DISCUSSION
This study describes the population characteristics and the results of all sialendoscopies (272) that were performed in our department between March 2008 until and June 2020. Furthermore, we evaluated the long-term outcome for the 146 sialendoscopies performed in 116 patients who returned their questionnaire. Median time between the intervention and completing the questionnaire was 37 months.
Regarding the total of 272 sialendoscopies performed, stone removal was successful in 111 out of 138 lithiasis patients (80.4%). This is in the range of successful endoscopic stone removal in the literature (70-90%).2 All but two retrieval failures were in fixed stones (L2 or greater; stones larger than the ductal diameter). Indeed, in our cohort, in line with the literature, L2b and L3b classification stones implied lower rates of successful retrieval and higher rates of recurrence.15-18 The mean diameter of the submandibular duct is 1.5 mm and the diameter of the parotid duct is 1.2 to 1.4 mm.13 When there is a stone, usually the duct gets dilated proximally of the obstruction. A stone that moves more proximally in the duct, can become a “floating stone”, even if the stone itself is e.g. 2 or 3 mm in diameter. Those are the stones that can be trapped in a basket and pulled toward the distally located, smaller diameter. It is at this point that an incision needs to be made to deliver the stone. Preferably this is at the level of the papilla, but if the basket-trapped stone gets stuck more proximally, a floor of mouth incision for submandibular stones, and a semilunar incision in the cheek or an external approach to the parotid may be needed. This concept explains that stones up to 4 mm in the submandibular duct and up to 3 mm in the parotid duct, (measured perpendicular to the duct) can be successfully extracted with baskets, as reported by Marchal et al17 and confirmed by Walvekar et al.18 For larger stones, fragmentation of the stone is necessary before basket retrieval becomes an option.17
We observed that, in the majority of patients in the lithiasis group (55.1%), sialendoscopy alone was not sufficient, and a combined approach was needed to retrieve the stone. In non-lithiasis pathologies, a combined approach was needed in only 15.7% (p<0.01). Gillespie et al10 found similar results, with, likewise, a combined approach rate needed in the majority of patients in their lithiasis group (83%). Conversely, the majority of their patients with non-lithiasis pathology (77%) could be helped by sialendoscopy alone, and we found similar results (82.8%).10
Regarding the results of the questionnaire, symptom improvement was found in 83.6% and salivary glands could be preserved in 89%. This is comparable with results described in the literature. The study of Gillespie et al10 reported symptom improvement in 89% and retention of the salivary gland in 92%.
Another study on 1,154 patients with lithiasis reported that 82% of them had no residual symptoms and salivary glands could be preserved in 96%.19 These results are similar to our findings in the lithiasis subgroup where we found no residual symptoms in 83.8% and preservation of the salivary gland in 85.3%. A systematic review showed success rates of 86% for sialendoscopy alone and 93% when combined with a minimally invasive surgical approach in case of obstructive disease.20
In case of stenosis and RPC, the rates of residual complaints were higher with respectively 44.6% and 81.3%. However, patients in the RPC group all reported an improvement of their symptoms with a median score of 8 on a scale from 0 to 10. When comparing these results with the literature, we see a higher rate of recurrence compared to previous studies on the effectiveness of sialendoscopy in RPC cases. The systematic review of Garavello et al21 reported a common rate of recurrence of only 25.8%, however none of the 19 studies were randomized, and most studies were case series or case reports with short follow-up. Even more, the definitions used for RPC varied widely, with some studies including children with only one or two episodes, causing an overestimation of the real success rate. We only proceeded to sialendoscopy in children with at least 6 consecutive episodes with intervals of less than 2 months. A small recent study on 33 patients with non-lithiasis related sialadenitis showed that only 59.3% was symptom-free after sialendoscopy.22 As previously reported by Gillespie et al10, the higher rate of clearance of symptoms in the lithiasis group can be explained by the impression that an obstruction that can be removed is more likely to lead to a better outcome compared to an obstruction due to inflammation or scar tissue, that can be improved but not completely resolved.10 This conclusion was also made by Aubin-Pouliot et al9 who found better outcomes for the lithiasis group as well, as compared to non-lithiasis related obstructive disease.
Sialendoscopies of the parotid gland were associated with a significantly higher percentage of residual symptoms compared to the submandibular gland (p= 0.012). This can probably be explained by the fact that stenosis and RPC was more often observed in the parotid gland than the prognostically better lithiasis-related pathologies.
Like also observed by Gillespie et al10, we found that bilateral sialendoscopic interventions were associated with a significant higher rate of residual symptoms on univariate analysis (p<0.01). This is most likely due to the underlying pathology (inflammatory causes of chronic sialadenitis), which is related to a worse outcome.10
When we compared the recurrence of symptoms at the first post-operative check-up and the long-term follow-up with the questionnaire, we notice a progressive and significant (p<0.01) reappearance of complaints at long-term follow-up. This suggests that being initially symptom-free does not necessarily predict a favorable long-term outcome. Gillespie et al10 compared follow-up time for less than their median 17 months with a time greater than 17 months and did not notice a difference in symptom deterioration with longer follow-up time. However, they did not compare the long-term outcome with the first post-operative results, like we did in this study.
When comparing demographic features, pathology distribution and early postoperative results among the responders and non-responders, only for age, a statistical difference was found. Non-responders were significantly younger than responders; an observation that we cannot explain. Nevertheless, given the absence of other differences between the two groups, we believe it is fair to say that the results of the questionnaire are probably representative for the entire patient group.
In general, as compared to external gland removing surgery, sialendoscopy is a safe procedure. In our study complications occurred in only 11 out of the 272 procedures (4%). Nahlieli et al23 found a 3.23% complication rate in a study with 526 patients, treated with interventional sialendoscopy or endoscopy assisted intraoral extraductal interventions. The complications they found were strictures (1.7%), traumatic ranulas (1.7%) and lingual nerve paresthesia (0.72%).23 The rate of complications after traditional surgery is higher. Following parotid or submandibular gland removal, there is a scar and contour deformity. On top of that, according to Capaccio et al24, the most important complications after parotidectomy are temporary (2%-76%) and permanent facial nerve injury (1%-3%), sensory damage resulting from the greater auricular nerve (2%-100%) and Frey syndrome (8%-33%). After removal of the submandibular gland, they reported permanent marginal mandibular nerve injury (1%-8%), hypoglossal nerve injury (3%), and lingual nerve injury (2%).24
The limitations of this study are related to its retrospective nature and the suboptimal response rate. For example, Gillespie et al10 obtained survey data in 67%, compared to 52.5% in our study. Another similar study previously done, showed a lower response rate than the current study (43%).9 The suboptimal response rate is due to the long follow-up time (about 5 years longer than in the study of Gillespie et al10) and therefore difficulties for the patient to recall the intervention (recall bias) and the impossibility to reach patients who were treated several years ago (information bias). This can lead to an over- or underrepresentation of dissatisfied patients. This study doesn’t address the issue of cost-effectiveness of sialendoscopy. In times of savings in health care this aspect must be studied in the future.32