Introduction

Voice and swallowing impairment, the so-called post-thyroidectomy syndrome, occurs frequently even without recurrent laryngeal nerve (RLN) palsy [1,2,3,4,5,6,7,8,9,10,11]. Voice symptoms after thyroidectomy include voice fatigue, roughness, difficulties with high frequencies and singing, and volume reduction [1, 3]. The possible causes of voice impairment after thyroidectomy without RLN palsy include injury to the external branch of the superior laryngeal nerve (EBSLN), cricothyroid muscle impairment, laryngotracheal fixation, endotracheal intubation, vocal cord edema or altered vascular supply and venous or lymphatic drainage, injury to the fine anastomotic branches connecting the RLN and EBSLN, laryngopharyngeal reflux, severe retractile cervical scar, local neck pain, and psychotic reaction and stress [6, 7, 12,13,14,15,16,17].

Voice quality is a matter of concern for most patients who undergo thyroidectomy. In general, subjective voice symptoms increase immediately after thyroidectomy and decrease gradually to pre-operative levels after about 3–6 months [10, 11, 18,19,20,21]. However, most previous evaluations of voice outcomes after thyroidectomy involved only short-term follow-up, usually of 3 or 6 months, and long-term voice outcomes after thyroidectomy have not yet been well evaluated, with only a few long-term follow-up studies. In addition, the effect of endotracheal intubation on voice outcome after thyroidectomy is not investigated well although endotracheal intubation is considered as one of the reasons of voice impairment after thyroidectomy, and the significant changes of jitter and shimmer values after endotracheal intubation due to vocal cord edema was reported [14].

The aim of this study was to examine long-term subjective and objective functional voice outcomes after thyroidectomy in patients without RLN palsy and in a control group of patients undergoing parotidectomy, and also to evaluate the effect of endotracheal intubation on post-thyroidectomy voice impairment.

Methods

Patients

We prospectively enrolled 159 patients with thyroid nodules who underwent the conventional trans-cervical thyroidectomy with or without central neck dissection between March 2013 and March 2015, and who agreed to undergo serial voice assessment before and after surgery (the case group) in a tertiary hospital. Sixty-nine patients with benign parotid tumors who underwent parotidectomy over the same period and who also agreed to undergo serial voice assessment before and after surgery served as the control group, because anterior neck structures such as the strap muscles and RLN are not affected by parotidectomy. Exclusion criteria were < 18 and > 70 years of age, vocal mucosal lesions such as vocal nodules, polyps, sulcus vocalis, and Reinke’s edema, pre-operative RLN paralysis, vocal cord movement disorder, and severe post-operative hematoma or hypocalcemia. Also excluded were patients who underwent completion thyroidectomy or lateral neck dissection simultaneously with thyroidectomy, patients with a history of neck irradiation or previous neck surgery, and patients with apparent injury related to endotracheal intubation. Patients with post-operative RLN paralysis or apparent injury to the EBSLN after thyroidectomy were excluded. Patients who underwent selective neck dissection along with parotidectomy or had facial nerve paralysis after parotidectomy were also excluded from the control group. All patients provided informed consent for the voice assessments and serial follow-up. The Institutional Review Board of Hanyang University Hospital approved the study.

The conventional trans-cervical thyroidectomy was performed using a transverse incision two finger breadths above the suprasternal notch. The sternothyroid muscles were not cut routinely. The RLN was identified and preserved routinely using a loupe. Intraoperative RLN and EBSLN neuromonitoring was not performed routinely. Parotidectomy was performed via a modified facelift approach preserving the facial nerve.

Functional voice outcomes were evaluated serially using a self-assessment voice questionnaire and simultaneous objective acoustic voice analysis and aerodynamic study on the day before surgery, and 1 day, 3 days, 1 week, 1 month, 3, 6, 12, 18, and 24 months after surgery, where possible. However, in the control (parotidectomy) group, subjective and objective voice analysis was evaluated pre-operatively and post-operatively only up to 12 months, because further follow-up was very difficult. All voice questionnaire and acoustic voice analysis was performed by an expert speech-language pathologist. Videolaryngostroboscopy with an EndSTROB DX (Sion, Berlin, Germany) and fiberoptic flexible laryngoscopy were also routinely performed the day before surgery and the day after surgery in all patients to assess movement of the vocal folds, vocal mucosal lesions, arytenoid asymmetry, vocal cord bowing, degree of glottis closure, and mucosal wave characteristics.

Voice symptom scores (VSS)

Subjective voice outcomes were assessed using a self-assessment voice questionnaire developed specifically by this group [22, 23]. The questionnaire consists of five questions concerning changes in voice pitch, range, intensity, fatigability, and singing quality: (1) Do you have vocal fatigue? (2) Do you have a hoarse voice? (3) Is it difficult to produce a high pitch? (4) Is your voice weak or breathy? (5) Do you have difficulty singing? Each question was scored on a 0–4 symptom grading scale: 0 (no symptoms), 1 (mild), 2 (moderate), 3 (severe), and 4 (very severe), and the voice symptom score (VSS) is defined as the sum of the scores on the five questions. It, therefore, ranges from a minimum of 0 to a maximum of 20.

Acoustic analysis and aerodynamic study

The acoustic voice analysis was performed using a Multidimensional Voice Program (MDVP) and Voice Range Profile (VRP) in the Computerized Speech Lab software (model 4150B, Kay PENTAX, Lincoln Park, NJ, USA). For the MDVP analysis, after three training emissions, a consistent vowel /a/ produced at a comfortable pitch and loudness level for at least 5 s was recorded. We chose an interval of 3 s from the mid-portion of each vowel. The following parameters were analyzed in the MDVP: fundamental frequency (F0, Hz), jitter (%), shimmer (%), and noise-to-harmonic ratio (NHR, dB). In the VRP analysis, patients were asked to phonate a consistent vowel /a/ as loudly and as softly as possible from the lowest to highest frequencies. Lowest (F-low) and highest (F-high) frequencies (Hz), frequency range (Hz), and intensity range (dB) were measured in the VRP analysis.

Maximum phonation time (MPT) was measured aerodynamically. It was obtained by having patients sustain the vowel /a/ for as long as possible on a single breath. The longest of three attempts was recorded as the MPT.

Statistical analysis

Statistical analysis was performed with SPSS 20 (SPSS Inc., Chicago, IL, USA). Categorical variables were evaluated by the Chi-square test or Fisher’s exact test in cases with small cell sizes. Continuous variables were analyzed by t tests or the Mann–Whitney U test in cases with small cell sizes. Paired t tests were used to compare post-operative VSS, acoustic parameters, and MPT with pre-operative variables. A p value less than 0.05 was considered significant.

Results

Clinicopathological characteristics

Of 159 thyroidectomy patients enrolled as the case group, three (1.9%) had temporary RLN palsies after the operation and were excluded from the study, and one patient who underwent reoperation under general anesthesia to remove a post-operative hematoma was also excluded. Finally, 155 thyroidectomy cases were included in the study. Table 1 presents the clinicopathologic characteristics of the case and control groups. In the thyroidectomy (case) group, there were 109 women and 46 men, and the mean age was 48.8 ± 17.1 years. There were 30 (19%) benign thyroid nodules and 125 (81%) malignancies including papillary and follicular carcinomas. Lobectomy was performed in 34 (22%) patients and total thyroidectomy in 121 (78%). The control group consisted of 54 women and 15 men. There were no significant differences in sex, age, and operative time between the case and control groups. Among the risk factors that can affect voice quality, smoking and alcohol consumption did not differ between the groups. However, laryngopharyngeal reflux disease (LPRD) was significantly more common in the thyroidectomy group than the control group.

Table 1 Clinicopathologic characteristics of the thyroidectomy (case) and parotidectomy (control) groups
Full size table

Changes of voice symptom scores (VSS)

The questionnaire about subjective voice impairment was completed by all 224 patients in the case and control groups 1 day before surgery and 1 day, 3 days, and 1 week post-operatively, and by 209 patients after 1 month, 196 patients after 3 months, 165 patients after 6 months, and 129 patients after 12 months. In the case group, 103 patients completed the voice questionnaire at 18 months post-operatively and 86 at 24 months.

The changes of VSS are shown in Table 2. In the thyroidectomy group, VSS increased immediately after the surgery and did not return to the pre-operative value until 24 months after surgery. In the control group, VSS also increased significantly on days 1 and 3 after surgery, but it returned to the pre-operative level within a week. The post-operative VSS of the case group was significantly higher than that of the control group from day 1 to 12 months post-operatively, the last time VSS was evaluated in the control group. The pre-operative VSS of the thyroidectomy group was also significantly higher than that of the control group.

Table 2 Pre- and post-operative voice symptom scores (VSS) in the thyroidectomy (case) and parotidectomy (control) groups
Full size table

In a stratification analysis according to the extent of thyroidectomy, the VSS of the total thyroidectomy patients was higher than that of the lobectomy group from 1 day to 1 week after surgery (data not shown). However, the difference disappeared after 1 month.

Changes of acoustic parameters and MPT

The acoustic analysis and MPT were completed in all 224 patients on pre-operative day 1 and 1 day, 3 days, and 1 week post-operatively, and by 209 patients after 1 month, 189 patients after 3 months, 158 patients after 6 months, and 120 patients after 12 months. In the case group, 99 patients completed the acoustic analysis and MPT 18 months post-operatively and 80 patients 24 months post-operatively.

We compared the acoustic parameters and MPT according to gender, because the fundamental frequency is different in females and males. The changes of acoustic parameters and MPT in females are shown in Table 3. Fundamental frequency (F0), jitter and NHR did not change significantly after surgery in either the case or control groups, except at occasional time points. Shimmer was significantly worse for 12 months after thyroidectomy, except in the 1st week post-operatively. F0, jitter, shimmer, and NHR did not differ significantly between the case and control groups at most time points.

Table 3 Comparison of pre- and post-operative acoustic parameters and MPT in females in the thyroidectomy and parotidectomy groups
Full size table

Highest frequency (F-high) decreased significantly for 6 months after surgery in the thyroidectomy group, and returned to the pre-operative level by 12 months after surgery. It decreased significantly on days 1 and 3 after surgery in the control group. F-high was significantly lower in the case group than the control group for the first post-operative month. Low frequency (F-low) decreased significantly for the first 3 months after surgery in the thyroidectomy group, but did not change in the control group. F-low did not differ significantly between the case and control groups.

Frequency range decreased significantly for a month after surgery in the thyroidectomy group and for 3 days after surgery in the control group. The frequency range of the thyroidectomy group was significantly lower than that of the control group for a month after surgery. Intensity range decreased for a month after surgery in the thyroidectomy group, and for 3 days in the control group. It did not differ significantly between the case and control groups before or after surgery. MPT decreased significantly for 18 months after surgery in the thyroidectomy group, and for only 1 day in the control group. It did not differ significantly between the case and control groups.

The results for the acoustic parameters and MPT in males are presented in Table 4. F0, jitter, shimmer, NHR, F-low, and MPT were not significantly affected at most times after surgery in either the case or control groups, and they did not differ between the groups.

Table 4 Comparison of pre- and post-operative acoustic parameters and MPT in males in the thyroidectomy and parotidectomy groups
Full size table

F-high decreased significantly for the first 12 months (except at 1 week) after surgery and frequency range decreased significantly for 3 months (except for week 1) after surgery in the thyroidectomy group. F-high and frequency range did not differ significantly between case and control groups. Intensity range decreased for 3 months in the thyroidectomy group. It also decreased for the first week post-operatively in the control group. There was no significant difference in intensity range between the case and control groups.

In a stratification analysis according to the extent of thyroidectomy, the acoustic parameters and MPT did not differ between the total thyroidectomy and lobectomy cases in either females or males (data not shown).

Discussion

In this study, we evaluated long-term functional voice outcomes after thyroidectomy without RLN palsy serially for up to 2 years. A few previous studies have evaluated outcomes for more than a year. In the study, self-assessment VSS decreased immediately after thyroidectomy and the decrease was maintained for 18 months post-operatively. Highest frequency also worsened immediately after thyroidectomy, and recovered by 12 months and 18 months post-operatively in females and males, respectively, whereas they recovered to pre-operative levels by 1 week after surgery in the control group. It is noteworthy that impairment of subjective voice capacities and F-high continued for more than 18 months and 6–12 months, respectively, after thyroidectomy, quite long compared to the previous studies. In many such studies, subjective voice outcomes and acoustic parameters recovered within 3 months after surgery, or at most 6 months [11, 19,20,21]. In the few articles that analyzed voice outcomes after thyroidectomy for up to 1 year, subjective voice impairment resolved within 6 months or 12 months, and only a few patients complained of voice impairment 1 year after surgery [6, 24, 25]. We need to recognize the possibility of long-term voice impairment after thyroidectomy and should inform patients of it pre-operatively.

The most commonly affected acoustic parameters after thyroidectomy are highest frequency and frequency range. In the previous studies, acoustic parameters including highest frequency and frequency range usually declined after thyroidectomy and recovered within 3 months [1, 4, 9]. However, as shown in this study, impairment of F-high can continue for 6 or 12 months after thyroidectomy. This provides objective evidence for long-term post-thyroidectomy voice impairment. F-high in the thyroidectomy groups was significantly worse than in the control group in females. In males, F-high in the thyroidectomy cases was also worse than in the control group, but the effect did not reach statistical significance, perhaps, due to the relatively small sample size.

In a meta-analysis, F0, shimmer and MPT significantly worsened in the early period after thyroidectomy, whereas jitter and NHR remained unchanged throughout [26]. In the present study, F0 decreased only on day 1 and shimmer increased for 12 months after thyroidectomy in females, whereas neither parameter changed after surgery in males. MPT also decreased after thyroidectomy for 18 months in females.

Laryngotracheal fixation is considered an important cause of post-thyroidectomy voice impairment [1] as is endotracheal intubation [7, 13, 27]. In this study, we investigated the effect of endotracheal intubation on post-thyroidectomy voice outcomes by analyzing self-reported voice symptoms and acoustic parameters. Impairment of VSS, F-high and frequency range continued for 1–18 months after thyroidectomy, compared with only 3 days after parotidectomy. These findings suggest that the early post-operative voice impairment in the parotidectomy group is associated with endotracheal intubation, while the long-term voice impairment after thyroidectomy is caused by the thyroidectomy itself.

Interestingly, pre-operative VSS in the thyroidectomy group was significantly worse than in the control group. The standard deviation of pre-operative VSS was significantly larger than the mean value. We do not know the exact cause of this effect. However, the worse pre-operative VSS in the thyroidectomy group might be related to compressive symptoms of thyroid nodule, LPRD or psychological attitude. Actually, thyroid pathologies could cause compressions on the larynx, the respiratory and/or digestive system, and voice and swallowing complaints are common in pre-operative thyroidectomy patients [28, 29]. The incidence of LPRD was significantly higher in the thyroidectomy group and LPRD can cause hoarseness, globus symptoms, throat discomfort, and throat clearing like post-thyroidectomy syndrome. In addition, patients with thyroid nodules may generally be more concerned or anxious about their throat and voice than patients with benign parotid tumors.

This study has several limitations. First, we did not perform laryngeal electromyography (EMG) to exclude EBSLN paralysis definitively. However, videolaryngostroboscopic examination was performed in all patients before and after thyroidectomy and there was no evidence of EBSLN paralysis in any of them. Second, the voice questionnaire for VSS developed by us was not validated officially, although it was used in our previous studies [16, 17]. However, it includes questions about all five typical symptoms, and it is easy and quick to use. Third, the incidence of LPRD was significantly higher in the thyroidectomy group than the control group and LPRD could affect voice outcomes. Fourth, we did not compare voice outcomes in the case and control groups at 18 and 24 months after surgery. However, the results at post-operative 18 and 24 months in the control group would probably not have differed from pre-operative levels because VSS and all acoustic parameters returned to pre-operative levels within 1 week of surgery.

Conclusion

Self-assessed voice symptoms and effects on objective acoustic parameters can persist for up to 18 months after thyroidectomy. This possibility should be explained to patients and discussed in pre-operative counseling. Endotracheal intubation can affect voice function adversely for 1 week post-operatively.