Phonetic Analysis of Snoring: A Novel Approach to Assessing Pediatric Palatine Tonsil Hypertrophy

Prashant Sharma^1*, Nufra Senopher Mohamed Sarfraz², Vinod Singhal³

1,2,3. Prime Hospital, Dubai UAE.

*Correspondence to: Prashant Sharma, Prime Hospital, Dubai UAE.

Copyright

© 2025 Prashant Sharma. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Received: 18 February 2025

Published: 24 February 2025

DOI: https://doi.org/10.5281/zenodo.14963939

Abstract

Background: Obstructive sleep apnea (OSA) in children is primarily caused by upper airway obstruction, with palatine tonsil and adenoid hypertrophy being the most common contributors. While polysomnography is the gold standard for diagnosis, its high cost limits its accessibility, necessitating alternative screening methods. This study investigates the correlation between commonly used subjective grading of palatine tonsil hypertrophy, actual tonsil volume, and phonetic characteristics of snoring to explore the potential of snoring analysis as a diagnostic tool.

Methods: A total of 21 pediatric patients (aged 4–15) undergoing tonsillectomy due to snoring symptoms were enrolled. Preoperative assessment included medical history collection, physical examination, and lateral neck X-rays. Palatine tonsil hypertrophy was evaluated using the Brodsky tonsil grade scale, and tonsil volume and weight were measured postoperatively. Snoring sounds were recorded preoperatively and analyzed for intensity and formant frequencies using Praat software. Statistical analyses, including linear and multiple regression models, were performed to determine correlations between tonsil size, snoring characteristics, and BMI.

Results:Although tonsil volume showed a general increasing trend with Brodsky tonsil grade, the correlation was not statistically significant. Snoring intensity did not correlate with tonsil volume, weight, or adenoid index, but a significant positive correlation with BMI was observed (β=0.465, p=0.025). Notably, formant frequency analysis revealed a significant inverse correlation between palatine tonsil volume and the first (F1) and second (F2) formants (p=0.011 and p=0.002, respectively), reflecting anatomical airway changes due to hypertrophy.

Conclusion:This study highlights the potential of phonetic analysis of snoring, particularly formant frequency analysis, as a screening tool for pediatric snoring patients. While subjective grading of tonsil hypertrophy did not show strong statistical correlation with actual volume, formant frequency analysis provided valuable insights into airway structure changes. Further studies with larger sample sizes and controlled variables are needed to validate the clinical utility of snoring analysis as a non-invasive screening method for pediatric OSA.

Introduction

The main cause of obstructive sleep apnea is upper airway obstruction that occurs when the tension of the upper airway muscles decreases during sleep. The incidence of obstructive sleep apnea in children is 2%-4% [1,2], and the most common cause is hypertrophy of the palatine tonsils and adenoids, and tonsillectomy and adenoidectomy are accepted as the primary treatment for pediatric sleep apnea [3,4]. The final diagnosis of pediatric sleep apnea is polysomnography, but it is not commonly used due to cost issues [5], and various diagnostic techniques are used to supplement this. In the case of children, it is important to understand the medical history reported by the guardian and the child's nocturnal symptoms (habitual snoring, apnea, abnormal sleeping posture, etc.). The sensitivity of the sleep history and snoring history reported by the guardian tends to be consistent with the obstructive sleep apnea diagnosed by polysomnography in children, and in particular, the snoring history is an important clue in deciding whether to perform tonsillectomy in children [6,7]. Physical examination to evaluate the size of the palatine tonsils and adenoids is also important. The size of the palatine tonsils is closely related to obstructive sleep apnea and is known to affect sleep apnea symptoms including snoring. However, there is insufficient evidence to determine whether the subjective measurement method of palatine tonsil enlargement, which is commonly used today, accurately reflects the actual palatine tonsil size and volume, or whether the tonsil size is clearly related to sleep apnea symptoms such as snoring. Therefore, the authors conducted this study to determine whether the currently commonly used grading method of palatine tonsil enlargement is related to the actual palatine tonsil volume. In addition, the snoring sounds that are different for each patient were analyzed phonetically to determine whether the snoring sound is louder as the palatine tonsil volume increases, and whether it is related to the actual volume of the palatine tonsils and adenoids, and whether the phonetic analysis of snoring sounds is useful as a screening test for pediatric snoring patients.

Subjects and Methods

Subjects

This study was conducted on children aged 4 to 15 years who visited our hospital for tonsillectomy due to snoring symptoms. Medical history was collected from the guardians of the target patients, and physical examinations and lateral neck X-rays were performed before surgery. There were 21 patients, 14 boys and 7 girls, with an average age of 6.71 years and an average body mass index (BMI) of 19.27 kg/m2 (Table 1).

Table 1: Summary of Patient Characteristics, Tonsil Grading, and Tonsil Volume

Method

Measurement of tonsils and adenoids

The degree of hypertrophy of the palatine tonsils and adenoids was measured using subjective and objective methods. To reduce inter-examiner errors, the same examiner performed the tests, and the palatine tonsil hypertrophy was first evaluated using the Brodsky tonsil grade scale, which is commonly used before surgery [8].

For objective measurement of palatine tonsils, palatine tonsils obtained through intra-compartmental resection (en block) during surgery were weighed using the same scale and placed in a mass cylinder to measure the volume (Fig. 1). Adenoids were evaluated by the adenoid index, which is a value measured by the adenoid-nasopharynx ratio proposed by Fujioka et al. [9] on lateral neck radiographs (Fig. 2).

Fig 1: Measurement of tonsillar weight and volume after tonsillectomy. We measured the weight by placing the surgically removed tonsil on an electronic balance (A) and tonsil volume by putting the tonsil tissue removed by surgery into a measuring cylinder containing water (B).

Fig 2: Adenoid (A) was measured along the line perpendicular to point A’ (the maximal convexity along the inferior margin of the adenoid shadow) to its intersection, with line B drawn along the straight part of the anterior margin of the basiocciput. The nasopharyngeal space (N) was measured as the distance between the posterosuperior edge of the hard palate and anteroinferior edge of the sphenobasioccipital synchondrosis. The adenoid-nasopharynx ratio was calculated by dividing A by N.

Recording and phonetic analysis of snoring sounds

Snoring sounds were recorded by guardians after the patient fell asleep the night before tonsil adenoidectomy. To exclude differences between recording devices, a recorder of the same product (ICD SX2000; SONY, Tokyo, Japan) was used, and the recorder was placed within 50 cm of the patient's shoulder just before the patient fell asleep. Snoring sounds were recorded for at least 1 hour while maintaining ambient noise below 5 dB. For audio analysis of snoring sounds, the most prominent snoring sounds recorded during sleep were selected for three sections of 10 seconds or more, and the acoustic characteristics of all three sounds were analyzed using Praat software (ver. 5.2.16) and the average value was calculated. The sound sampling rate was 44 kHz, and sound intensity (dB) and formant frequency (Hz) were measured. Statistical analysis

A linear regression model was used to determine the relationship between quantitative values ??of the palatine tonsil and adenoid (tonsil volume, weight, adenoid index) and snoring voice analysis indices (formant frequency F1, F2, F3, F4, and sound intensity), and a dummy variable was used to determine the relationship between the categorical variable Brodsky tonsil grade and voice analysis indices. In addition, multiple regression analysis was performed by adjusting age and BMI for palatine tonsil volume, weight, and adenoid index, and a p value of 0.05 or less was considered statistically significant.

Results

Measurement of tonsil and adenoid

The average Brodsky tonsil grade and adenoid index on physical examination were 3.05 and 0.55, respectively, and the average palatine tonsil volume measured after surgery was 7.32 mL and the average weight was 14.58 g. When looking at the correlation between Brodsky tonsil grade and actual palatine tonsil volume, the average volume of extracted palatine tonsils in grade II patients was 6.83 mL, grade III was 7.02 mL, and grade IV was 8.75 mL. There was a general tendency to increase, but it was not statistically significant.

Phonetic analysis of snoring sound

There was no statistically significant correlation between snoring sound intensity and palatine tonsil volume, weight, Brodsky tonsil grade, and adenoid index. In multiple regression analysis, a significant positive correlation was observed only between snoring sound intensity and BMI (β=0.465, p=0.025).

As the volume of the palatine tonsil increased, the formant frequency of snoring significantly decreased for the first formant (F1) and the second formant (F2) at p;0.011 and 0.002, respectively. The same statistical significance was observed in the multiple regression analysis results adjusted for BMI and age (β=-0.603, p=0.002 for F1, β=-0.644, p=0.002 for F2). However, palatine tonsil weight and Brodsky tonsil grade were not significantly correlated with formant frequency and the adenoid index was not significantly correlated with formant frequency except for F4.

Fig 3: 1Bar Chart: Shows the average tonsil volume for different Brodsky tonsil grades with error bars.

2. Scatter Plot: Displays the relationship between BMI and snoring sound intensity.

3. Scatter Plot: Illustrates how formant frequencies (F1 and F2) change with increasing tonsil volume.

Discussion and Conclusion

The human voice is created through the larynx, which is the acoustic source of airflow from the lungs, and the vocal tract, which acts as a filter. When the vocal cords vibrate in the larynx, the airflow that becomes the source of sound is created, and the vocal tract shape is changed depending on the position of the tongue, the degree of jaw opening, and the shape of the lips, etc., to suppress sound energy transmission at some frequencies and allow maximum energy at other frequencies, thereby creating various sounds. The formant represents the specific energy distribution of the sound created in this way, and is called the first (F1), second (F2), and third (F3) formants from the lowest frequencies. Therefore, the formant acts as a fingerprint representing a specific sound and is affected by the vocal tract, and among them, the oral cavity and larynx have a great influence on the formant because individuals can freely change the space. Therefore, by measuring and analyzing the formant of the sound, the anatomical characteristics or deformations of the vocal tract that created the sound can be predicted [10]. Snoring is a noise that occurs when respiratory air passes through the airway during sleep, causing vibrations in the soft tissues such as the pharynx, soft palate, and uvula. Since it is modified by the nasal cavity and resonated by the oral cavity, it varies depending on the anatomical conditions of the upper airway for each patient, and its acoustic characteristics are clearly different from those of other types of sounds such as respiratory sounds [11]. Therefore, snoring can act as a unique characteristic value for each individual, and phonetic analysis of snoring can be a good diagnostic tool for predicting the anatomical shape of the upper airway. The history of phonetic analysis of snoring can be traced back to 1993, when Perez-Padilla et al. [12] first studied the differences in snoring between patients with simple snoring and those with obstructive sleep apnea, and in 1995, McCombe et al. [13] devised an acoustic index that can screen for obstructive sleep apnea. In 1996, Fiz et al. [14] attempted to analyze the spectrum of snoring sounds, and in 1999, Itasaka et al. [15] conducted a study on the intensity of snoring. There have been efforts to diagnose snoring through phonetic analysis. However, these concepts have not yet become widespread, and it is thought that more research will be needed in the future. The frequency of habitual snoring is reported to be 2.4%-34.5% in children [16], and the differences between studies are due to differences in the definition of habitual snoring, age groups of the target patients, and differences in snoring collection. There are still many differences in opinions between studies regarding differences by gender, and the authors plan to collect snoring through more cases and subdivide it by age, gender, etc. The definitive diagnosis of pediatric obstructive sleep apnea is polysomnography, but it is performed in approximately 20% of cases due to time-consuming and cost-intensive problems [5], and most cases depend on medical history and physical examination of local findings, so there is a need for a more rational and quantifiable screening test. Various methods have been proposed to clinically evaluate and quantify palatine tonsil and adenoid hypertrophy. Brodsky tonsil grade, Friedman scale, and modified 3-point/5-point scales are mainly used to measure palatine tonsil hypertrophy, and among them, Brodsky tonsil grade shows the highest inter-examiner agreement in measuring palatine tonsil hypertrophy [17]. The authors used Brodsky tonsil grade as a subjective test to evaluate palatine tonsil hypertrophy, and directly measured the volume and weight of palatine tonsils removed surgically as an objective test. In the case of adenoids, since they are often cauterized with a coblator rather than an en-bloc, the exact volume or weight cannot be measured, so the adenoid index was evaluated using a lateral cervical X-ray. Although the measurement of the volume of palatine tonsillary hypertrophy and adenoid hypertrophy has not been widely used, several studies have confirmed that the measurement of palatine tonsillary volume after surgery has a higher correlation with symptoms of sleep-disordered breathing than the measurement of subjective tonsillary size [18]. In the authors' case, there was a statistically significant result between palatine tonsillary volume and phonetic analysis of snoring, which suggests that the actual palatine tonsillary volume is a variable more highly correlated with the phonetic analysis of snoring than the subjective grade. In the authors' study, the actual palatine tonsillary volume tended to increase as the Brodsky tonsil grade increased, but the correlation between the two variables was not statistically significant. This is because it is difficult to accurately measure the anterior-posterior length of the palatine tonsilla, and it is difficult to accurately measure the size when it is deeply embedded in the tonsillary fossa. It is also necessary to consider that the observer's subjective thoughts may intervene when evaluating the grade.

The results of multiple regression analysis showed that the intensity of snoring was not statistically significantly related to the volume, weight, and adenoid index of the palatine tonsils, except for BMI. According to a study by Hwang et al. [19], the size of the palatine tonsils and adenoids was not significantly related to the severity of obstructive sleep apnea, which is consistent with the results of several previous studies that the size of the palatine tonsils and adenoids did not predict the severity of obstructive sleep apnea. In this regard, it can be assumed that the loudness of snoring in children is related to various factors such as obesity, lingual tonsil hypertrophy, allergic rhinitis, and craniofacial deformations related to the position and shape of the tongue and mandible in addition to the size of the palatine tonsils or adenoids [20]. While the formant frequency of snoring was not significantly related to the weight of the palatine tonsil and the adenoid index, as the volume of the palatine tonsil increased, F1 and F2 decreased statistically significantly (p = 0.011 and 0.002, respectively). The results of multiple regression analysis adjusted for BMI and age also showed statistical significance, which can be seen as a result that well reflects the changes in the vocal tract due to palatine tonsillary hypertrophy. F1 reflects the width of the vocal tract, and F2 reflects the length and position of the tongue, and both are inversely proportional. In general, as palatine tonsillary hypertrophy becomes more severe, there is a tendency to open the mouth and stick out the tongue to secure the airway during sleep, so it can be inferred that F1 and F2 decreased as the vocal tract widened and the tongue lengthened. In summary, the physical examination of palatine tonsillary hypertrophy tended to be somewhat consistent with the actual palatine tonsillary volume, but there was no statistical significance in this study. In addition, there was no correlation between snoring intensity and the actual palatine tonsillary volume, weight, or adenoid index. When snoring was analyzed phonetically, a significant negative correlation was confirmed between the actual palatine tonsillary volume and F1 and F2, which can be seen as a result that well reflects the vocal tract changes in patients with palatine tonsillary hypertrophy. These results confirmed the possibility of predicting palatine tonsillary volume and vocal tract changes due to palatine tonsillary hypertrophy through phonetic analysis of pediatric snoring, especially formant frequency analysis of snoring sounds. However, because the number of cases in this study was relatively small, it is thought that further statistical verification through more studies should be conducted in the future. In addition, as a limitation of this study, variables that may occur during the collection of snoring sounds, such as sleep stages, sleeping positions, sleeping environments, and surrounding noises, could not be systematically controlled, and I think this is a task that must be solved in order to establish it as a screening test in the future. With the recent trend of increasing interest in smartphone app development and remote diagnosis, it is thought that the utility of snoring voice analysis can be increased through more cases and studies in the future.

References

1. Shine NP, Coates HL, Lannigan FJ. Obstructive sleep apnea, morbid obesity, and adenotonsillar surgery: A review of the literature. Int J Pediatr Otorhinolaryngol. 2005; 69(11):1475–82.

2. Ali NJ, Pitson D, Stradling JR. The prevalence of snoring, sleep disturbance and sleep related breathing disorders and their relation to daytime sleepiness in 4-5 year old children. Am Rev Respir Dis. 1991; 143:A381.

3. Brouillette RT, Fernbach SK, Hunt CE. Obstructive sleep apnea in infants and children. J Pediatr. 1982; 100(1):31–40.

4. Frank Y, Kravath RE, Pollak CP, Weitzman ED. Obstructive sleep apnea and its therapy: Clinical and polysomnographic manifestations. Pediatrics. 1983; 71(5):737–42.

5. Acar M, Yaz?c? D, Bayar Muluk N, Hanc? D, Seren E, Cingi C. Is there a relationship between snoring sound intensity and frequency and OSAS severity? Ann Otol Rhinol Laryngol. 2016; 125(1):31–6.

6. Montgomery-Downs HE, O’Brien LM, Holbrook CR, Gozal D. Snoring and sleep-disordered breathing in young children: Subjective and objective correlates. Sleep. 2004; 27(1):87–94.

7. Brietzke SE, Gallagher D. The effectiveness of tonsillectomy and adenoidectomy in the treatment of pediatric obstructive sleep apnea/hypopnea syndrome: A meta-analysis. Otolaryngol Head Neck Surg. 2006; 134(6):979–84.

8. Brodsky L. Modern assessment of tonsils and adenoids. Pediatr Clin North Am. 1989; 36(6):1551–69.

9. Fujioka M, Young LW, Girdany BR. Radiographic evaluation of adenoidal size in children: Adenoidal-nasopharyngeal ratio. AJR Am J Roentgenol. 1979; 133(3):401–4.

10. Lee JH, Koo GJ, Koo HE, Kim YC, Koo SK, Lee SH, et al. An acoustic and radiologic study on voice change after tonsillectomy and adenotonsillectomy. Korean J Otorhinolaryngol-Head Neck Surg. 1999; 42(6):762–9.

11. Hoffstein V. Snoring. Chest. 1996; 109(1):201–22.

12. Perez-Padilla JR, Slawinski E, Difrancesco LM, Feige RR, Remmers JE, Whitelaw WA. Characteristics of the snoring noise in patients with and without occlusive sleep apnea. Am Rev Respir Dis. 1993; 147(3):635–44.

13. McCombe AW, Kwok V, Hawke WM. An acoustic screening test for obstructive sleep apnoea. Clin Otolaryngol Allied Sci. 1995; 20(4):348–51.

14. Fiz JA, Abad J, Jané R, Riera M, Mañanas MA, Caminal P, et al. Acoustic analysis of snoring sound in patients with simple snoring and obstructive sleep apnoea. Eur Respir J. 1996; 9(11):2365–70.

15. Itasaka Y, Miyazaki S, Ishikawa K, Togawa K. Intensity of snoring in patients with sleep-related breathing disorders. Psychiatry Clin Neurosci. 1999; 53(2):299–300.

16. Lumeng JC, Chervin RD. Epidemiology of pediatric obstructive sleep apnea. Proc Am Thorac Soc. 2008; 5(2):242–52.

17. Kumar DS, Valenzuela D, Kozak FK, Ludemann JP, Moxham JP, Lea J, et al. The reliability of clinical tonsil size grading in children. JAMA Otolaryngol Head Neck Surg. 2014; 140(11):1034–7.

18. Cahali MB, Soares CF, Dantas DA, Formigoni GG. Tonsil volume, tonsil grade and obstructive sleep apnea: Is there any meaningful correlation? Clinics (Sao Paulo). 2011; 66(8):1347–52.

19. Hwang SH, Guilleminault C, Park CS, Kim TW, Hong SC. Usefulness of adenotonsillar size for prediction of severity of obstructive sleep apnea and flow limitation. Otolaryngol Head Neck Surg. 2013; 149(2):326–34.

20. Gulotta G, Iannella G, Vicini C, Polimeni A, Greco A, de Vincentiis M, et al. Risk factors for obstructive sleep apnea syndrome in children: State of the art. Int J Environ Res Public Health. 2019; 16(18):3235.