In this case series of PBD, ocular manifestations were severe, with comparable phenotypes to what is usually encountered in adult BD patients [1]. All patients had posterior segment involvement, including retinal vasculitis, and required intensive systemic immunosuppressive treatments in association with systemic corticosteroids.

The prevalence of ocular manifestations in both BD and PBD is highly variable. It is reported in 9 to 76% of patients in large PBD series [5, 7, 11,12,13,14,15] and in 18 to 75% of adult BD series [2, 14, 16, 17].

Regarding the ocular phenotypes, the ICBD study reported that retinal vasculitis, posterior uveitis, and anterior uveitis occurred in 23%, 38%, and 40% of adult cases, respectively which seems to be comparable with what is reported in PBD [5, 7, 11, 12]. Papilledema may also be present, as in our cases. The latter can be caused by either inflammation, ischemia, or intracranial hypertension secondary to cerebral venous thrombosis [18,19,20,21]. As in adult BD, various unusual ocular presentations have been reported in PBD such as recurrent neuroretinitis [22] or immune keratitis [23]. In both adult and pediatric BD patients, bilateral panuveitis is likely to be the most frequent presentation, occurring in up to half of the patients [19, 24,25,26]. In both adults and children (as in our series) the macula is frequently involved [19, 24,25,26]. Ocular complications seem comparable in PBD and BD ocular involvement cohorts: cataract, optic atrophy and posterior synechiae are the most frequent [19, 24,25,26], occurring approximatively in a third to half of the patients, followed by rarer complications, such as intraocular pressure elevation, retinal detachment, neovascular glaucoma, phthisis bulbi and band keratopathy (Tables 3 and 4) [19, 24,25,26].

Age at onset of symptoms also varies among the different PBD cohorts. Koné-Paut et al. and Nanthapisal et al.reported a mean age at onset of BD of 4.9-year-old (0.1 to 15.7 y.o.) and 7.4 ± 4.2 y.o., respectively [2, 7]. In these two studies, family history was presentin 17% and 24% of patients, respectively, a higher rate than in adult BD cohorts, suggesting a stronger familial aggregation in children [2, 7, 11, 14]. Female/male ratio seems similar in BD and PBD [2, 5, 7, 9, 11,12,13,14,15,16,17, 13, 14, 16, 27]. Geographical location and ethnic origins may be involved in clinical phenotypes seen in PBD cohorts [2]. For instance, the prevalence of ocular involvement reported in Iran and Egypt(66% and 76% respectively) [5, 14], seems higher than what was reported in Italy, Taiwan, China and UK (44%, 27%, 9% and 9% respectively) [11,12,13, 15]. BD is typically considered as a polygenic disorder. Zhou et al. reported 6 families with early onset auto-inflammatory disease caused by heterozygous mutation in the TNFAIP3 gene coding for A20 proteins with clinical manifestations mimicking BD. However, a mutation in A20 protein was identified in only one patient in a large BD genome-wide association study [28].

As seen in our study, delay of diagnosis may be an important issue in PBD, especially when ocular manifestations reveal the disease, as children may not complain in cases of unilateral and painless ocular involvement. As in other inflammatory diseases, diagnostic delay may impact the severity of clinical presentation. In published literature, diagnostic delays for PBD vary from 2.9 ± 3.6 years [12], to 6.0 ± 3.5 years [7]. However, this long diagnostic delay is also found in adults [16].

In our study, three patients met the PEDBD criteria, while one patient was diagnosed on the combination of typical neurological features (cerebellar peduncle) and retinal vasculitis, both of which are highly suggestive of Behçet's disease [29]. However, three patients met the ICBD criteria designed for adult BD [2]. As in our series, in previously published PBD cohorts, ICBD criteria are constantly more sensitive than PEDBD criteria (71 to 97% versus 36 to 69%, respectively) [5, 12, 13, 15]. ICBD and PEDBD criteria comprise 7 and 6 items, respectively, that are weighted in ICBD classification while they’re not in PEDBD criteria [2, 7]. Thus, the presence of uveitis and oral aphtosis is sufficient to diagnose PBD with ICBD but not with PEDBD criteria. ICBD criteria have been developed from larger cohorts comprising both adult and children originating from various countries, with the underlying hypothesis that adult and pediatric disease are similar [2]. While in most of the cases, oral or genital ulcerations are the first symptoms of the disease, scarce information is available about the referring symptom in neither PBD studies nor BD studies [2, 5, 7, 11,12,13,14,15,16,17, 19, 24,25,26].

In cases of isolated ophthalmic or atypical presentations, other cause of posterior uveitis may be considered in the pediatric context, including infectious diseases, such as toxoplasmosis, toxocarosis, and herpesviridae infections, that can be associated with vitritis and retinal necrosis [30]; inflammatory diseases such as (but not limited to) sarcoidosis or Blau Syndrome that can cause posterior uveitis with retinal vasculitis [31]. Uveitis associated with tubulointerstitial nephritis and uveitis syndrome and juvenile idiopathic arthritis are usually limited to the anterior segment [32, 33].

Contrarily to adult BD, there is no recommendation available to treat ocular manifestation in PBD. As in our series, other teams used European alliance of associations for rheumatology (EULAR) recommendation derived treatment protocols [11,12,13,14, 34]. Intravenous high dose corticosteroids are helpful in the acute phase, but long-term oral steroids are used with caution in children essentially because of their impact on children’s growth [35]. Immunosuppressive therapies are recommended in combination with steroids in posterior uveitis [34], but data on their efficacy in PBD are scarce. Some small case series reported the efficacy of treatments such as methotrexate, cyclosporine, chlorambucil, azathioprine, sulfasalazine, cyclophosphamide or thalidomide in this context [11,12,13, 25, 36]. Regarding biotherapies, anti-TNF α, mainly infliximab, which is recommended in EULAR guidelines for severe cases has been reported in PBD series [11, 12, 30, 37]. Serious side effects with long-term anti-TNF α use in non-infectious pediatric uveitis remains rare, infliximab antibodies is the most frequent issue [30, 38].

Interferon have been reported in PBD case reports or small case series in association with conventional immunosuppressive treatments for severe and relapsing forms, but adverse effects are frequent (flu-like syndrome especially) and may be severe (lymphopenia, neutropenia, depression) [36, 30, 39, 40]. In our experience, in addition to the choice of the drug itself, the most important elements in managing PBD and other severe inflammatory ocular conditions in children are i) a close collaboration between pediatricians and ophthalmologists to manage systemic manifestations, optimize systemic treatment while monitoring side effects and adherence, and ii) to provide educational support to patients and their family, in order for them to be involved in the therapeutic project [41].

As in adults, visual prognosis of ocular involvement in PBD is guarded. In a cohort of 36 PBD patients with 66 eyes affected by uveitis, and a mean follow up of 84.7 ± 91.5, Tugal-Tuktun et al. reported 23% of eyes with a severe and irreversible visual loss (BCVA below 20/200), and 17% of patients with legal blindness (BCVA below 20/200 on the best eye) [25]. Koné-Paut et al. reported a BCVA < 20/200 in 19% of PBD patients with uveitis in at least one eye, and 3% of patients with legal blindness [42]. In recent adult BD cohorts with uveitis, the proportion of patient with at least one eye with BCVA < 20/200 varies between 28 and 30% [43, 44]. In PBD, the most frequent cause of legal blindness.