Discussion

The aetiology of vocal process granulomas is not completely understood. The generally accepted view is that laryngeal-pharynx reflux plays a significant role in the cause of vocal process granulomas. Reflux is seen in 30 to 76 per cent of cases of vocal process granuloma.^{Reference Havas, Priestley and Lowinger19} Lei et al.^{Reference Lei, Yang, Zhang and Ren14} reported that 27 of 46 (58.7 per cent) patients with vocal process granulomas had acid reflux based on diagnosis of Reflux Symptom Index and Reflux Finding Score. A total of 47 per cent of vocal process granulomas were accompanied with acid reflux in the present study. However, Wani et al.^{Reference Wani and Woodson20} reported that laryngopharyngeal reflux was confirmed in only 24 per cent of vocal process granulomas. Some studies suggest that vocal process granulomas could be related to arytenoid adduction asymmetry.^{Reference Mathew and Menon5} Carroll et al.^{Reference Carroll, Gartner-Schmidt, Statham and Rosen3} reported that 53 per cent of patients had underlying glottal insufficiency. We observed that 5 of 16 (31.2 per cent) cases had underlying glottal insufficiency in this study. Others believe that vocal process granulomas can be associated with unilateral superior laryngeal nerve paresis, potentially related to altered contact points between the vocal processes of the arytenoids.^{Reference Halum, Miller and Early21} We believe that this cause may well explain the formation of vocal process granulomas. A suddenly hyperfunctional voice could result in the abnormal contact of one side of the vocal process and contralateral arytenoid cartilage body for the patients with unilateral superior laryngeal nerve paresis, thereby causing trauma and chronic inflammation of the arytenoid cartilage body and gradual formation of vocal process granulomas. This is the reason why the occurrence of vocal process granulomas in the arytenoid cartilage body was significantly higher than for the vocal process in this study.

Lee et al.^{Reference Lin, Cheng and Su22} explained that laryngeal denervation (paresis or paralysis) resulting in aerodynamic incompetence and supraglottic compensatory activity can also lead to the posterior vocal process granuloma formation, often at the supraglottic level (higher on the arytenoids). If superior laryngeal nerve paresis was recovered and voice rest was maintained, the abnormal contact would be improved and the granuloma would go into spontaneous remission. In addition, similar to another study, the occurrence of vocal process granulomas on the left side was significantly higher than that of the right side.^{Reference Hillel, Lin, Samlan, Starmer, Leahy and Flint16,Reference Lee, Yoon, Lee and Lim23} Idiopathic vocal fold paralysis is more common on the left side compared with the right; similarly, superior laryngeal nerve paresis is more common on the left side compared with the right. This may explain the reason why the occurrence of vocal process granulomas was the most common on the left side.

On the contrary, it is difficult to explain granulomas being caused by acid reflux. Many patients with hiatal hernias and laryngopharyngeal reflux do not have contact granulomas.^{Reference Ward, Zwitman, Hanson and Berci4} Some cases with vocal process granulomas and acid reflux failed to respond to proton pump inhibitor treatment. In our study, only 50 per cent of cases with vocal process granulomas and laryngopharyngeal reflux achieved complete remission. de Lima Pontes et al.^{Reference de Lima Pontes, De Biase and Gadelha24} reported a success rate of 75 per cent using anti-reflux treatment for only those patients with proven gastroesophageal reflux disease. Although some patients had a good response for proton pump inhibitors, it was difficult to explain whether the effect was because of proton pump inhibitor treatment or other treatments. Other studies reported that a few patients with vocal process granulomas went into spontaneous complete remission after simple observation.^{Reference Wang, Ko, Wang, Hu and Hsiao6,Reference Lee, Hong, Choi, Sun, Park and Lee10} Lee et al.^{Reference Lee, Hong, Choi, Sun, Park and Lee10} reported that 20.5 per cent of patients had spontaneous remission without any therapy. One explanation is that acid reflux should theoretically result in equal opportunity of bilateral sides or right dominance (most people sleep supine on the right). Nevertheless, the exact mechanisms of vocal process granulomas remain to be further studied.

A reliable system for grading vocal process granulomas is lacking at present. Although Farwell et al.^{Reference Farwell, Belafsky and Rees17} presented an endoscopic grading system, it was only suitable for the assessment of granuloma size. The grading system was based on the degree of granuloma extending past the vocal process and the midline of the airway but ignored the granuloma extending to posterior arytenoid cartilage body or corniculate cartilage. The granulomas were classified as vocal process type, arytenoid cartilage body type and associated area type based on their origin in this study, and thereby the relation between proton pump inhibitor treatment effect and origin was investigated. In addition, the granulomas were classified as medial type, surface type and mixed type based on whether the granulomas were on the surface or margin of vocal process and arytenoid cartilage body. Theoretically, the marginal granuloma could be related to hyperfunctional voice use and abuse, which would improve by voice rest or correction of voice use. Granulomas of the arytenoid cartilage body and superficial granulomas of the vocal process could be related to motion abnormality of arytenoid cartilage. Thus, these classifications were useful for evaluating the efficacy of proton pump inhibitors and voice therapy on the idiopathic vocal process granulomas. In addition, granulomas were subjectively classified as ulceration type, cyst type and granulation type based on the endoscopic appearance. This classification may be used for preliminary evaluation of granulation morphology and efficacy; however, further classification remains to be further studied.

Proton pump inhibitor therapy for 2–4 months was recommended by most scholars.^{Reference Emami, Morrison, Rammage and Bosch25} Large dosage and prolonged period of treatment may result in osteoporosis.^{Reference Chou, Jiang, Chen, Ho and Lee26–Reference Park, Lee, Yu, Jung, Han and Kim28} The treatment of 10–20 weeks in this study was consistent with the suggestion of most studies; thus, its efficacy may be compared with the literature. The complete remission rate was 31.25 per cent following proton pump inhibitor treatment in this study, similar to other studies, which reported a complete remission rate of 41.3–48.6 per cent.^{Reference Lemos, Sennes, Imamura and Tsuji1,Reference Lee, Hong, Choi, Sun, Park and Lee10,Reference Ma, Xiao, Ye, Yang and Wang15} However, the complete remission rate found by Hillel et al.^{Reference Hillel, Lin, Samlan, Starmer, Leahy and Flint16} was 69 per cent and by Wani et al.^{Reference Wani and Woodson20} was 77.8 per cent. Although Lei et al.^{Reference Lei, Yang, Zhang and Ren14} reported that esomeprazole plus mosapride citrate were used and achieved the complete remission rate of 95 per cent, the reliability of this study is questionable.

In this study, complete remission was defined as no recurrence after three months’ follow up after the resolution of granuloma. Although the follow-up time of three months is short, it is well beyond the typically reported recurrence window of six weeks.^{Reference Emami, Morrison, Rammage and Bosch25} We evaluated the correlation between complete remission of granulomas and related factors. Although multiple regression analysis of the related factors was not performed because of the small sample size, we determined the factors that influence complete remission. This initial study showed that the complete remission rate of vocal process granulomas was 31.25 per cent and was only significantly related to the position and size of vocal process granulomas but was not significantly associated with other factors. The complete remission rate of vocal process granulomas of medial type was significantly higher compared with the surface type.

Pham et al.^{Reference Pham, Campbell, Mattioni and Sataloff11} reported that all the granulomas were located on the surface of the arytenoid cartilage body for two patients with recurrence of the granuloma following intralesional steroid injection and surgical excision. Hamdan et al.^{Reference Hamdan, Khalifee, Jaffal and Ghanem29} and Fink et al.^{Reference Fink, Achkar, Franco and Song30} each reported that eight patients with granuloma resistant to anti-reflux therapy were also located on the surface of the arytenoid cartilage body. In addition, the complete remission the rate of vocal process granulomas of grade I was significantly higher compared with grades III and IV. This finding may be easily understood. Similar to a vocal cord polyp, some polyps of the margin of the vocal fold may completely disappear after voice rest, whereas the polyps on the surface of vocal fold usually require surgical excision. We speculate that the aetiology and pathogenesis of vocal process granulomas of medial type and surface type could be different. The drawback of this study was a small sample size and the lack of 24-hour pH monitoring. In addition, all the classifications of position, location and morphology were based on endoscopic observation by naked eye.