AB0634 Variability in Phenotype Clusters in Behçet’s Syndrome: A Systematic Review

Abstract

BackgroundThe presence of distinct clinical phenotypes with clustering of certain organ manifestations is well-recognized Behçet’s syndrome (BS). Differences in demographic features, treatment response, and possibly inflammatory pathways involved in the pathogenesis of different phenotypes have been proposed. However, studies from different BS cohorts have shown variability in the phenotypes that were defined.ObjectivesWe aimed to explore the causes of variability in clinical phenotype clustering across different countries and cohorts.MethodsAn electronic search was carried out in PubMed to find articles published in or before November 2021, using the keywords of Behcet, cluster and factor analysis. Two reviewers independently performed a screening of titles, abstracts, and full-texts .ResultsAmongst 496 articles searched, 30 full-texts were assessed, and 10 studies were relevant for data extraction. Ten articles studied 12 different cohorts, 3 from China, 3 from Turkey, 2 from Japan, 1 from South Korea, 1 from Israel, 1 from Greece, and 1 from Italy. 9 out of 10 studies demonstrated clustering of organ manifestations (11 cohorts). There were important differences between the clusters that were identified in these studies (Table 1). Clusters including skin and mucosa manifestations were present in all cohorts, but the skin and mucosa manifestations that clustered together differed from cohort to cohort. Uveitis stood by itself in some studies, whereas it clustered with vascular and central nervous system (CNS) involvement in some cohorts, and certain skin and mucosa lesions in another. Papulopustular lesions (PPL) and arthritis showed a positive correlation in 4 cohorts whereas these manifestations were negatively correlated in 1 cohort. Moreover, no clusters were identified in 1 study. Potential causes of differences in clusters that we have identified in these studies were: study design (database vs multicenter vs single-centre cohort), statistical analysis method (hierarchical cluster vs factor analysis) patient population (pediatric vs adult vs late onset), setting (dermatology vs rheumatology), diagnostic criteria (ISG vs ICBD), disease duration, definition of organ involvement (such as PPL vs folliculitis, or CNS involvement vs dural sinus thrombosis, ascertainment of manifestations (confirmed gastrointestinal involvement vs any diarrhea, lack of ascertainment in diagnosis of nodular lesions as erythema nodosum vs superficial thrombophlebitis), time interval (manifestations throughout the disease course vs manifestations that were active during the last 3 months), and change in the natural history of BS over decades.Table 1.Clinical phenotype clustering across cohortsAuthor, yearnCluster 1Cluster 2Cluster 3Cluster 4Cluster 5Cluster 6Zou,202169MC (G, EN, PPL)U, V, NBSGIZou ,2021860MC (G, EN, PPL)UGIJCVS, NBSZou ,2021152MC (G, EN, PPL)U, V, NBSGIJSoejima, 2021657MC(O, G and Skin) w/o JUNegative correlation of GI &UMC (O, G and Skin inv.), JNBSSoejima, 20216754MC (O, G and Skin) w/ o JU, O, G and Skin inv.Negative correlation of GI &UMC (O, G and Skin), JNBSU, O, SkinKrause 199968MC(G, PPL)Negative correlation of STM and ENGI, PPLJDVT, NBSChung 2021338EN dominantMixt EN and PPLPPL dominantKaraca 2012186MC(G, EN)USTM, DVTJ, PPL, OKaraca 2012221MC(O, G, EN)USTM, DVTJ, PPLTunc, 2002272MC (O, G, EN)USTM DVTJ, PPLSota, 2020396MC (O, G, EN, PF)UNegative correlation of J and PPLArida, 2009142No cluster was foundCVS: Cardiovascular, DVT: deep vein thrombosis; EN: erythema nodosum; G: genital ulcers; GI: gastrointestinal J: joint; MC: mucocutaneous; NBS: Neuro-Behçet syndrome, O: oral ulcers; PF: pseudofolliculitis; PPL: papulopustular lesions; STM: superficial thrombophlebitis; U: uveitis; V: vascularConclusionDifferences in phenotype clusters may result from differences in study characteristics rather than real geographic or ethnic differences. A large multi-national study with uniform inclusion criteria is needed to better understand phenotype clusters and their implication towards management strategies in BS.Disclosure of InterestsBetul Macit: None declared, Kevser Akyuz: None declared, Sinem Nihal Esatoglu Speakers bureau: Dr Esatoglu received honorariums for presentations from UCB Pharma, Roche, Pfizer, and Merck Sharp Dohme., Gulen Hatemi Speakers bureau: Dr Hatemi served as a speaker for AbbVie, Celgene, Novartis, and UCB Pharma., Grant/research support from: Dr Hatemi has received grant/research support from Celgene.