The Laryngoscope C 2015 The American Laryngological, V
Rhinological and Otological Society, Inc.
Systematic Review
A Systematic Review and Meta-Analysis on Acoustic Voice Parameters After Uncomplicated Thyroidectomy Brian H. H. Lang, MS, FRACS; Carlos K. H. Wong, PhD; Estella P. M. Ma, PhD Background: Postthyroidectomy voice changes are common even without apparent laryngeal nerve injury. Our study evaluated the impact of open cervical thyroidectomy on five acoustic voice parameters in the early (< 3 months) and late ( 3 months) postoperative periods. Methods: A systematic review was performed to identify studies that quantitatively assessed voice quality by acoustic voice analysis before and after thyroidectomy. Parameters included average fundamental frequency (F0, Hz), jitter (%), shimmer (%), noise-to-harmonic ratio (NHR), and maximum phonation time (MPT) (in secs). Meta-analysis was performed using both fixed- and random-effects models. Results: A total of 896 patients were analyzed. Relative to baseline, F0 significantly worsened in the early period (from 194.9 6 34.9 Hz to 188.0 6 34.0 Hz, P 5 0.001). This was equivalent to a quarter-tone loss (P 5 0.004). Shimmer (from 3.15 6 1.59% to 3.19 6 1.70%, P 5 0.040) and MPT (from 17.9 secs to 16.7 secs, P 5 0.046) also worsened in the early period, whereas jitter and NHR remained unchanged in the early and late periods. Males suffered greater deterioration in F0 (from 120.6 6 18.8 Hz to 111.0 6 18.5 Hz, P 5 0.048) and in NHR (from 0.12 6 0.02 to 0.16 6 0.03, P 5 0.019) than females in the early period. Four of the five acoustic parameters (F0, jitter, shimmer, and NHR) significantly worsened after total thyroidectomy (TT) and not after lesser resection. Conclusion: F0, shimmer, and MPT significantly worsened in the early and not in the late postoperative period. F0 impairment was perceptually significant. Males and those undergoing TT suffered greater voice impairment than their counterparts during the early period. Laryngoscope, 126:528–537, 2016
INTRODUCTION Thyroidectomy is one of the most common surgical procedures undertaken worldwide.1 Although it is associated with a low morbidity when done by high-volume ( 100 cases per year) surgeons, debilitating postoperative voice changes occur not uncommonly without apparent recurrent laryngeal nerve (RLN) injury.2,3 Presumed causes include injury to external branch of superior laryngeal nerve (EBSLN), orotracheal intubation, surgical adhesions, strap muscles denervation, and pain or psychological distress. A recent survey of > 200 postthyroidectomy patients found that over a third of them reported persistent vocal changes and felt these
Additional Supporting Information may be found in the online version of this article. From the Department of Surgery (B.H.H.L.); Department of Family Medicine and Primary Care, 3/F Ap Lei Chau Clinic, (C.K.H.W.); and Division of Speech and Hearing Sciences (E.P.M.M.), the University of Hong Kong, Hong Kong SAR, the People’s Republic of China. Editor’s Note: This Manuscript was accepted for publication on May 19, 2015. The authors have no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Dr. Brian H.H. Lang, Department of Surgery, Queen Mary Hospital, 102 Pokfulam Road, Hong Kong SAR, People’s Republic of China. E-mail: [email protected] DOI: 10.1002/lary.25452
Laryngoscope 126: February 2016
528
changes more concerning than traditional complications such as hypocalcemia or unsightly cervical scar.4 Common voice changes including upper pitch loss, difficulty speaking aloud, and voice huskiness could be objectively quantified by a computerized acoustic voice program such as the multidimensional voice program (MDVP).2,3 Five objective parameters have generally been used to measure voice quality: 1) average fundamental frequency (F0), 2) jitter, 3) shimmer, 4) noise-to-harmonic ratio (NHR), and 5)maximum phonation time (MPT). Fundamental frequency is a measurement of pitch, whereas jitter and shimmer measure pitch and amplitude variability, respectively.3,5 Noise-to-harmonic ratio measures the amount of noise present, whereas MPT measures voice capacity (i.e., efficiency of voice production).3,5 Although numerous studies reported results of acoustic voice analysis in postthyroidectomy patients, findings were inconsistent. Furthermore, due to the difficulty with patient recruitment and compliance, few studies had sufficient power to quantify longer-term (more than 3 months) acoustic changes. Therefore, it remains unclear whether acoustic voice impairments persist. Our study aimed to summarize the findings from the literature and evaluate the impact of open cervical thyroidectomy on acoustic voice quality parameters in the early (< 3 months) and late periods ( 3 months) by pooling data from the current literature.
Lang et al.: Voice Impairment After Thyroidectomy Appears Temporary
MATERIALS AND METHODS This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses statement.6
Search Strategy Studies evaluating voice quality or voice-related outcomes after open cervical thyroidectomy were retrieved from the Scopus, Medline (PubMed), and Cochrane Library electronic databases on December 2, 2014. We used the following free text search terms in “All fields”: #1: #2: #3: #4:
‘voice’ ‘voice quality’ ‘thyroidectomy’ #1 OR #2 AND #3
There was no language restriction or methodological filters. The bibliography of two recent reviews on voice after thyroidectomy was searched for additional relevant references.2,7
Study Selection All titles identified by the search strategy were independently screened by three authors (B.H.H.L., CK.H.W., E.P.M.M.). Search results were compared, and disagreements were resolved by consensus. Abstracts of potentially relevant titles were then reviewed for eligibility, and full-length articles were selected for closer examination. Any studies that quantitatively assessed voice quality by acoustic voice analysis under a controlled environment were considered. Eligible studies had to report data on at least one of the five acoustic parameters (F0, jitter, shimmer, NHR, and MPT) generated from MDVP before and after an open cervical thyroidectomy. Controlled environment meant that all audio recordings were done in a quiet room, with the patient sitting 15- to 20-cm away from a high-quality dynamic microphone, and then digitized and analyzed using MDVP (Computerized Speech Lab Model, Kay Elemetrics, Lincoln Park, NJ, USA). For MPT, the longest of three attempts was generally taken. Studies that did not specify the exact time of the postoperative testing or did not separate the results of thyroidectomy from other nonthyroid operations (such as parathyroidectomy) or noncervical approaches were excluded. Case reports, editorials, expert opinions, reviews without original data, and studies on pediatric population were excluded. Multiple reports of the same dataset were assessed, and only the most representative or updated report of a study was included.
Data Extraction All data were extracted onto a standardized form. The primary data extracted from each article included: type or design of study, type of surgical services (low- or high-volume center), first authorship, country of origin, year of publication, demographics, extent of surgery (total thyroidectomy [TT] or less than TT [LTT]), thyroid pathology, and type and timing of postoperative acoustic voice assessments. A high-volume center was defined as one that performed over 100 cases per year.8 Because there was no standardization on the timing of acoustic assessments, they were categorized into early and late periods. Early period was defined as within 3 months of thyroidectomy, whereas late was defined as 3 months or more after thyroidectomy. If two or more assessments were performed within the early or late period (e.g., at 4 and 10 weeks), the average of these assessments was taken. Total thyroidectomy comprised near-TT and TT, whereas LTT comprised lobectomy and subtotal thyroidectomy. Fundamental frequency was expressed in Hz,
Laryngoscope 126: February 2016
Fig. 1. A flow diagram for study selection.
whereas jitter and shimmer were expressed in percentage. Noise-to-harmonic ratio was a ratio and thus no unit was given. Maximum phonation time was expressed in seconds.
Statistical Analysis Paired t test was used for comparison of five acoustic parameters between preoperative and postoperative assessments. To make it more clinically relevant, if significant, comparison was repeated with both preoperative and postoperative F0 data being converted into semitones by logarithmic scale similar to a piano scale. The conversion equation used was log10 (the highest frequency in Hz/the lowest frequency in Hz)/log102 3 12). All the individual outcomes were integrated with the meta-analysis software Review Manager Software 5.0 (Cochrane Collaborative, Oxford, England). Standardized mean differences (SMD) were calculated for significant findings. Results were aggregated and analyzed using both fixed- and random-effect models. Publication bias was estimated by Begg’s rank correlation test and Egger’s regression test.9,10 These meta-analyses were conducted using IBM SPSS 20.0 for Windows (IBM Corp., Armonk, NY) and Comprehensive Meta-Analysis Version 2.2.064 (Biostat, Inc., Englewood, NJ, USA).
RESULTS Of the 950 titles initially identified from the database search (Fig. 1), 36 full-length articles were assessed for inclusion. Twenty were excluded, and 16 studies3,5,11–24 were determined to be eligible for inclusion. Appendix 1 lists these 20 articles25–44 and the reason for their exclusion. No additional study was found from our search of the two bibliographies in previous reviews.2,7
Patient Selection Consecutive patients undergoing thyroidectomy were recruited in all the studies examined. All studies
Lang et al.: Voice Impairment After Thyroidectomy Appears Temporary
529
TABLE I. Patient Baseline Characteristics and Types of Voice Assessment in the 16 Included Studies. Extent of Surgery
Study Design
No. of Patients*
Sex (M:F)
Mean Age 6 SD (range)
PS
54
9:45
43 (13–67)
46
Debruyne11 (1997)
PS
47
0:47
Not available
Aluffi12 (2001)
RS
42
6:36
52 (33–76)
Stojadinovic3 (2002)
PS
50
10:40
Sinagra13 (2004) de Pedro Netto14 (2006)
PS PS
46 88
3:43 11:77
First Author (year) 5
Hong (1997)
TT
Postoperative Follow-up Visits
Underlying Thyroid Pathology
Subjective/ Perceptual Assessment
8
1m, 3m, 6m
Benign 1 malignant
47
0
D4,2w,3m
Benign 1 malignant
Symptoms, questionnaire None
9
33
12–18m
Benign
Symptoms
53 (22–77)
21
29
2w,3m
Benign 1 malignant
Symptoms, GRBAS
43 (20–70) 46 (17–79)
0 35
46 53
2m, 4m, 6m 2w
Benign 1 malignant Benign 1 malignant
– GRBAS
LTT
Soylu15 (2007)
PS
48
6:42
47 (19–72)
8
40
D2, 3m
Benign 1 malignant
Symptoms
Akyildiz16 (2008) Lombardi17 (2009)
PS PS
36 110
9:27 13:97
50 (17–74) 46.5 6 13.2
27 0
9 110
1w 3m, 12m
Benign 1 malignant Benign 1 malignant
– Symptoms, questionnaire
Van Lierde18 (2010)
RS
44
12:32
49 (17–72)
7
37
1w, 6w, 3m
Benign 1 malignant
GRBAS, questionnaire
Li19 (2012)
RS
32
8:24
48 6 9
32
0
1w, 1m, 3m
Not available
Symptoms
Lombardi20 (2012)
PS
32
7:25
44.5 6 10.7
0
32
3m
Benign 1 malignant
Tae21 (2012)
PS
61
0:61
54.4 6 10.6
10
51
1w, 1m, 3m, 6m
Benign 1 malignant
Symptoms, questionnaire Symptoms, questionnaire
Lee22 (2012)
PS
46
6:40
38.5 6 68
11
35
1w, 3m
Malignant
GRBAS
Ryu23 (2013)
PS
74
33:41
44.8 6 8.5† 50.8 6 11.3†
33
41
1m, 6m, 12m
Malignant
Symptoms, GRBAS
Jung24 (2013)
PS
86
12:30
48.0 6 8.7‡
22
20
2w, 3m
Malignant
Questionnaire
6:38
51.8 6 8.7‡
22
22
151:745
47.1 6 24.0
330
566
–
–
Overall
–
896
–
D 5 day; GRBAS scale 5 grade, rough, breathy, asthenic, strained; LTT 5 less than total thyroidectomy; m 5 month; PS 5 prospective study; RS 5 retrospective study; TT 5 total thyroidectomy; w 5 week. *After excluding those with incomplete follow-up or postoperative vocal cord paresis. † Data for LTT and TT. ‡ Data for subplatysmal and subfascial groups.
considered previous laryngeal diseases, vocal cord paralysis, and incomplete follow-up as exclusion criteria. Some studies also imposed an age exclusion criterion.17,20–24
Baseline Characteristics and Subjective Voice Assessments Table I shows a summary of the baseline characteristics and type of subjective voice assessments in the 16 eligible studies (3 retrospective and 13 prospective). All 16 studies were conducted in high-volume centers. In the two robotic-thyroidectomy series, only data from the open group were used.21,23 A total of 896 patients were analyzed. The cohort comprised predominantly females (83.1%) with a mean (6 standard deviation) age of 47.1 6 24.0 years old. The LTT:TT ratio was 1 to 1.7. Three (18.8%) studies exclusively analyzed TT,13,17,20 whereas two (12.5%) studies only analyzed LTT.12,19 Twelve (75.0%) studies performed acoustic voice analysis two or more times in the postoperative period.3,5,11,13,15,17–19,21–24 Eleven (68.8%) studies analyzed benign and malignant cases together,3,5,11,13–18,20,21 whereas three (18.9%) Laryngoscope 126: February 2016
530
included malignant cases only 22–24 and one (6.3%) had benign case only.12 One (6.3%) study did not mention underlying thyroid pathology.19 Subjective voice changes were commonly reported after thyroidectomy, and the incidence ranged between 29.7% to 87.0%, depending on definition and timing of the assessment.13,14,17 “Voice fatigue” was the most common complaint, and this was followed by “loss of high pitch” and “difficulty of loud voice.”3,5,12,13 One study found there was a significant relationship between the extent of surgery and voice change15 whereas another study did not.16 Nevertheless, these voice complaints tended to improve over time.15,19 Changes in voice quality by perceptual rating using the GRBAS (grade of hoarseness, roughness, breathiness, asthenic and strain) scale was reported in several studies.3,14,18,22,23 Significant increases in GRBAS score were found in the first 2 weeks but less so after 3 to 6 months.3,14,18 Postoperative videolaryngostroboscopy (VLS) was done in all of the studies to exclude any postoperative vocal cord abnormalities and RLN injury, but only three studies performed routine postoperative cricothyroid electromyography (CT EMG) in
Lang et al.: Voice Impairment After Thyroidectomy Appears Temporary
Laryngoscope 126: February 2016
Lang et al.: Voice Impairment After Thyroidectomy Appears Temporary
531
183.1 6 35.8
TT (n 5 41)
Lee22 (2012) Ryu23 (2013)
196 6 35
188.1 6 34.8*
180.00 6 31.30
180.4 6 30.5 192.7 6 35.6
169.80 6 43.40
169.7 6 47.9
174.10 6 38.10
189.70 6 45.00 187.70 6 50.50
201.40 6 34.90
231.80 6 44.75 –
117.43 6 16.74
104.75 6 17.62 189.10 6 27.30
–
118.06 6 28.20 215.80 6 24.70
221.64 6 26.30
197.80 6 27.10 217.88 6 71.80
110.30 6 7.50
168.07 6 45.50 176.00 6 19.00
170.82 6 44.40
205.4 6 19.7 –
1.08 6 0.76 0.90 6 0.62
190.5 6 38.1†
1.01 6 0.79
1.49 6 0.87
1.37 6 0.40 1.39 6 0.97
1.31 6 0.89
1.65 6 1.14 0.63 6 0.22
1.22 6 0.85
1.0 6 0.61 1.0 6 0.59
0.5 6 0.3
0.70 6 0.25 1.45 6 1.03
0.397 6 0.15
0.5 6 0.3 0.605 6 0.39
0.4 6 0.2
0.68 6 0.39 –
0.55 6 0.23
1.14 6 0.89 0.64
0.44 6 0.10
0.99 6 0.72†
0.98 6 1.04
0.99 6 1.05
1.98 6 1.21
1.38 6 0.45 1.42 6 0.59
1.10 6 0.70
1.27 6 1.21 –
1.29 6 0.99
1.25 6 0.79 1.19 6 0.67
–
0.91 6 0.81 0.99 6 0.76
0.70 6 0.16
0.50 6 0.30 0.77 6 0.56
0.60 6 0.30
0.71 6 0.40 –
0.56 6 0.46
1.13 6 0.89 –
0.41 6 0.06
Rhinological and Otological Society, Inc.
Systematic Review
A Systematic Review and Meta-Analysis on Acoustic Voice Parameters After Uncomplicated Thyroidectomy Brian H. H. Lang, MS, FRACS; Carlos K. H. Wong, PhD; Estella P. M. Ma, PhD Background: Postthyroidectomy voice changes are common even without apparent laryngeal nerve injury. Our study evaluated the impact of open cervical thyroidectomy on five acoustic voice parameters in the early (< 3 months) and late ( 3 months) postoperative periods. Methods: A systematic review was performed to identify studies that quantitatively assessed voice quality by acoustic voice analysis before and after thyroidectomy. Parameters included average fundamental frequency (F0, Hz), jitter (%), shimmer (%), noise-to-harmonic ratio (NHR), and maximum phonation time (MPT) (in secs). Meta-analysis was performed using both fixed- and random-effects models. Results: A total of 896 patients were analyzed. Relative to baseline, F0 significantly worsened in the early period (from 194.9 6 34.9 Hz to 188.0 6 34.0 Hz, P 5 0.001). This was equivalent to a quarter-tone loss (P 5 0.004). Shimmer (from 3.15 6 1.59% to 3.19 6 1.70%, P 5 0.040) and MPT (from 17.9 secs to 16.7 secs, P 5 0.046) also worsened in the early period, whereas jitter and NHR remained unchanged in the early and late periods. Males suffered greater deterioration in F0 (from 120.6 6 18.8 Hz to 111.0 6 18.5 Hz, P 5 0.048) and in NHR (from 0.12 6 0.02 to 0.16 6 0.03, P 5 0.019) than females in the early period. Four of the five acoustic parameters (F0, jitter, shimmer, and NHR) significantly worsened after total thyroidectomy (TT) and not after lesser resection. Conclusion: F0, shimmer, and MPT significantly worsened in the early and not in the late postoperative period. F0 impairment was perceptually significant. Males and those undergoing TT suffered greater voice impairment than their counterparts during the early period. Laryngoscope, 126:528–537, 2016
INTRODUCTION Thyroidectomy is one of the most common surgical procedures undertaken worldwide.1 Although it is associated with a low morbidity when done by high-volume ( 100 cases per year) surgeons, debilitating postoperative voice changes occur not uncommonly without apparent recurrent laryngeal nerve (RLN) injury.2,3 Presumed causes include injury to external branch of superior laryngeal nerve (EBSLN), orotracheal intubation, surgical adhesions, strap muscles denervation, and pain or psychological distress. A recent survey of > 200 postthyroidectomy patients found that over a third of them reported persistent vocal changes and felt these
Additional Supporting Information may be found in the online version of this article. From the Department of Surgery (B.H.H.L.); Department of Family Medicine and Primary Care, 3/F Ap Lei Chau Clinic, (C.K.H.W.); and Division of Speech and Hearing Sciences (E.P.M.M.), the University of Hong Kong, Hong Kong SAR, the People’s Republic of China. Editor’s Note: This Manuscript was accepted for publication on May 19, 2015. The authors have no funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Dr. Brian H.H. Lang, Department of Surgery, Queen Mary Hospital, 102 Pokfulam Road, Hong Kong SAR, People’s Republic of China. E-mail: [email protected] DOI: 10.1002/lary.25452
Laryngoscope 126: February 2016
528
changes more concerning than traditional complications such as hypocalcemia or unsightly cervical scar.4 Common voice changes including upper pitch loss, difficulty speaking aloud, and voice huskiness could be objectively quantified by a computerized acoustic voice program such as the multidimensional voice program (MDVP).2,3 Five objective parameters have generally been used to measure voice quality: 1) average fundamental frequency (F0), 2) jitter, 3) shimmer, 4) noise-to-harmonic ratio (NHR), and 5)maximum phonation time (MPT). Fundamental frequency is a measurement of pitch, whereas jitter and shimmer measure pitch and amplitude variability, respectively.3,5 Noise-to-harmonic ratio measures the amount of noise present, whereas MPT measures voice capacity (i.e., efficiency of voice production).3,5 Although numerous studies reported results of acoustic voice analysis in postthyroidectomy patients, findings were inconsistent. Furthermore, due to the difficulty with patient recruitment and compliance, few studies had sufficient power to quantify longer-term (more than 3 months) acoustic changes. Therefore, it remains unclear whether acoustic voice impairments persist. Our study aimed to summarize the findings from the literature and evaluate the impact of open cervical thyroidectomy on acoustic voice quality parameters in the early (< 3 months) and late periods ( 3 months) by pooling data from the current literature.
Lang et al.: Voice Impairment After Thyroidectomy Appears Temporary
MATERIALS AND METHODS This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses statement.6
Search Strategy Studies evaluating voice quality or voice-related outcomes after open cervical thyroidectomy were retrieved from the Scopus, Medline (PubMed), and Cochrane Library electronic databases on December 2, 2014. We used the following free text search terms in “All fields”: #1: #2: #3: #4:
‘voice’ ‘voice quality’ ‘thyroidectomy’ #1 OR #2 AND #3
There was no language restriction or methodological filters. The bibliography of two recent reviews on voice after thyroidectomy was searched for additional relevant references.2,7
Study Selection All titles identified by the search strategy were independently screened by three authors (B.H.H.L., CK.H.W., E.P.M.M.). Search results were compared, and disagreements were resolved by consensus. Abstracts of potentially relevant titles were then reviewed for eligibility, and full-length articles were selected for closer examination. Any studies that quantitatively assessed voice quality by acoustic voice analysis under a controlled environment were considered. Eligible studies had to report data on at least one of the five acoustic parameters (F0, jitter, shimmer, NHR, and MPT) generated from MDVP before and after an open cervical thyroidectomy. Controlled environment meant that all audio recordings were done in a quiet room, with the patient sitting 15- to 20-cm away from a high-quality dynamic microphone, and then digitized and analyzed using MDVP (Computerized Speech Lab Model, Kay Elemetrics, Lincoln Park, NJ, USA). For MPT, the longest of three attempts was generally taken. Studies that did not specify the exact time of the postoperative testing or did not separate the results of thyroidectomy from other nonthyroid operations (such as parathyroidectomy) or noncervical approaches were excluded. Case reports, editorials, expert opinions, reviews without original data, and studies on pediatric population were excluded. Multiple reports of the same dataset were assessed, and only the most representative or updated report of a study was included.
Data Extraction All data were extracted onto a standardized form. The primary data extracted from each article included: type or design of study, type of surgical services (low- or high-volume center), first authorship, country of origin, year of publication, demographics, extent of surgery (total thyroidectomy [TT] or less than TT [LTT]), thyroid pathology, and type and timing of postoperative acoustic voice assessments. A high-volume center was defined as one that performed over 100 cases per year.8 Because there was no standardization on the timing of acoustic assessments, they were categorized into early and late periods. Early period was defined as within 3 months of thyroidectomy, whereas late was defined as 3 months or more after thyroidectomy. If two or more assessments were performed within the early or late period (e.g., at 4 and 10 weeks), the average of these assessments was taken. Total thyroidectomy comprised near-TT and TT, whereas LTT comprised lobectomy and subtotal thyroidectomy. Fundamental frequency was expressed in Hz,
Laryngoscope 126: February 2016
Fig. 1. A flow diagram for study selection.
whereas jitter and shimmer were expressed in percentage. Noise-to-harmonic ratio was a ratio and thus no unit was given. Maximum phonation time was expressed in seconds.
Statistical Analysis Paired t test was used for comparison of five acoustic parameters between preoperative and postoperative assessments. To make it more clinically relevant, if significant, comparison was repeated with both preoperative and postoperative F0 data being converted into semitones by logarithmic scale similar to a piano scale. The conversion equation used was log10 (the highest frequency in Hz/the lowest frequency in Hz)/log102 3 12). All the individual outcomes were integrated with the meta-analysis software Review Manager Software 5.0 (Cochrane Collaborative, Oxford, England). Standardized mean differences (SMD) were calculated for significant findings. Results were aggregated and analyzed using both fixed- and random-effect models. Publication bias was estimated by Begg’s rank correlation test and Egger’s regression test.9,10 These meta-analyses were conducted using IBM SPSS 20.0 for Windows (IBM Corp., Armonk, NY) and Comprehensive Meta-Analysis Version 2.2.064 (Biostat, Inc., Englewood, NJ, USA).
RESULTS Of the 950 titles initially identified from the database search (Fig. 1), 36 full-length articles were assessed for inclusion. Twenty were excluded, and 16 studies3,5,11–24 were determined to be eligible for inclusion. Appendix 1 lists these 20 articles25–44 and the reason for their exclusion. No additional study was found from our search of the two bibliographies in previous reviews.2,7
Patient Selection Consecutive patients undergoing thyroidectomy were recruited in all the studies examined. All studies
Lang et al.: Voice Impairment After Thyroidectomy Appears Temporary
529
TABLE I. Patient Baseline Characteristics and Types of Voice Assessment in the 16 Included Studies. Extent of Surgery
Study Design
No. of Patients*
Sex (M:F)
Mean Age 6 SD (range)
PS
54
9:45
43 (13–67)
46
Debruyne11 (1997)
PS
47
0:47
Not available
Aluffi12 (2001)
RS
42
6:36
52 (33–76)
Stojadinovic3 (2002)
PS
50
10:40
Sinagra13 (2004) de Pedro Netto14 (2006)
PS PS
46 88
3:43 11:77
First Author (year) 5
Hong (1997)
TT
Postoperative Follow-up Visits
Underlying Thyroid Pathology
Subjective/ Perceptual Assessment
8
1m, 3m, 6m
Benign 1 malignant
47
0
D4,2w,3m
Benign 1 malignant
Symptoms, questionnaire None
9
33
12–18m
Benign
Symptoms
53 (22–77)
21
29
2w,3m
Benign 1 malignant
Symptoms, GRBAS
43 (20–70) 46 (17–79)
0 35
46 53
2m, 4m, 6m 2w
Benign 1 malignant Benign 1 malignant
– GRBAS
LTT
Soylu15 (2007)
PS
48
6:42
47 (19–72)
8
40
D2, 3m
Benign 1 malignant
Symptoms
Akyildiz16 (2008) Lombardi17 (2009)
PS PS
36 110
9:27 13:97
50 (17–74) 46.5 6 13.2
27 0
9 110
1w 3m, 12m
Benign 1 malignant Benign 1 malignant
– Symptoms, questionnaire
Van Lierde18 (2010)
RS
44
12:32
49 (17–72)
7
37
1w, 6w, 3m
Benign 1 malignant
GRBAS, questionnaire
Li19 (2012)
RS
32
8:24
48 6 9
32
0
1w, 1m, 3m
Not available
Symptoms
Lombardi20 (2012)
PS
32
7:25
44.5 6 10.7
0
32
3m
Benign 1 malignant
Tae21 (2012)
PS
61
0:61
54.4 6 10.6
10
51
1w, 1m, 3m, 6m
Benign 1 malignant
Symptoms, questionnaire Symptoms, questionnaire
Lee22 (2012)
PS
46
6:40
38.5 6 68
11
35
1w, 3m
Malignant
GRBAS
Ryu23 (2013)
PS
74
33:41
44.8 6 8.5† 50.8 6 11.3†
33
41
1m, 6m, 12m
Malignant
Symptoms, GRBAS
Jung24 (2013)
PS
86
12:30
48.0 6 8.7‡
22
20
2w, 3m
Malignant
Questionnaire
6:38
51.8 6 8.7‡
22
22
151:745
47.1 6 24.0
330
566
–
–
Overall
–
896
–
D 5 day; GRBAS scale 5 grade, rough, breathy, asthenic, strained; LTT 5 less than total thyroidectomy; m 5 month; PS 5 prospective study; RS 5 retrospective study; TT 5 total thyroidectomy; w 5 week. *After excluding those with incomplete follow-up or postoperative vocal cord paresis. † Data for LTT and TT. ‡ Data for subplatysmal and subfascial groups.
considered previous laryngeal diseases, vocal cord paralysis, and incomplete follow-up as exclusion criteria. Some studies also imposed an age exclusion criterion.17,20–24
Baseline Characteristics and Subjective Voice Assessments Table I shows a summary of the baseline characteristics and type of subjective voice assessments in the 16 eligible studies (3 retrospective and 13 prospective). All 16 studies were conducted in high-volume centers. In the two robotic-thyroidectomy series, only data from the open group were used.21,23 A total of 896 patients were analyzed. The cohort comprised predominantly females (83.1%) with a mean (6 standard deviation) age of 47.1 6 24.0 years old. The LTT:TT ratio was 1 to 1.7. Three (18.8%) studies exclusively analyzed TT,13,17,20 whereas two (12.5%) studies only analyzed LTT.12,19 Twelve (75.0%) studies performed acoustic voice analysis two or more times in the postoperative period.3,5,11,13,15,17–19,21–24 Eleven (68.8%) studies analyzed benign and malignant cases together,3,5,11,13–18,20,21 whereas three (18.9%) Laryngoscope 126: February 2016
530
included malignant cases only 22–24 and one (6.3%) had benign case only.12 One (6.3%) study did not mention underlying thyroid pathology.19 Subjective voice changes were commonly reported after thyroidectomy, and the incidence ranged between 29.7% to 87.0%, depending on definition and timing of the assessment.13,14,17 “Voice fatigue” was the most common complaint, and this was followed by “loss of high pitch” and “difficulty of loud voice.”3,5,12,13 One study found there was a significant relationship between the extent of surgery and voice change15 whereas another study did not.16 Nevertheless, these voice complaints tended to improve over time.15,19 Changes in voice quality by perceptual rating using the GRBAS (grade of hoarseness, roughness, breathiness, asthenic and strain) scale was reported in several studies.3,14,18,22,23 Significant increases in GRBAS score were found in the first 2 weeks but less so after 3 to 6 months.3,14,18 Postoperative videolaryngostroboscopy (VLS) was done in all of the studies to exclude any postoperative vocal cord abnormalities and RLN injury, but only three studies performed routine postoperative cricothyroid electromyography (CT EMG) in
Lang et al.: Voice Impairment After Thyroidectomy Appears Temporary
Laryngoscope 126: February 2016
Lang et al.: Voice Impairment After Thyroidectomy Appears Temporary
531
183.1 6 35.8
TT (n 5 41)
Lee22 (2012) Ryu23 (2013)
196 6 35
188.1 6 34.8*
180.00 6 31.30
180.4 6 30.5 192.7 6 35.6
169.80 6 43.40
169.7 6 47.9
174.10 6 38.10
189.70 6 45.00 187.70 6 50.50
201.40 6 34.90
231.80 6 44.75 –
117.43 6 16.74
104.75 6 17.62 189.10 6 27.30
–
118.06 6 28.20 215.80 6 24.70
221.64 6 26.30
197.80 6 27.10 217.88 6 71.80
110.30 6 7.50
168.07 6 45.50 176.00 6 19.00
170.82 6 44.40
205.4 6 19.7 –
1.08 6 0.76 0.90 6 0.62
190.5 6 38.1†
1.01 6 0.79
1.49 6 0.87
1.37 6 0.40 1.39 6 0.97
1.31 6 0.89
1.65 6 1.14 0.63 6 0.22
1.22 6 0.85
1.0 6 0.61 1.0 6 0.59
0.5 6 0.3
0.70 6 0.25 1.45 6 1.03
0.397 6 0.15
0.5 6 0.3 0.605 6 0.39
0.4 6 0.2
0.68 6 0.39 –
0.55 6 0.23
1.14 6 0.89 0.64
0.44 6 0.10
0.99 6 0.72†
0.98 6 1.04
0.99 6 1.05
1.98 6 1.21
1.38 6 0.45 1.42 6 0.59
1.10 6 0.70
1.27 6 1.21 –
1.29 6 0.99
1.25 6 0.79 1.19 6 0.67
–
0.91 6 0.81 0.99 6 0.76
0.70 6 0.16
0.50 6 0.30 0.77 6 0.56
0.60 6 0.30
0.71 6 0.40 –
0.56 6 0.46
1.13 6 0.89 –
0.41 6 0.06
