JRP is defined as recurrent parotid inflammation of a non-obstructive, non-suppurative nature in a child aged 1 to 16 years. It presents as unilateral or bilateral parotid gland inflammation with 2 or more episodes occurring before puberty. It is an uncommon condition and its etiology, which is likely to be multifactorial, is unknown. Historically, JRP has been described in association with Sjögren's syndrome, hypogammaglobulinemia, IgG3 deficiency, IgA deficiency and as a frequent manifestation in patients with HIV infection. At present, JRP can be considered a sentinel sign of other diseases of immunologic/autoimmune etiology whose early diagnosis, follow-up and treatment can improve prognosis [5, 40, 41].

Originally, JRP was attributed to congenital dilatations and malformations and/or recurrent infections [42], but nowadays a multifactorial approach to etiology is more accepted. Genetically, JRP presents an autosomal dominant pattern with incomplete penetrance and variable expressions [43]. The higher concentrations of Streptococcus pneumoniae and Haemophilus influenzae isolated in the saliva of JRP patients may support an infectious etiology [31, 44]. Currently, the main cause postulated to explain its pathogenesis is decreased salivary production with insufficient salivary flow through the ductal system, which favors ascending infections of the salivary glands through the oral cavity [45].

The diagnosis of JRP is based on the clinical picture and can be confirmed by ultrasonography. In this diagnostic procedure, typical findings are: enlarged and heterogeneous parotid gland, with unilateral or bilateral involvement, and hypoechogenic areas of 2–4 mm and/or hyperemia on Doppler ultrasound indicative of sialectasia or lymphocytic infiltration [46]. The noninvasive nature of ultrasound makes it an ideal imaging modality for children. Expected ultrasound findings in JRP include scattered hypoechoic foci (referred to as "Swiss cheese" or "moth-eaten") [31].

Our demographic results are consistent with previous reports of patients with JRP. JRP most frequently affected male children. The ages of JRP onset were bimodally distributed, with a primary peak between the ages of 4 and 8 years [31, 39]. Although the criteria for JRP diagnosis postulated by Garavello et al. [5] included patients younger than 16 years, most of the reviewed papers included patients younger than 18 years.

The main criteria to establish the severity of the disease are the frequency of recurrences, the duration of the event, the severity of inflammatory symptoms and the importance of glandular alterations [3]. None of the papers in the present review specifically contemplates the influence of these variables on the results of sialoendoscopic treatment. In our review and based on the number of episodes of parotid swelling before sialoendoscopy, we observed the existence of a statistical relationship between the number of episodes/year and recurrences. A higher number of swelling episodes in a shorter time presents a higher probability of recurrence after sialoendoscopy. However, the specific number of episodes from which the different authors proceed to perform sialoendoscopy is not clearly defined in many publications. The frequency of these acute episodes is variable and ranges from 2 to > 10 per year. Studies with a more specific design would be necessary to analyze the possible influence of the number of episodes of inflammation and the efficacy of sialoendoscopy. However, in an exploratory manner, the present review points to a possible relationship.

The mean percentage of recurrences observed in our systematic review was 25.1% (95% CI 23.6%-26.6%). This result is similar to that reported in previous studies. In 179 children included in 10 studies, Canzi et al. [38] observed complete evanescence of symptoms after sialendoscopic treatment in 78% of patients and partial regression in 22% of cases. In the review by Ramakrishna et al. [39], based on 7 studies with 120 patients and 165 glands, the primary success rate for interventional sialoendoscopy was 73% (95% CI 64–82). The review by Garavello et al. [5] of 336 children showed that only 25.8% (95% CI 21.5–30.8) of treated children had further recurrences. Nevertheless, there is little information on the number of sialoendoscopies that should be performed to achieve clinical resolution of the pathology. Some authors reported that even one sialendoscopic session may be sufficient to cure the patient [11, 15, 36], while others observed an improvement, and not a cure. In this regard, the serie of 17 patients by Benaim et al. [31], provides interesting information. In their study, the success rate after the first sialoendoscopy was 47.1%, after the second sialoendoscopy it was 17.6% and after the third it was 11.8%. In short, for these authors, an overall success rate of 76.5% was only achieved after three sialoendoscopies. Generically, more than one sialoendoscopy would be necessary to obtain complete resolution ("cured"). The mean value of repeat procedures observed in the present study was 14%, with a range of 0% [24] al 25% [30].

In the present study, the most frequently sialoendoscopic finding was the presence of stenosis, followed by pale ductus, debris and mucous plug. These results are not in agreement with previous studies. In the study by Canzy et al. [38] the most relevant and recognized sialoendoscopic finding was the white appearance of the wall and the lack of vascularity in the ductal layer (mean 75%). Confined or diffuse stenosis and multiple fibrinous debris/mucous plugs were observed in a high percentage of children (mean 56% and 45%, respectively) [7, 9, 25, 38]. Nevertheless, the percentage of dilatations reported by different authors during sialoendoscopy was lower than the recognized percentage of stenosis [1, 9, 11, 30].

Histologically in patients with JRP there are intraductal cystic dilatations of peripheral ducts with periductal lymphocytic infiltration, called as sialectasis. The ecstatic ducts are usually 1-2 mm in diameter and typical have a white appearance of the ductal layer without the healthy blood vessel coverage, when compared with a normal gland [9, 45].

Although an international consensus on the classification of parotid duct stenosis has not yet been achieved, recent publications suggest that stenoses can be classified into up to three groups: inflammatory (type 1); fibrous, associated with circular or web-like intraductal inclusions, often with only a moderate degree of luminal narrowing (type 2); and fibrous, affecting the entire ductal wall, almost always with high-grade to complete obstruction (type 3) [47]. This diagnostic information could be reached indirectly with different imaging methods such as ultrasound or sialography. In adult patients, sialendoscopy is considered to be performed in 33.3% of type 1 stenoses, 52.9% of type 2 stenoses, and 77.1% of type 3 stenoses. Cortisone lavage guided by sialendoscopy was sufficient in 73% of cases of type 1 stenosis. Interventional sialendoscopy with instrumental dilation was successful in more than 47.1% of cases of type 2 and 3 stenosis [48].

In children, at present, ultrasound would be the best imaging option for suspected inflammatory ductal pathology or ductal stenosis; it can provide a diagnosis in the parotid gland in most cases and could contribute to the indication for sialendoscopy and its control efficacy. Recently Goncalves et al. [49] observed that parotid glands with normal sialendoscopic findings had a duct diameter of 0.3 mm (0–2.7 mm) and homogeneous hyperechoic parenchyma on ultrasound in 98.7%. Ductal inflammation/sialodochitis on sialendoscopy had significantly larger ductal diameter of 0.7 mm (0–4.3 mm) and hypoechoic parenchyma in 78.%. Parotid glands with stenosis had hypoechoic parenchyma in 52.6% and a ductal diameter of 4.1 mm (0–19.0 mm). The ductal diameter was ≥ 2.7 mm in 95.6% of stenoses. Nonetheless, to our knowledge, there are no studies using preoperative and postoperative ultrasound to compare the efficacy of sialoendoscopy.

Treatment of JRP in the acute phase is based on a combination of sialogogues, parotid gland massage and antibiotics [5]. In the serie by Schneider et al. [19], thirty-six patients were treated over a period of 79 months, 15 with salivary endoscopy with cortisone irrigation and 21 with antibiotic therapy alone. A significant reduction in recurrent episodes and pain intensity after therapy was observed in both groups. With respect to these two outcomes, the comparison showed two therapeutic options of equal marketability. However, patients with JRP who underwent sialendoscopy had significantly higher costs of care during the observation period compared to those who did not undergo the procedure, with no statistically significant difference in outcomes [19]. Subsequent work, also with a limited number of patients, reported similar results [50].

Interestingly, diagnostic sialography was also found to have a therapeutic effect, which has been attributed to the irrigation effect and potential antibacterial activity of the iodine-based contrast material [2, 51]. However in children, sialoendoscopy would avoid the radiation of sialography.

Different authors have evaluated the efficacy of sialendoscopy associated with lavage for the prevention of recurrence, as well as the efficacy of the various lavage solutions [52]. Lavage seems to break the vicious circle of decreased secretion, stasis, and infection by evacuating mucus plugs and intraductal debris. The optimal lavage solution and dilation site have not yet been defined. In our review, the various intraductal lavage solutions (corticosteroids, antibiotics, or saline) appear to be effective, and no one solution has been shown to be superior to another. Direct lavage through the parotid duct also appears to be effective and remains a treatment option after confirmation of the diagnosis by ultrasound and/or MR sialography [53].

It has recently been published that irrigation of the affected gland with 3–10 ml saline solution without any type of anesthesia is a reasonable, simple, and minimally invasive treatment alternative for JRP. Nonetheless, it is a retrospective study that only included 11 boys (age 3.3–11 years) [51]. For the authors of this study, the effect of mechanical manipulation by introduction and advancement of the endoscope remains unknown. Touching the walls of the inflamed duct with the relatively sharp tip of the endoscope may have no relevance, but could theoretically lead to scar formation. Canzi et al. [38] stated that possible side effects of sialendoscopy were ductal breach (up to 8%), proximal duct stenosis (up to 66%), and upper airway obstruction. In contrast, the intravenous catheter used by these authors was soft and flexible and was only introduced into the most distal part of the duct. Complications related to sialendoscopy were minor, but were reported in 4.4% of the procedures in our review. Some authors reported upper airway obstruction in 0.1% of patients due to parotid inflammation of the pharyngeal portion of the gland. In all cases, these events were self-limited and resolved spontaneously within 24 h [20, 26].

In our review all studies demonstrated the diagnostic value of sialoendoscopy by visualizing stenosis, hypovascularization and intraductal whitish debris. However, these sialoendoscopic findings do not allow to reach a differential diagnosis between JRP and other autoimmune parotitis of childhood such as Sjogren's syndrome [52].

Different authors point out that many children with JRP or persistent salivary gland enlargement of unknown etiology are likely to be diagnosed with Sjogren's syndrome (SS) after appropriate testing. Nevertheless, failure to meet the existing criteria for SS in adults does not exclude the diagnosis of SS. In the case of these children, continued observation with periodic repetition of tests (imaging, serological, functional) is crucial to assess progression to SS, but sialendoscopy has no diagnostic value [54,55,56].

Direct hospital costs one year before and after the sialendoscopy procedure in children were recently collected and analyzed. To estimate the cost of care, we obtained direct hospital costs per clinical encounter (pediatric otolaryngology, emergency room and primary care provider visits), imaging modality, outpatient antibiotic prescriptions, and for the sialendoscopy procedure, anesthesia, and post-anesthesia care unit costs from institutional administrative sources. Costs of the sialendoscopy and related expenses in patients with JRP including anesthesia and post-anesthesia care unit cost were $13,506. Mean total hospital costs were significantly higher in patients with JRP one year before and after the sialendoscopy ($4308.8 vs. $3330) compared to patients with sialolithiasis [27].

Previously, other authors compared the mean costs of care for patients with JRP treated with sialendoscopy compared to those treated conservatively. Mean costs were much higher in the sialendoscopy group ($31,338 per patient vs. $698 per patient), although treatment outcomes did not differ significantly [50]. However, in adults, the costs of sialendoscopy are usually lower than those derived from other more aggressive surgical techniques [57].

The limitations of this study include results based on case series in the absence of a control group and randomization (Level of evidence: 4). Although the number of procedures reviewed was high (646 sialendoscopies), different variables were not collected or were not uniform across the different series consulted. Due to the recurrent nature of JRP, it is possible that some patients experience a recurrence of symptoms and have not yet followed up. In addition, the different studies were often performed by the same teams, which may be a source of bias. In some studies, sialography was performed by sialendoscopy, which could lead to overestimation of the efficacy of sialendoscopy. The mean value of the 27 studies reviewed on the Newcastle–Ottawa scale is low.