and denervation duration using multivariate logistic
regression analysis method in a large series of 349
UVFP patients who underwent delayed laryngeal rein-
nervation. We also performed further stratification anal-
ysis aimed at one of the identified significant variables—
denervation duration—in order to investigate the effect
of denervation duration on the surgical outcome of
laryngeal reinnervation.
MATERIALS AND METHODS
Patient Characteristics
Our study was approved by the institutional review board
of Second Military Medical University, Shanghai, People’s
Republic of China. The medical records of 349 UVFP patients
(94 males and 255 females; mean age 44.0 years, ranging 17–69
years) who underwent anastomosis of the main branch of the
ansa cervicalis to the RLN between January 1996 and January
2011, and who were followed for at least 2 years, were reviewed.
The etiology of UVFP in this series of patients was RLN injury
during thyroid surgery. Informed consent was obtained from all
patients involved in this study. Patients who were lost to follow-
up were excluded. There was a minimum waiting period of 6
months following onset of RLN injury to allow for possible spon-
taneous recovery or compensation. The median denervation
course was 16.1 months (range, 6–45 months). When stratified
by denervation duration, the number of patients in each dener-
vation duration group was: 172 (49.3%) patients with a dener-
vation duration 6 to 12 months (group A); 108 (30.9%) patients
with a denervation duration 12 to 24 months (group B), and 69
(19.8%) patients with a denervation duration
>
24 months
(group C). The median follow-up period after laryngeal reinner-
vation was 70.8 months (range, 24–156 months).
Surgical Procedure
The surgical procedure has been elaborated in our previ-
ous report.
4
Briefly, under general anesthesia, the ipsilateral
ansa cervicalis was explored, and the main branch was trans-
ected at the bifurcation and freely mobilized for preparation of
anastomosis. The RLN was dissected at a point sufficiently far
from the injury site to provide a tension-free anastomosis and
then transected. Under an operating microscope, the distal
RLN stump was anastomosed to the main branch of the ansa
cervicalis using nylon 11-0 thread in three to five epineural
sutures.
Videostroboscopy
All patients were observed via a videostroboscope (RICH-
ARD WOLF GmbH, model 5570, Knittlingen, Germany) during
“eee” phonation at a comfortable loudness and pitch for as long
as possible, and dynamic videos were recorded preoperatively
and postoperatively. Three experienced laryngologists who had
not performed any of the surgeries reviewed all of the videos.
The videos were randomized, and the reviewers were blinded to
whether the videos were preoperative or postoperative. Visual
laryngeal analysis included glottal closure (0, complete; 1,
slightly incomplete; 2, moderately incomplete; 3, severely
incomplete), vocal fold position, vocal fold edge of paralyzed
side, phrase symmetry, and regularity. Consensus of the
reviewers was reached on the visual appearance of the larynx.
Our previous studies demonstrated that the above parameters
were consistent in presenting reinnervation outcome of vocal
fold paralysis, among which the parameter glottal closure was
the most representative one
4
; therefore, only the parameter
“glottal closure” was included when performing statistical anal-
ysis using a univariable analysis and multivariable logistic
regression analysis.
Vocal Function Assessment
Vocal function assessment included perceptual evaluation,
acoustic analysis, and maximum phonation time (MPT) mea-
surement. Preoperative and postoperative voice samples con-
taining sustained vowels /a/ and connected speech samples were
used for perceptual evaluation and acoustic analysis. The
recording equipment consisted of a digital audiotape recorder
and a dynamic microphone (Tiger Electronics Inc., North Read-
ing, MA). Five laryngologists who had been trained in grade,
roughness, breathiness, asthenia, and strain (GRBAS) rating
performed voice perceptual evaluation using a perceptual rating
scale (GRBAS) for voice quality and characteristics. The ratings
were accomplished in a blinded fashion, with patient voice sam-
ples arranged in a random manner. Each listener was asked to
score connected speech samples for overall grade, roughness,
breathiness, asthenia, and strain using a voice-quality scale for
each parameter (0, normal; 1, mild; 2, moderate; 3, severe). The
values were averaged among the five listeners. Our previous
studies demonstrated that the interrater and intrarater reliabil-
ity was acceptable (interrater reliability
>
0.76; intrarater
reliability
>
0.81).
4,9
In addition, the above five parameters of
perceptual evaluation were consistent in presenting vocal out-
come of vocal fold paralysis, among which the parameter overall
grade was the most representative one.
4
Therefore, only the
parameter “overall grade” was included when performing statis-
tical analysis using a univariable analysis and multivariable
logistic regression analysis.
The acoustic parameters of sustained vowel /a/ were eval-
uated using Praat software (Boersma, Paul & Weenink, David
(2011). Praat: doing phonetics by computer [Computer pro-
gram]. Version 5.1.12, retrieved from
http://www.praat.org/).
The acoustic parameters were mean noise-to-harmonics ratio
(NHR) and measures of phonatory stability—jitter (local) and
shimmer (local). MPT was defined as the duration of sustained
phonation of the vowel /a/ after maximum inspiration and was
measured preoperatively and postoperatively.
4
Laryngeal Electromyography
A four-channel electromyograph and concentric needle
electrodes (Dantec Counterpoint, Copenhagen, Denmark) were
used for the laryngeal electromyography (EMG) recordings. To
test for proper needle position, the unaffected vocal fold was
examined first. The electromyographic activity of the bilateral
thyroarytenoid (TA) muscles was recorded during the following
two stages: while breathing quietly when relaxed, and while
pronouncing the vowel /eee/ with the greatest exertion, then
sniff. One board-certified otolaryngologist performed the EMG,
and a neurologist operated the EMG machine and interpreted
the EMG results. The neurologist rated the VMUR using the
following scale: 0, full interference; 1, mixed interference; 2,
simple interference; and 3, without motor unit potential.
4
Statistical Analysis
The perceptual evaluation, acoustic analysis, and MPT
data did not follow normal distribution and were presented as
median (low quartile, upper quartile). We sought to evaluate
influencing factors for the surgical outcome of laryngeal rein-
nervation using multivariable logistic regression methods.
Potential influencing factors were examined in univariable
Laryngoscope 124: August 2014
Li et al.: Denervated Duration on Reinnervation for UVFP
6