Very long‐term Voice Handicap Index Voice Outcomes after Montgomery Thyroplasty: A cross‐sectional study

The aim of this multicentric cross‐sectional study was to examine the permanency of Montgomery thyroplasty (MTIS) results from a patient's perspective.

Fourteen months was chosen as this was found to be the cut-off between short and long-term post-op according to the literature. In case of multiple postoperative early VHI-30 assessments, the earliest data were considered for the study. The VHI-30 must have been validated and published for the language of use.
A careful electronic medical record (EMR) review excluded deceased patients and patients that had to undergo further voice surgeries after the MTIS.
This study is a cross-sectional, survey based, multicentric study filed on NIH's ClinicalTrials.gov under the number NCT02969993 (first received November 18th 2016).
The study consisted of collecting VHI-30 questionnaires from patients who had been previously been operated with MTIS at least 2 years before.
The questionnaires were obtained in two different ways. In case of centres 1 and 3, VHI-30 questionnaires were sent out by mail accompanied by an introduction letter and an informed consent that had to be returned. The patient certified having filled in the VHI-30 personally, with no assistance whatsoever. Furthermore, patients were asked to disclose possible further voice surgeries, new disease or conditions that could affect their voice quality. It was also possible for relatives to communicate about the death of the patient by returning the mail.
In case of centre 2, patients were invited by phone, to come to the hospital to fill in the VHI-30 along with an informed consent.
Questions about possible further voice-related surgeries or conditions that could affect their voice quality were checked by phone.
Patients were isolated alone in a room to personally fill in the VHI-30 questionnaire.
Deaths reported by relatives and patients presenting with post-MTIS surgeries, diseases or conditions that could affect voice quality, were excluded and not considered for the study.
All 3 centres entered their data in one Excel data bank. The data included anonymised identification of the patient, date of birth, gender, side and size of the prosthesis, date of the MTIS procedure, date and value of preoperative VHI-30, date and value of early postoperative VHI-30 and date and value of very long-term postoperative VHI-30. Inclusion and exclusion data were also checked for each centre.
Once data were obtained, very long-term (>2 years) postoperative data were compared with the previously acquired preoperative and early postoperative VHI results. Influence of factors: age, gender, size/side of the prosthesis, centre where surgery has Keypoints • Montgomery thyroplasty offers good results over a long period of time.
• Gender, age and prostheses size/side do not influence very long term results.
• Montgomery thyroplasty is a good standard of care when long term results are sought.
been performed, and length of VHI-30 follow-up was also analysed.
Continuous variables were summarised using medians, interquartile ranges and ranges. To analyse the evolution of VHI in time (repeated measurements: preoperative, early and very long-term postoperative), Friedman's rank sum test, a nonparametric alternative to the parametric test one-way repeated measures ANOVA, was used. A pairwise comparison using Nemenyi multiple comparison test was then applied to identify significant differences between time period groups. Variables associated with the decrease in VHI over time were assessed using random effects-expectation maximisation (RE-EM) tree, a mixed effects model for longitudinal data, to take into account repeated measurements. Time period and patient identifier were inserted as random effects to take into account repeated measures for each patient. The centre was assessed as a fixed effect in the model because of the low number of categories (3 centres).

| RESULTS
Eighty-two (82) patients were considered for the study, as they had undergone a thyroplasty using MTIS for UVFP, as a sole procedure, that had been performed at least 2 years earlier. The number of deceased patients according to the data available in their EMR was 13, 1 and 0 for centres 1, 2 and 3, respectively. In none of the cases, the cause of death was related to the MTIS. The EMR review revealed one patient that underwent a complimentary injection laryngoplasty (IL) after MTIS because of unsatisfactory voice result. This additional procedure was performed within 2 years postoperatively. Therefore, in total, 15 of the initial 82 patients were excluded from the study.
The 67 remaining patients were asked to participate in the study.
The VHI-30 score along with an accompanying letter and an informed consent form, to be returned, was sent to 33 patients from centre 1 and 16 patients from centre 3. Eighteen patients from centre 2 were invited by phone to fill in the VHI-30 questionnaire at the hospital.
Relatives informed authors of the death of 2 patients from centre 1. One patient from centre 1 informed authors about the occurrence of a voice impairing disease (the recurrence of metastatic thyroid papillary carcinoma for which he was treated with external radiotherapy). These patients were categorised as non-responders and excluded from the study. Finally, 49 patients were included in the study (24 from centre 1, 16 from centre 2 and 9 from centre 3).    Four variables, possibly influencing these results, were analysed through the RE-EM tree technique. Results for age differences, gender differences and size/side differences of the prosthesis are displayed in Figure 2. None of these variables showed a significant effect on previously mentioned results.
A possible "centre" effect was also statistically ruled-out.  Finally, as centres 1 and 2 had a response rate above the 80%, and centre 3 a response rate of 56%, this difference could also represent a possible bias. We do not have an explanation for this difference in responder's rate except for the fact that some of this centre's patients had been operated by a surgeon who left the institution. Some patients of centre 3 might have been less motivated by a request for participation signed by an unknown surgeon. Nevertheless, despite this caveat, the overall response rate (78%) was 10% higher than the response rate of Hogikyan et al. (68%) 12 and was thus considered valid for interpretation.

| CONCLUSIONS
This study shows that MT performed with the MTIS offers permanency of voice improvement from the patient's perspective. Age, gender, side of procedure and size of MTIS implants do not influence very long-term results in terms of VHI-30. Therefore, MT overall and MTIS, in particular, should be considered as a possible standard of care for UVFP when permanency of voice results is sought.