The vocal tract includes various structures that begin from the vocal folds passing through the oropharynx ending at the lips, and the nasal cavity ending at the nostrils. Hence, theoretically, changes in any of these structures may affect the acoustic properties of the radiated sound signal.1,2
Nasal obstruction is one of the most common symptoms seen in a general otorhinolaryngology clinic. This symptom may be associated with multiple pathologies, including septal deviation or nasal polyposis.3–6 The shape of the septal cartilage, its curvature and relation to the vomer, may alter the pass of air entering through each nostril, but also the inverse pass of air from the velum to the fossa nasal. This deformity offers resistance to the airflow, which leads to respiratory difficulties, objectified by the patient. However, it can also produce a modification of nasal resonance during the production of some speech sounds.5,7 Septoplasty is a very frequent surgery, consisting of an intranasal correction of these deformities. It may modify the anatomy of the vocal tract, and thus affect the objective and subjective properties of the patient's voice, changing the nasal perception of voice, so called, nasality, as shown by different authors.7–15
Nasalance is calculated as the division between nasal acoustic energy and oral acoustic energy.8 It can be evaluated with the nasometer, which measures the nasal percentage, providing values close to 0% for those cases with strong hyponasalance, and thus, great nasal block. Depending on the values of nasalance, patients may have hypernasalance, hyponasalance, and mixed resonance. The 1st one is characteristic of patients with cleft palate, due to inadequate velopharyngeal closure during phonation. The hyponasalance, on the other hand, is conceptually correlated with the blocked nose.16 Hence, obstructive pathologies such as septal deviation can produce nasal block during the production of nasal consonants, causing hyponasalance.8,17
Following the well-known source-filter theory,1,18 the sound of a vowel is produced due to the vibration of the vocal folds. This produces a sinusoidal waveform with a discrete spectrum constituted by lines representing the harmonics, all of them separated from the fundamental frequency (F0) measured in Hz, and with an envelope that decreases by approximately 6 dB per octave. This sound travels through the supraglottic tract, where it undergoes a series of changes due to the resonances of the vocal tract, which acts like a filter. These changes enhance certain frequencies leading to the formants, and attenuate others, producing the antiformants, thus changing the envelope of the spectrum, but not its harmonic structure.19
The aim of this study is to evaluate the subjective nasality and objective nasalance changes, and the acoustic analysis in a group of patients before and after a septoplasty, using a control group to compare. This will allow to evaluate the impact of the surgery on the quality of the voice, the perception of the patient, and the acoustic characterization of phonation.
A prospective study was performed between January 2017 and June 2017 including patients operated on septoplasty and a control group, at a University Hospital. This study was approved by the ethics committee of the hospital and all participants signed informed consent.
Two groups of patients were defined: those having nasal obstruction and undergoing septoplasty (septoplasty group), and a control group without nasal obstruction (control group).
All patients were evaluated at baseline, 2 weeks after surgery, and 3 months later.
Exclusion criteria were any type of organic or neurological pathology related to phonation or speech or having undergone any previous otorhinolaryngological surgical procedure. The inclusion of patients was prospective, fulfilling the following criteria of inclusion: age over 18 years, with symptoms of nasal obstruction, with a diagnosis of severe septal deviation by rhinoscopy and nasal endoscopy. In these patients, septoplasty was performed under general anaesthesia, by the same surgeon. The surgery performed was following Cottle principles, designed to correct osteocartilaginous deviations, in all patients.
The controls were patients included in the surgical waiting list for some intervention with local anaesthesia that did not affect the upper airway.
Our protocol included demographic data (age, gender) main symptoms, weight and height, tobacco use, and the professional use of their voice (i.e. whether they have or not an intensive use of the voice in their professional activity).
Every visit, a GRBAS (Grade, Roughness, Breathiness, Asthenia, Strain) evaluation20 was also performed as a subjective assessment of the quality of the patient's voice. Patients with pathological GRBAS or with evident pathology in the vocal folds observed in the nasofibroscopy were excluded from the study.
Every visit, a nasality questionnaire published by Liapi et al10 and previously validated into Spanish, was used with the goal to measure the degree of subjective nasality. It consists of 13 questions about nasal symptoms, valued from 0 to 4 points for each item, providing a final score, in which the higher, the more degree of nasal obstruction the patient has.
Also, at every visit, nasalance was also measured with Nasometer II model 6450.16 It has a headset, that is placed between the nose and mouth of the patient, which gives a nasal percentage, as reported elsewhere.17
Finally, at every visit, voice recording was made in a room conditioned with soundproofing, an AKG C420 headset microphone, and a Soundblaster Live 24 bits sound card connected to a personal computer equipped with the PRAAT software. The recording protocol was the same in the 3 sessions, collecting 3 utterances of the vowel /a/ with a duration of at least 3 seconds, and then 3 more of the remaining Spanish vowels with a duration of 1 second per vowel (/e/, /i/, /o/, /u/), with pauses for breathing between each of the vowels. The voice was recorded at comfortable pitch and loudness.
The acoustic analysis of the vowels was made using the tools provided in the WPCVox package and using ad-hoc algorithms developed to run in a Matlab environment. The WPCVox21,22 is an acoustic analysis software oriented to the field of voice quality measurement in clinical environments, professional voice and scientific research. The following parameters were extracted using this software: fundamental frequency (F0), Jitter, Shimmer, Harmonic to Noise Ratio (HNR) and Noise to Harmonic Ratio (NHR). The Matlab-based software has been used to obtain different parameters related to the resonance of the vocal tract, following the algorithms described by Mehta et al in 2012.23 The parameters obtained are: F1, F2, and F3 formants; antiF1, antiF2, and antiF3 antiformants; and antiF1_BW, antiF2_BW, and antiF3_BW bandwidth antiformants. All these parameters have been calculated for all vowels (/a/, /e/, /i/, /o/, /u/).
Subsequently, the review of the calculated parameters and those measured for the 3 sessions was carried out, with a statistical analysis study performed with STATA 11.1 software. The amplitude and frequency perturbations, noise parameters, formant and antiformants values (antiF1-F3) and their bandwidths (antiF1-F3 BW), nasalance questionnaire, nasalance measured with the nasometer, and GRBAS were also analysed, comparing both quantitative and qualitative data between the different samples obtained. The GRBAS index was measured as a continuous quantitative variable in Table 1 for the pre-surgical comparison between groups, but then analysed as a categorical variable (adding the total of the values in each GRBAS category and giving a value that in the series ranged from 0 to 3, at most). We applied Wilcoxon's test to quantitative variable with no normal distribution or Fisher's exact test for categorical variables. Hence, we used ANOVA to variables with homogeneous variances or Kruskal Wallis's test to variables with heterogeneous variances. Holm-Bonferroni correction was used when multiple repeated measures were done. This P-value was calculated by Fisher's Exact Test.
A total of 71 patients distributed in 2 groups were assessed: those operated on septoplasty; and the control group. Throughout the study there were a total of 13 patients that were lost, due to lack of follow-up by the patient or alterations in the recording, mainly sound interferences that invalidated the sample (9 in the control group, and 4 in the septoplasty group). Finally, a total of 58 patients had all visits and recordings available to be evaluated, 31 in the septoplasty group, and 27 in the control group. After the septoplasty, only in 3 patients there was a small osteocartilaginous foot as a complication. In these patients, a rhinomanometry was performed to evaluate its importancee, being normal in all 3 cases, so they remained in the study, without confunding the observed data.
The demographic data are shown in Table 1 (See Supplemental Digital Content, Table 1, http://links.lww.com/SCS/A449). Differences can be observed in gender and age, because this surgery is more frequent in men and young people. Due to these differences in both groups, it was necessary to perform an analysis of variance (ANOVA) to determine the influence of these factors on the data. After this ANOVA, these differences were not significantly influential.
In Table 1 we observed baseline data, showing significant differences in GRBAS between both groups (1.32 mean in septoplasty group versus 0.63 mean in control group). In the nasality questionnaire and nasometry we also found significant differences between the septoplasty group and the control group: subjective nasality in questionnaires is higher in the septoplasty group than control group, whereas nasometry values are lower in the septoplasty group, both showing that nasal perception of voice is higher in patients with nasal obstruction.
The data obtained comparing the 2 groups in the 3 sessions (baseline, 2 weeks and 3 months), concerning GRBAS, subjective nasalance and objective nasalance, are shown in Table 2 (See Supplemental Digital Content, Table 2, http://links.lww.com/SCS/A450). A significant change for both GRBAS, subjective nasality and objective nasalance in the septoplasty group, was found, whereas there were no significant changes in the control group with an increase in patients with scores of 0 was observed, and a decrease in the rest.
In respect of the nasality questionnaire, a decrease in the score was seen, revealing a subjective improvement of the patients operated, if compared with the control group (Fig. 1A). Regarding objective nasalance, an increase of nasometry percentage was seen at 2 weeks after surgery and a return to normal values before 3 months (Fig. 1B).
With reference to the acoustic analysis, neither F0, HNR, NHR, Jitter nor Shimmer varied in any cohort. However, differences were found in other values less studied in the literature, such as in the antiformants, especially in antiF3, antiF2, and the bandwidth of the antiformants.
The study of the vocal tract and its modification and influence in GRBAS, nasalance and acoustic parameters has been published by different authors, with a wide controversy in their results.
Some studies have used the GRBAS scale to evaluate the voice after nasal procedures.8 It provides a general overview of the voice quality. However, nasal aspects of the voice are not considered in the scale.20 Nevertheless, in our study, GRBAS evaluation revealed in the septoplasty group a significant change between baseline and the 2 postoperative measures, showing an increase of the values 0 (improvement of the voice subjectively), with a decrease of the rest. Liapi et al. introduced the nasality questionnaire.10 These authors measured and analysed each question before and after septoplasty, and compared it with a control group, with significant results, as an alternative to the Voice Handicap Index (VHI) for this type of patients. The VHI was not used in our study because it is considered a questionnaire of vocal quality, more focused on the study of vocal folds pathologies, not considering aspects of resonances and/or nasality. In our work, for the nasality questionnaire, once translated into Spanish, a significant difference was found in the septoplasty group, with an improvement of the pre-surgical results in comparison to those 3 months later, contrasting with the control group. These results are expected after a septoplasty without complications, since it does improve the nasal symptoms, and the sensation of breathing better, obtaining a lower punctuation on the test.
Nevertheless, with reference to nasalance, measured with the nasometer, an increase in their values was observed 2 weeks after surgery, decreasing after 3 months with significant differences in the septoplasty group (Fig. 1B). This finding has already been observed in other studies, such as Hong et al24 who reported on patients operated on polyposis whose nasalance increased from 45.7 to 57.8%, at 3 weeks after surgery. Similarly, Soneghet et al4 found an increase of nasalance from 38 to 42.9%, at 5 weeks after nasal surgery. On the contrary, Kim et al8 also found an increase in nasalance a month after surgery, but a decrease to the preoperative values at 3 and 6 months to the surgery, as we found in our study. Amer et al25 have also found a similar transitory increase in nasalance after septoplasty that resolves after 3 months. This increase of nasalance after the 1st 2 weeks can result from an intensification of the nasal acoustic energy, which is due to the decrease in the resistance to the airflow; but then it tends to return to its preoperative value, due to normalization of the mucosa.
Many authors have studied acoustic changes after other surgeries of the vocal tract, with contradictory results again. Mora et al7 concluded that there is a significant difference in several acoustic parameters (F0, Jitter, Shimmer, and others), between a group who had a septoplasty done and a control group, a month after surgery. Liapi et al10 conducted a study on 15 patients who underwent septoplasty, comparing them with a control group, measuring F0, jitter, shimmer, and other parameters, without finding significant differences. Later, Atan et al13 performed a study on 43 patients operated on septoplasty. They concluded that in patients with severe septal deviation, after surgery, there is a significant change in the Fo. In none of these studies, formants or antiformants were analysed.
Formants were analysed by Gulec et al14 on 50 patients who underwent septoplasty comparing them with a control group. They concluded that there were no significant changes in any acoustic parameters studied, such as the F0 or formants. Ozbal Koc et al9 also studied acoustic parameters, including formants, without finding significant differences in F0, Jitter, Shimmer or formants (F1-F4). Antiformants were not analysed in these works.
In our study, no significant differences are found in the perturbation and noise parameters, following the idea that no changes have appeared in the source due to the effects of the septoplasty. In addition, no changes are found in the formants. Regarding the antiformants (antiF1-F3) or the antiformants bandwidth (antiF1-F3 BW), which are not widely used in the literature and are important indicators of nasalization, some pre and post-surgical differences have appeared. We have observed that antiF3 varies in all the vowels studied. However, the modifications seen in the antiformants and their bandwidth, even when they are significant from the statistical point of view, they are not significant from the physical and/or clinical perspective since the deviations found are minimal (antiF3 is modified less than 1%).
It is important to highlight that in the Spanish language, no nasal vowels exist. As we have done in our study, the nasalization effect of a vowel is only visible as a result of coarticulation between vowels and adjacent nasal consonants (e.g. in the /na/ syllable), due to the velar lowering gesture associated with the nasal consonant, which overlaps with the vowel. Thus, a Spanish patient uttering a sustained vowel at comfortable pitch and loudness and with no other specific production requirements produces an oral vowel, which means that the acoustic material recorded for this study does not contain any contribution of the nasal tract.
As limitations, it is a before and after study on a nested cohort, not being an experimental study. In the same way, differences were found in the age, gender of both groups and GRBAS evaluation. Due to the differences between both groups, it was necessary to perform an analysis of variance (ANOVA) to determine the influence of these factors on the data, not being significant. Furthermore, more studies on antiformants and nasalized vowels are necessary using an acoustic material which would ensure nasalization by introducing a coarticulation with a nasal consonant. In addition, although longer than in other studies, our follow-up of 3 months may be not enough to provide certain evidences. As a conclusion, septoplasty produces changes in the vocal tract, with an increase in initial nasalance but with a subsequent normalization. However, regarding the acoustic analysis, although changes could be expected, these are minor, and we cannot conclude that they are due to the septoplasty.
1. Stevens K, Kenneth N. Acoustic phonetics. 2000; Cambridge (MA): MIT Press, pp. 30.
2. Hirano M. Clinical examination of voice. 1981; Viena: Springer-Verlag, pp. 100.
3. Hosemann W, Göde U, Dunker JE, et al. Influence of endoscopic sinus surgery on voice quality. Eur Arch Otorhinolaryngol
4. Soneghet R, Santos RP, Behlau M, et al. Nasalance
changes after functional endoscopic sinus surgery. J Voice
5. Behrman A, Shikowitz MJ, Dailey S. The effect of upper airway surgery on voice. Otolaryngol Head Neck Surg
6. Acar A, Cayonu M, Ozman M, et al. Changes in acoustic parameters of voice after endoscopic sinus surgery in patients with nasal polyposis. Indian J Otolaryngol Head Neck Surg
7. Mora R, Jankowska B, Dellepiane M, et al. Acoustic features of voice after septoplasty
. Med Sci Monit
8. Kim YH, Lee SH, Park CW, et al. Nasalance
change after sinonasal surgery: analysis of voice after septoturbinoplasty and endoscopic sinus surgery. Am J Rhinol Allergy
9. Ozbal Koc EA, Koc B, Ercan I, et al. Effects of septoplasty
on speech and voice. J Voice
10. Liapi A, Hirani S, Rubin J. Changes in nasal resonance following septoplasty
in adults: acoustic and perceptual characteristics. Logoped Phoniatr Vocol
11. Kim SD, Park HJ, Kim GH, et al. Changes and recovery of voice quality after sinonasal surgery. Eur Arch Otorhinolaryngol
12. Arslan F, Polat B, Durmaz A, et al. Effects of nasal obstruction due to nasal polyposis on nasal resonance and voice perception. Folia Phoniatr Logop
13. Atan D, Özcan KM, Gürbüz ABT, et al. The effect of septoplasty
on voice performance in patients with severe and mild nasal septal deviation. J Craniofac Surg
14. Gulec S, Kulahli I, Ilhan M, et al. Effect of septoplasty
on voice quality: a prospective-controlled trial. Clin Exp Otorhinolaryngol
15. Göker AE, Aydoğdu I, Saltürk Z, et al. Comparison of voice quality between patients who underwent inferior turbinoplasty or radiofrequency cauterization. J Voice
16. Awan SN, Virani A. Nasometer 6200 Versus Nasometer II 6400: Effect on measures of nasalance
. Cleft Palate-Craniofacial J
17. Mayo R, Floyd LA, Warren DW, et al. Nasalance
and nasal area values: cross-racial study. Cleft Palate-Craniofacial J
18. Gray SD, Hirano MSK. Titze L. Mollecular and cellular structure of vocal fold tissue. Vocal fold physiology: frontiers in basic science
. San Diego: Singular Publishing Group, Inc; 1993. 1–36.
19. Warren DW, Dalston RM, Mayo R. Huffman MK, Krakow RA. Aerodynamics of nasalization. Nasals, nasalization and the velum
. San Diego: Academic Press; 1993. 119–147.
20. Núñez Batalla F, Corte Santos P, Sequeiros Santiago G, et al. Evaluación perceptual de la disfonía: correlación con los parámetros acústicos y fiabilidad. Acta Otorrinolaringol Esp
21. Godino-Llorente JI, Saenz-Lechon N, Osma-Ruiz VVV, et al. An integrated tool for the diagnosis of voice disorders. Med Eng Phys
22. Godino-Llorente JI, Osma-Ruiz V, Sáenz-Lechón N, et al. Acoustic analysis
of voice using WPCVox: a comparative study with Multi Dimensional Voice Program. Eur Arch Oto-Rhino-Laryngology
23. Mehta DD, Rudoy D, Wolfe PJ. Kalman-based autoregressive moving average modeling and inference for formant and antiformant tracking. J Acoust Soc Am
24. Hong KH, Kwon SH, Jung SS. The assessment of nasality with a nasometer and sound spectrography in patients with nasal polyposis. Otolaryngol Head Neck Surg
25. Amer HS, Elaassar AS, Anany AM, et al. Nasalance
changes following various endonasal surgeries. Int Arch Otorhinolaryngol