Gore-Tex Medialization Thyroplasty: Objective and Subjective Evaluation

INTRODUCTION 

The first surgical treatment of unilateral paralytic dysphonia in the modern era was Bruning's intracordal injection of paraffin in 1911.1 In 1915, Payr introduced the anteriorly based thyroid cartilage flap.2 Each procedure produced only limited success, and perhaps as a result, surgical treatment of glottic insufficiency was limited through the mid 1950s. Teflon for vocal fold injection in the 1960s was the first synthetic material that produced improved voice consistently through vocal fold medialization.3

For a variety of reasons since its introduction, otolaryngologists have looked for alternatives to Teflon injections. Implant migration, granuloma formation, vocal fold scarring and bowing have all factored into the decreasing popularity of Teflon use for vocal fold mediziation.4 In the early 1970s, Isshiki described the first external medialization procedure that produced positive results consistently without interrupting the relationship of the vocalis muscle to the mucosal cover.5 With the evolution of stroboscopy and its incorporation into the otolaryngologic practice in the 1980s, the importance of the independent and fluid movement of the leading edge of the vocal fold mucosa became recognized widely by clinicians. Discoveries relating to vocal fold anatomy and physiology supported arguments in favor of medialization via external laryngeal framework surgery over injections into the submucosa.

Since that time, both external medialization and intracordal injection techniques have been refined greatly. Materials for injection have included Teflon, Gelfoam, autologous fat, autologous and allogeneic collagen, bovine collagen, and autologous fascia. Type I thyroplasty has proven consistently effective in improving voice outcomes, and the procedure has been modified by many surgeons using various prostheses. Silastic blocks, described by Isshiki5 can be carved by the surgeon or purchased prefabricated and are placed usually between the inner and outer perichondrium. When carved intraoperatively, the Silastic block allows medialization to be customized to the specifications of the preoperative glottic gap. In 1992, the Montgomery stent was introduced to improve closure of the posterior commisure.6., 7. It locks into the window in the thyroid cartilage, allowing easy accessibility and the possibility of exchange or removal. It is currently in use as an alternative to self-carved Silastic implants. Hydroxyapatite prosthetics provide yet another option for medialization.8

One of the most recent additions to medialization implant surgery is the use of Gore-Tex ribbon placed in a traditional or smaller, modified window in the thyroid cartilage.9 There are some notable advantages to the Gore-Tex ribbon method. The flexibility of the ribbon allows the surgeon to distribute the degree of medialization differentially along the length of the vocal fold, thus allowing finely tuned, intraoperative adjustments that do not involve removal and replacement of the entire prosthesis. This flexibility also allows the surgeon to fit the ribbon through a small cartilage fenestration. The senior author (RTS) uses a 4 mm thyrotomy window created with a diamond burr. The Gore-Tex implant requires no carving, is relatively easy to place, and its malleability permits contouring of the surrounding tissue. Greater pliability also may decrease extrusion potential and make Gore-Tex a more “natural” implant for vocal fold augmentation. Because of these unique properties inherent to the material itself, and the ease of surgical placement, indications for thyroplasty may be expanded to include almost any anatomic defect at the glottic level that leads to aerodynamic glottic insufficiency.

Objective acoustic, aerodynamic and stroboscopic data, as well as subjective ratings, have been assessed widely, in multiple institutions, for most implant techniques. However, as far as the authors are aware, there are no data currently available on a broad spectrum of objective and subjective parameters for Type I thyroplasties using Gore-Tex, subjected to rigorous statistical analysis. We provide such data and analysis, with randomized pre- and postoperative stroboscopic recordings, randomized voice recordings, and blinded evaluations on patients for whom Gore-Tex prostheses have been placed.

MATERIALS AND METHODS 

RESULTS 

The authors performed multiple logistic regression analysis with vocal fold amplitude symmetry as the outcome variable, and evaluator, preoperative, and postoperative status as the dependent variables. The odds ratio shows a 2.29 times greater chance of amplitude symmetry after thyroplasty, although, due to the small sample size, these results were not statistically significant (P=0.177, Table 1). There was no effect based on different evaluators (P=0.189).

TABLE 1. Stroboscopic Outcome Measures After Gore-Tex Thyroplasty

	Vocal Fold Measure	Odds Ratio	P-Value
	Amplitude Symmetry	2.29	0.177
	Wave Form Symmetry	1.73	0.123
	Glottic Closure	5.74	0.004

Next, the authors performed multiple logistic regression using waveform symmetry as the outcome variable, to see if either the evaluator or the procedure had an effect on the outcome. The odds ratio was 1.73, meaning that patients had 1.73 times greater chance of wave form symmetry after thyroplasty (P=0.123). Again, the evaluator did not significantly affect the outcome of the test.

Finally, the authors looked at the effect of evaluator and pre/postoperative status on glottic closure using a multiple logistic regression model. In this case, Gore-Tex thyroplasty showed a significant effect on glottic closure based on the ratings. Odds ratio was 5.74, meaning that there was a 5.74 times greater chance of achieving glottic closure after Gore-Tex thyroplasty (P=0.004). In this model, there was no evaluator effect (P=0.736).

To evaluate the effect of Gore-Tex thyroplasty on objective voice measurements, the authors used a Wilcoxon sign–rank test because of the possibility that the paired data were not normally distributed. The parameters evaluated were jitter, noise-to-harmonic ratio (NHR), maximum phonation time (MPT), and mean flow rate (MFR). The authors found each of these parameters to have changed, but none of these changes was statistically significant. Mean jitter was shown to decrease from 290.0 to 202.8 (P=0.1823, Table 2). NHR decreased from .255 to .227 (P=0.5829). Mean MPT increased by 1.3 seconds (P=0.5940), and mean MFR decreased from 430.4 ml/s to 367.3 ml/s (P=0.678). Most of these changes favored Gore-Tex thyroplasty having a beneficial effect, except for MFR. It should be noted, however, that there was some degree of difficulty obtaining accurate measurements for this parameter on some patients because of varying abilities to follow instructions of the evaluator.

TABLE 2. Acoustical Analysis, Phonatory Ability, and Aerodynamic Outcome Variables

	Voice Measure	Preoperative Mean	Postoperative Mean	P-Value
	Jitter	290.0	202.8	0.183
	Noise-Harmonic Ratio	0.255	0.227	0.583
	Maximum Phonation Time	8.0 seconds	9.3 seconds	0.594
	Mean Flow Rate	430.4 mL/s	367.3 mL/s	0.678

Psychoacoustic evaluation 

To measure grade, roughness, breathiness, asthenia, and strain, examiners listened to a sustained /s/ and then /a/, and checked on a 100 mm line, how well the patient performed for each outcome. A one hundred was the maximum grade, strain, and so forth, and 0 was none at all (see Figure 1). The multiple examiners, and multiple outcomes were fitted into a random effects model based on preoperative and postoperative test status for the sounds /s/ and /a/ The author (JS) constructed a linear mixed effects model to model the effect of surgery on voice quality, controlling for evaluator effect. The dependent variable was the square root of distance to the mark (in mm). The square root transformation was used to meet the assumption of normality of the dependent variable. The covariates were evaluator, test type (/a/ or /s/), evaluation (roughness, etc), test (pre/post), and subject. Subject was a random effect, while all others were fixed. It turned out that there was a great deal of interaction between the covariates in this model, making it difficult to interpret. Therefore, four different mixed models were applied, one for each type of evaluation. This approach is not ideal, as it does not account for the high degree of correlation between the outcomes for the four different evaluation types. It seemed that this approach would yield the most interpretable results. It was found that there was a mixing of effects between evaluators test status, and outcome for each measure, and in no case was an effect found to be significant. The mean and standard deviations for each evaluator are shown for both /s/ and /a/ (see Table 3 and 4, respectively).

View Large Image Download to PowerPoint

Figure 1. Psychoacoustic analysis pre- and post-thyroplasty of three examiners, “00,” “01,” “10.”

TABLE 3. Psychoacoustic Analysis of 3 Evaluators on /s/, Pre and Post-Thyroplasty

	Test status	Voice Measure	Evaluator 1 Mean (mm)	Evaluator 1	Evaluator 2	Evaluator 2	Evaluator 3	Evaluator 3
	Preoperative /s/	Grade	43.90	22.60	55.90	27.45	54.91	26.21
		Strain	19.64	13.92	51.80	30.15	43.82	30.83
		Breathiness	21.27	13.98	39.18	23.26	21.27	20.81
		Roughness	32.73	25.79	46.64	26.34	49.45	28.77
	Postoperative /s/	Grade	39.00	27.46	48.40	30.57	45.82	30.93
		Strain	27.18	33.90	47.36	28.86	33.82	29.45
		Breathiness	20.91	22.71	29.55	28.73	25.64	31.93
		Roughness	36.64	26.98	47.18	25.56	39.73	34.82

TABLE 4. Psychoacoustic Analysis of 3 Evaluators on /a/, Pre and Post-Thyroplasty

	Test status	Voice Measure	Evaluator 1 Mean (mm)	Evaluator 1 SD	Evaluator 2 Mean	Evaluator 2	Evaluator 3 Mean	Evaluator 3
	Preoperative /a/	Grade	53.27	29.93	54.10	22.11	56.55	25.42
		Strain	41.64	32.11	24.30	11.56	32.27	28.98
		Breathiness	37.55	25.69	36.82	17.61	42.27	27.54
		Roughness	39.82	30.58	35.70	27.93	51.64	26.98
	Postoperative /a/	Grade	42.55	31.84	53.91	25.53	59.18	23.02
		Strain	41.18	31.33	33.91	20.42	29.18	31.19
		Breathiness	18.45	16.69	32.18	20.42	49.00	29.10
		Roughness	42.45	31.88	49.09	31.75	52.64	29.83

DISCUSSION 

Because of the number of different procedures available for laryngeal framework surgery, it has become especially important that each procedure be subjected to rigorous testing and analysis so that results can be compared in meaningful ways. In the literature, there is a wealth of studies performed, but a dearth of statistical analysis to validate the results. Some interesting observations can be made based on the analysis of our 14 patients who underwent Gore-Tex thyroplasty.

First, stroboscopic analysis demonstrated that the odds ratios for each of the outcomes, vocal fold symmetry, vocal fold amplitude, and glottic closure ranged from around two to over five. This can be interpreted to mean that patients who underwent Gore-Tex medialization thyroplasty were between two and five times more likely to have greater amplitude, symmetry, and closure after surgery than before the surgery. Even though these odds ratios are very high and because the sample size is so small, the results are not statistically significant. This should not, however, detract from the clear effect shown by each parameter independently.

For acoustic analysis, phonatory ability, and aerodynamic studies, the trends were in the direction of improvement for all parameters studied with the exception of MPT. Again, although the small sample size prevented these results from attaining statistical significance, the trend is encouraging.

With respect to psychoacoustic analysis of grade, strain, breathiness, and roughness, the sample size was more than sufficient to show an effect because of the statistical power in the random effects model, the multiple examiners, and the two different sounds used. The lack of an effect, and particularly the very high standard deviations, suggest that the instrument used failed to reflect the reality of surgical results. Further substantiating this suspicion was the lack of interrater reliability for the outcomes measured, as can be seen from the data presented in TABLE 3, TABLE 4. If all the evidence about surgical results is viewed together, we have the following picture: patient satisfaction with voice improvement is almost universal and is statistically significant, and aerodynamic, acoustic, and phonatory results all trend in the direction of voice improvement, and stroboscopic analysis indicates that parameters improved after thyroplasty. On the other hand, psychoacoustic evaluation shows no consistent effect of surgery, and no reliable trends among evaluators of the same passages.

The fact that the psychoacoustic instrument does not reflect improvements seen by other tests, indicates that this may be a poor reflection of surgical outcome. This test might be further evaluated in order to determine in a more systematic way, whether it is an appropriate instrument to evaluate thyroplasty outcomes.

CONCLUSION 

The authors have had good experience with the use of Gore-Tex thyroplasty. The technique has advantages over more traditional prostheses. Gore-Tex is malleable, adjustable within the cartilaginous window, and requires only a small thyroplasty that can be created easily without disrupting the strap muscles. Results with a small group of patients are promising. It is difficult to compare the results here with other methods of surgery because of differences in reporting methods, but the authors look forward to greater efforts toward standardizing outcome evaluation with the hopes of stronger tools for comparison among techniques.