525595
research-article2014
AORXXX10.1177/0003489414525595Annals of Otology, Rhinology & LaryngologyBrandt et al
Article Annals of Otology, Rhinology & Laryngology 2014, Vol. 123(8) 564–570 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0003489414525595 aor.sagepub.com
Impact of Nasal Surgery on Speech Resonance Michael G. Brandt, BSc, MD1,2, Brian W. Rotenberg, MD, MPH2, Corey C. Moore, MD, MSc2, Catherine C. Bornbaum, PhD3, Agnieszka Dzioba, PhD3, Jordan T. Glicksman, BSc, MD3, and Philip C. Doyle, PhD2,3
Abstract Objectives: The nose and paranasal sinuses contribute to speech resonance and changes to these structures may alter speech nasality. This change may influence one’s vocational and social functioning and quality of life. Our investigation explored objective and subjective changes in nasality following nasal surgery in a prospective and longitudinal fashion. Methods: Recordings of sustained vowel and sentence stimuli and voice-related quality of life measurements were obtained preoperatively and at 2, 4, 8, and 24 weeks postoperatively from individuals undergoing nasal and/or sinus surgery. Objective measures of fundamental frequency, jitter, shimmer, and harmonic to noise ratio (HNR) were determined. Preand postoperative speech samples were assessed by 15 naïve listeners. Results: In all, 15 subjects completed the study. Neither speakers nor listeners perceived a subjective change in nasality following surgery. No statistically significant change in microacoustic measures were identified. Although nasal sentences did not reveal differences for 3 microacoustic measures, a difference in HNR was identified. Conclusions: Patients undergoing nasal surgery did not exhibit subjective changes in resonance postoperatively. Aside from a difference in HNR for the nasal sentence, objective microacoustics remained unchanged. These results demonstrate the stability of oranasal resonance despite nasal surgery and provide valuable data for patient informed decision-making. Keywords nasality, resonance, acoustics, nasal surgery, endoscopic sinus surgery, rhinoplasty To date only a limited number of studies have sought to evaluate the impact of nasal surgery on the speech signal.1-5 Although these investigations have elucidated some relationships, the overall impact of nasal procedures on speech quality remains poorly understood. Of specific interest is the potential influence of nasal surgery on speech resonance—defined as “the quality of the voice that is determined by the balance of sound vibration in the oral, nasal, and pharyngeal cavities during speech.”6 In general, the literature has suggested a surgical effect on nasalized vowels1,2,5 and nasal consonants,1,4,5 focusing on the measurement of nasalance (the ratio of nasal to oral acoustic energy),3,4 and the overall resonance of the vocal tract.1,2,5 Acoustic analysis has involved the assessment of frequency and amplitude differences between the first resonant frequency (first “formant”) which is an important acoustic descriptor of vowels, and the peak or murmur associated with nasal resonance amplitudes. An excellent review by Behrman and colleagues5 highlights the outcomes of the aforementioned studies and goes on to demonstrate similar findings of increased nasal resonance for both consonants
and vowels following nasal surgery. This information supports the notion that changes in nasal resonance can be objectively evaluated. Consequently, the ability to track acoustic changes and their potential variability holds great promise at both basic and applied levels of evaluation. Many articles comment on the subjective identification of speech/voice “quality” changes among patients post– nasal surgery.1-3,5 Of these reports, only 1 evaluated the impact of these changes on voice-related quality of life 1
Department of Otolaryngology–Head and Neck Surgery, Division of Facial Plastic & Reconstructive Surgery, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada 2 Rehabilitation Sciences, Faculty of Health Sciences, Western University, London, Ontario, Canada 3 Department of Otolaryngology–Head and Neck Surgery, Schulich School of Medicine and Dentistry, Western University, London, Ontario, Canada Corresponding Author: Brian W. Rotenberg, MD, MPH, St. Joseph’s Healthcare Centre, 268 Grosvenor St, London, Ontario, Canada, N6A 4V2. Email: [email protected]
565
Brandt et al (QOL) using the Voice Handicap Index (VHI).5,7 While the VHI was originally designed specifically to assess the impact of vocal/laryngeal based abnormalities on perceived QOL, it is not uncommon for patients to refer more generically to “voice” or “speech” problems as opposed to using the more technical terms of “resonance” or “nasality”; thus, use of the VHI or similar instruments to evaluate changes in those undergoing nasal surgery was not inappropriate despite its primary design and clinical use. Using the VHI, the aforementioned study indicated that 20% of the participants noted a positive change in their voice postoperatively.5 These results were found with combined procedures to the upper airway and did not specifically focus on nasal surgery; thus, it is difficult to extrapolate these results and reliably differentiate between specific surgical approaches and any resultant changes secondary to nasal surgery. Although some patients may perceive postoperative changes to the quality of their speech as an interesting curiosity, the professional voice user or the nonprofessional individual who experiences high demand for voice use may find even minor modulations to their trademark or everyday voice to have disastrous personal and social implications. Thus, it is imperative to better define the true impact of nasal surgery on voice/speech quality so as to better inform patients and predict alterations that may occur as a consequence of changes in resonance and associated nasality. Information obtained may also help to reduce patient fears or concerns specific to a variety of domains of interest such as acoustic, auditory-perceptual, or changes that influence self-assessed QOL. Thus, the objective of our study was to empirically determine the influence of nasal surgery on speech resonance in a prospective and longitudinal fashion.
Methods Institutional ethics approval was obtained from the University of Western Ontario Research Ethics Board (HSREB# 12639E).
Participants Participants included any individual presenting to a tertiary care Otolaryngology–Head and Neck Surgery practice considering rhinoplasty, septoplasty, and/or endoscopic sinus surgery for chronic sinusitis. Patients with sinonasal disease other than chronic rhinosinusitis (ie, pituitary tumor, papilloma, etc) were excluded given the increased variability of their surgical procedure. Only those participants fluent in written and spoken English were voluntarily recruited to participate. Participants were excluded from study involvement if they had a preexisting speech abnormality or preexisting anatomic laryngeal abnormality so as to prevent any confounding influence on signal analysis and perceptual assessment. Patients with nasal polyposis underwent CT
imaging and had their sinus opacification documented and graded (Lund–Mackay stage)8 to control for their influence on resonance outcomes. Patients with nasal obstruction due to causes not pertaining to chronic rhinosinusitis (eg, nasal fracture) did not undergo Lund–Mackay staging.
Research Protocol Patients meeting the inclusion criteria and agreeing to participate were assessed at 5 predetermined evaluation intervals: (1) preoperatively, (2) 2 weeks postoperatively, (3) 4 weeks postoperatively, (4) 8 weeks postoperatively, and (5) 24 weeks (6 months) postoperatively. At each assessment the participant was asked to a. Complete the Voice-Related Quality of Life (V-RQOL)9 questionnaire b. Complete a full speech recording protocol
Data Acquisition—Speech Stimuli and Recording Each participant was recorded at multiple points over the course of the study. Recordings were obtained prior to surgery, then at 2, 4, and 8 weeks postsurgery, with a final measure occurring 6 months postsurgery. During each recording session, participants were asked to perform 3 specific tasks. First, participants were asked to produce 5 samples each of 3 vowels, /a/, /i/, and /u/. Because of their relatively static nature in the context of a quasiperiodic vocal signal, vowels provide an ideal opportunity to evaluate microacoustic characteristics of the vocal signal. Participants were instructed to sustain 3 samples of each of these vowels as steadily as possible at their conversational pitch and loudness level for approximately 6 to 8 seconds. Participants were permitted time to practice the task prior to recording of the vowel stimuli with immediate and direct feedback provided as necessary by the individual gathering recordings (C.C.B., A.D., P.C.D.). Following the sustained vowel task, participants were asked to read aloud the Rainbow Passage10 and 2 sets of sentences (one of which contained a high number of nasal consonants in the sentence, and the other without any nasal consonants). Again, participants were instructed to speak at a conversational pitch and loudness level; participants were able to familiarize themselves with the passage prior to formal recording. Two recordings of the Rainbow Passage were gathered along with 2 productions each of sentence stimuli (both nasal and non-nasal sentences). The third standardized vocal task11 consisted of recording a short monologue (approximately 30 seconds in length), produced once again at a typical level of pitch and loudness. Although the content of this monologue was not controlled, all participants were asked to speak on the general topic of a hobby or special interest that they enjoy and/
566 or participate in. At the conclusion of each recording session, participants were asked to grossly assess their selfperceived level of the “acceptability” of their voice/speech quality using a simple visual analogue scale; this assessment was obtained at the end of all recording sessions. This task served solely as an index of a potentially self-perceived change that could lead the investigators to more extensive evaluation of a given individual’s data. All recordings were gathered in a sound-treated environment, free of ambient noise. The full recording procedure for any given session required approximately 20 minutes. Following completion of the recording, each participant was asked to complete the 10-item V-RQOL9 measure. Briefly, the V-RQOL asks the respondent to answer each question regarding specific issues of voice use by rating it from 1 (not a problem) to 5 (problem as bad as it can be). These scores were then transformed according the V-RQOL algorithm and used as a self-rating measure for each participant at the time of all recordings
Instrumentation All speech recordings were acquired digitally using the Kay Sona-Speech II System—Model 3650 (KayPentax, Montvale, New Jersey, USA). Samples were recorded using a unidirectional microphone that was positioned at a fixed distance of 6.0 inches from the participant’s mouth. The mouth to microphone distance was rechecked before and after every experimental task to ensure consistency. The microphone was routed to a computer through a microphone preamplifier (Kay Elemetrics Model 3708, KayPentax, Mississauga, Canada) allowing input levels to be adjusted such that signal strength reached approximately 10 to 13 LEDS on the Multi-Dimensional Voice Profile system. All recordings were sampled at 44 KHz and stored directly to the computer hard drive for later conversion and acoustic analysis.
Acoustic Analysis of Data All vowel data were evaluated for basic microacoustic measures of fundamental frequency, frequency perturbation (jitter), amplitude perturbation (shimmer), and harmonic to noise ratio (HNR), and these were gathered from the extracted segments of the steady-state vowel entities. This involved the analysis of a 2-second segment extracted from the midpoint of each vowel sample. Care was taken to extract a vowel segment from the middle of samples to eliminate acoustic variation due to vocal onset and offset. Acoustic measures were collected for each vowel independently using a PRAAT (version 5.3.29) an open-source acoustical analysis program (http://www.fon.hum.uva.nl/ praat/). Once all analyses of each participant’s vowels were completed, the middle frequency value for each of the 3
Annals of Otology, Rhinology & Laryngology 123(8) samples of each vowel was then used for statistical analysis. For the running speech stimuli, only sentence stimuli were acoustically analyzed. The second sentence of the Rainbow Passage was extracted from the entire passage and was then used in the perceptual phase of this project. All stimuli were acoustically evaluated independently and in a blinded fashion by a single researcher to maintain consistency in measurement procedures for each of the experimental tasks.
Auditory-perceptual Evaluation of Speech Stimuli The second sentence of the Rainbow Passage (12 words) was submitted to perceptual evaluation by a group of 15 naïve, normal-hearing young adult listeners. None of the listeners had any formal exposure to or educational exposure to the area of voice, speech, or resonance disorders. The assessment of the Rainbow Passage stimuli involved an assessment of potential changes in resonance, and more specifically in nasality. To gather these auditory-perceptual data, a scaling procedure designed specifically for this study was used. This involved rating each sample using a scale that ranged from hyponasal to hypernasal, thus, providing the opportunity to identify a range of potential changes in nasal resonance characteristics outside of the normal expectation. The midpoint of the scale was identified to represent “normal nasality” (ie, a normal balance of oral and nasal resonance) to permit the listener to make judgments of both types of abnormal nasality using a single visual analogue (VA) scale. Our decision to use a VA scale and include the entire continuum including the hyponasal side of nasality in this perceptual scaling procedure was done to ensure that any potential change in resonance (hypo or hypernasal variation) could be accounted for over the course of the study. In addition, by using a VA scale, concerns related to whether the nasality continua are linear or curvilinear could be accounted for accordingly.12 Based on this VA scale, a marking to the far left of the scale would represent a numeric value of 1 (hyponasality), with a score to the far right being equal to a numeric value of 100 (hypernasality). Thus, as rater markings on the VA scale moved toward the middle of the scale from either end of the scale, they would represent increasing levels of “normal” resonance. Conversely, as scales scores departed from the middle of the scale, they would represent increased levels of either hypo or hyper nasal resonance. Our design and use of this unique scale was undertaken in an effort to gather perceptual data that spanned the continuum of nasality. Operational definitions of each anchor term were provided to the listeners, as were recorded examples of extreme hypernasality and hyponasality. Following completion of this assessment, the numerical value for each rating was determined from direct measurement in millimeters of the VA scale. These perceptual data were then submitted to statistical analysis.
567
Brandt et al
Data Acquisition—V-RQOL The V-RQOL is a standardized 10-item questionnaire initially developed by Hogikyan and Sethuraman9 to quantify potential vocal difficulties specific to a range of situations. This assessment instrument is well validated and directly focuses on the impact of voice use on QOL with specific attention directed at physical functioning concerns (eg, capacity for loudness, etc) and social-emotional impact relative to one’s voice use (eg, restriction of activities, etc). Given the impact of nasal resonance on QOL demonstrated in previous reports,5 we chose to include this measure to objectively evaluate the “voice-related” impact of nasal surgery among our participants over the course of this study. Similar to the collection of speech recordings, each participant was asked to complete the V-RQOL presurgically, then again at 2, 4, and 8 week postsurgically, then finally at 6 months postoperatively.
Data Acquisition—CT Scan Evaluation Those patients undergoing endoscopic sinus surgery for chronic rhinosinusitis underwent standard preoperative CT imaging. These scans were evaluated and scored using the standardized Lund–Mackay staging system for chronic rhinosinusitis by a single observer familiar with the Lund– Mackay staging system (J.T.G.).8
Statistical Analysis Prior to commencement of this study, we determined that a total sample size (N) of 14 individuals would be sufficient to detect the hypothesized effect (r2 = .20) of a 2-level within-subject independent variable 83.2% of the time using a .05 alpha level, assuming a within-subject correlation of .30. This calculation would apply to all independent measures gathered (acoustic, auditory-perceptual, and V-RQOL data). All statistical analyses were performed using SPSS version 18 (Minneapolis, Minnesota, USA). Acoustic measures obtained from vowel samples were compared independently for samples of /a/, /i/, and /u/ using an analysis of variance (ANOVA) with repeated measures across the preoperative and 4 postoperative time periods (2, 4, and 8 weeks and 6 months). All remaining acoustic measures were evaluated using a between-groups ANOVA with repeated measures over the 4 data collection periods. Similar analyses for both the physical, socialemotional and total V-RQOL scores were also conducted. An a priori probability level of .05 was employed. Correlational analyses were performed for the V-RQOL measures and the perceptual assessment data obtained from naïve listeners, and also evaluated these measures against the preoperative Lund–Mackay sinus opacification score. Measurement reliability and consistency was
Table 1. Demographics and Procedures Performed Among the Patient Population Evaluated for Postoperative Voice Changes. Mean age, y Sex Male Female Diagnosis Nasal obstruction and deformity Nasal obstruction CRS with polyps Presented with nasal obstruction Prior sinonasal operation Procedure Septorhinoplasty Septorhinoplasty and turbinatoplasty FESS and septoplasty FESS Septoplasty and turbinatoplasty Smoking history Active smoker Past smoking history LMS average among CRS patients (n = 7)
52 (SD, 16.8; range, 21-83) 11 4 4 4 7 15 2 3 3 1 6 2 1 2 15.1 (95% CI, 13.8-16.5)
Abbreviations: CI, confidence interval; CRS, chronic rhinosinusitis; FESS, functional endoscopic sinus surgery; LMS, Lund–Mackay score; SD, standard deviation.
determined via the use of Cronbach’s alpha and assessment of point-by-point agreement.
Results Fifteen patients met all study criteria and took part in the study to its conclusion. Their demographic information and related findings are summarized in Table 1. Of the 15 patients, there were 11 males and 4 females. The mean age was 52 years (SD = 16.8; range, 21 to 83). The presenting symptom among all 15 patients was nasal obstruction; with 4 having an external nasal deformity, and 7 demonstrating chronic rhinosinusitis with nasal polyposis. Operative procedures included septorhinoplasty with and without turbinoplasty (cautery and out-fracture of the inferior turbinates) (n = 3 and n = 3), endoscopic sinus surgery (ESS) with and without septoplasty (n = 1 and n = 6, respectively) and septoplasty with turbinoplasty (n = 2). The studied intervention was the first sinonasal operation in 13 of the patients studied. Procedures were performed on 1 active smoker and 2 former smokers. The average Lund–Mackay score among patients with polyposis was 15.1 (n = 7, 95% CI = 13.816.5). None of the patients reported any subjective change in their speech at any time postoperatively.
568
Figure 1. Harmonic to noise ratios over time per group.
Acoustic Analysis—Vowels Full microacoustic measures (F0, jitter, shimmer, and HNR) were obtained independently for each vowel (/a/, /i/, /u/) and then submitted to statistical analysis. Analyses were performed on each vowel independently for women and men. For all 4 microacoustic measures for each of the 3 vowels assessed, no significant difference was observed over time due to surgical intervention for either male or female participants.
Acoustic Analysis—Rainbow Passage The Rainbow Passage allowed for analysis of fundamental frequency, jitter, shimmer, and HNR scores. Mauchly’s test was used to evaluate jitter scores and indicated that the assumption of sphericity was not violated, χ2(9) = 10.404, P = .323. Thus, the results suggest that there was no significant change in the jitter scores over time following nasal surgery, F(4, 56) = 1.556, P = .199. Relative to the shimmer scores, Mauchly’s test indicated that the assumption of sphericity had been violated, χ2(9) = 19.655, P < .05; therefore degrees of freedom were corrected using Greenhouse–Geisser estimates of sphericity (ε = .576). The results were then evaluated and identified that there was no significant change in shimmer when comparing preoperative to postoperative voice samples, F(2.305, 32.264) = 0.265, P = .789. Once again, Mauchly’s test was used to evaluate HNR scores and indicated that the assumption of sphericity was not violated, χ2(9) = 11.995, P = .217. These results indicate that there was no significant change in HNR scores across the preoperative and postoperative samples, F(4, 56) = 1.400, P = .246. Figure 1 shows the results over time.
Acoustic Analysis—Nasal Sentence Similar to the Rainbow Passage stimuli, an analysis of fundamental frequency, jitter, shimmer, and HNR scores were
Annals of Otology, Rhinology & Laryngology 123(8) performed on the nasal sentence samples. Relative to the jitter scores, Mauchly’s test indicated that the assumption of sphericity was not violated, χ2(9) = 12.145, P = .209. The results indicate that there was no significant change in jitter scores following nasal surgery, F(4, 56) = 1.296, P = .283. Relative to the shimmer scores, Mauchly’s test indicated that the assumption of sphericity was not violated, χ2(9) = 8.244, P = .514. The results indicate that there was no significant change in shimmer scores during healing from nasal surgery, F(4, 56) = 2.192, P = .082. Finally, with respect to the HNR scores, Mauchly’s test indicated that the assumption of sphericity was not violated, χ2(9) = 13.082, P = .162. These results indicate that there was a significant change in HNR scores when comparing pre- and postsurgical voice samples, F(4, 56) = 4.571, P = .003.
Auditory-Perceptual Evaluation of Rainbow Passage Preoperative and 6-month postoperative voice recordings were subjectively evaluated and rated by naïve listeners. Pre- and postoperative subjective ratings were then statistically evaluated using a paired t test. Results of this analysis revealed that no significant difference in voice quality was demonstrated based on listener judgments of voice samples (t = 1.678, df 14, P = .116).
VRQOL With respect to the social-emotional (SE) subscore, Mauchly’s test indicated that the assumption of sphericity had been violated, χ2(9) = 54.496, P < .05; therefore, degrees of freedom were corrected using Greenhouse– Geisser estimates of sphericity (ε = .367). The results indicate that there was no significant difference within the SE subscore of the VRQOL throughout the postoperative period, F(1.469, 20.563) = 1.683, P = .212. Hence, nasal surgery did not significantly influence one’s perceived SE score on the V-RQOL. Regarding the physical domain, Mauchly’s test indicated that the assumption of sphericity had been violated, χ2(9) = 27.700, P < .05; therefore, degrees of freedom were corrected using Greenhouse– Geisser estimates of sphericity (ε = .596). The results indicate that nasal surgery did not produce any significant change in physical functioning throughout the postoperative period, F(2.384, 33.377) = 3.059, P = .052. These results suggest that the nasal surgery did not significantly influence one’s perceived score related to physical factors. Relative to the overall V-RQOL scores, Mauchly’s test indicated that the assumption of sphericity had been violated, χ2(9) = 38.772, P < .05; therefore degrees of freedom were corrected using Greenhouse–Geisser estimates of sphericity (ε = .496). The results indicate that there was
Brandt et al no significant change in V-RQOL scores over time following nasal surgery, F(1.985, 27.788) = 2.963, P = .069.
Discussion This study explored objective and subjective changes in resonance following nasal surgery in a comprehensive, prospective, and longitudinal fashion. In seeking to meet this objective, a comprehensive, longitudinal evaluation of resonance and nasality occurred over a 6-month time period. This included obtaining microacoustic measures of speech signals, auditory-perceptual stimuli, and self-report information on QOL using the V-RQOL. With a single exception, the results of this investigation indicate that no significant change in any acoustic measure obtained occurred by the end of the study. The sole difference was observed for the HNR measure. However, no additional differences were noted across all other micracoustic data (frequency, jitter, and shimmer), nor for perceptually scaled assessments of sentence samples by naïve listeners for the dimension of nasality, and finally, for either the subscore or total scores of the V-RQOL. This finding supports the notion that regardless of sinonasal pathology or nasal surgery, resonance characteristics remain remarkably stable for the objective acoustic measures generated. This interpretation is further supported via the auditory-perceptual and self-report measures that were gathered as part of this investigation. The finding of a difference in HNR measures indicates that some collective change in the relative ratio of quasiperiodicity and the inherent noise floor in the vocal signal did occur. Although a significant different in HNR was detected, this suggests that while the ratio of “harmonics to noise” in the samples may have existed for the vowel samples, this difference does not extend to naïve listener judgments of more natural, running speech that is represented by the sentence from the Rainbow Passage. Thus, in spite of an objective acoustic change noted in this single instance, no other significant change occurred based on either the subjective impression of listener participants or the self-reports of the patient participants studied. This finding suggests that gestalt perceptual evaluations of nasality obtained herein may not account for statistically significant alterations in HNR—an academically curious finding that may not demonstrate substantial clinical utility. As noted in the literature, the ability to link acoustic signal characteristics with auditory-perceptual assessment13 is an important area for continuing study; hence, further research that is directed toward such a relationship remains timely. Similarly, the ability to utilize subjective perceptual measures as a viable means of documenting changes in resonance cannot be discounted as a important clinical outcome measure.14 When all acoustic data are evaluated together, it is clear that a consistent trend of no difference in the measures
569 obtained was identified. This finding provides support that differences of any appreciable magnitude did not exist between pre- and postsurgery points of measurement. This finding remains consistent in spite of a heterogeneous group with various pathology and degrees of intervention to the nasal airway. As a result, and based on collective evaluation, these data provide support for the assumption that nasal surgery (rhinoplasty, septoplasty, and/or ESS) do not result in any significant change in multiple measurement parameters of resonance—acoustic, perceptual, or via self-assessment. While the present data cannot be generalized in their entirety, the consistency of the present findings provide information that allows the surgeon and the patient to better understand outcomes following rhinoplasty, septoplasty, and/or endoscopic sinus surgical procedures. While the present findings do not eliminate the potential for such disruptions in resonance and the patient’s functional capacity, they do offer evidence that suggests these areas may be of insignificant concern. Nevertheless, care should always be taken by the surgeon to adequately inform the patient who may be considering surgery as to the possible “speechrelated” (ie, potential changes in nasal resonance) consequences of nasal surgery procedures such as those of the reported participants. One of the critical issues to consider in acoustic analysis is the variation that will always exist between speakers. That is, each speaker will exhibit a unique F0 and, thus, the need to carefully assess changes must be specific to that individual. Group comparisons under these circumstances are likely to obscure potential changes. However, 1 of the key elements with such a concern rests with auditory-perceptual judgments of both the speaker and the listener. That is, if changes are not identified by either, the opportunity to infer that no substantial change has taken place in the quality of one’s speech resonance from the point prior to surgery to any point after surgery becomes more meaningful. In the present study we noted no significant change in either the auditory-perceptual evaluations of sentence sample, nor in the subjects’ self-perceived assessments of their oral communication via use of the V-RQOL. Consequently, our data provide evidence that nasal surgery did not impact resonance in the patients studied. Although the present results are consistent across a heterogeneous group, the heterogeneity of this group also serves as a limitation of this study. While a post hoc subgroup analysis to evaluate the influence of particular sinonasal pathology or interventions (ie, nasal polyposis vs septal deviation, septoplasty vs ESS) would have been of interest, small and unequal subgroup sample sizes and heterogeneity of pathology preclude meaningful conclusions. Hence the results of this study must be interpreted within the context of its heterogeneous group. Conclusions cannot be drawn for those individuals with significantly obstructive
570 sinonasal conditions such as advanced polyposis. Future investigations would benefit from a similar longitudinal study design, but more restrictive evaluations of individuals undergoing homogenous surgical interventions for variable sinonasal pathology (ie, ESS for chronic rhinosinusitis with endoscopically graded polyposis). A further limitation to this study may stem from the subjective evaluation of the patients by naïve, normal hearing, young-adult listeners using the visual-analog scale. Perhaps trained listeners may be able to discern subtle changes to nasal resonance. These trained listeners may in fact be those professional voice users most concerned about undergoing sinonasal surgery. Thus in spite of the consistency of the results obtained, appropriate patient counseling with reference to the potential for subtle changes must take place. Taken within the context of the heterogeneous group studied and the aforementioned limitations, the longitudinal and comprehensive evaluation performed provides data confirming that resonance changes secondary to nasal surgery are minimal. While this does not exclude the potential that real and substantial changes can occur at any or all levels of evaluation (ie, acoustic, perceptual, and/or self-evaluation), the current data suggest that concerns about such changes are minimized considerably.
Conclusion This study has addressed the impact of nasal surgery on resonance via the evaluation of acoustic, perceptual, and selfreport measures. With a single exception, the present data did not identify differences in nasal resonance occurring as a consequence of nasal surgery. Conclusions cannot be drawn for those individuals with significantly obstructive sinonasal conditions such as advanced polyposis. Although further study is warranted, the level of consistency in the present work, not only for the acoustic measures gathered but also for the auditory-perceptual and patient self-report measures, adds considerable strength to the suggestion that alteration in the resonance of speech is an unlikely complication of nasal surgery. Thus, the present work provides valuable information for patients considering nasal surgery. Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Annals of Otology, Rhinology & Laryngology 123(8) Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by a Physician Services Incorporated (PSI) Foundation Grant.
References 1. Chen MY, Metson R. Effects of sinus surgery on speech. Arch Otolaryngol Head Neck Surg. 1997;123(8):845-852. 2. Hosemann W, Göde U, Dunker JE, Eysholdt U. Influence of endoscopic sinus surgery on voice quality. Eur Arch Otorhinolaryngol. 1998;255(10):499-503. 3. Soneghet R, Santos RP, Behlau M, Habermann W, Friedrich G, Stammberger H. Nasalance changes after functional endoscopic sinus surgery. J Voice. 2002;16(3):392-397. 4. 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. 1997;117(4): 343-348. 5. Behrman A, Shikowitz MJ, Dailey S. The effect of upper airway surgery on voice. Otolaryngol Head Neck Surg. 2002;127(1):36-42. 6. Kummer AW. Resonance disorders and nasal emission. evaluation and treatment using “low tech” and “no tech” procedures. ASHA Leader. 2006. Available at: http://www.asha. org/Publications/leader/2006/060207/060207c/. 7. Jacobson BH, Johnson A, Grywalski C, et al. The Voice Handicap Index (VHI): Development and validation. Am J Speech Lang Pathol. 1997;6(3):66. 8. Lund VJ, Mackay IS. Staging in rhinosinusitus. Rhinology. 1993;31(4):183-184. 9. Hogikyan ND, Sethuraman G. Validation of an instrument to measure voice-related quality of life (V-RQOL). J Voice. 1999;13(4):557-569. 10. Fairbanks G. Voice and Articulation Drillbook. New York, NY: Harper; 1960. 11. Britto AI, Doyle PC. A comparison of habitual and derived optimal voice fundamental frequency values in normal young adult speakers. J Speech Hear Disord. 1990;55:476-484. 12. Stevens SS. Psychophysics: Introduction to Its Perceptual, Neural, and Social Prospects. New Brunswick, NJ: Transaction; 1975. 13. Eadie TL, Doyle PC. Direct magnitude estimation and interval scaling of pleasantness and severity in dysphonic and normal speakers. J Acoust Soc Am. 2002;112(6):3014-3021. 14. Eadie TL, Doyle PC. Classification of dysphonic voice: acoustic and auditory-perceptual measures. J Voice. 2005;19: 1-14.
research-article2014
AORXXX10.1177/0003489414525595Annals of Otology, Rhinology & LaryngologyBrandt et al
Article Annals of Otology, Rhinology & Laryngology 2014, Vol. 123(8) 564–570 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0003489414525595 aor.sagepub.com
Impact of Nasal Surgery on Speech Resonance Michael G. Brandt, BSc, MD1,2, Brian W. Rotenberg, MD, MPH2, Corey C. Moore, MD, MSc2, Catherine C. Bornbaum, PhD3, Agnieszka Dzioba, PhD3, Jordan T. Glicksman, BSc, MD3, and Philip C. Doyle, PhD2,3
Abstract Objectives: The nose and paranasal sinuses contribute to speech resonance and changes to these structures may alter speech nasality. This change may influence one’s vocational and social functioning and quality of life. Our investigation explored objective and subjective changes in nasality following nasal surgery in a prospective and longitudinal fashion. Methods: Recordings of sustained vowel and sentence stimuli and voice-related quality of life measurements were obtained preoperatively and at 2, 4, 8, and 24 weeks postoperatively from individuals undergoing nasal and/or sinus surgery. Objective measures of fundamental frequency, jitter, shimmer, and harmonic to noise ratio (HNR) were determined. Preand postoperative speech samples were assessed by 15 naïve listeners. Results: In all, 15 subjects completed the study. Neither speakers nor listeners perceived a subjective change in nasality following surgery. No statistically significant change in microacoustic measures were identified. Although nasal sentences did not reveal differences for 3 microacoustic measures, a difference in HNR was identified. Conclusions: Patients undergoing nasal surgery did not exhibit subjective changes in resonance postoperatively. Aside from a difference in HNR for the nasal sentence, objective microacoustics remained unchanged. These results demonstrate the stability of oranasal resonance despite nasal surgery and provide valuable data for patient informed decision-making. Keywords nasality, resonance, acoustics, nasal surgery, endoscopic sinus surgery, rhinoplasty To date only a limited number of studies have sought to evaluate the impact of nasal surgery on the speech signal.1-5 Although these investigations have elucidated some relationships, the overall impact of nasal procedures on speech quality remains poorly understood. Of specific interest is the potential influence of nasal surgery on speech resonance—defined as “the quality of the voice that is determined by the balance of sound vibration in the oral, nasal, and pharyngeal cavities during speech.”6 In general, the literature has suggested a surgical effect on nasalized vowels1,2,5 and nasal consonants,1,4,5 focusing on the measurement of nasalance (the ratio of nasal to oral acoustic energy),3,4 and the overall resonance of the vocal tract.1,2,5 Acoustic analysis has involved the assessment of frequency and amplitude differences between the first resonant frequency (first “formant”) which is an important acoustic descriptor of vowels, and the peak or murmur associated with nasal resonance amplitudes. An excellent review by Behrman and colleagues5 highlights the outcomes of the aforementioned studies and goes on to demonstrate similar findings of increased nasal resonance for both consonants
and vowels following nasal surgery. This information supports the notion that changes in nasal resonance can be objectively evaluated. Consequently, the ability to track acoustic changes and their potential variability holds great promise at both basic and applied levels of evaluation. Many articles comment on the subjective identification of speech/voice “quality” changes among patients post– nasal surgery.1-3,5 Of these reports, only 1 evaluated the impact of these changes on voice-related quality of life 1
Department of Otolaryngology–Head and Neck Surgery, Division of Facial Plastic & Reconstructive Surgery, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada 2 Rehabilitation Sciences, Faculty of Health Sciences, Western University, London, Ontario, Canada 3 Department of Otolaryngology–Head and Neck Surgery, Schulich School of Medicine and Dentistry, Western University, London, Ontario, Canada Corresponding Author: Brian W. Rotenberg, MD, MPH, St. Joseph’s Healthcare Centre, 268 Grosvenor St, London, Ontario, Canada, N6A 4V2. Email: [email protected]
565
Brandt et al (QOL) using the Voice Handicap Index (VHI).5,7 While the VHI was originally designed specifically to assess the impact of vocal/laryngeal based abnormalities on perceived QOL, it is not uncommon for patients to refer more generically to “voice” or “speech” problems as opposed to using the more technical terms of “resonance” or “nasality”; thus, use of the VHI or similar instruments to evaluate changes in those undergoing nasal surgery was not inappropriate despite its primary design and clinical use. Using the VHI, the aforementioned study indicated that 20% of the participants noted a positive change in their voice postoperatively.5 These results were found with combined procedures to the upper airway and did not specifically focus on nasal surgery; thus, it is difficult to extrapolate these results and reliably differentiate between specific surgical approaches and any resultant changes secondary to nasal surgery. Although some patients may perceive postoperative changes to the quality of their speech as an interesting curiosity, the professional voice user or the nonprofessional individual who experiences high demand for voice use may find even minor modulations to their trademark or everyday voice to have disastrous personal and social implications. Thus, it is imperative to better define the true impact of nasal surgery on voice/speech quality so as to better inform patients and predict alterations that may occur as a consequence of changes in resonance and associated nasality. Information obtained may also help to reduce patient fears or concerns specific to a variety of domains of interest such as acoustic, auditory-perceptual, or changes that influence self-assessed QOL. Thus, the objective of our study was to empirically determine the influence of nasal surgery on speech resonance in a prospective and longitudinal fashion.
Methods Institutional ethics approval was obtained from the University of Western Ontario Research Ethics Board (HSREB# 12639E).
Participants Participants included any individual presenting to a tertiary care Otolaryngology–Head and Neck Surgery practice considering rhinoplasty, septoplasty, and/or endoscopic sinus surgery for chronic sinusitis. Patients with sinonasal disease other than chronic rhinosinusitis (ie, pituitary tumor, papilloma, etc) were excluded given the increased variability of their surgical procedure. Only those participants fluent in written and spoken English were voluntarily recruited to participate. Participants were excluded from study involvement if they had a preexisting speech abnormality or preexisting anatomic laryngeal abnormality so as to prevent any confounding influence on signal analysis and perceptual assessment. Patients with nasal polyposis underwent CT
imaging and had their sinus opacification documented and graded (Lund–Mackay stage)8 to control for their influence on resonance outcomes. Patients with nasal obstruction due to causes not pertaining to chronic rhinosinusitis (eg, nasal fracture) did not undergo Lund–Mackay staging.
Research Protocol Patients meeting the inclusion criteria and agreeing to participate were assessed at 5 predetermined evaluation intervals: (1) preoperatively, (2) 2 weeks postoperatively, (3) 4 weeks postoperatively, (4) 8 weeks postoperatively, and (5) 24 weeks (6 months) postoperatively. At each assessment the participant was asked to a. Complete the Voice-Related Quality of Life (V-RQOL)9 questionnaire b. Complete a full speech recording protocol
Data Acquisition—Speech Stimuli and Recording Each participant was recorded at multiple points over the course of the study. Recordings were obtained prior to surgery, then at 2, 4, and 8 weeks postsurgery, with a final measure occurring 6 months postsurgery. During each recording session, participants were asked to perform 3 specific tasks. First, participants were asked to produce 5 samples each of 3 vowels, /a/, /i/, and /u/. Because of their relatively static nature in the context of a quasiperiodic vocal signal, vowels provide an ideal opportunity to evaluate microacoustic characteristics of the vocal signal. Participants were instructed to sustain 3 samples of each of these vowels as steadily as possible at their conversational pitch and loudness level for approximately 6 to 8 seconds. Participants were permitted time to practice the task prior to recording of the vowel stimuli with immediate and direct feedback provided as necessary by the individual gathering recordings (C.C.B., A.D., P.C.D.). Following the sustained vowel task, participants were asked to read aloud the Rainbow Passage10 and 2 sets of sentences (one of which contained a high number of nasal consonants in the sentence, and the other without any nasal consonants). Again, participants were instructed to speak at a conversational pitch and loudness level; participants were able to familiarize themselves with the passage prior to formal recording. Two recordings of the Rainbow Passage were gathered along with 2 productions each of sentence stimuli (both nasal and non-nasal sentences). The third standardized vocal task11 consisted of recording a short monologue (approximately 30 seconds in length), produced once again at a typical level of pitch and loudness. Although the content of this monologue was not controlled, all participants were asked to speak on the general topic of a hobby or special interest that they enjoy and/
566 or participate in. At the conclusion of each recording session, participants were asked to grossly assess their selfperceived level of the “acceptability” of their voice/speech quality using a simple visual analogue scale; this assessment was obtained at the end of all recording sessions. This task served solely as an index of a potentially self-perceived change that could lead the investigators to more extensive evaluation of a given individual’s data. All recordings were gathered in a sound-treated environment, free of ambient noise. The full recording procedure for any given session required approximately 20 minutes. Following completion of the recording, each participant was asked to complete the 10-item V-RQOL9 measure. Briefly, the V-RQOL asks the respondent to answer each question regarding specific issues of voice use by rating it from 1 (not a problem) to 5 (problem as bad as it can be). These scores were then transformed according the V-RQOL algorithm and used as a self-rating measure for each participant at the time of all recordings
Instrumentation All speech recordings were acquired digitally using the Kay Sona-Speech II System—Model 3650 (KayPentax, Montvale, New Jersey, USA). Samples were recorded using a unidirectional microphone that was positioned at a fixed distance of 6.0 inches from the participant’s mouth. The mouth to microphone distance was rechecked before and after every experimental task to ensure consistency. The microphone was routed to a computer through a microphone preamplifier (Kay Elemetrics Model 3708, KayPentax, Mississauga, Canada) allowing input levels to be adjusted such that signal strength reached approximately 10 to 13 LEDS on the Multi-Dimensional Voice Profile system. All recordings were sampled at 44 KHz and stored directly to the computer hard drive for later conversion and acoustic analysis.
Acoustic Analysis of Data All vowel data were evaluated for basic microacoustic measures of fundamental frequency, frequency perturbation (jitter), amplitude perturbation (shimmer), and harmonic to noise ratio (HNR), and these were gathered from the extracted segments of the steady-state vowel entities. This involved the analysis of a 2-second segment extracted from the midpoint of each vowel sample. Care was taken to extract a vowel segment from the middle of samples to eliminate acoustic variation due to vocal onset and offset. Acoustic measures were collected for each vowel independently using a PRAAT (version 5.3.29) an open-source acoustical analysis program (http://www.fon.hum.uva.nl/ praat/). Once all analyses of each participant’s vowels were completed, the middle frequency value for each of the 3
Annals of Otology, Rhinology & Laryngology 123(8) samples of each vowel was then used for statistical analysis. For the running speech stimuli, only sentence stimuli were acoustically analyzed. The second sentence of the Rainbow Passage was extracted from the entire passage and was then used in the perceptual phase of this project. All stimuli were acoustically evaluated independently and in a blinded fashion by a single researcher to maintain consistency in measurement procedures for each of the experimental tasks.
Auditory-perceptual Evaluation of Speech Stimuli The second sentence of the Rainbow Passage (12 words) was submitted to perceptual evaluation by a group of 15 naïve, normal-hearing young adult listeners. None of the listeners had any formal exposure to or educational exposure to the area of voice, speech, or resonance disorders. The assessment of the Rainbow Passage stimuli involved an assessment of potential changes in resonance, and more specifically in nasality. To gather these auditory-perceptual data, a scaling procedure designed specifically for this study was used. This involved rating each sample using a scale that ranged from hyponasal to hypernasal, thus, providing the opportunity to identify a range of potential changes in nasal resonance characteristics outside of the normal expectation. The midpoint of the scale was identified to represent “normal nasality” (ie, a normal balance of oral and nasal resonance) to permit the listener to make judgments of both types of abnormal nasality using a single visual analogue (VA) scale. Our decision to use a VA scale and include the entire continuum including the hyponasal side of nasality in this perceptual scaling procedure was done to ensure that any potential change in resonance (hypo or hypernasal variation) could be accounted for over the course of the study. In addition, by using a VA scale, concerns related to whether the nasality continua are linear or curvilinear could be accounted for accordingly.12 Based on this VA scale, a marking to the far left of the scale would represent a numeric value of 1 (hyponasality), with a score to the far right being equal to a numeric value of 100 (hypernasality). Thus, as rater markings on the VA scale moved toward the middle of the scale from either end of the scale, they would represent increasing levels of “normal” resonance. Conversely, as scales scores departed from the middle of the scale, they would represent increased levels of either hypo or hyper nasal resonance. Our design and use of this unique scale was undertaken in an effort to gather perceptual data that spanned the continuum of nasality. Operational definitions of each anchor term were provided to the listeners, as were recorded examples of extreme hypernasality and hyponasality. Following completion of this assessment, the numerical value for each rating was determined from direct measurement in millimeters of the VA scale. These perceptual data were then submitted to statistical analysis.
567
Brandt et al
Data Acquisition—V-RQOL The V-RQOL is a standardized 10-item questionnaire initially developed by Hogikyan and Sethuraman9 to quantify potential vocal difficulties specific to a range of situations. This assessment instrument is well validated and directly focuses on the impact of voice use on QOL with specific attention directed at physical functioning concerns (eg, capacity for loudness, etc) and social-emotional impact relative to one’s voice use (eg, restriction of activities, etc). Given the impact of nasal resonance on QOL demonstrated in previous reports,5 we chose to include this measure to objectively evaluate the “voice-related” impact of nasal surgery among our participants over the course of this study. Similar to the collection of speech recordings, each participant was asked to complete the V-RQOL presurgically, then again at 2, 4, and 8 week postsurgically, then finally at 6 months postoperatively.
Data Acquisition—CT Scan Evaluation Those patients undergoing endoscopic sinus surgery for chronic rhinosinusitis underwent standard preoperative CT imaging. These scans were evaluated and scored using the standardized Lund–Mackay staging system for chronic rhinosinusitis by a single observer familiar with the Lund– Mackay staging system (J.T.G.).8
Statistical Analysis Prior to commencement of this study, we determined that a total sample size (N) of 14 individuals would be sufficient to detect the hypothesized effect (r2 = .20) of a 2-level within-subject independent variable 83.2% of the time using a .05 alpha level, assuming a within-subject correlation of .30. This calculation would apply to all independent measures gathered (acoustic, auditory-perceptual, and V-RQOL data). All statistical analyses were performed using SPSS version 18 (Minneapolis, Minnesota, USA). Acoustic measures obtained from vowel samples were compared independently for samples of /a/, /i/, and /u/ using an analysis of variance (ANOVA) with repeated measures across the preoperative and 4 postoperative time periods (2, 4, and 8 weeks and 6 months). All remaining acoustic measures were evaluated using a between-groups ANOVA with repeated measures over the 4 data collection periods. Similar analyses for both the physical, socialemotional and total V-RQOL scores were also conducted. An a priori probability level of .05 was employed. Correlational analyses were performed for the V-RQOL measures and the perceptual assessment data obtained from naïve listeners, and also evaluated these measures against the preoperative Lund–Mackay sinus opacification score. Measurement reliability and consistency was
Table 1. Demographics and Procedures Performed Among the Patient Population Evaluated for Postoperative Voice Changes. Mean age, y Sex Male Female Diagnosis Nasal obstruction and deformity Nasal obstruction CRS with polyps Presented with nasal obstruction Prior sinonasal operation Procedure Septorhinoplasty Septorhinoplasty and turbinatoplasty FESS and septoplasty FESS Septoplasty and turbinatoplasty Smoking history Active smoker Past smoking history LMS average among CRS patients (n = 7)
52 (SD, 16.8; range, 21-83) 11 4 4 4 7 15 2 3 3 1 6 2 1 2 15.1 (95% CI, 13.8-16.5)
Abbreviations: CI, confidence interval; CRS, chronic rhinosinusitis; FESS, functional endoscopic sinus surgery; LMS, Lund–Mackay score; SD, standard deviation.
determined via the use of Cronbach’s alpha and assessment of point-by-point agreement.
Results Fifteen patients met all study criteria and took part in the study to its conclusion. Their demographic information and related findings are summarized in Table 1. Of the 15 patients, there were 11 males and 4 females. The mean age was 52 years (SD = 16.8; range, 21 to 83). The presenting symptom among all 15 patients was nasal obstruction; with 4 having an external nasal deformity, and 7 demonstrating chronic rhinosinusitis with nasal polyposis. Operative procedures included septorhinoplasty with and without turbinoplasty (cautery and out-fracture of the inferior turbinates) (n = 3 and n = 3), endoscopic sinus surgery (ESS) with and without septoplasty (n = 1 and n = 6, respectively) and septoplasty with turbinoplasty (n = 2). The studied intervention was the first sinonasal operation in 13 of the patients studied. Procedures were performed on 1 active smoker and 2 former smokers. The average Lund–Mackay score among patients with polyposis was 15.1 (n = 7, 95% CI = 13.816.5). None of the patients reported any subjective change in their speech at any time postoperatively.
568
Figure 1. Harmonic to noise ratios over time per group.
Acoustic Analysis—Vowels Full microacoustic measures (F0, jitter, shimmer, and HNR) were obtained independently for each vowel (/a/, /i/, /u/) and then submitted to statistical analysis. Analyses were performed on each vowel independently for women and men. For all 4 microacoustic measures for each of the 3 vowels assessed, no significant difference was observed over time due to surgical intervention for either male or female participants.
Acoustic Analysis—Rainbow Passage The Rainbow Passage allowed for analysis of fundamental frequency, jitter, shimmer, and HNR scores. Mauchly’s test was used to evaluate jitter scores and indicated that the assumption of sphericity was not violated, χ2(9) = 10.404, P = .323. Thus, the results suggest that there was no significant change in the jitter scores over time following nasal surgery, F(4, 56) = 1.556, P = .199. Relative to the shimmer scores, Mauchly’s test indicated that the assumption of sphericity had been violated, χ2(9) = 19.655, P < .05; therefore degrees of freedom were corrected using Greenhouse–Geisser estimates of sphericity (ε = .576). The results were then evaluated and identified that there was no significant change in shimmer when comparing preoperative to postoperative voice samples, F(2.305, 32.264) = 0.265, P = .789. Once again, Mauchly’s test was used to evaluate HNR scores and indicated that the assumption of sphericity was not violated, χ2(9) = 11.995, P = .217. These results indicate that there was no significant change in HNR scores across the preoperative and postoperative samples, F(4, 56) = 1.400, P = .246. Figure 1 shows the results over time.
Acoustic Analysis—Nasal Sentence Similar to the Rainbow Passage stimuli, an analysis of fundamental frequency, jitter, shimmer, and HNR scores were
Annals of Otology, Rhinology & Laryngology 123(8) performed on the nasal sentence samples. Relative to the jitter scores, Mauchly’s test indicated that the assumption of sphericity was not violated, χ2(9) = 12.145, P = .209. The results indicate that there was no significant change in jitter scores following nasal surgery, F(4, 56) = 1.296, P = .283. Relative to the shimmer scores, Mauchly’s test indicated that the assumption of sphericity was not violated, χ2(9) = 8.244, P = .514. The results indicate that there was no significant change in shimmer scores during healing from nasal surgery, F(4, 56) = 2.192, P = .082. Finally, with respect to the HNR scores, Mauchly’s test indicated that the assumption of sphericity was not violated, χ2(9) = 13.082, P = .162. These results indicate that there was a significant change in HNR scores when comparing pre- and postsurgical voice samples, F(4, 56) = 4.571, P = .003.
Auditory-Perceptual Evaluation of Rainbow Passage Preoperative and 6-month postoperative voice recordings were subjectively evaluated and rated by naïve listeners. Pre- and postoperative subjective ratings were then statistically evaluated using a paired t test. Results of this analysis revealed that no significant difference in voice quality was demonstrated based on listener judgments of voice samples (t = 1.678, df 14, P = .116).
VRQOL With respect to the social-emotional (SE) subscore, Mauchly’s test indicated that the assumption of sphericity had been violated, χ2(9) = 54.496, P < .05; therefore, degrees of freedom were corrected using Greenhouse– Geisser estimates of sphericity (ε = .367). The results indicate that there was no significant difference within the SE subscore of the VRQOL throughout the postoperative period, F(1.469, 20.563) = 1.683, P = .212. Hence, nasal surgery did not significantly influence one’s perceived SE score on the V-RQOL. Regarding the physical domain, Mauchly’s test indicated that the assumption of sphericity had been violated, χ2(9) = 27.700, P < .05; therefore, degrees of freedom were corrected using Greenhouse– Geisser estimates of sphericity (ε = .596). The results indicate that nasal surgery did not produce any significant change in physical functioning throughout the postoperative period, F(2.384, 33.377) = 3.059, P = .052. These results suggest that the nasal surgery did not significantly influence one’s perceived score related to physical factors. Relative to the overall V-RQOL scores, Mauchly’s test indicated that the assumption of sphericity had been violated, χ2(9) = 38.772, P < .05; therefore degrees of freedom were corrected using Greenhouse–Geisser estimates of sphericity (ε = .496). The results indicate that there was
Brandt et al no significant change in V-RQOL scores over time following nasal surgery, F(1.985, 27.788) = 2.963, P = .069.
Discussion This study explored objective and subjective changes in resonance following nasal surgery in a comprehensive, prospective, and longitudinal fashion. In seeking to meet this objective, a comprehensive, longitudinal evaluation of resonance and nasality occurred over a 6-month time period. This included obtaining microacoustic measures of speech signals, auditory-perceptual stimuli, and self-report information on QOL using the V-RQOL. With a single exception, the results of this investigation indicate that no significant change in any acoustic measure obtained occurred by the end of the study. The sole difference was observed for the HNR measure. However, no additional differences were noted across all other micracoustic data (frequency, jitter, and shimmer), nor for perceptually scaled assessments of sentence samples by naïve listeners for the dimension of nasality, and finally, for either the subscore or total scores of the V-RQOL. This finding supports the notion that regardless of sinonasal pathology or nasal surgery, resonance characteristics remain remarkably stable for the objective acoustic measures generated. This interpretation is further supported via the auditory-perceptual and self-report measures that were gathered as part of this investigation. The finding of a difference in HNR measures indicates that some collective change in the relative ratio of quasiperiodicity and the inherent noise floor in the vocal signal did occur. Although a significant different in HNR was detected, this suggests that while the ratio of “harmonics to noise” in the samples may have existed for the vowel samples, this difference does not extend to naïve listener judgments of more natural, running speech that is represented by the sentence from the Rainbow Passage. Thus, in spite of an objective acoustic change noted in this single instance, no other significant change occurred based on either the subjective impression of listener participants or the self-reports of the patient participants studied. This finding suggests that gestalt perceptual evaluations of nasality obtained herein may not account for statistically significant alterations in HNR—an academically curious finding that may not demonstrate substantial clinical utility. As noted in the literature, the ability to link acoustic signal characteristics with auditory-perceptual assessment13 is an important area for continuing study; hence, further research that is directed toward such a relationship remains timely. Similarly, the ability to utilize subjective perceptual measures as a viable means of documenting changes in resonance cannot be discounted as a important clinical outcome measure.14 When all acoustic data are evaluated together, it is clear that a consistent trend of no difference in the measures
569 obtained was identified. This finding provides support that differences of any appreciable magnitude did not exist between pre- and postsurgery points of measurement. This finding remains consistent in spite of a heterogeneous group with various pathology and degrees of intervention to the nasal airway. As a result, and based on collective evaluation, these data provide support for the assumption that nasal surgery (rhinoplasty, septoplasty, and/or ESS) do not result in any significant change in multiple measurement parameters of resonance—acoustic, perceptual, or via self-assessment. While the present data cannot be generalized in their entirety, the consistency of the present findings provide information that allows the surgeon and the patient to better understand outcomes following rhinoplasty, septoplasty, and/or endoscopic sinus surgical procedures. While the present findings do not eliminate the potential for such disruptions in resonance and the patient’s functional capacity, they do offer evidence that suggests these areas may be of insignificant concern. Nevertheless, care should always be taken by the surgeon to adequately inform the patient who may be considering surgery as to the possible “speechrelated” (ie, potential changes in nasal resonance) consequences of nasal surgery procedures such as those of the reported participants. One of the critical issues to consider in acoustic analysis is the variation that will always exist between speakers. That is, each speaker will exhibit a unique F0 and, thus, the need to carefully assess changes must be specific to that individual. Group comparisons under these circumstances are likely to obscure potential changes. However, 1 of the key elements with such a concern rests with auditory-perceptual judgments of both the speaker and the listener. That is, if changes are not identified by either, the opportunity to infer that no substantial change has taken place in the quality of one’s speech resonance from the point prior to surgery to any point after surgery becomes more meaningful. In the present study we noted no significant change in either the auditory-perceptual evaluations of sentence sample, nor in the subjects’ self-perceived assessments of their oral communication via use of the V-RQOL. Consequently, our data provide evidence that nasal surgery did not impact resonance in the patients studied. Although the present results are consistent across a heterogeneous group, the heterogeneity of this group also serves as a limitation of this study. While a post hoc subgroup analysis to evaluate the influence of particular sinonasal pathology or interventions (ie, nasal polyposis vs septal deviation, septoplasty vs ESS) would have been of interest, small and unequal subgroup sample sizes and heterogeneity of pathology preclude meaningful conclusions. Hence the results of this study must be interpreted within the context of its heterogeneous group. Conclusions cannot be drawn for those individuals with significantly obstructive
570 sinonasal conditions such as advanced polyposis. Future investigations would benefit from a similar longitudinal study design, but more restrictive evaluations of individuals undergoing homogenous surgical interventions for variable sinonasal pathology (ie, ESS for chronic rhinosinusitis with endoscopically graded polyposis). A further limitation to this study may stem from the subjective evaluation of the patients by naïve, normal hearing, young-adult listeners using the visual-analog scale. Perhaps trained listeners may be able to discern subtle changes to nasal resonance. These trained listeners may in fact be those professional voice users most concerned about undergoing sinonasal surgery. Thus in spite of the consistency of the results obtained, appropriate patient counseling with reference to the potential for subtle changes must take place. Taken within the context of the heterogeneous group studied and the aforementioned limitations, the longitudinal and comprehensive evaluation performed provides data confirming that resonance changes secondary to nasal surgery are minimal. While this does not exclude the potential that real and substantial changes can occur at any or all levels of evaluation (ie, acoustic, perceptual, and/or self-evaluation), the current data suggest that concerns about such changes are minimized considerably.
Conclusion This study has addressed the impact of nasal surgery on resonance via the evaluation of acoustic, perceptual, and selfreport measures. With a single exception, the present data did not identify differences in nasal resonance occurring as a consequence of nasal surgery. Conclusions cannot be drawn for those individuals with significantly obstructive sinonasal conditions such as advanced polyposis. Although further study is warranted, the level of consistency in the present work, not only for the acoustic measures gathered but also for the auditory-perceptual and patient self-report measures, adds considerable strength to the suggestion that alteration in the resonance of speech is an unlikely complication of nasal surgery. Thus, the present work provides valuable information for patients considering nasal surgery. Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Annals of Otology, Rhinology & Laryngology 123(8) Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by a Physician Services Incorporated (PSI) Foundation Grant.
References 1. Chen MY, Metson R. Effects of sinus surgery on speech. Arch Otolaryngol Head Neck Surg. 1997;123(8):845-852. 2. Hosemann W, Göde U, Dunker JE, Eysholdt U. Influence of endoscopic sinus surgery on voice quality. Eur Arch Otorhinolaryngol. 1998;255(10):499-503. 3. Soneghet R, Santos RP, Behlau M, Habermann W, Friedrich G, Stammberger H. Nasalance changes after functional endoscopic sinus surgery. J Voice. 2002;16(3):392-397. 4. 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. 1997;117(4): 343-348. 5. Behrman A, Shikowitz MJ, Dailey S. The effect of upper airway surgery on voice. Otolaryngol Head Neck Surg. 2002;127(1):36-42. 6. Kummer AW. Resonance disorders and nasal emission. evaluation and treatment using “low tech” and “no tech” procedures. ASHA Leader. 2006. Available at: http://www.asha. org/Publications/leader/2006/060207/060207c/. 7. Jacobson BH, Johnson A, Grywalski C, et al. The Voice Handicap Index (VHI): Development and validation. Am J Speech Lang Pathol. 1997;6(3):66. 8. Lund VJ, Mackay IS. Staging in rhinosinusitus. Rhinology. 1993;31(4):183-184. 9. Hogikyan ND, Sethuraman G. Validation of an instrument to measure voice-related quality of life (V-RQOL). J Voice. 1999;13(4):557-569. 10. Fairbanks G. Voice and Articulation Drillbook. New York, NY: Harper; 1960. 11. Britto AI, Doyle PC. A comparison of habitual and derived optimal voice fundamental frequency values in normal young adult speakers. J Speech Hear Disord. 1990;55:476-484. 12. Stevens SS. Psychophysics: Introduction to Its Perceptual, Neural, and Social Prospects. New Brunswick, NJ: Transaction; 1975. 13. Eadie TL, Doyle PC. Direct magnitude estimation and interval scaling of pleasantness and severity in dysphonic and normal speakers. J Acoust Soc Am. 2002;112(6):3014-3021. 14. Eadie TL, Doyle PC. Classification of dysphonic voice: acoustic and auditory-perceptual measures. J Voice. 2005;19: 1-14.
