Comprehensive Outcome Researches of Intralesional Steroid Injection on Benign Vocal Fold Lesions *,†,‡Chi-Te Wang, †,§Mei-Shu Lai, and ‡Tzu-Yu Hsiao, *yzxTaipei, Taiwan Summary: Objective. This study investigated multidimensional treatment outcomes, including prognostic factors and side effects of vocal fold steroid injection (VFSI). Methods. We recruited 126 consecutive patients, including patients with 49 nodules, 47 polyps, and 30 mucus retention cysts. All the patients received VFSI under local anesthesia in the office settings. Treatment outcomes were evaluated 1 and 2 months after the procedure, including endoscopic evaluation, perceptual voice quality (GRB scores), acoustic analysis, and 10-item Voice Handicap Index (VHI-10). Results. More than 80% of the patients reported subjective improvements after VFSI. Objective measurements revealed significant improvements from baseline in most of the outcome parameters (P < 0.05). Higher occupational vocal demands and fibrotic vocal nodules were significantly associated with poorer clinical responses as measured by the VHI-10 and GRB scores, respectively. For vocal polyps, dysphonia for more than 12 months were significantly associated with higher postoperative VHI-10 scores, whereas patients with laryngopharyngeal reflux (LPR) showed significantly poor postoperative voice quality as measured by GRB scores. Side effects after VFSI included hematoma (27%), triamcinolone deposits (4%), and vocal atrophy (1%), which resolved spontaneously within 1–2 months. Presentation with vocal fold ectasias/varicosities and higher vocal demands were significantly correlated with postoperative vocal hematoma. Conclusions. This study demonstrated significant improvements after VFSI in vocal nodules, polyps, and cysts. Occupational vocal demand and subtypes of vocal nodules are closely related to the treatment outcomes after VFSI, whereas symptom duration and LPR were significant prognostic factors for VFSI treatment outcomes in vocal polyps. Side effects after receiving VFSI were mostly self-limited without sequel, whereas the incidence rates might be varied by the injection approach and the timing for postoperative follow-up. Key Words: Nodules–Polyp–Cyst–Triamcinolone–Dexamethasone–Occupation–Vocal demand–Laryngopharyngeal reflux–Hematoma–Atrophy–Ectasias–Varicosity. INTRODUCTION Benign vocal fold lesions, such as vocal nodules, polyp, and cyst, refer to unilateral or bilateral lesions of the midmembranous portion of the vocal fold that lie within the superficial lamina propria (ie, the Reinke’s space).1 Common clinical presentations of these lesions include dysphonia, voice fatigue, dryness or tightness of voice, narrowed vocal range, and deteriorated voice quality (husky or breathy character). Because most of these patients have an occupational dependence on their voices, noisy working environments, or inappropriate phonation habits; in most circumstances, behavioral modification remains the first-line management, followed by phonomicrosurgery.2,3 However, when compliance with conservative management is poor, when the risk from general anesthesia is high, when the patient is unwilling to have an operation, or in case of recurrence after surgery, very few options remain for patients. Intralesional steroid injection has been advocated in the management of inflammatory laryngeal diseases such as sarcoidAccepted for publication November 5, 2014. From the *Department of Otolaryngology, Far Eastern Memorial Hospital, Taipei, Taiwan; yGraduate Institute of Epidemiology and Preventive Medicine, College of Public Health, National Taiwan University, Taipei, Taiwan; zDepartment of Otolaryngology, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan; and the xCenter of Comparative Effectiveness Research, National Center of Excellence for Clinical Trial and Research, National Taiwan University Hospital, Taipei, Taiwan. Address correspondence and reprint requests to Tzu-Yu Hsiao, Department of Otolaryngology, National Taiwan University Hospital and National Taiwan University College of Medicine, 7 Chung-Shan South Road, Taipei 100, Taiwan. E-mail: [email protected] Journal of Voice, Vol. -, No. -, pp. 1-10 0892-1997/$36.00 Ó 2015 The Voice Foundation http://dx.doi.org/10.1016/j.jvoice.2014.11.002
osis, systemic lupus erythematosus, and Wegener’s granulomatosis.4–6 A recent study demonstrated that officebased laryngeal procedures are well tolerated in the outpatient department with limited impact on the hemodynamic profiles.7 Compared with surgeries in the operating room under general anesthesia, office-based procedures not only avoid potential risks and discomforts associated with general anesthesia but also offer the advantages of real-time monitoring of vocal function and considerable savings of medical expenses.8–10 A number of studies have advocated the potential role of vocal fold steroid injection (VFSI) for the treatment of benign vocal fold lesions.11–13 However, few studies have investigated the treatment outcomes between different disease entities. Moreover, the prognostic factors of VFSI (eg, laryngopharyngeal reflux [LPR] and occupational vocal demand) have rarely been reported in the literature.14 Although side effects after VFSI have been noted in the literature, the incidence rates varied widely.14 Therefore, the objectives of this study are to conduct comprehensive researches that reports multidimensional treatment outcomes and prognostic factors between different categories of vocal lesions and to investigate the incidence rates and risk factors of side effects after VFSI. METHODS Study subjects This study prospectively recruited consecutive patients who had received VFSI from January 2012 to December 2013 at a tertiary teaching hospital. The inclusion criteria include vocal nodules, polyp, and mucus retention cyst. Clinical diagnoses
2 were based on the appearance of the vocal fold edge and the vibratory property of the mucosa revealed by videolaryngostroboscopy (VLS), based on the published nomenclature and diagnostic paradigm.1 Vocal nodules were classified as (1) soft (ie, mid-membranous thickening without fibrotic change of the epithelium or limitation of the mucosal wave) or (2) hard (ie, hyperkeratosis of the epithelium, fibrotic appearance with some limitation of mucosal wave propagation). Vocal polyps were classified as hemorrhagic (vascular), fusiform (wide-based), pedunculated, and fibrous.15 The presence of ectasias (defined as spheroidal, tangible, subepithelial hemangioma) and varices (defined as enlarged, aberrant, and tortuous vessels) were recorded as well.16,17 Patients with keratin cysts were excluded. Patients who had received previous phonomicrosurgery were also excluded from the study. The study protocol has been approved by the Research Ethics Review Committee. All of the clinical and demographical factors were collected at the initial clinical visit, including age, gender, duration of dysphonic symptoms, smoking, occupational vocal demand, as well as the 10-item Voice Handicap Index (VHI-10),18 and Reflux Symptom Index (RSI).19 Subjects with RSI scores of more than 13 points were defined as positive for LPR. We reviewed the occupations of each patient and classified them into (1) professional voice users (eg, singers, actor/actress, radio broadcast, and singing student); (2) high voice users (eg, teachers, clergy, lecturer, sales, tour guide, aerobic exercise coach, and patients who work in noisy environment; or (3) routine voice users, as suggested in the literature.20 Professional and high voice users were labeled as having ‘‘high’’ occupational vocal demand, and the others were labeled as having ‘‘routine’’ occupational vocal demand. Vocal Fold Steroid Injection The VFSI was performed under local anesthesia of the pharynx and larynx,21 and the patient was instructed to pull out and hold the tongue, while the surgeon uses the nondominant hand to operate the rigid laryngoscope. Under visual guidance, 0.1 mL, 1:1 mixture of triamcinolone acetonide (10 mg/mL) and dexamethasone sodium phosphate (5 mg/mL) was injected by curved injection needle (Model 16–50050, Medtronic Xomed Jacksonville, Florida, USA) into the Reinke’s space of the vocal lesions.22 For patients who were intolerant of the above transoral procedure,23 transnasal injection via the working channel of a flexible nasopharyngoscope could be applied with an endoscopic injection apparatus (Model NM-101C-0427 Olympus, Aomori, Japan.), which includes a reusable metallic external sheath (MAJ-655) and a disposable flexible needle tract with a 27G rigid tip (MAJ-656).24 All the patients received postoperative voice rest for 3 days. Outcome evaluation Treatment outcomes were evaluated before, 1 month, and 2 months after the procedure. Perceptual evaluation of voice quality was performed using the GRB scales and scored as 0: normal, 1: slightly deviated, 2: moderately deviated, and 3: extremely deviated.25 Scores from the three subscales (grade,
Journal of Voice, Vol. -, No. -, 2015
roughness, and breathiness) were summed up for statistical inferences. Acoustic analysis was conducted by recording a 3-second sample of the sustained vowel sound /a:/ at a comfortable level of loudness, with a microphone-to-mouth distance of approximately 15–20 cm, using a high-quality microphone (Model: SM58, SHURE Niles, Illinois, USA.) with a digital amplifier (Model: X2u, SHURE). Analyzed parameters include jitter (ie, cycle-to-cycle frequency perturbation), shimmer (ie, cycle-to-cycle amplitude perturbation), and noise-to-harmonic ratio (NHR; MDVP, Model 4500; Kay Elematrics, Corp., Lincoln Park, NJ, USA). Under similar settings, we measured maximal phonation time (MPT) by instructing the patient to produce the /a:/ sound for as long as possible after deep inspiration and at a spontaneous, comfortable pitch and loudness level for three consequent trials. The patient-reported outcomes included a Visual Analog Scale (VAS) of voice quality (ranging from zero [worst] to 10 [best]),26 subjectively perceived resolution of clinical symptoms (categorized as complete remission, more than 50% improvement, 50% improvement or less, no effect, or worsening), and Mandarin-Chinese version of VHI10.27 The vibratory capacity of the vocal fold was evaluated using VLS, which was conducted by instructing the patient to phonate a sustained /ee/ sound with habitual pitch and intensity, using a 70 rigid endoscope and a three-chip CCD camera (Model 2706CA and 20222120; KARL STORZ, Germany) or a digital laryngoscope with the corresponding video processor (VNL1590 STi and EPK-i; PENTAX). Each session of the VLS test was digitally recorded onto a portable hard disk using a computerized video processor (NHX-B10; Grass Valey, Inc.). Reduction of the lesion size was subjectively evaluated by the primary surgeon (C-.T.W.) during the clinical follow-ups. Intrarater reliability was tested by randomly re-examining 50% of the original video recordings, in a blinded fashion. Inter-rater reliability was examined in the same method, by an experienced speech therapist (Feng-Chuan Lin, MSc). We also objectively measured the normalized vibration amplitude of the membranous vocal fold by: (1) capturing video clips of maximally opened and closed phases during phonation with correction to the vocal fold length (ie, normalized glottal gap area [NGGA], Image J software, version 1.44; National Institute of Mental Health, Bethesda, MD) and (2) subtract the open phase from the closed phase NGGA (Figure 1).28,29 Side effects after VFSI were identified by reviewing postoperative video recordings and included (1) vocal fold hematoma surrounding the injection sites; (2) deposits of injected material, which was defined as the presence of whitish plaque over Reinke’s space, with minimal-to-moderate influence of the vibration of mucosal wave; and (3) vocal fold atrophy, which was defined as ‘‘bowing’’ of the vocal fold(s) with the presence of glottic gap and atrophic thyroarytenoid muscle. Statistics Treatment outcomes measured at 1 and 2 months postoperatively were compared with baseline measurements using paired t tests for the different diagnoses. Intra- and inter-rater reliabilities were examined by Cronbach’s alpha coefficient. To
Chi-Te Wang, et al
Treatment Outcomes of VFSI
3
FIGURE 1. The normalized vibration amplitude is calculated by subtracting the maximally opened glottic area (A) from the maximally closed glottic area (B) during the phonation cycle, and corrected by the length of the vocal fold, measured from the anterior commissure to the vocal process (C).
investigate the prognostic factors of the treatment outcomes, we defined ‘‘responders’’ to VFSI as: (1) postoperative VHI-10 scores 10 points, which denoted the upper limit of scores in the normal population27 and (2) GRB scores 1, which represents an objective rating of normal voice quality.30 Analyzed prognostic factors include gender, duration of clinical symptoms (>12 months vs 12 months), occupational vocal demand (high vs routine), cigarette smoking, LPR, and subtypes of vocal lesions.20,31,32 Univariate analyses for prognostic factors for treatment outcomes and risk factors for postoperative side effects were conducted by Chi-square tests. Multivariate analyses were subsequently conducted via multiple logistic regression model. A P value less than 0.05 indicated statistical significance. All statistical analyses were conducted by SAS software, version 9.3 (SAS Institute, Inc., Cary, NC). RESULTS This study recruited 126 patients with benign vocal lesions, including 19 men and 107 women. The mean age was 40 years (standard deviation: 10 years; range: 18–69 years). Vocal nodules were diagnosed in 49 patients, whereas polyps and cysts were diagnosed in the other 47 and 30 patients, respectively. Subtypes of vocal nodules include soft nodules in 32 patients and hard, fibrotic nodules in the other 17 patients. Subtypes of vocal polyps included hemorrhagic polyp in three, fusiform (wide-based) in 35, pedunculated in three, and fibrotic in six patients. High occupational vocal demands were reported in 75 patients, consisting mainly of teachers (26%), sales people (7%), customer service agents (25%), performing artists (6%), and others. Mean clinical duration of dysphonic symptoms was 15 months (standard deviation: 25 months). A total of 13 and 28 patients attended pre- and postoperative voice therapy during the treatment course, respectively. More than 80% of the patients with vocal nodules reported subjective improvements in their clinical symptoms 1 and 2 months after receiving VFSI (Table 1, left column). Similarly, subjective improvements in dysphonia were reported in up to 90% of the patients with vocal polyps and cysts (Table 1, middle and right column). The VLS examinations also revealed
decreased lesion sizes for more than 80% of the vocal lesions (Figure 2), as well as significantly improved vibration amplitudes (P < 0.01, paired t test, Table 1). Intra- and inter-rater reliabilities of VLS evaluations were 0.82 and 0.74, respectively (Cronbach’s alpha coefficient), compatible with a recent study, which indicated that VLS rating of the static lesion size is a reliable method.33 Treatment outcomes demonstrated significant improvements in VHI-10, VAS, and GRB scores after VFSI for all three diagnoses (P < 0.01, paired t test, Figure 3). Although significant improvements in MPT 1 month postoperatively were revealed in all three disease categories (P < 0.05, paired t test), the 2month results did not reach statistical significance (Table 1). Other parameters, including jitter, shimmer, and NHR, showed more universal improvements in vocal cysts, followed by partial improvements in vocal polyps and minimal improvements for vocal nodules. Because the etiology and contributing factors for developing various vocal lesions are different, we subsequently compared the prognostic factors for better clinical responses after VFSI between each of the three disease entities. For vocal nodules, patients with higher occupational vocal demands were associated with a poorer clinical response as measured by the VHI10 (P ¼ 0.03, Chi-square test, Table 2). Multivariate analyses also demonstrated significant correlation between vocal demand and treatment outcomes (adjusted odds ratio [aOR]: 4.32, 95% confidence interval [CI]: 1.06–17.6, P ¼ 0.04, multiple logistic regression). Additionally, hard and fibrotic vocal nodules, compared with soft nodules, had a lower chance of recovering normal perceptual voice quality as measured by GRB scores (P ¼ 0.01, Chi-square test). Multivariate analyses also validated this significant correlation (aOR: 9.27, 95% CI: 1.4–61.3, P ¼ 0.02). Univariate analyses of the prognostic factors for vocal polyps after VFSI showed significantly higher VHI-10 scores among the subsets of patients with dysphonia for more than 12 months (P ¼ 0.03, Chi-square test, Table 3). Subsequent multivariate analyses also demonstrated this association with borderline significance (aOR: 4.20, 95% CI: 0.91–19.3,
4
TABLE 1. Treatment Outcomes of Vocal Fold Steroid Injection in the 126 Patients of Vocal Nodules, Polyp, and Cyst Nodules Outcome Parameters Subjective evaluation, n (%) Complete remission Much improvement (>50%) Some improvement (50%) No effect Endoscopic evaluation, n (%) Complete remission Much improvement Some improvement No effect Normalized vibration amplitude Acoustic analysis MPT Jitter Shimmer NHR
Polyp
Cyst
Pre (N ¼ 49) 1 mo (N ¼ 43) 2 mo (N ¼ 33) Pre (N ¼ 47) 1 mo (N ¼ 42) 2 mo (N ¼ 29) Pre (N ¼ 30) 1 mo (N ¼ 28) 2 mo (N ¼ 21) 0 (0) 23 (54) 15 (34) 5 (12)
2 (6) 17 (52) 10 (30) 4 (12)
6 (14) 24 (57) 7 (17) 5 (12)
4 (14) 16 (55) 8 (28) 1 (3)
6 (21) 14 (50) 6 (21) 2 (7)
6 (29) 9 (43) 6 (29) 0 (0)
10.5 ± 4.7
6 (14) 28 (65) 7 (16) 2 (5) 13.4 ± 5.2*
4 (12) 15 (46) 8 (24) 6 (18) 14.9 ± 6.5*
10.6 ± 8.1
8 (19) 23 (55) 10 (24) 1 (2) 15.3 ± 10.7*
4 (14) 14 (48) 9 (31) 2 (7) 12.0 ± 7.9*
9.2 ± 5.1
6 (21) 16 (57) 3 (11) 3 (11) 11.5 ± 6.1*
6 (29) 10 (76) 4 (14) 1 (14) 11.9 ± 7.2
10.2 ± 4.1 1.6 ± 1.0 3.2 ± 1.5 0.2 ± 0.6
12.6 ± 5.1* 1.4 ± 0.7 2.9 ± 1.0 0.1 ± 0.02
10.8 ± 4.2 1.3 ± 0.9* 3.6 ± 3.4 0.1 ± 0.1
9.9 ± 4.7 1.9 ± 1.4 4.5 ± 2.8 0.1 ± 0.1
11.5 ± 4.9* 1.4 ± 0.9y 3.1 ± 1.2* 0.1 ± 0.1
11.4 ± 4.5 1.4 ± 1.3 3.8 ± 2.5 0.2 ± 0.2
12.0 ± 4.7 1.8 ± 1.2 4.6 ± 3.1 0.1 ± 0.1
13.8 ± 4.3y 1.3 ± 0.9y 3.5 ± 2.4y 0.1 ± 0.1
13.4 ± 4.6 1.1 ± 0.9y 3.3 ± 2.1y 0.1 ± 0.04 Journal of Voice, Vol. -, No. -, 2015
Abbreviations: MPT, maximal phonation time; NHR, noise-to-harmonic ratio. Notes: Data are expressed as mean ± standard deviation. * P < 0.01, paired t test, compared with baseline. y P < 0.05, paired t test, compared with baseline.
Chi-Te Wang, et al
Treatment Outcomes of VFSI
5
FIGURE 2. (A). A 47-year-old female patient who suffered from a mucus retention cyst on her left vocal fold. After transnasal VFSI (B), the lesion resolved completely within 1 month (C). P ¼ 0.07). Alternatively, we observed that patients who presented with LPR showed significantly poor postoperative voice quality based on univariate and multivariate analyses
(P ¼ 0.01, Chi-square test; aOR: 14.5, multiple logistic regression). Subsequent univariate and multivariate evaluations of vocal fold cysts did not reveal a significant correlation
FIGURE 3. Treatment outcomes of VFSI in patients with vocal nodules, polyp, and mucus retention cyst, measured by VHI-10, VAS, and GRB scores.
6
TABLE 2. Univariate and Multivariate Analyses of Prognostic Factors of VFSI Among Patients With Vocal Nodules Outcomes Parameters
Univariate
VHI-10 Gender Duration of symptoms (mo) Vocal demand Smoking Reflux Nodular type
Male/female >12/12 High/routine Active/nil Presence/absence Hard/soft
GRB scores Gender Duration of symptoms (mo) Vocal demand Smoking Reflux Nodular type
Male/female >12/12 High/routine Active/nil Presence/absence Hard/soft
Multivariate
Responder (10), n (%)
Nonresponder (>10), n (%)
P Value*
aOR (95% CI)
P Value
2 (10)/19 (90) 9 (43)/12 (57) 7 (33)/14 (67) 2 (10)/19 (90) 6 (29)/15 (71) 6 (29)/15 (71)
1 (4)/27 (96) 8 (29)/20 (71) 18 (64)/10 (36) 2 (7)/26 (93) 13 (46)/15 (54) 11 (39)/17 (61)
0.39 0.30 0.03 0.76 0.20 0.44
14.3 (0.50–411) 0.25 (0.05–1.16) 4.32 (1.06–17.6) 2.27 (0.18–28.1) 1.76 (0.45–6.92) 1.99 (0.47–8.43)
0.12 0.08 0.04 0.52 0.42 0.35
Responder (1), n (%)
Nonresponder (>1), n (%)
P Value*
1 (6)/16 (94) 6 (35)/11 (65) 9 (53)/8 (47) 3 (18)/14 (82) 5 (29)/12 (71) 2 (12)/15 (88)
2 (6)/30 (94) 11 (34)/21 (66) 16 (50)/16 (50) 1 (3)/31 (97) 14 (44)/18 (56) 15 (47)/17 (53)
0.96 0.95 0.84 0.11 0.33 0.01
aOR (95% CI) 0.64 (0.02–17.6) 1.30 (0.26–6.45) 0.45 (0.11–1.87) 0.09 (0.01–1.56) 3.37 (0.72–15.7) 9.27 (1.40–61.3)
P Value 0.79 0.75 0.27 0.10 0.12 0.02
Notes: aOR (95% CI): Adjusted odds ratio (95% confidence intervals) by multiple logistic regression model Bolded values represents statistical significances. * Chi-square test.
TABLE 3. Univariate and Multivariate Analyses of Prognostic Factors of VFSI Among Patients With Vocal Polyp Outcomes Parameters
Univariate
VHI-10 Male/female >12/12 High/routine Active/nil Presence/absence Fibrous/others
GRB scores Gender Duration of symptoms (mo) Vocal demand Smoking Reflux Polyp type
Male/female >12/12 High/routine Active/nil Presence/absence Fibrous/others
Responder (10), n (%)
Nonresponder (>10), n (%)
P Value*
3 (20)/12 (80) 3 (20)/12 (80) 6 (40)/9 (60) 3 (20)/12 (80) 6 (40)/9 (60) 1 (7)/14 (93)
3 (9)/29 (91) 17 (53)/15 (47) 15 (47)/17 (53) 3 (9)/29 (91) 14 (44)/18 (56) 5 (16)/27 (84)
0.31 0.03 0.66 0.31 0.81 0.39
Responder (1), n (%)
Nonresponder (>1), n (%)
P Value*
3 (27)/8 (73) 3 (27)/8 (73) 5 (45)/6 (55) 2 (18)/9 (82) 1 (9)/10 (91) 0 (0)/11 (100)
3 (8)/33 (92) 17 (47)/19 (53) 16 (44)/20 (56) 4 (11)/32 (89) 19 (53)/17 (47) 6 (17)/30 (83)
0.13 0.24 0.95 0.54 0.01 0.31
Notes: aOR (95% CI): Adjusted odds ratio (95% confidence intervals) by multiple logistic regression model Bolded values represents statistical significances. * Chi-square test. y Unable to estimate owing to the absence of subjects.
aOR (95% CI) 1.52 (0.19–11.9) 4.20 (0.91–19.3) 1.55 (0.38–6.26) 0.46 (0.06–3.64) 0.88 (0.21–3.70) 3.70 (0.29–47.9) aOR (95% CI) 3.00 (0.25–35.5) 1.33 (0.22–8.21) 1.79 (0.33–9.68) 0.31 (0.01–6.83) 14.5 (1.33–157) >1 (unavailabley)
P Value 0.69 0.07 0.54 0.46 0.86 0.32 P Value 0.38 0.76 0.50 0.46 0.03 —
Journal of Voice, Vol. -, No. -, 2015
Gender Duration of symptoms (mo) Vocal demand Smoking Reflux Polyp type
Multivariate
Chi-Te Wang, et al
Treatment Outcomes of VFSI
between clinical, endoscopic, or interventional factors and the treatment outcomes. After receiving VFSI, the most common side effects were hematoma (34 of 126 patients, 27%), followed by deposits of triamcinolone (5 cases, 4%; Figure 4A and B). Both conditions generally presented with dysphonia; increased phonation effort; and, occasionally, decreased voice quality compared with preoperative status. The VLS examination tended to reveal decreased amplitude and asymmetry of mucosal waves. Clinical observation showed that all side effects resolved spontaneously within 2 months, leaving no sequel. We observed one case with a breathy voice and increased phonation effort after VFSI. The VLS showed bowing of the vocal folds with prominent glottic gap and ventricles. Two months afterward, her symptoms resolved, and follow-up endoscopy showed recovery of her baseline vocal fold bulk (Figure 4C and D). The overall incidence rates for developing postoperative side effects were 32%, with no significant differences between the three disease entities. Univariate and multivariate analyses revealed that presentation with vocal fold ectasias and varicosities were significantly correlated with postoperative vocal hematoma (P < 0.05, Table 4). Meanwhile, multivariate analysis also identified that subjects with high occupational vocal demands were more likely to develop vocal hematoma after VFSI (P < 0.05, multiple logistic regression). Other clinical factors, including gender, smoking, hypertension, and LPR, all showed nonsignificant correlation with postoperative vocal hematomas (Table 4). Subsequent analyses of the potential risk factors for triamcinolone deposits and vocal atrophy were not performed because of the limited sample. DISCUSSION Corticosteroids are one of the most common anti-inflammatory agents in modern medicine, and they can be administered via multiple routes (eg, transoral, intramuscular, intravascular, or inhalational). Specifically, direct intralesional steroid injection has the advantage of maintaining a higher regional drug concentration with minimal systemic adverse effects. Yanagihara et al34 first applied direct intralesional steroid injection in the treatment of benign vocal fold disorders in 1964. However, VFSI as an office-based procedure to treat benign vocal fold disorders had not been widely accepted until the past decade, when Tateya et al35 made two major refinements to the proce-
7 dure, namely (1) enhanced precision of needle placement by flexible fiberscope guidance and (2) triamcinolone, a depot solution with longer effective duration, as the injected regimen. Since then, a number of published studies have demonstrated that VFSI can be applied to treat many structural vocal fold disorders, such as vocal nodules, polyp, cyst, scar, granuloma, and Reinke’s edema.11–13 The clinical effectiveness of VFSI was further documented in our recent study that demonstrated significantly higher lesion reduction rates after VFSI than following vocal hygiene education.22 This study investigated the multidimensional treatment outcomes among the most common vocal fold pathologies. For vocal nodules, we demonstrated that VFSI can result in significant subjective and objective improvements (Table 1, left column), which was comparable with previously published results.11,36 Moreover, we investigated the prognostic factors of vocal nodules after VFSI, which had not previously been done. The study results demonstrated that patients with higher vocal demand were more likely to perceive higher phonation discomforts as measured by their postoperative VHI-10 scores (Table 2, upper half). Because vocal nodules are closely related to occupational overuse of the voice, without adequate modification of phonation habits, the therapeutic effects of corticosteroids on vocal folds may gradually be countered by persistent voice overloads; if that is the case, dysphonic symptoms may persist.37 This study also found that subjects with thick, fibrotic nodules were less likely to regain normal voice quality after VFSI (Table 2, lower half). The histological features of vocal nodules may help to explain these findings. In soft nodules, which present with stroma edema, dilation, and the increased permeability of microvasculature among the superficial lamina propria,38 VFSI can be more effective than it is for fibrotic or thick nodules, which feature a thickened epithelium and the proliferation of fibroblasts with abundant fibrin and organized collagen.39,40 For patients with vocal polyps, this study demonstrated significant improvements in most of the outcome parameters, comparable with results from a previous study.13 Subsequent analyses revealed that patients with dysphonic symptoms lasting for more than 1 year were more likely to have higher VHI-10 scores postoperatively (Table 3, upper half). It is likely that chronic vocal polyps may present with a higher degree of fibrin deposition, inflammation, and amyloid-like material
FIGURE 4. Side effects after VFSI, including hematoma (A), triamcinolone deposits (B, arrow) and bowing atrophic vocal folds (C, arrowhead). Atrophic vocal folds recovered completely after 2 months (D).
8
Journal of Voice, Vol. -, No. -, 2015
TABLE 4. Univariate and Multivariate Analyses of Risk Factors of post-VFSI Vocal Hematoma Analyses
Univariate
Parameters Gender Smoking Hypertension Vocal demand Reflux Ectasia/varicosity
Male/female Active/nil Presence/absence High/routine Presence/absence Presence/absence
Multivariate
Hematoma (N ¼ 34)
No Hematoma (N ¼ 92)
P Value*
aOR (95% CI)
P Value
3 (9)/31 (91) 1 (3)/33 (97) 4 (12)/30 (88) 18 (53)/16 (47) 11 (32)/23 (68) 13 (38)/21 (62)
16 (17)/76 (83) 15 (16)/77 (84) 7 (8)/85 (92) 33 (36)/59 (64) 38 (41)/54 (59) 18 (20)/74 (80)
0.23 0.07 0.46 0.08 0.36 0.03
1.60 (0.38–6.68) 0.11 (0.01–1.10) 1.40 (0.35–5.67) 2.74 (1.11–6.75) 0.54 (0.21–1.37) 4.28 (1.55–11.9)
0.52 0.06 0.64 0.03 0.20 0.01
Notes: aOR (95% CI): Adjusted odds ratio (95% confidence intervals) by multiple logistic regression model Bolded values represents statistical significances. * Chi-square tests.
accumulation in Reinke’s space,40 which may be less reversible after one course of VFSI. Meanwhile, chronic vocal polyps may also more likely to cause reactive nodules on the contralateral vocal folds, which can also contribute to the less favorable postoperative recovery. A recent study found a 6% coexistence rate of vocal polyps with sulcus vocalis, which may negatively affect surgical outcomes.41 In this study, we also identified two cases of vocal polyps with underlying sulcus; both patients presented with symptom duration longer than 12 months and a less favorable response to VFSI. These comparable findings may also help to explain the association between the chronicity of diseases and treatment outcomes. Our study results found that patients with vocal polyp comorbid with LPR tended to present with poorer voice quality after VFSI compared with the patients without LPR (Table 3, lower half). Although a previous study had already recognized the strong relationship between LPR and vocal fold disorders,42 this study provides additional clinical evidence of LPR and the treatment outcomes. Persistent inflammation surrounding the glottis caused by LPR (eg, laryngeal edema and increased phlegm) may explain the fair recovery of voice quality after VFSI.43,44 Mucus retention cysts are typically located at the subepithelial plane of the vocal folds that indicate an obstructed excretory duct with retained mucus of the affected gland.15 Histologically, mucus retention cysts consist of an outer layer of cuboidal cells with an internal layer of ciliated columnar cells.45 Except for a small number of reported cases,12 this is the first clinical series to document the treatment effectiveness of VFSI for mucus retention cysts of vocal folds. Based on our preliminary experience, the needle tip can sometimes directly puncture through the cyst wall and drain the cystic content during the injection process, which might explain the higher remission rate and rapid clinical response of mucus retention cysts after VFSI comparing with vocal nodules and polyps (Table 1). Previous studies had already reported potential adverse events after VFSI, with a highly variable range of incident rates.14 For example, although triamcinolone’s depot nature can result in longer duration of effectiveness within the injected site,13 transient impairment of vocal fold vibration might occur
after thick plaque formation of the injected triamcinolone, with incidence rates varying from 2.5% to nearly 100%.36,46,47 Such inconsistent results might be explained by different time frames for postoperative visits. Because the triamcinolone is mostly absorbed gradually, it was likely that subjects returning to the clinic earlier were more likely to reveal such adverse effects than subjects who return to the clinic after 1–2 months. To reduce the incidence rates of whitish plaque formation while maintaining the effectiveness of corticosteroid, we modified the steroid regimen by mixing triamcinolone with dexamethasone by half (50:50). Our study results demonstrated treatment outcomes comparable with the findings in the literature,14 with the incidence rate of triamcinolone deposits limited to 3.9%. The balance between benefits and harm may help to validate the rationale of applying modified corticosteroid regimen in this study. The most common complication after VFSI in this series is vocal fold hematoma (34 cases, 27%), much higher than that reported in previous studies (1%).13,46 Compared with the transcutaneous approach in which the needle migrates in subepithelial direction without interrupting the epithelium of the vocal folds,36,46 the transoral and transnasal injection approaches performed in this series require direct puncture of the injection needle through the upper epithelial cover of the vocal fold, which may result in higher rates of postoperative hematoma. Similarly, time between intervention to clinical follow-up might also have altered the incidence rates of vocal hematoma because these adverse effects tended to regress spontaneously within 1–2 months. In other words, if the surgeon choose to follow-up the patient long after the procedure, vocal fold hematoma might not have a chance to be identified. We noticed that the presence of vocal varices was the most significant risk factor for postoperative hematoma (Table 4); otolaryngologists may thus consider controlling varices before VFSI to prevent injury to these fragile neovasculatures during the injection process, especially for those with minimal tolerance of postoperative side effects (eg, singers).17,48 Another potential risk factor for post-VFSI vocal hematoma is high occupational vocal demand (Table 4). Although this study did not investigate the compliance rate of voice rest after VFSI,
Chi-Te Wang, et al
9
Treatment Outcomes of VFSI
we presumably suspected that patients with high occupational vocal demands might be less able to maintain 3 days of voice rest after VFSI than were patients with routine vocal demands,22 which might result in oozing before the injection wound was completely healed. A previous study observed a 5% incidence rate of vocal fold atrophy after VFSI.36 In contrast, this study revealed a much lower incidence rate (only one of 126 cases). One reasonable explanation for this discrepancy is the different injection approaches. As mentioned earlier, the transcutaneous approach requires subepithelial needle migration without a directly visible needle tip;36 therefore, misplacement of the steroid around the vocal ligament and vocalis muscle could occur. In contrast, transoral and transnasal needle placements were guided by direct, magnified endoscopic view; thus, surgeons could be more confident about the precise placement of the needle tip into the targeted Reinke’s space, which might reduce the rate of vocal atrophy as found in this study and previous studies.11,12 However, because the improvement of MPT in this series was less prominent than were other parameters (Table 1), one should be cautious that some patients with subclinical atrophic change might not be adequately detected by the routine VLS examinations. Based on the present study, the clinical roles of VFSI can be more clearly proposed. For vocal nodules, voice therapy is generally indicated with potentially longer-lasting effects in adherent cases.37,49 In patients who cannot comply with voice conservation and phonation habit modification, VFSI may be applied for symptomatic relief. However, with only one-third of the recruited patients having attended the voice therapy before or after VFSI in this series (41 of 126 patients), complete resolution of the nodules were noted for less than 15% (Table 1). Further interdisciplinary practice between otolaryngologist and speech therapist shall improve the treatment outcomes of VFSI, especially for hard/fibrotic nodules and the patients with higher occupational vocal demand (Table 1).50 For similar reasons, VFSI may be applied on vocal polyps to postpone or avoid (but not replace) phonosurgery,12 because only a fraction of the patients experienced complete remission postoperatively (
osis, systemic lupus erythematosus, and Wegener’s granulomatosis.4–6 A recent study demonstrated that officebased laryngeal procedures are well tolerated in the outpatient department with limited impact on the hemodynamic profiles.7 Compared with surgeries in the operating room under general anesthesia, office-based procedures not only avoid potential risks and discomforts associated with general anesthesia but also offer the advantages of real-time monitoring of vocal function and considerable savings of medical expenses.8–10 A number of studies have advocated the potential role of vocal fold steroid injection (VFSI) for the treatment of benign vocal fold lesions.11–13 However, few studies have investigated the treatment outcomes between different disease entities. Moreover, the prognostic factors of VFSI (eg, laryngopharyngeal reflux [LPR] and occupational vocal demand) have rarely been reported in the literature.14 Although side effects after VFSI have been noted in the literature, the incidence rates varied widely.14 Therefore, the objectives of this study are to conduct comprehensive researches that reports multidimensional treatment outcomes and prognostic factors between different categories of vocal lesions and to investigate the incidence rates and risk factors of side effects after VFSI. METHODS Study subjects This study prospectively recruited consecutive patients who had received VFSI from January 2012 to December 2013 at a tertiary teaching hospital. The inclusion criteria include vocal nodules, polyp, and mucus retention cyst. Clinical diagnoses
2 were based on the appearance of the vocal fold edge and the vibratory property of the mucosa revealed by videolaryngostroboscopy (VLS), based on the published nomenclature and diagnostic paradigm.1 Vocal nodules were classified as (1) soft (ie, mid-membranous thickening without fibrotic change of the epithelium or limitation of the mucosal wave) or (2) hard (ie, hyperkeratosis of the epithelium, fibrotic appearance with some limitation of mucosal wave propagation). Vocal polyps were classified as hemorrhagic (vascular), fusiform (wide-based), pedunculated, and fibrous.15 The presence of ectasias (defined as spheroidal, tangible, subepithelial hemangioma) and varices (defined as enlarged, aberrant, and tortuous vessels) were recorded as well.16,17 Patients with keratin cysts were excluded. Patients who had received previous phonomicrosurgery were also excluded from the study. The study protocol has been approved by the Research Ethics Review Committee. All of the clinical and demographical factors were collected at the initial clinical visit, including age, gender, duration of dysphonic symptoms, smoking, occupational vocal demand, as well as the 10-item Voice Handicap Index (VHI-10),18 and Reflux Symptom Index (RSI).19 Subjects with RSI scores of more than 13 points were defined as positive for LPR. We reviewed the occupations of each patient and classified them into (1) professional voice users (eg, singers, actor/actress, radio broadcast, and singing student); (2) high voice users (eg, teachers, clergy, lecturer, sales, tour guide, aerobic exercise coach, and patients who work in noisy environment; or (3) routine voice users, as suggested in the literature.20 Professional and high voice users were labeled as having ‘‘high’’ occupational vocal demand, and the others were labeled as having ‘‘routine’’ occupational vocal demand. Vocal Fold Steroid Injection The VFSI was performed under local anesthesia of the pharynx and larynx,21 and the patient was instructed to pull out and hold the tongue, while the surgeon uses the nondominant hand to operate the rigid laryngoscope. Under visual guidance, 0.1 mL, 1:1 mixture of triamcinolone acetonide (10 mg/mL) and dexamethasone sodium phosphate (5 mg/mL) was injected by curved injection needle (Model 16–50050, Medtronic Xomed Jacksonville, Florida, USA) into the Reinke’s space of the vocal lesions.22 For patients who were intolerant of the above transoral procedure,23 transnasal injection via the working channel of a flexible nasopharyngoscope could be applied with an endoscopic injection apparatus (Model NM-101C-0427 Olympus, Aomori, Japan.), which includes a reusable metallic external sheath (MAJ-655) and a disposable flexible needle tract with a 27G rigid tip (MAJ-656).24 All the patients received postoperative voice rest for 3 days. Outcome evaluation Treatment outcomes were evaluated before, 1 month, and 2 months after the procedure. Perceptual evaluation of voice quality was performed using the GRB scales and scored as 0: normal, 1: slightly deviated, 2: moderately deviated, and 3: extremely deviated.25 Scores from the three subscales (grade,
Journal of Voice, Vol. -, No. -, 2015
roughness, and breathiness) were summed up for statistical inferences. Acoustic analysis was conducted by recording a 3-second sample of the sustained vowel sound /a:/ at a comfortable level of loudness, with a microphone-to-mouth distance of approximately 15–20 cm, using a high-quality microphone (Model: SM58, SHURE Niles, Illinois, USA.) with a digital amplifier (Model: X2u, SHURE). Analyzed parameters include jitter (ie, cycle-to-cycle frequency perturbation), shimmer (ie, cycle-to-cycle amplitude perturbation), and noise-to-harmonic ratio (NHR; MDVP, Model 4500; Kay Elematrics, Corp., Lincoln Park, NJ, USA). Under similar settings, we measured maximal phonation time (MPT) by instructing the patient to produce the /a:/ sound for as long as possible after deep inspiration and at a spontaneous, comfortable pitch and loudness level for three consequent trials. The patient-reported outcomes included a Visual Analog Scale (VAS) of voice quality (ranging from zero [worst] to 10 [best]),26 subjectively perceived resolution of clinical symptoms (categorized as complete remission, more than 50% improvement, 50% improvement or less, no effect, or worsening), and Mandarin-Chinese version of VHI10.27 The vibratory capacity of the vocal fold was evaluated using VLS, which was conducted by instructing the patient to phonate a sustained /ee/ sound with habitual pitch and intensity, using a 70 rigid endoscope and a three-chip CCD camera (Model 2706CA and 20222120; KARL STORZ, Germany) or a digital laryngoscope with the corresponding video processor (VNL1590 STi and EPK-i; PENTAX). Each session of the VLS test was digitally recorded onto a portable hard disk using a computerized video processor (NHX-B10; Grass Valey, Inc.). Reduction of the lesion size was subjectively evaluated by the primary surgeon (C-.T.W.) during the clinical follow-ups. Intrarater reliability was tested by randomly re-examining 50% of the original video recordings, in a blinded fashion. Inter-rater reliability was examined in the same method, by an experienced speech therapist (Feng-Chuan Lin, MSc). We also objectively measured the normalized vibration amplitude of the membranous vocal fold by: (1) capturing video clips of maximally opened and closed phases during phonation with correction to the vocal fold length (ie, normalized glottal gap area [NGGA], Image J software, version 1.44; National Institute of Mental Health, Bethesda, MD) and (2) subtract the open phase from the closed phase NGGA (Figure 1).28,29 Side effects after VFSI were identified by reviewing postoperative video recordings and included (1) vocal fold hematoma surrounding the injection sites; (2) deposits of injected material, which was defined as the presence of whitish plaque over Reinke’s space, with minimal-to-moderate influence of the vibration of mucosal wave; and (3) vocal fold atrophy, which was defined as ‘‘bowing’’ of the vocal fold(s) with the presence of glottic gap and atrophic thyroarytenoid muscle. Statistics Treatment outcomes measured at 1 and 2 months postoperatively were compared with baseline measurements using paired t tests for the different diagnoses. Intra- and inter-rater reliabilities were examined by Cronbach’s alpha coefficient. To
Chi-Te Wang, et al
Treatment Outcomes of VFSI
3
FIGURE 1. The normalized vibration amplitude is calculated by subtracting the maximally opened glottic area (A) from the maximally closed glottic area (B) during the phonation cycle, and corrected by the length of the vocal fold, measured from the anterior commissure to the vocal process (C).
investigate the prognostic factors of the treatment outcomes, we defined ‘‘responders’’ to VFSI as: (1) postoperative VHI-10 scores 10 points, which denoted the upper limit of scores in the normal population27 and (2) GRB scores 1, which represents an objective rating of normal voice quality.30 Analyzed prognostic factors include gender, duration of clinical symptoms (>12 months vs 12 months), occupational vocal demand (high vs routine), cigarette smoking, LPR, and subtypes of vocal lesions.20,31,32 Univariate analyses for prognostic factors for treatment outcomes and risk factors for postoperative side effects were conducted by Chi-square tests. Multivariate analyses were subsequently conducted via multiple logistic regression model. A P value less than 0.05 indicated statistical significance. All statistical analyses were conducted by SAS software, version 9.3 (SAS Institute, Inc., Cary, NC). RESULTS This study recruited 126 patients with benign vocal lesions, including 19 men and 107 women. The mean age was 40 years (standard deviation: 10 years; range: 18–69 years). Vocal nodules were diagnosed in 49 patients, whereas polyps and cysts were diagnosed in the other 47 and 30 patients, respectively. Subtypes of vocal nodules include soft nodules in 32 patients and hard, fibrotic nodules in the other 17 patients. Subtypes of vocal polyps included hemorrhagic polyp in three, fusiform (wide-based) in 35, pedunculated in three, and fibrotic in six patients. High occupational vocal demands were reported in 75 patients, consisting mainly of teachers (26%), sales people (7%), customer service agents (25%), performing artists (6%), and others. Mean clinical duration of dysphonic symptoms was 15 months (standard deviation: 25 months). A total of 13 and 28 patients attended pre- and postoperative voice therapy during the treatment course, respectively. More than 80% of the patients with vocal nodules reported subjective improvements in their clinical symptoms 1 and 2 months after receiving VFSI (Table 1, left column). Similarly, subjective improvements in dysphonia were reported in up to 90% of the patients with vocal polyps and cysts (Table 1, middle and right column). The VLS examinations also revealed
decreased lesion sizes for more than 80% of the vocal lesions (Figure 2), as well as significantly improved vibration amplitudes (P < 0.01, paired t test, Table 1). Intra- and inter-rater reliabilities of VLS evaluations were 0.82 and 0.74, respectively (Cronbach’s alpha coefficient), compatible with a recent study, which indicated that VLS rating of the static lesion size is a reliable method.33 Treatment outcomes demonstrated significant improvements in VHI-10, VAS, and GRB scores after VFSI for all three diagnoses (P < 0.01, paired t test, Figure 3). Although significant improvements in MPT 1 month postoperatively were revealed in all three disease categories (P < 0.05, paired t test), the 2month results did not reach statistical significance (Table 1). Other parameters, including jitter, shimmer, and NHR, showed more universal improvements in vocal cysts, followed by partial improvements in vocal polyps and minimal improvements for vocal nodules. Because the etiology and contributing factors for developing various vocal lesions are different, we subsequently compared the prognostic factors for better clinical responses after VFSI between each of the three disease entities. For vocal nodules, patients with higher occupational vocal demands were associated with a poorer clinical response as measured by the VHI10 (P ¼ 0.03, Chi-square test, Table 2). Multivariate analyses also demonstrated significant correlation between vocal demand and treatment outcomes (adjusted odds ratio [aOR]: 4.32, 95% confidence interval [CI]: 1.06–17.6, P ¼ 0.04, multiple logistic regression). Additionally, hard and fibrotic vocal nodules, compared with soft nodules, had a lower chance of recovering normal perceptual voice quality as measured by GRB scores (P ¼ 0.01, Chi-square test). Multivariate analyses also validated this significant correlation (aOR: 9.27, 95% CI: 1.4–61.3, P ¼ 0.02). Univariate analyses of the prognostic factors for vocal polyps after VFSI showed significantly higher VHI-10 scores among the subsets of patients with dysphonia for more than 12 months (P ¼ 0.03, Chi-square test, Table 3). Subsequent multivariate analyses also demonstrated this association with borderline significance (aOR: 4.20, 95% CI: 0.91–19.3,
4
TABLE 1. Treatment Outcomes of Vocal Fold Steroid Injection in the 126 Patients of Vocal Nodules, Polyp, and Cyst Nodules Outcome Parameters Subjective evaluation, n (%) Complete remission Much improvement (>50%) Some improvement (50%) No effect Endoscopic evaluation, n (%) Complete remission Much improvement Some improvement No effect Normalized vibration amplitude Acoustic analysis MPT Jitter Shimmer NHR
Polyp
Cyst
Pre (N ¼ 49) 1 mo (N ¼ 43) 2 mo (N ¼ 33) Pre (N ¼ 47) 1 mo (N ¼ 42) 2 mo (N ¼ 29) Pre (N ¼ 30) 1 mo (N ¼ 28) 2 mo (N ¼ 21) 0 (0) 23 (54) 15 (34) 5 (12)
2 (6) 17 (52) 10 (30) 4 (12)
6 (14) 24 (57) 7 (17) 5 (12)
4 (14) 16 (55) 8 (28) 1 (3)
6 (21) 14 (50) 6 (21) 2 (7)
6 (29) 9 (43) 6 (29) 0 (0)
10.5 ± 4.7
6 (14) 28 (65) 7 (16) 2 (5) 13.4 ± 5.2*
4 (12) 15 (46) 8 (24) 6 (18) 14.9 ± 6.5*
10.6 ± 8.1
8 (19) 23 (55) 10 (24) 1 (2) 15.3 ± 10.7*
4 (14) 14 (48) 9 (31) 2 (7) 12.0 ± 7.9*
9.2 ± 5.1
6 (21) 16 (57) 3 (11) 3 (11) 11.5 ± 6.1*
6 (29) 10 (76) 4 (14) 1 (14) 11.9 ± 7.2
10.2 ± 4.1 1.6 ± 1.0 3.2 ± 1.5 0.2 ± 0.6
12.6 ± 5.1* 1.4 ± 0.7 2.9 ± 1.0 0.1 ± 0.02
10.8 ± 4.2 1.3 ± 0.9* 3.6 ± 3.4 0.1 ± 0.1
9.9 ± 4.7 1.9 ± 1.4 4.5 ± 2.8 0.1 ± 0.1
11.5 ± 4.9* 1.4 ± 0.9y 3.1 ± 1.2* 0.1 ± 0.1
11.4 ± 4.5 1.4 ± 1.3 3.8 ± 2.5 0.2 ± 0.2
12.0 ± 4.7 1.8 ± 1.2 4.6 ± 3.1 0.1 ± 0.1
13.8 ± 4.3y 1.3 ± 0.9y 3.5 ± 2.4y 0.1 ± 0.1
13.4 ± 4.6 1.1 ± 0.9y 3.3 ± 2.1y 0.1 ± 0.04 Journal of Voice, Vol. -, No. -, 2015
Abbreviations: MPT, maximal phonation time; NHR, noise-to-harmonic ratio. Notes: Data are expressed as mean ± standard deviation. * P < 0.01, paired t test, compared with baseline. y P < 0.05, paired t test, compared with baseline.
Chi-Te Wang, et al
Treatment Outcomes of VFSI
5
FIGURE 2. (A). A 47-year-old female patient who suffered from a mucus retention cyst on her left vocal fold. After transnasal VFSI (B), the lesion resolved completely within 1 month (C). P ¼ 0.07). Alternatively, we observed that patients who presented with LPR showed significantly poor postoperative voice quality based on univariate and multivariate analyses
(P ¼ 0.01, Chi-square test; aOR: 14.5, multiple logistic regression). Subsequent univariate and multivariate evaluations of vocal fold cysts did not reveal a significant correlation
FIGURE 3. Treatment outcomes of VFSI in patients with vocal nodules, polyp, and mucus retention cyst, measured by VHI-10, VAS, and GRB scores.
6
TABLE 2. Univariate and Multivariate Analyses of Prognostic Factors of VFSI Among Patients With Vocal Nodules Outcomes Parameters
Univariate
VHI-10 Gender Duration of symptoms (mo) Vocal demand Smoking Reflux Nodular type
Male/female >12/12 High/routine Active/nil Presence/absence Hard/soft
GRB scores Gender Duration of symptoms (mo) Vocal demand Smoking Reflux Nodular type
Male/female >12/12 High/routine Active/nil Presence/absence Hard/soft
Multivariate
Responder (10), n (%)
Nonresponder (>10), n (%)
P Value*
aOR (95% CI)
P Value
2 (10)/19 (90) 9 (43)/12 (57) 7 (33)/14 (67) 2 (10)/19 (90) 6 (29)/15 (71) 6 (29)/15 (71)
1 (4)/27 (96) 8 (29)/20 (71) 18 (64)/10 (36) 2 (7)/26 (93) 13 (46)/15 (54) 11 (39)/17 (61)
0.39 0.30 0.03 0.76 0.20 0.44
14.3 (0.50–411) 0.25 (0.05–1.16) 4.32 (1.06–17.6) 2.27 (0.18–28.1) 1.76 (0.45–6.92) 1.99 (0.47–8.43)
0.12 0.08 0.04 0.52 0.42 0.35
Responder (1), n (%)
Nonresponder (>1), n (%)
P Value*
1 (6)/16 (94) 6 (35)/11 (65) 9 (53)/8 (47) 3 (18)/14 (82) 5 (29)/12 (71) 2 (12)/15 (88)
2 (6)/30 (94) 11 (34)/21 (66) 16 (50)/16 (50) 1 (3)/31 (97) 14 (44)/18 (56) 15 (47)/17 (53)
0.96 0.95 0.84 0.11 0.33 0.01
aOR (95% CI) 0.64 (0.02–17.6) 1.30 (0.26–6.45) 0.45 (0.11–1.87) 0.09 (0.01–1.56) 3.37 (0.72–15.7) 9.27 (1.40–61.3)
P Value 0.79 0.75 0.27 0.10 0.12 0.02
Notes: aOR (95% CI): Adjusted odds ratio (95% confidence intervals) by multiple logistic regression model Bolded values represents statistical significances. * Chi-square test.
TABLE 3. Univariate and Multivariate Analyses of Prognostic Factors of VFSI Among Patients With Vocal Polyp Outcomes Parameters
Univariate
VHI-10 Male/female >12/12 High/routine Active/nil Presence/absence Fibrous/others
GRB scores Gender Duration of symptoms (mo) Vocal demand Smoking Reflux Polyp type
Male/female >12/12 High/routine Active/nil Presence/absence Fibrous/others
Responder (10), n (%)
Nonresponder (>10), n (%)
P Value*
3 (20)/12 (80) 3 (20)/12 (80) 6 (40)/9 (60) 3 (20)/12 (80) 6 (40)/9 (60) 1 (7)/14 (93)
3 (9)/29 (91) 17 (53)/15 (47) 15 (47)/17 (53) 3 (9)/29 (91) 14 (44)/18 (56) 5 (16)/27 (84)
0.31 0.03 0.66 0.31 0.81 0.39
Responder (1), n (%)
Nonresponder (>1), n (%)
P Value*
3 (27)/8 (73) 3 (27)/8 (73) 5 (45)/6 (55) 2 (18)/9 (82) 1 (9)/10 (91) 0 (0)/11 (100)
3 (8)/33 (92) 17 (47)/19 (53) 16 (44)/20 (56) 4 (11)/32 (89) 19 (53)/17 (47) 6 (17)/30 (83)
0.13 0.24 0.95 0.54 0.01 0.31
Notes: aOR (95% CI): Adjusted odds ratio (95% confidence intervals) by multiple logistic regression model Bolded values represents statistical significances. * Chi-square test. y Unable to estimate owing to the absence of subjects.
aOR (95% CI) 1.52 (0.19–11.9) 4.20 (0.91–19.3) 1.55 (0.38–6.26) 0.46 (0.06–3.64) 0.88 (0.21–3.70) 3.70 (0.29–47.9) aOR (95% CI) 3.00 (0.25–35.5) 1.33 (0.22–8.21) 1.79 (0.33–9.68) 0.31 (0.01–6.83) 14.5 (1.33–157) >1 (unavailabley)
P Value 0.69 0.07 0.54 0.46 0.86 0.32 P Value 0.38 0.76 0.50 0.46 0.03 —
Journal of Voice, Vol. -, No. -, 2015
Gender Duration of symptoms (mo) Vocal demand Smoking Reflux Polyp type
Multivariate
Chi-Te Wang, et al
Treatment Outcomes of VFSI
between clinical, endoscopic, or interventional factors and the treatment outcomes. After receiving VFSI, the most common side effects were hematoma (34 of 126 patients, 27%), followed by deposits of triamcinolone (5 cases, 4%; Figure 4A and B). Both conditions generally presented with dysphonia; increased phonation effort; and, occasionally, decreased voice quality compared with preoperative status. The VLS examination tended to reveal decreased amplitude and asymmetry of mucosal waves. Clinical observation showed that all side effects resolved spontaneously within 2 months, leaving no sequel. We observed one case with a breathy voice and increased phonation effort after VFSI. The VLS showed bowing of the vocal folds with prominent glottic gap and ventricles. Two months afterward, her symptoms resolved, and follow-up endoscopy showed recovery of her baseline vocal fold bulk (Figure 4C and D). The overall incidence rates for developing postoperative side effects were 32%, with no significant differences between the three disease entities. Univariate and multivariate analyses revealed that presentation with vocal fold ectasias and varicosities were significantly correlated with postoperative vocal hematoma (P < 0.05, Table 4). Meanwhile, multivariate analysis also identified that subjects with high occupational vocal demands were more likely to develop vocal hematoma after VFSI (P < 0.05, multiple logistic regression). Other clinical factors, including gender, smoking, hypertension, and LPR, all showed nonsignificant correlation with postoperative vocal hematomas (Table 4). Subsequent analyses of the potential risk factors for triamcinolone deposits and vocal atrophy were not performed because of the limited sample. DISCUSSION Corticosteroids are one of the most common anti-inflammatory agents in modern medicine, and they can be administered via multiple routes (eg, transoral, intramuscular, intravascular, or inhalational). Specifically, direct intralesional steroid injection has the advantage of maintaining a higher regional drug concentration with minimal systemic adverse effects. Yanagihara et al34 first applied direct intralesional steroid injection in the treatment of benign vocal fold disorders in 1964. However, VFSI as an office-based procedure to treat benign vocal fold disorders had not been widely accepted until the past decade, when Tateya et al35 made two major refinements to the proce-
7 dure, namely (1) enhanced precision of needle placement by flexible fiberscope guidance and (2) triamcinolone, a depot solution with longer effective duration, as the injected regimen. Since then, a number of published studies have demonstrated that VFSI can be applied to treat many structural vocal fold disorders, such as vocal nodules, polyp, cyst, scar, granuloma, and Reinke’s edema.11–13 The clinical effectiveness of VFSI was further documented in our recent study that demonstrated significantly higher lesion reduction rates after VFSI than following vocal hygiene education.22 This study investigated the multidimensional treatment outcomes among the most common vocal fold pathologies. For vocal nodules, we demonstrated that VFSI can result in significant subjective and objective improvements (Table 1, left column), which was comparable with previously published results.11,36 Moreover, we investigated the prognostic factors of vocal nodules after VFSI, which had not previously been done. The study results demonstrated that patients with higher vocal demand were more likely to perceive higher phonation discomforts as measured by their postoperative VHI-10 scores (Table 2, upper half). Because vocal nodules are closely related to occupational overuse of the voice, without adequate modification of phonation habits, the therapeutic effects of corticosteroids on vocal folds may gradually be countered by persistent voice overloads; if that is the case, dysphonic symptoms may persist.37 This study also found that subjects with thick, fibrotic nodules were less likely to regain normal voice quality after VFSI (Table 2, lower half). The histological features of vocal nodules may help to explain these findings. In soft nodules, which present with stroma edema, dilation, and the increased permeability of microvasculature among the superficial lamina propria,38 VFSI can be more effective than it is for fibrotic or thick nodules, which feature a thickened epithelium and the proliferation of fibroblasts with abundant fibrin and organized collagen.39,40 For patients with vocal polyps, this study demonstrated significant improvements in most of the outcome parameters, comparable with results from a previous study.13 Subsequent analyses revealed that patients with dysphonic symptoms lasting for more than 1 year were more likely to have higher VHI-10 scores postoperatively (Table 3, upper half). It is likely that chronic vocal polyps may present with a higher degree of fibrin deposition, inflammation, and amyloid-like material
FIGURE 4. Side effects after VFSI, including hematoma (A), triamcinolone deposits (B, arrow) and bowing atrophic vocal folds (C, arrowhead). Atrophic vocal folds recovered completely after 2 months (D).
8
Journal of Voice, Vol. -, No. -, 2015
TABLE 4. Univariate and Multivariate Analyses of Risk Factors of post-VFSI Vocal Hematoma Analyses
Univariate
Parameters Gender Smoking Hypertension Vocal demand Reflux Ectasia/varicosity
Male/female Active/nil Presence/absence High/routine Presence/absence Presence/absence
Multivariate
Hematoma (N ¼ 34)
No Hematoma (N ¼ 92)
P Value*
aOR (95% CI)
P Value
3 (9)/31 (91) 1 (3)/33 (97) 4 (12)/30 (88) 18 (53)/16 (47) 11 (32)/23 (68) 13 (38)/21 (62)
16 (17)/76 (83) 15 (16)/77 (84) 7 (8)/85 (92) 33 (36)/59 (64) 38 (41)/54 (59) 18 (20)/74 (80)
0.23 0.07 0.46 0.08 0.36 0.03
1.60 (0.38–6.68) 0.11 (0.01–1.10) 1.40 (0.35–5.67) 2.74 (1.11–6.75) 0.54 (0.21–1.37) 4.28 (1.55–11.9)
0.52 0.06 0.64 0.03 0.20 0.01
Notes: aOR (95% CI): Adjusted odds ratio (95% confidence intervals) by multiple logistic regression model Bolded values represents statistical significances. * Chi-square tests.
accumulation in Reinke’s space,40 which may be less reversible after one course of VFSI. Meanwhile, chronic vocal polyps may also more likely to cause reactive nodules on the contralateral vocal folds, which can also contribute to the less favorable postoperative recovery. A recent study found a 6% coexistence rate of vocal polyps with sulcus vocalis, which may negatively affect surgical outcomes.41 In this study, we also identified two cases of vocal polyps with underlying sulcus; both patients presented with symptom duration longer than 12 months and a less favorable response to VFSI. These comparable findings may also help to explain the association between the chronicity of diseases and treatment outcomes. Our study results found that patients with vocal polyp comorbid with LPR tended to present with poorer voice quality after VFSI compared with the patients without LPR (Table 3, lower half). Although a previous study had already recognized the strong relationship between LPR and vocal fold disorders,42 this study provides additional clinical evidence of LPR and the treatment outcomes. Persistent inflammation surrounding the glottis caused by LPR (eg, laryngeal edema and increased phlegm) may explain the fair recovery of voice quality after VFSI.43,44 Mucus retention cysts are typically located at the subepithelial plane of the vocal folds that indicate an obstructed excretory duct with retained mucus of the affected gland.15 Histologically, mucus retention cysts consist of an outer layer of cuboidal cells with an internal layer of ciliated columnar cells.45 Except for a small number of reported cases,12 this is the first clinical series to document the treatment effectiveness of VFSI for mucus retention cysts of vocal folds. Based on our preliminary experience, the needle tip can sometimes directly puncture through the cyst wall and drain the cystic content during the injection process, which might explain the higher remission rate and rapid clinical response of mucus retention cysts after VFSI comparing with vocal nodules and polyps (Table 1). Previous studies had already reported potential adverse events after VFSI, with a highly variable range of incident rates.14 For example, although triamcinolone’s depot nature can result in longer duration of effectiveness within the injected site,13 transient impairment of vocal fold vibration might occur
after thick plaque formation of the injected triamcinolone, with incidence rates varying from 2.5% to nearly 100%.36,46,47 Such inconsistent results might be explained by different time frames for postoperative visits. Because the triamcinolone is mostly absorbed gradually, it was likely that subjects returning to the clinic earlier were more likely to reveal such adverse effects than subjects who return to the clinic after 1–2 months. To reduce the incidence rates of whitish plaque formation while maintaining the effectiveness of corticosteroid, we modified the steroid regimen by mixing triamcinolone with dexamethasone by half (50:50). Our study results demonstrated treatment outcomes comparable with the findings in the literature,14 with the incidence rate of triamcinolone deposits limited to 3.9%. The balance between benefits and harm may help to validate the rationale of applying modified corticosteroid regimen in this study. The most common complication after VFSI in this series is vocal fold hematoma (34 cases, 27%), much higher than that reported in previous studies (1%).13,46 Compared with the transcutaneous approach in which the needle migrates in subepithelial direction without interrupting the epithelium of the vocal folds,36,46 the transoral and transnasal injection approaches performed in this series require direct puncture of the injection needle through the upper epithelial cover of the vocal fold, which may result in higher rates of postoperative hematoma. Similarly, time between intervention to clinical follow-up might also have altered the incidence rates of vocal hematoma because these adverse effects tended to regress spontaneously within 1–2 months. In other words, if the surgeon choose to follow-up the patient long after the procedure, vocal fold hematoma might not have a chance to be identified. We noticed that the presence of vocal varices was the most significant risk factor for postoperative hematoma (Table 4); otolaryngologists may thus consider controlling varices before VFSI to prevent injury to these fragile neovasculatures during the injection process, especially for those with minimal tolerance of postoperative side effects (eg, singers).17,48 Another potential risk factor for post-VFSI vocal hematoma is high occupational vocal demand (Table 4). Although this study did not investigate the compliance rate of voice rest after VFSI,
Chi-Te Wang, et al
9
Treatment Outcomes of VFSI
we presumably suspected that patients with high occupational vocal demands might be less able to maintain 3 days of voice rest after VFSI than were patients with routine vocal demands,22 which might result in oozing before the injection wound was completely healed. A previous study observed a 5% incidence rate of vocal fold atrophy after VFSI.36 In contrast, this study revealed a much lower incidence rate (only one of 126 cases). One reasonable explanation for this discrepancy is the different injection approaches. As mentioned earlier, the transcutaneous approach requires subepithelial needle migration without a directly visible needle tip;36 therefore, misplacement of the steroid around the vocal ligament and vocalis muscle could occur. In contrast, transoral and transnasal needle placements were guided by direct, magnified endoscopic view; thus, surgeons could be more confident about the precise placement of the needle tip into the targeted Reinke’s space, which might reduce the rate of vocal atrophy as found in this study and previous studies.11,12 However, because the improvement of MPT in this series was less prominent than were other parameters (Table 1), one should be cautious that some patients with subclinical atrophic change might not be adequately detected by the routine VLS examinations. Based on the present study, the clinical roles of VFSI can be more clearly proposed. For vocal nodules, voice therapy is generally indicated with potentially longer-lasting effects in adherent cases.37,49 In patients who cannot comply with voice conservation and phonation habit modification, VFSI may be applied for symptomatic relief. However, with only one-third of the recruited patients having attended the voice therapy before or after VFSI in this series (41 of 126 patients), complete resolution of the nodules were noted for less than 15% (Table 1). Further interdisciplinary practice between otolaryngologist and speech therapist shall improve the treatment outcomes of VFSI, especially for hard/fibrotic nodules and the patients with higher occupational vocal demand (Table 1).50 For similar reasons, VFSI may be applied on vocal polyps to postpone or avoid (but not replace) phonosurgery,12 because only a fraction of the patients experienced complete remission postoperatively (
