorders (these patients are rarely screened) or patients with suspicious symptomatology but no risk factors or suggestive history (screening questionnaires are often negative in these cases). We propose two different thresholds depending on the screening objective. The 3 threshold has a sensitivity of 61% and a specificity of 80%; the 5.7 threshold has a sensitivity of 71% and a specificity of 63%. The tongue base evaluation (lingual tonsils and muscle hypertrophy) does not appear to correlate with the severity of OSAHS. Our results align with those of Tang and Friedman [29]. On the other hand, our results revealed an association between the presence of a mega‐epiglottis and moderate/severe OSAHS. The role of the epiglottis in obstructive phenomena of the UA has been underestimated and seems to play a major role in the disease. At this level, anteroposterior collapse of the UA predominates over tongue base collapse, pushing the epiglottis toward the posterior pharyngeal wall [15], and the epiglottis size could facilitate this obstruction [16, 30]. Our study did not find an association between the shape or position of the epiglottis and the severity of OSAHS. These results are consistent with those of Sung et al., who highlighted the difficulty of demonstrating epiglottis collapse under endoscopy during natural sleep [31]. In 1987, Fujita's classification [32] was the first standardized UA for assessing obstructive sites in the oropharynx and hypopharynx. This classification was modified by Moore and Phillips [33] and focused on evaluating the retrobasilingual and hypopharyngeal space using nasofibroscopy and cephalometry. These classifications are primarily used to characterize obstructive sites and adapt surgical management. Since the advent of induced sleep endoscopy, new classifications have been developed to refine identifying obstructive sites and their mechanisms: VOTE classification [34] and NOHL classification [35]. Dynamic assessment of obstructive sites using induced sleep endoscopy [36] or induced sleep videofluoroscopy [37] has the advantage of assessing obstructive sites in conditions closer to the patient's sleep, with similar results [38]. This still requires a specific technical platform and qualified personnel. These assessments take a long time to perform and are not available in all centers. Furthermore, these classification systems are not comparable to UA assessments in the vigorous seated state. In addition, these systems do not assess in detail the anatomical structures involved in forming the lumen of the UA. Finally, our study contains some biases. First, there is a selective bias due to the exclusive recruitment of patients with suspected OSAHS. These results need to be confirmed by further studies comparing these factors with a nonsuspect population. Second, some clinical parameters predictive of OSAHS could not be demonstrated in this study (the male sex, BMI, modified Cormack–Lehane score), likely due to the insufficient number of patients recruited or the characteristics of the patients included in this study (anatomical risk factors). Third, the performance of the screening model does not appear to be optimal in terms of sensitivity and specificity. Consequently, further studies are needed to improve this model. Within the limits, two sleep tests must have been performed for diagnosis (polysomnography and ventilatory polygraphy). The use of two sleep tests is explained because some patients referred for consultation had already undergone polysomnography. For the others, ventilatory polygraphy was performed. Conclusions Routine ENT examination coupled with fiberoptic pharyngolaryngoscopy in awake patients can identify predictive factors of moderate/severe OSAHS. The pharyngolaryngeal tract is an underestimated obstructive site that must be considered when identifying patients at risk of OSAHS. These new anatomical criteria need to be confirmed in a population not suspected of having OSAHS before they can be used as a screening tool in the general population. Funding The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest.