Discussion

Two of the most commonly performed surgical procedures to correct glottic incompetence are medialisation thyroplasty and arytenoid adduction.^{Reference Woo10} Medialisation thyroplasty with Silastic®, which principally improves glottic closure by altering the membranous glottis, is currently the commonest surgical modality available for patients with unilateral vocal fold paralysis.^{Reference Wong, Smith, Stone, Palme, Smith and Riffat11} However, this procedure plays no role in repositioning the malpositioned arytenoid to its normal physiological position, which is a basic requisite for normal voice production. This was cited as the primary cause for this procedure yielding an unsatisfactory result in numerous cases, when performed alone.^{Reference Blitzer, Zeitels, Netterville, Meyer, Smith, Blitzer, Brin and Ramig6}

Arytenoid adduction entails the surgically simulated pull of the lateral cricoarytenoid muscle by strategically placed and anchored sutures, thereby repositioning the arytenoid on the side of paralysis.^{Reference Isshiki, Tanabe and Sawada8} The decision regarding when to perform an arytenoid adduction remains highly contentious. Certain otolaryngologists consider it to be a routine part of medialisation surgery, while others believe in performing it for its classically accepted indications only.^{Reference Rubinstein, Courey, Amin and Johns7} However, the common consensus in current practice advocates that arytenoid adduction be considered as an adjunct to medialisation thyroplasty when suboptimal voice is achieved with implant placement alone.^{Reference Daniero, Garrett and Francis9} Surgical intervention is hence preferably performed under local anaesthesia (sedoanalgesia) in order to facilitate this on-table decision and to enable minor adjustments or fine tuning of the voice for an optimal result via feedback.^{Reference Isshiki, Tanabe and Sawada8,Reference McCulloch, Hoffman, Andrews and Karnell12}

The disadvantages of arytenoid adduction include its technically challenging nature, its significant learning curve due to surgical complexity, higher complication rates and prolonged operation time with consequent patient discomfort when combined with a type 1 thyroplasty under local anaesthesia.^{Reference Rubinstein, Courey, Amin and Johns7,Reference McCulloch, Hoffman, Andrews and Karnell12,Reference Chhetri, Gerratt, Kreiman and Berke13} One case with injury to the mucosa of the pyriform fossa was documented in our study (in group 2), and none of the cases required revision surgery. Rosen observed a higher complication rate in surgeons who performed fewer than two procedures per year or had an experience of 10 total cases.^{Reference Rosen14} The success of laryngeal framework surgery is supremely dependant on the experience and skill of the surgeon as it exemplifies the amalgamation of both art and science.^{Reference Daniero, Garrett and Francis9} Cases in our study were all performed by the same chief surgeon (first author) with experience and expertise in laryngeal framework surgical procedures, spanning over two decades. Further, the entire surgical team comprised laryngologists with additional training in phoniatrics, who were able to assist in accurate intra-operative assessment and appropriate recommendations.

In cases of unilateral vocal fold paralysis, the maximum phonation time is reduced to less than 10 seconds and certain literature has stated that a pre-operative maximum phonation time of less than 5 seconds indicates severe uncompensated unilateral vocal fold paralysis.^{Reference Simpson, Cheung, Sulica and Blitzer15} Isshiki has stated that maximum phonation time can be used intra-operatively and after the medialisation procedure as it is roughly in inverse relation to the mean flow rate during phonation.^{Reference Isshiki16} The grade, roughness, breathiness, asthenia, strain scale is a simple, reliable and established method of perceptual voice assessment.^{Reference Nemr, Simoes-Zenari, Cordeiro, Tsuji, Ogawa and Ubrig17} However, intra-operative phonation is affected by factors such as the level of sedation, position and oedema, which makes on-table phonatory feedback unreliable^{Reference Rubinstein, Courey, Amin and Johns7,Reference Lau, Loizou, Palme and Riffat18} (hence the relevance of this study in identifying an independent pre-operative determinant for arytenoid adduction).

In order to help improve functional and anatomical optimisation, intra-operative flexible laryngoscopic examination has been advocated to confirm vocal fold alignment,^{Reference Lau, Loizou, Palme and Riffat18} as was performed in our institution. In our study, both groups showed a significant improvement in maximum phonation time and grade, roughness, breathiness, asthenia, strain scale score following surgical intervention as outcome measures for a satisfactory result that was comparable with literature.^{Reference McCulloch, Hoffman, Andrews and Karnell12,Reference Woodson and Murry19} As was seen in our study, Daniero et al. inferred that the evidence-based finding of equivalence of post-intervention maximum phonation time and grade, roughness, breathiness, asthenia, strain scale score in both groups demonstrated the usefulness of arytenoid adduction in patients who may otherwise have had a suboptimal result with medialisation alone.^{Reference Daniero, Garrett and Francis9}

In our study, 47 per cent of cases required only type 1 thyroplasty, whereas 53 per cent required combined type 1 thyroplasty and arytenoid adduction. This is a greater incidence than seen in literature wherein only about 30 per cent of type 1 thyroplasty procedures also needed arytenoid adduction.^{Reference Nerurkar, Menon, Mohan, Young and Judd2,Reference McCulloch, Hoffman, Andrews and Karnell12} This may be rationalised by the fact that in our country, for a myriad of reasons, patients infrequently undergo an injection laryngoplasty within the first few months of onset of unilateral vocal fold paralysis and thereby presumably adopt various abnormal compensatory laryngeal postures to optimise the vocal quality, which in turn adversely affects the pattern of glottic closure as seen at the time of late presentation. Phonosurgery is relatively in its infancy in our country, with the first publication on type 1 thyroplasty as recent as in 1997.^{Reference Phaniendrakumar, Reddy and Das20} As our centre was a tertiary centre with an exclusive laryngology department, patients are most often referred by their primary care practitioners after a prolonged period of conservative or expectant management, which is one of the many reasons for late presentation.

In a first of its kind study that is distinct from currently available literature, our study has objectivised arytenoid asymmetry and quantified the degree of this asymmetry by the ‘arytenoid asymmetry angle’ formed between incongruent glottic and interarytenoid axes which have been defined. In group 1, the mean angle was 21.34 ± 8.54⁰ while in group 2 the mean angle was more than double (i.e. 46.60 ± 8.97⁰), with the difference being highly statistically significant (p < 0.0001).Woodson and Murry noted that if the paralysed vocal fold was not near the midline, the angle between the membranous and cartilaginous segments of the vocal fold decreased from the normal 180⁰, resulting in a posterior gap that could not be closed by contralateral hyperadduction, thereby necessitating arytenoid adduction.^{Reference Woodson and Murry19} An arytenoid asymmetry angle measuring 33.9⁰ has been calculated in this study as the ‘critical angle’ above which arytenoid adduction is indicated.

Although not quantified, a subjective observation of post-operative reduction in arytenoid asymmetry was noted in patients who had undergone arytenoid adduction in our study. This led us to the interpretation that perfect alignment of the glottic and interarytenoid axes was not required for a satisfactory voice outcome to be obtained. In a study by Chhetri et al., no significant improvement in vocal cord symmetry was found in spite of a significant improvement in voice quality following arytenoid adduction, and this observation was attributed to the limited motion of the paralysed vocal fold due to the adduction sutures.^{Reference Chhetri, Gerratt, Kreiman and Berke13} The fact that 47 per cent of our patients with unilateral vocal fold paralysis who required only type 1 thyroplasty also had a certain degree of arytenoid asymmetry (5⁰ to 33.8⁰) coupled with the incidental finding of arytenoid asymmetry by Lindestad et al. in a large proportion (70 per cent) of the non-dysphonic or normophonic population with normal vocal fold mobility, ^{Reference Lindestad, Hertegård and Björck21} further supported our interpretation. However, studies quantifying the arytenoid asymmetry angle post-operatively in patients who have undergone arytenoid adduction for unilateral vocal fold paralysis and also in normophonic individuals with normal vocal fold mobility are required for accurate analysis and better understanding.

A wide posterior phonatory gap is an indication for arytenoid adduction.^{Reference Isshiki, Tanabe and Sawada8} In our study, the mean posterior phonatory gap was significantly greater by 1.7 times in group 2 than in group 1 (p = 0.001). In corroboration, a significant positive linear correlation was observed between the arytenoid asymmetry angle and the width of the posterior glottic gap.

A vertical level difference between the two vocal folds was observed in 32.9 per cent of our cases. On group analysis, a vertical level difference was seen in 17.5 per cent of cases in group 1 and 46.7 per cent of cases in group 2. This difference was statistically significant (p = 0.004), thereby concurring with the literature which accepts vertical level differences between the two vocal folds as an indication for arytenoid adduction.^{Reference Isshiki, Tanabe and Sawada8}

Identification of a level difference between the two vocal folds is subjective and is also affected by perceptual differences in images obtained during a bird's eye view of the larynx during fibre-optic laryngoscopy.^{Reference Wong, Smith, Stone, Palme, Smith and Riffat11} This may explain the low incidence of this finding among those who required arytenoid adduction and also the presence of this finding among those who did not require arytenoid adduction in our study. In the latter scenario, it has also been suggested that absolute perfect approximation with respect to the vertical plane may not be essential for normal voice production and that the opposite cord may compensate to some degree for vertical misalignment.^{Reference Woodson, Sulica and Blitzer22} In support of this view, Wong et al., who compared arytenoid vertical height discrepancy between normal individuals and those with unilateral vocal fold paralysis on high resolution computerised tomogram of the larynx, concluded that arytenoid vertical height discrepancy was less than or equal to 2 mm in normal individuals and a mean of 2.39 mm in patients with unilateral vocal fold paralysis.^{Reference Wong, Smith, Tjahjono, Stone, Sritharan and Palme23} A significantly greater arytenoid asymmetry angle was observed among those with vertical level differences in our study, compared with those in whom it was absent (p = 0.007) thereby further corroborating the correlation between arytenoid asymmetry and the need for arytenoid adduction.

From our study, we therefore inferred, based on the p-value, that arytenoid asymmetry (p < 0.0001) followed by posterior phonatory gap (p = 0.001) and vertical level difference between the vocal folds (p = 0.004), in descending order, are indications or predictive parameters for arytenoid adduction. Arytenoid asymmetry has thus far not been described in literature as an indication for arytenoid adduction, unlike posterior phonatory gap and vertical level differences, which are historically established and accepted indications for arytenoid adduction.^{Reference Isshiki, Tanabe and Sawada8} From this study, we can deduce that an arytenoid asymmetry angle of more than or equal to 33.9⁰ is not only an indication, but also the most predictive pre-operative parameter for arytenoid adduction in unilateral vocal fold paralysis.

Patients in our study who underwent arytenoid adduction, were significantly younger by almost a decade than those who required only type 1 thyroplasty. Bielamowicz et al. stated that only younger individuals with a wide posterior glottic gap were subjected to arytenoid adduction in their study as they feared airway obstruction secondary to over-adduction or oedema in older individuals.^{Reference Bielamowicz, Berke and Gerratt5} McCulloch et al., while studying arytenoid adduction combined with Gore-Tex® medialisation thyroplasty, observed that the mean age in the Gore-Tex plus arytenoid adduction group was younger, which they explained as a likely surgeon bias towards younger patients who would better tolerate the extended length of the combined procedure and in whom the best possible long-term voice result was hoped for.^{Reference McCulloch, Hoffman, Andrews and Karnell12} However, the presence of a greater degree of arytenoid asymmetry in younger individuals in our study indicates a true need for arytenoid adduction in younger individuals and not surgeon bias as was presumed in literature. A greater arytenoid asymmetry angle among younger individuals may be explained by the hypothesis that greater vocal demand in younger individuals brings about greater compensatory mechanisms and hence greater asymmetry in cases with unilateral vocal fold paralysis.

McCulloch et al. observed a male:female ratio of 1:0.9 in the medialisation thyroplasty group and 1:1.1 in the medialisation thyroplasty plus arytenoid adduction group,^{Reference McCulloch, Hoffman, Andrews and Karnell12} which concurred with the finding in our study that the number of females was greater in group 2 than in group 1 (1.05:1 vs 1.35:1). Similarly, Niimi et al. who studied 60 patients undergoing arytenoid adduction, observed a male:female ratio of 1.06:1.^{Reference Niimi24}

Niimi et al. also noted that the interval from onset of paralysis to arytenoid adduction was on average 32.5 months, ranging from 3 months to 19 years.^{Reference Niimi24} In our study, the duration since onset of unilateral vocal fold paralysis was found to be slightly longer in group 2 than in group 1 (19.98 ± 17.04 months vs 16.28 ± 13.21 months), with a positive linear correlation being identified between the arytenoid asymmetry angle and the duration since onset of unilateral vocal fold paralysis. However, these results were statistically insignificant. This observation can be similarly explained by the fact that a longer duration of unilateral vocal fold paralysis leads to greater compensatory mechanisms and hence a greater degree of arytenoid asymmetry.

A RLN injury is the most common traumatic neuro-laryngological lesion according to literature.^{Reference Maamary, Cole, Darveniza, Pemberton, Brake and Tisch25} In our study, although no statistically significant inter-group difference was identified with respect to aetiology, idiopathic cause (50 per cent) in group 1 and surgical trauma (55.6 per cent) in group 2 were the most commonly identified causes. Thyroidectomy and skull base surgery, accounting for 48 per cent and 28 per cent, respectively, of cases in the surgical trauma sub-classification in group 2, were seen most frequently. The mean arytenoid asymmetry angle was greatest among those with unilateral vocal fold paralysis caused by surgical trauma, followed by idiopathic cause. However, there was no statistically significant correlation between the aetiology and the degree of arytenoid asymmetry. Similarly, in the study by Niimi et al., surgical trauma (68 per cent) was the most common cause for unilateral vocal fold paralysis among those undergoing arytenoid adduction.^{Reference Niimi24} Comparable results were noted in studies by McCulloch et al. and Al-Khtoum et al.^{Reference McCulloch, Hoffman, Andrews and Karnell12,Reference Al-Khtoum, Shawakfeh, Al-Safadi, Al-Momani and Hamasha26} Maamary et al. found that RLN lesions were more commonly (not in all cases) associated with a laterally displaced paralysed vocal fold.^{Reference Maamary, Cole, Darveniza, Pemberton, Brake and Tisch25} In contrast, Woodson, in her study comparing the glottic configuration in patients with unilateral vocal fold paralysis due to lesions of either the RLN or the vagus nerve, concluded that vagus nerve lesions had a statistically insignificant tendency for a more lateral vocal fold position.^{Reference Woodson27} Similarly, McCulloch et al. found that more than 50 per cent of those requiring medialisation thyroplasty plus arytenoid adduction had surgically induced nerve injury at the level of the skull base (i.e. high vagal lesion) demonstrating a wide posterior chink with a vertical level mismatch (owing to the position of the arytenoid on its arc near its posterior superior limit because of lack of pull from the cricothyroid muscle) and were thus better treated with a combined approach.^{Reference McCulloch, Hoffman, Andrews and Karnell12}

In spite of a positive correlation between factors such as the arytenoid asymmetry angle and the type of surgical intervention required, with variables such as gender, duration and aetiology in unilateral vocal fold paralysis, the lack of a statistically significant result may be attributed to a relatively small sample size in this study. Meta-analysis is required for further subgroup analysis to ascertain the significance of the relationship between arytenoid asymmetry and other variables such as gender, duration since onset of unilateral vocal fold paralysis and aetiology for unilateral vocal fold paralysis.