Abstract
OBJECTIVE: Medialization thyroplasty (MT) and arytenoid adduction (AA) are effective treatments for medializing the paralyzed vocal cord, but indications and benefits of each procedure are controversial. We performed a formal evidence-based review to answer this question: In patients with unilateral vocal cord paralysis, does AA combined with MT significantly improve subjective and objective voice outcomes compared with MT alone?
STUDY DESIGN: We performed a MEDLINE literature search using specific search terms to identify pertinent articles, which were reviewed and graded according to the evidence quality.
RESULTS: We identified 219 potentially pertinent articles; detailed review yielded 10 studies for further analysis. Only 3 studies directly compared MT with AA plus MT: overall, there was no clear benefit in subjective or objective outcomes for AA plus MT.
CONCLUSIONS: Very limited grade C evidence indicates no clear benefit, or lack of benefit, in subjective or objective outcomes if AA is added to MT, compared with MT alone.
A second procedure developed by Isshiki attempts to address this deficiency. AA uses a suture placed through the muscular process of the arytenoid to produce rotation of the arytenoid cartilage and medialization of the true vocal fold, simulating the action of the lateral cricoarytenoid muscle. 2 In theory, this should close any posterior glottic chink and allow the mobile vocal cord to meet the paralyzed vocal cord in the midline at the same level. When added to anterior glottic closure (ie, MT), phonation should be improved and aspiration diminished. However, the increased tissue dissection resulting from the need to access the arytenoid process is associated with a small risk of significant complications. 3 Furthermore, the AA procedure may require an additional night's stay in the hospital due to the need to monitor for airway edema and drain output. 3
Although both AA and MT each have distinct advantages, in practice there is no clear consensus on the unique role of each procedure. Many otolaryngologists use thyroplasty alone and seldom or never perform AA, and they report very good outcomes. Others are proponents of AA in selected patients, whereas others perform both AA and thyroplasty together fairly routinely. Because of this clinical uncertainty concerning the role and benefit of each procedure, we performed an evidence-based review to evaluate the published literature.
In an evidence-based review, there are 4 steps: 1) ask a focused clinical question; 2) systematically search the published literature; 3) critically evaluate the literature for methodology and quality; and 4) make recommendations based on the quality of the evidence and the results of the studies. 4 Then, in the practice of evidence-based medicine, the clinician should take the results of the evidence-based review and integrate that with their own clinical experience and the wishes of each individual patient. 4,5
In this review, our question concerns the role of AA and MT combined and the potential outcome improvement. If a significant improvement in voice outcomes is found when the procedures are combined, then the increased risk of airway complications and lengthened hospital stay may be warranted.
METHODS
Our focused clinical question was the following: In patients with unilateral vocal cord paralysis, does AA combined with MT significantly improve subjective and objective voice outcomes compared with MT alone?
To address this question, we performed a MEDLINE literature search using the following methodology. The limits were the English-language literature, and MEDLINE entry dates from 1996 to July 2002. The following were not MeSH (Medical Subject Heading) terms and were therefore used as free text terms in individual searches: “thyroplasty,” “medialization laryngoplasty,” “laryngeal framework surgery,” “Silastic medialization,” and “arytenoid adduction.” The following MeSH terms were used in individual searches: “voice quality,” “voice,” “treatment outcome,” “voice disorders,” “vocal cord paralysis,” “glottis,” “thyroid cartilage,” “arytenoid cartilage,” and “laryngeal muscle.” Various combinations of the free text searches were performed along with MeSH term searches, and the combination that yielded the greatest number of relevant citations included the 5 free-text searches combined with “vocal cord paralysis.” Search results were further reviewed to identify articles pertinent to the focused questions, and each pertinent article was assigned an evidence level based on the research methodology and data reporting, as described by the Centre for Evidence-Based Medicine. 6
RESULTS
We retrieved 219 citations using our MEDLINE search. The title and abstract of each citation were then reviewed to identify relevant articles for further review—based on our focused clinical question. The majority of those articles discussed the benefits of MT or AA as a single intervention for treating patients with unilateral vocal cord paralysis. Some articles described indications for each procedure being used in different patients and then combined the results of the 2 procedures to determine whether laryngeal framework surgery considered as a whole improved subjective and objective voice outcomes. Other articles focused on the complications encountered with each procedure or on the different possible combinations of laryngeal framework surgery.
Ten of the 219 articles seemed to be pertinent to our focused question, and they were reviewed in detail. 3,7-15 On detailed review, only 3 of these articles directly evaluated the voice outcomes of MT plus AA versus MT alone, 3,7,8 although 3 other articles reported limited data comparing those procedures. 9-11 One article evaluated patients who had primarily undergone both AA and MT, 12 and 1 article directly compared AA outcomes with MT outcomes. 13 Finally, 1 article assessed AA alone compared with AA plus MT, 14 which was interesting but not pertinent to our focused question. Another article compared complications between the 2 procedures but did not report other subjective or objective outcomes. 15 These articles are summarized in the Evidence Table.
Evidence table
One study 3 reported outcomes data in terms of percent with individual symptom improvement, and then the authors calculated a total percent with “overall” symptom improvement. Those data indicated that 93% had overall symptom improvement with MT plus AA, versus 94% with MT alone. The authors also reported that the visually estimated posterior glottic gap was significantly larger (P = 0.005) for MT patients (compared with AA plus MT), but raw data and the method of data measurement were not reported. In addition, the larger posterior glottic gap did not mean poorer subjective outcomes. Objective data such as jitter or maximum phonation time were not reported, and the difference in complication rates (MT plus AA = 19% versus MT alone = 14%) did not reach statistical significance (P = 0.14).
Another study 7 reported only mean group data with no statistical analysis. For example, mean subjective rating of severity (by a nonblinded speech pathologist using a 0-to-6 scale) improved from 4.5 to 2.8 in the MT group but from 4.2 to 1.6 in the AA plus MT group. Similarly, mean phonatory time increased from 8.3 to 9.9 in the MT group and from 6.9 to 16.7 in the MT plus AA group. Both of these differences appear to be significant, but without statistical analysis (or the raw data so the analysis could be performed), we cannot be sure. Similar evaluation of the mean data indicates that overall vocal effort, roughness, breathiness, and strain were very similar in the 2 groups. No objective outcome data were reported.
The Thompson study presented only graphical data showing the direction of change from preoperative to postoperative, with no individual or group mean data. The authors concluded from their own experience that MT alone—either with or without AA—is adequate treatment for unilateral vocal cord paralysis. Visually, their data seem to support this opinion because there is no clear advantage seen by adding AA, but more substantive analysis is impossible. Furthermore, even if raw data were available, the sample size of this study was very small. There were 3 other studies 9-11 in which different procedures were performed, but they did not report comparison data—only pooled group data, making comparison of MT plus AA versus MT alone groups not possible.
DISCUSSION
The literature is replete with studies concerning MT and/or AA. However, only a few studies have been reported that compare subjective or objective voice outcomes in patients who have undergone MT plus AA versus MT alone. We believe this is the most important question, because the role and benefits of MT have been clearly reported and established. 1,16,17 Therefore the question of added benefit (and risk) of adding AA is worthy of review and study. Other questions, such as AA alone versus MT alone are interesting but probably less pertinent to the practicing otolaryngologist because MT alone is so effective and widely used.
In our overall review of the literature, we found that most studies addressing both procedures combined the 2 together and compared preoperative and postoperative laryngeal function but did not compare the 2 different techniques or their outcomes. These studies were excluded from further analysis and discussion.
We did identify a handful of studies that directly or indirectly addressed our focused question. All of those were case series, some with comparisons between groups and some with no comparisons, and none had any form of randomization. Most studies were retrospective. Because they were case series, all were level IV evidence. Some case series were graded level IV(—) because of incomplete data reporting or incomplete statistical analysis. Studies without an internal control group can show improved outcomes; however, only studies with a control group or comparison group can demonstrate effectiveness, and only randomization between groups can demonstrate true efficacy. 5
Another problem we found is the lack of consistent outcomes measure reporting and the lack of detailed data reporting from different groups of patients. Without consistent outcomes measures, data from different studies cannot be pooled to take advantage of the larger sample size and power of pooled analysis. Furthermore, without more clearly defined outcome data, the “number needed to treat” (NNT) value is difficult to calculate. The NNT is helpful in understanding the magnitude of outcome differences between treatments. For example, say there was agreement on a threshold of “excellent” outcome (ie, score of 85 on a 100-point outcome scale), and 90% of patients undergoing MT plus AA reached that outcome level, but only 80% of patients undergoing MT alone reached that level. Then, the NNT could be calculated as (100/90 — 80) = 10 patients. In other words, in this hypothetical example, 10 patients would have to receive MT plus AA to spare 1 patient a poorer outcome using MT alone. That information could be used by the clinician in weighing potential benefits and risks of the more extensive procedure. However, unfortunately, the lack of consistent reporting standards means that the NNT for AA cannot be calculated.
Although the data from the Abraham study were not amenable to further analysis, if we use the “overall” symptom improvement data, the outcome differences were trivial, and the estimated NNT would be 100 patients. In addition, the patients in the Abraham study 3 were cancer patients with generally poor pulmonary and nutritional status. Therefore, the results of this study may not be generalizable to all patients with unilateral vocal cord paralysis. The McCulloch study 7 presented mean group data only, with no statistical comparisons but a fairly good sample size. Although overall subjective voice rating and mean phonation time appeared to be a bit better with the combined procedure, individual subjective voice characteristics did not appear to be different. The data reported do not allow an NNT analysis. There are even fewer comparative data reported in the Thompson study, 8 and further analysis or calculation of NNT is not possible. Although it is not technically part of an evidence-based review, it is interesting to note that the authors of 2 of the 3 comparative reports 3,8 did report their opinion that there did not seem to be any consistent outcome benefit by performing AA plus MT compared with MT alone.
Two studies did compare the complication rates of MT and AA. 3,15 In one study, complication rates were not significantly different between the AA plus MT group compared with the MT group; however, the only 2 patients who did require tracheotomy had both undergone AA plus MT. 3 In addition, that study found that mean time of surgery and the mean length of hospital stay were significantly increased in the AA plus MT group. 3 In another study, 3.5% of patients undergoing AA had airway compromise requiring tracheostomy 15 ; however, no statistical analysis was provided to compare the incidence of tracheostomy between AA and MT. Neither study showed a statistically significant increase in total complications after AA.
There are several interesting additional questions regarding the role of AA in revision laryngoplasty procedures such as failed MT or the use of AA in only selected patients. These questions were not addressed in this evidence review because we were reviewing the evidence supporting our focused clinical question. Certainly we found no evidence to indicate that AA results in increased surgical risk or poor outcomes, but this evidence review does not address the overall effectiveness of AA or its use for specific indications or patients.
In summary, the overall grade of the literature evidence to answer the question, “In patients with unilateral vocal cord paralysis, does arytenoid adduction combined with medialization thyroplasty significantly improve subjective and objective voice outcomes compared with medialization thyroplasty alone?” is grade C. Controlled studies, or ideally randomized studies, would be very helpful in answering this clinical question with stronger evidence. From the studies that are available, however, there is no clear evidence that adding AA to MT provides any benefit, or lack of benefit, in subjective or objective voice outcomes. However, there is also no evidence that AA adds significant risk compared with the MT procedure or that AA causes poorer outcomes.