Voice Rehabilitation by Voice Prostheses After Total Laryngectomy: A Systematic Review and Network Meta-Analysis for 11,918 Patients.

Purpose Our aim was to assess the different voice prostheses (VPs) to identify the most efficient, safest, patient-tailored, longest lifetime, and inexpensive VPs and assess the different factors affecting their quality. Method In September 2017, 15 databases were searched to include all randomized controlled trials. A new search was done in May 2019 to include all other study design articles, which include all the new-era VPs subtypes. Network meta-analysis (NMA) was applied to all 27 outcomes, besides NMA overall and partial order setting was done by using Hasse scatter plots. p values were used in NMA, where the best VPs are approaching one and the least approaches zero. Meta-analysis was done for the rest of the outcomes. Results Two hundred one articles were eligible for inclusion in our study (N = 11,918). Provox-2 was significantly the most efficient and safest device concerning the most patient preference (odds ratio [OR] = 33.88 [0.65, 1762.24]; p = .92), the least dislodgement (risk ratio [RR] = 0.27 [0.13, 0.57]; p = .79), the least airflow resistance (RR = 0.42 [0.08, 2.11]; p = .84), the least granulation formation (RR = 0.73 [0.02, 26.32]; p = .60), and the least VPs' inaccurate size (RR = 0.77 (0.23, 2.61); p = .66). Heat and moisture exchanger addition showed a significant increase in maximum phonation time and breathing experience, with p values (1 and .59), respectively. While heat and moisture exchanger addition showed a significant decline in stoma cleaning frequency, coughing frequency, forced expectoration, sputum production, sleeping problems, and loosening of adhesive, with p values (.99, .72, .69, .96, 1, and 0.96), respectively, Groningen low resistance and Nijdam were considered the worst devices with both overall mean p value of .44. Conclusions Provox-2 is considered the best choice as being the most preferable for patients, with the least airflow resistance, dislodgment, granulation formation, and prosthesis inaccurate size. Groningen low resistance and Nijdam were considered the worst devices according to our analysis. Supplemental Material https://doi.org/10.23641/asha.14802903.