A Multivariate Analysis of Objective Voice Changes After Thyroidectomy Without Laryngeal Nerve Injury

Objective  To evaluate the impact of thyroidectomy and the possible effects of factors such as patient sex, operation type, and surgeon experience on objective voice parameters of patients undergoing thyroidectomy without laryngeal nerve injury.

Design  Prospective study.

Setting  University hospital.

Patients  Thirty-six patients undergoing primary thyroidectomy because of thyroid disease.

Main Outcome Measures  The effect of thyroidectomy on voice was examined by recording the voices of the patients before and 1 week after thyroidectomy. The Multi-Dimensional Voice Program was used for capturing and analyzing the voice samples.

Results  On postoperative examination of objective voice changes, thyroidectomy had no multivariate effect on the combination of voice parameters. Patient sex, type of surgery, and surgeon experience had no effect on the combination of voice parameters before and after thyroidectomy. Regardless of within-patient factors (type of surgery, patient sex, and surgeon experience), 4 acoustic parameters (highest fundamental frequency, standard deviation of average fundamental frequency, phonatory average fundamental frequency range in semitones, and degree of subharmonics) significantly decreased after thyroidectomy (P < .05). Although they tended to be worse, none of the acoustic parameters showed significant changes in male patients. However, significant changes in some of the acoustic parameters of female patients were observed. Highest fundamental frequency, standard deviation of average fundamental frequency, phonatory average fundamental frequency range in semitones, absolute jitter, relative average perturbation, pitch perturbation quotient, shimmer in decibels, percentage of shimmer, amplitude perturbation quotient, noise to harmonic ratio, and degree of subharmonics values were all lower in female patients after thyroidectomy (P < .05).

Conclusions  Voice changes may occur after thyroidectomy without any evident laryngeal injury, and deterioration and amelioration of acoustic parameters can be observed to occur differently among male and female patients. Preoperative and postoperative objective voice analyses may be helpful in documenting voice changes.

Voice alterations may occur after thyroidectomy and are usually due to the injury of laryngeal nerves. Inferior laryngeal nerve injury has been a well-known cause, and many surgical techniques to avoid this problem have already been described in detail. More recently, the palsy of the external branch of the superior laryngeal nerve (EBSLN) has been increasingly recognized, and techniques have been described to reduce the risk of injury to this nerve.^1-6 However, it is still unclear whether the identification of the EBSLN is necessary to avoid damage to this nerve because it occasionally runs distally through the pharyngeal constrictor muscle. This necessitates intramuscular dissection for identification in the area around the superior thyroid pole, which seems to be inadvisable.² Moreover, it may not always be possible to visualize the complete course of the superior laryngeal nerve, and insistence on dissection in such cases may result in unwanted manipulation of this nerve. Some researchers do not pursue further dissection when the EBSLN cannot be readily identified. There is some evidence that even if the EBSLN cannot be identified, accurate distal ligation of the branches of the superior thyroid artery is a safe technique to prevent damage to this nerve.^7,8 The symptoms of EBSLN injury can be nonspecific, and laryngoscopic manifestations are often subtle even when there is some voice impairment, which makes prevention crucial.

Injury of the laryngeal nerves may not be the only cause of voice changes. Other possible causes include injury of the prethyroid strap muscles and cricothyroid muscles or impairment of laryngotracheal movement due to wound contracture after surgical trauma of the soft tissues.^7-9 A delicate surgical technique may prevent such complications, but it remains to be clarified whether voice alterations may occur after thyroidectomy without any laryngeal nerve injury. Computerized acoustic analysis of the patients undergoing thyroidectomy without laryngeal nerve injury may help determine possible voice changes objectively.^7,9-13

The aim of this prospective study was to evaluate the impact of thyroid surgery on objective voice parameters of patients without laryngeal nerve injury. Voices of patients were recorded preoperatively and postoperatively, and objective acoustic parameters were evaluated using a computer.

Patients who were scheduled to undergo primary thyroidectomy because of benign thyroid disease or differentiated thyroid carcinomas managed by standard thyroidectomy procedures were included in this study. Patients with advanced malignant diseases and anaplastic or medullary carcinoma of the thyroid were excluded because of the need for extensive surgical procedures, resulting in major trauma to the soft tissues of the neck. Other exclusion criteria included having any history of voice disease or previous neck surgery, having symptoms of upper respiratory tract involvement or being premenstrual at voice recording, and having hearing impairment.

The preoperative examination consisted of fiberoptic laryngoscopy or laryngostroboscopy and a voice recording on the computer program. None of the patients had vocal fold motion disorder. No typical sign of palsy of the superior laryngeal nerve was observed. The same procedures were performed again 1 week after surgery, and patients who showed any sign of laryngeal nerve damage were excluded from the study. Patients with any lesion of the endolarynx or upper aerodigestive tract that might be the result of a trauma (eg, intubation or suction) were not enrolled.

All patients underwent lobectomy and subtotal or total thyroidectomy as indicated by the primary condition. Surgery was performed by the third- to fifth-year residents under the supervision and direct assistance of one of the institution's general surgeons or by the surgeon (M.A.). The operators' surgical experience was noted as “more” for surgeons or “less” for residents to evaluate the impact of surgical technique. During surgery, the strap muscles were only retracted but not cut. Recurrent laryngeal nerves were routinely identified and protected. The superior thyroid artery and vein were individually ligated on the thyroid capsule to prevent EBSLN damage. When the nerve was not identified easily, no further dissection was performed to avoid inadvertent nerve injury. The cricothyroid muscle was carefully protected from injury due to electrocoagulation or manual retraction.

The phonatory evaluation consisted of recording the voice on a computer program (Multi-Dimensional Voice Program model 5105; Kay Elemetrics Corporation, Lincoln Park, New Jersey) with an industry standard, high-quality, 24-bit sound card (Creative Audigy; Creative Labs Inc, Milpitas, California) and an omnidirectional microphone (Sennheiser Electronic Corporation, Old Lyme, Connecticut). The microphone was set at a distance of approximately 15 cm from the mouth so that any distortions or modifications in the recording could be avoided. After 3 training emissions, the patient was asked to sustain the vowel /a/ at a self-selected comfortable pitch and intensity as long as steadily possible, and at least a 5-second period of voice sample was recorded at a 44 100-Hz sampling rate. Recordings were begun after initiation of voicing and ended before the patient terminated voicing. All samples were recorded in a sound-proof room at the institution. An interval of 3 seconds from the midportion of each sample was selected for acoustic analysis. The fundamental frequency parameters used for objective analysis were as follows: average fundamental frequency, standard deviation of average fundamental frequency, highest fundamental frequency, lowest fundamental frequency, and phonatory average fundamental frequency range in semitones. The frequency perturbation parameters were as follows: absolute jitter, percentage of jitter, relative average perturbation, pitch perturbation quotient, smoothed pitch perturbation quotient, and fundamental frequency variation. The amplitude perturbation parameters were as follows: shimmer in decibels, percentage of shimmer, amplitude perturbation quotient, smoothed amplitude perturbation quotient, and peak-to-peak amplitude variation. The noise parameters were as follows: noise to harmonic ratio, Voice Turbulence Index, and Soft Phonation Index. The parameters of subharmonic components included degree of subharmonics. All patients were informed about thyroidectomy and voice recording procedures and were asked to sign a consent form.

Statistical analysis of preoperative and postoperative data was performed using SPSS statistical software (SPSS Inc, Chicago, Illinois). The level of statistical significance was set at .05 for all tests.

Thirty-six patients undergoing thyroidectomy were enrolled in the study: 9 (25%) were male and 27 (75%) were female (male to female ratio, 1:3). Ages ranged from 17 to 74 years (median, 50 years). Patient demographics and clinical and pathologic data are reported in Table 1. Most of the patients were operated on because of multinodular goiter (30 [83%]). Sixteen patients (44%) were operated on by more experienced hands (surgeons), whereas the remaining 20 (56%) were operated on by less experienced hands (residents). Subtotal thyroidectomy was the most common type of operation performed (64%), and total thyroidectomy was the second most common type of operation performed (25%). Finally, only 4 patients underwent lobectomy (11%). No intraoperative or postoperative complications occurred, and all of the patients were discharged on either the first or the second postoperative day.

A doubly multivariate analysis of variance was conducted to evaluate the voice changes after thyroidectomy and the possible impact of patient sex, type of operation, and surgeon experience on acoustic parameters. Table 2 gives the multivariate effects of thyroidectomy on combinations of acoustic parameters and the interaction effects of patient sex, type of operation, and surgeon experience. The linear combination of voice parameters after thyroidectomy did not significantly differ from the preoperative combination (F_20,7 = 0.90; P = .60). This finding indicates that thyroidectomy had no effect on the combination of voice parameters. Type of surgery, surgeon experience, and sex of the patient had no effect on the combination of voice parameters before and after thyroidectomy.

Follow-up univariate analyses of variances and Wilcoxon signed rank tests were conducted to evaluate the changes in each voice parameter separately. Table 3 gives the preoperative and postoperative values of various acoustic parameters and the significance levels of the effects of within-patient factors. Regardless of within-patient factors (type of surgery, patient sex, and surgeon experience), 4 acoustic parameters significantly changed after thyroidectomy. Highest fundamental frequency, standard deviation of average fundamental frequency, phonatory average fundamental frequency range in semitones, and degree of subharmonics values were lower after the operation (P < .05). When the interaction of patient sex was considered, absolute jitter and smoothed amplitude perturbation quotient changes after thyroidectomy were different in male patients than in female patients (P < .05). However, no significant effect was found of operation type and surgeon's experience on acoustic parameters of the patients undergoing thyroidectomy (P > .05). In other words, thyroidectomy had different effects on some of the acoustic parameters of the male patients compared with the female patients, but neither the type of thyroidectomy nor the surgeon's experience affected any of the acoustic parameters after thyroidectomy.

Because sex affected some of the acoustic parameters in univariate analyses of variance, Wilcoxon signed rank tests were performed to further evaluate the changes of acoustic parameters in each sex group. Table 4 gives the results of these tests. According to these tests, none of the acoustic parameters showed significant changes in male patients, whereas significant changes in some of the acoustic parameters of female patients were observed. Highest fundamental frequency, standard deviation of average fundamental frequency, phonatory average fundamental frequency range in semitones, absolute jitter, relative average perturbation, pitch perturbation quotient, shimmer in decibels, percentage of shimmer, amplitude perturbation quotient, noise to harmonic ratio, and degree of subharmonics were all decreased in female patients after thyroidectomy (P < .05). The decrease in all these acoustic parameters of the female patients seemed to reflect an amelioration of voice after thyroidectomy.

In this study, the impact of thyroid surgery and sex of the patient, surgeon experience, and type of operation on objective acoustic parameters of patients undergoing thyroidectomy without laryngeal nerve injury were examined. The study had no control group. The postoperative acoustic parameters of the patients were compared with their preoperative values, which served as their own controls.

The number of patients included in our study is relatively small (n = 36), which might be a major limitation of the study. Although some researchers have reported larger series,^8,13 some other noteworthy studies^7,9,10 had small numbers of patients, similar to our study. Including different types of thyroidectomy in the same study may be considered another limitation. Although a study population that consists of patients who have undergone the same type of thyroidectomy might be useful in evaluating the exact impact of a certain type of surgery on the voice, such a study design would not allow comparison of the effects of different thyroidectomy types. We elected to include different types of thyroidectomy in our study to evaluate the impact of thyroidectomy type on voice, and most of the reports^1,7,10-13 concerning postthyroidectomy voice changes included patients with different types of thyroidectomy. Another limitation of our study may be the absence of late postoperative voice recordings. We were able to gather voice data of the patients only 1 week after thyroidectomy. We tried to convince our patients to undergo laryngoscopic examination and voice recording 1 month or more postoperatively, but most of them were unwilling to participate in this unpleasant and time-consuming procedure. Such problems were mentioned by other researchers.⁷ Early postoperative voice disorders after thyroidectomy without laryngeal nerve injury were reported to disappear later, and long-term voice evaluation might be necessary to document this improvement.^7,9-11 However, even in the first postoperative week, voice analysis of our patients did not reveal significant deterioration of voice parameters, questioning the necessity and the benefit of the long-term voice evaluation in our series.

In multivariate analysis, the combination of acoustic parameters of male patients in general was significantly different from that of the female patients (P = .02). This finding was mostly due to the differences in fundamental frequency parameters (P < .001). The fundamental frequency depends on the size and tension of the speaker's vocal folds at any given instant, and the average size of vocal folds in men is larger than in women; hence, the average fundamental frequency of a man in speaking a given utterance generally is lower than of a woman.¹⁴

Reports have analyzed the objective voice parameters before and after thyroid surgery, but most of these reports^3,9-11,15 used univariate statistical methods and usually examined the preoperative and postoperative changes of acoustic parameters separately. However, Stojadinovic et al⁷ used repeated-measures analysis of variance to examine the changes in acoustic parameters collectively, and they did not find any significant changes after thyroidectomy. They reported that type of thyroidectomy did not affect the changes in voice parameters. We conducted a multivariate repeated-measures analysis of variance to evaluate the effect of thyroidectomy and the interaction effects of patient sex, type of surgery, and surgeon's experience on objective voice parameters. No significant multivariate effect of thyroidectomy was found on the combination of voice parameters. Also, type of surgery, surgeon experience, and sex of the patient had no effect on the combination of voice parameters before and after thyroidectomy. However, changes of separate voice parameters occurred.

When the acoustic parameters were evaluated univariately, highest fundamental frequency, standard deviation of average fundamental frequency, phonatory average fundamental frequency range in semitones, and degree of subharmonics values showed significant differences after thyroidectomy (Table 3). Highest fundamental frequency, standard deviation of average fundamental frequency, and phonatory average fundamental frequency range in semitones are all fundamental frequency parameters, and they were decreased after thyroid surgery. Highest fundamental frequency indicates the highest fundamental frequency for all extracted pitch periods, and phonatory average fundamental frequency range in semitones is the range between the highest and lowest fundamental frequency.¹⁶ Some researchers^7,9-11 observed deteriorations of fundamental frequency parameters in the early postoperative period of the patients undergoing thyroidectomy without evident laryngeal nerve injury, but they usually recovered by the 15th day to 6th month. In a study performed by Kark et al,¹ permanent voice problems were directly related to the extent of dissection of the thyroid, occurring twice as often after bilateral (26% [40/155]) as after unilateral (11% [19/170]) removal. Debruyne et al¹⁰ found that type of surgical intervention, bilateral or unilateral, had some effect on voice changes when significant alterations were observed only in bilateral interventions. However, type of surgery had no significant effect on voice parameters after thyroidectomy in our study.

Loss of upper pitch range in our study might be indicative of clinically undetected superior laryngeal nerve palsy, but it is impossible to prove this because cricothyroid electromyography was not performed in our study. Most of the patients in our study, especially those with no subjective voice changes, were reluctant to participate with an invasive test such as laryngeal electromyography. Aluffi et al³ found electromyographically documented superior laryngeal nerve palsy in 3 of 21 postoperative patients who were symptomatic. On the other hand, some researchers^7,17 discussed the potential of using electromyography after thyroidectomy and were content with videostroboscopic evaluation only. Furthermore, it recently has been demonstrated that the objective voice measures correlate well with electromyographic findings.¹⁸

In this study, superior laryngeal vessels were individually ligated on the thyroid capsule to prevent EBSLN damage. When the nerve was not identified easily, no further dissection was performed to avoid inadvertent nerve injury. Bellantone et al⁸ found that careful distal ligation of the superior thyroid vessels is a safe technique to preserve the EBSLN, making routine exposure unnecessary during thyroidectomy.

There is a prevailing belief that surgeon experience affects patient outcomes in thyroid surgery. It is reported that individual surgeon experience is significantly associated with complication rates and length of stay for thyroidectomy.¹⁹ Nerve injury was more common in less experienced surgeons. Surgeon experience may have some effect on objective acoustic parameters after thyroidectomy without laryngeal nerve injury; however, to our knowledge, no reports have mentioned this issue. More experienced surgeons may perform more delicate dissection techniques, providing less injury to laryngeal nerves or strap muscles. However, our results failed to reveal a significant effect of surgeon experience on objective voice parameters of patients undergoing thyroidectomy without laryngeal nerve injury.

Thyroidectomy may have different effects on voice parameters of female and male patients. Objective voice changes in female patients were more common than in male patients.^7,15 Similar results in our study showed that objective acoustic parameters of female patients were affected after thyroidectomy, whereas voices of male patients were not affected. The affected parameters in female patients were highest fundamental frequency, standard deviation of average fundamental frequency, phonatory average fundamental frequency range in semitones, absolute jitter, relative average perturbation, pitch perturbation quotient, shimmer in decibels, percentage of shimmer, amplitude perturbation quotient, noise to harmonic ratio, and degree of subharmonics (Table 4). Surprisingly, we found that those values were all lower after thyroidectomy compared with preoperative values, indicating an amelioration of voice in female patients. Improvement of objective voice parameters after thyroidectomy was observed by some researchers. Watt-Boolsen et al²⁰ reported that 12 patients with sporadic nontoxic goiter and 8 with medically pretreated toxic goiter had impaired voice function that improved after thyroid surgery in the nontoxic goiter group but remained unchanged in the toxic goiter group. Postthyroidectomy voice improvement was reported by McIvor et al¹² in 36% of patients. In the same study, 47% of patients had no change and 17% had worsening of their objective acoustic parameters. It was postulated that after removal of the thyroid gland a better vibration of the laryngotracheal complex and relief of a slight respiratory obstruction might be obtained, especially in patients with large or compressive goiters. Such patients should have preoperative voice abnormalities as mentioned in both series previously discussed. When the mean preoperative values of acoustic parameters of our patients were compared with preset normative thresholds of the Multi-Dimensional Voice Program, most of the parameters of male patients seemed to be below the threshold values, whereas most of the parameters of female patients were above the thresholds, indicating that female patients might have objective voice abnormalities but male patients might have normal voices preoperatively (Table 4). Some significantly improved parameters in female patients (eg, relative average perturbation, pitch perturbation quotient, and amplitude perturbation quotient) even fell below the threshold values, pointing out a voice correction.

Although not statistically significant, the acoustic parameters of the male patients tended to be worse after thyroidectomy. Some preoperatively normal parameters, such as amplitude perturbation quotient, smoothed amplitude perturbation quotient, and percentage of shimmer, were even above the threshold values postoperatively, indicating a worsening of the voices. The few male patients (n = 9) in this study might have prevented us from making a statistically significant conclusion, but others^7,9,10,15,17 have reported some voice deteriorations after thyroidectomy without any evident laryngeal nerve injury. Such changes were thought to be the result of the injury of the prethyroid strap muscles and cricothyroid muscles or impairment of laryngotracheal movement due to wound contracture after surgical trauma of the soft tissues, but they were usually temporary and resolved after a few weeks to months postoperatively.

Another issue that should be discussed may be the reason why female patients had preoperative voice abnormalities and male patients did not. As mentioned previously, some researchers^7,15 reported that objective voice changes in female patients were more common compared with male patients. Most patients were female in the other studies^9-11 that reported voice alterations without mentioning a sex effect. Mineralization of thyroid cartilage differs in male and female patients, and a preponderance of laryngeal cartilage ossification has been reported in men compared with women radiographically.²¹ In contrast to male thyroid cartilages, in which cartilage mineralization continued until advanced age followed by nearly complete ossification, the ventral half of the female thyroid cartilage remained unmineralized and did not ossify.²² A nonossified relatively soft female larynx may be more susceptible to mechanical effects preoperatively and postoperatively, such as compression by a large goiter before surgery or diminished mobility of muscles and soft tissues after thyroidectomy. Studying the thyroid cartilage ossification status with acoustic analysis of voice changes in patients undergoing thyroidectomy would be interesting, but such data were not recorded in our study.

In conclusion, postoperative voice changes may be an important part of the outcome of the patients undergoing thyroidectomy. Deterioration and amelioration of acoustic parameters can be observed, and those changes may occur differently among male and female patients. Physicians should take the patient's sex into consideration when informing the patient about possible voice alterations after thyroidectomy. Preoperative and post operative acoustic analyses may be helpful in revealing any voice abnormality already present before the surgery and any possible alteration occurring after the surgery. Even though type of thyroidectomy and surgeon's experience did not have a significant effect on voice parameters after thyroidectomy in the present study, further investigations based on a larger cohort series are necessary to determine the exact impact of these factors.

Correspondence: Serdar Akyildiz, MD, Department of Otolaryngology, Ege University Faculty of Medicine, Bornova, 35100 Izmir, Turkey (serdar@akyildiz.net).

Submitted for Publication: August 1, 2007; final revision received October 19, 2007; accepted October 24, 2007.

Author Contributions: Dr S. Akyldiz had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: S. Akyildiz and Ogut. Acquisition of data: S. Akyildiz, Ogut, and M. Akyildiz. Analysis and interpretation of data: S. Akyildiz and Engin. Drafting of the manuscript: S. Akyildiz and Engin. Critical revision of the manuscript for important intellectual content: S. Akyildiz, Ogut, and M. Akyildiz. Statistical analysis: S. Akyildiz and Engin. Administrative, technical, and material support: S. Akyildiz and Ogut. Study supervision: S. Akyildiz, Ogut, and M. Akyildiz.

Financial Disclosure: None reported.