Assessing disease phenotypes in Behçet’s syndrome: insights from a multiple correspondence analysis

Abstract

Introduction:

Behçet’s syndrome (BS) is a rare systemic vasculitis. Clinical manifestations in BS are frequently clustered rather than discrete, and the concept that distinct clinical phenotypes may exist in BS has recently emerged. The aim of the present work was to identify and analyze the disease phenotypes in a monocentric historical cohort of patients with BS.

Methods:

A total of 202 patients with BS diagnosis followed up at the Behçet Clinic of the Azienda Ospedaliero-Universitaria Pisana were identified, and demographics, clinical, and therapeutic data were retrospectively collected. Pairwise correlation among variables was evaluated by means of Pearson or Spearman correlation coefficient. A multiple correspondence analysis was performed to investigate the possible phenotypes resulting from the different patterns of associations among the demographic and clinical variables.

Results:

Most of the patients were female (67%), Caucasian (92%), and HLA-B51 carriers (65.5%). Mean age at disease onset was 30.06 ± 11.39 years, and oral ulcers (OU) and genital ulcers (GU) were the most common manifestations (96% and 61%, respectively). According to bivariate correlation analysis, significant positive correlations were observed between skin lesions and both OU (p = 0.005) and arthritis (p = 0.014), as well as pathergy (p = 0.001), gastrointestinal (GI) symptoms (p = 0.001), and other involvement (fever and serositis) (p = 0.015). Neurological involvement was significantly and positively associated with ocular lesions (p = 0.0114), GI symptoms (p = 0.030), pathergy (rho = 0.147, p = 0.037) and vein thromboses (p = 0.037). Despite the high heterogeneity, four disease phenotypes emerged from the MCA: (A) male Caucasians with greater age at onset and at diagnosis than the median values, with OU and GU, skin lesions, erythema nodosum (EN), arthritis, and GI symptoms; (B) co-existence of benign subset and pathergy; (C) orchitis/epididymitis associated with neurological involvement and ocular lesions; and (D) GI symptoms plus endoscopic lesions, large vessel disease (both arterial and venous), and other involvement.

Discussion:

This study provides valuable insights into the possible BS clinical phenotypes, and the results partially agree with previous association studies on European and extra-European cohorts. Observational comparative studies are warranted to assess the response of tailored phenotype-based therapeutic approaches.

Introduction

Behçet’s syndrome (BS) is a rare multi-system vasculitis with a relapsing–remitting course, usually beginning in young adulthood. Recurrent oral and genital ulcerations (OU and GU, respectively) are the clinical hallmarks of disease, along with posterior uveitis (1, 2). Other common manifestations are papulo-pustular skin lesions, large joints arthritis, and vascular, neuro-psychiatric, and gastrointestinal (GI) alterations (3–7). Such manifestations show variable prevalence according to gender, age, and ethnicity (4, 8). Moreover, disease severity and prognosis are also highly heterogeneous depending on how the single manifestations cluster together resulting in distinct clinical phenotypes. Indeed, it is discussed whether BS should be considered as a complex disorder of distinct clinical phenotypes rather than a single nosological entity (9–11). The existence of different disease phenotypes is currently accepted, and it is supported by the numerous immunopathogenic pathways underlying distinct clinical manifestations, the highly heterogeneous presentation among patients from different geographies, and the differing response of different clinical manifestations to one same drug (12–15).

In the framework of the clinical variability that BS may have, four main well-recognized clinical phenotypes have been first suggested: (1) muco-cutaneous phenotype; (2) papulopustular and articular phenotype; (3) vascular phenotype; and (4) ocular phenotype (10, 16–18). More recently, several cohort studies have analyzed other possible combinations of disease manifestations, but study designs are heterogeneous and results are conflicting (19–26). However, identifying and describing disease phenotypes could contribute to further explore the diverse immunopathogenesis of BS manifestations, thus promoting a wider understanding of disease and earlier diagnosis for patients with BS. Moreover, the definition of clinical outcomes and therapeutic response within each BS phenotype might lead to improved and tailored therapeutic strategies based on the disease phenotype rather than on the single disease manifestations (10, 19). Ultimately, deeper insights into disease phenotypes contribute to accomplish effective overall management and better quality of life for patients with BS (27, 28).

The aim of the present work was to identify disease phenotypes in a monocentric cohort of patients with BS evaluating the possible associations between clinical, epidemiological, and therapeutic variables.

Materials and methods

Patients and data collection

This study was conducted at the Rheumatology Unit of the Azienda Ospedaliero-Universitaria Pisana in Italy. In this historical cohort, study patients were selected if they fulfill the following criteria: (i) BS diagnosis according to the International Study Group (ISG) and/or the International Criteria for BS diagnosis (ICBS) (29, 30), and (ii) regular follow-up at our Behçet Clinic. A total of 202 patients were included for the analysis using the outpatient clinic database and the medical charts. Demographic information, previous organ involvements, and therapies at last evaluation were retrospectively reported. The presence of the major histocompatibility complex (MHC) locus—HLA-B51—was also reported, when available. Organ involvement was defined as the presence of suggestive clinical presentation confirmed by performing imaging tests and according to a specialistic evaluation. However, the presence of GI disturbances consistent with the disease allowed the definition of GI involvement even in the absence of typical endoscopic lesions, provided the agreement with the gastroenterologist specialist.

The disease subset was classified as either benign or severe according to the previous organ involvement. Specifically, the presence of ocular manifestations (OM), neuro-psychiatric symptoms, GI, and/or vascular disease classified the disease subset as severe. Conversely, the presence of skin and mucosal lesions, arthralgias/arthritis, erythema nodosum (EN), and pathergy defined a benign disease.

Statistical analysis

Quantitative variables were expressed as mean and standard deviation, while categorical variables were expressed as numbers (N) and percentages. The distribution of variables according to gender and disease course was evaluated considering the chi-square test or Fisher’s exact test as appropriate for categorical variables and using independent-samples t-test for quantitative variables. Pairwise correlation among variables was evaluated by means of Pearson or Spearman correlation coefficient as appropriate according to the type of variable.

To investigate possible distinct phenotypes resulting from different patterns of associations between the variables, a multiple correspondence analysis (MCA) was used. MCA is an explorative multivariate statistical method of dimension reduction for multiple categorical variables. It allows revealing the underlying structures in complex datasets - such as patterns of correlation among variables or similarities among observations, and providing a geometric interpretation through a spatial map of the data’s significant dimensions, where proximities between points and other geometric features indicate the associations among these dimensions (31). Patterns resulting from MCA are generally interpreted considering the first two dimensions extracted from the analysis, which represent most of the data variability displayed on two axes (the horizontal axis for the first dimension and the vertical axis for the second one). The graph allows for a visual assessment of the proximities between variables along each dimension, revealing existing patterns among them. Contributions of the different variables to each dimension were also obtained in order to understand the most influential variables contributing to the identification of the dimensions.

The advantage of MCA is that there is no need to meet assumption requirements.

All analyses were performed using the R package.

Results

Characteristics of patients according to gender and subset of disease are shown in Tables 1 and 2, respectively. The majority of patients were female and Caucasian. Mean age at disease onset was 30.06 ± 11.39 years, and mean age at the time of diagnosis was 37.39 ± 10.94 years. HLA-B51 was available for 119/202 patients, of whom 78 (65%) were HLA-B51+. OU and GU were the most common manifestations (96% and 61%, respectively), while venous thromboses and neurological and GI involvement were less frequent. Disease subset was benign for 94 (47%) patients and severe for 108 (53%) patients, according to their previous organ involvement. Colchicine was the most used medication (44%), and adalimumab was the most used biotechnological disease-modifying anti-rheumatic drug (bDMARD) (21%).

The bivariate correlation analysis showed that few variables reported statistically significant correlations, as shown in Figure 1.

Figure 1

The MCA highlights the high degree of variability among observed data with 10 dimensions extracted explaining approximately 80% of the overall variability, with the first 2 dimensions capturing 22% of overall variability. Percentages of explained variance of the dimensions are shown in the scree plot in Figure 2. In particular, benign subset, skin lesions, OM, GI symptoms plus endoscopic lesions, pathergy, both arterial and venous disease [large vessel involvement (LVI)], deep and superficial vein thromboses, neurological involvement, and other involvement (such as fever, constitutional symptoms, and serositis) significantly contributed to the first two dimensions; thus, these variables mostly determined the different profiles that could be deduced from the analysis (Table 3).

Figure 2

According to the results, and applying the MCA, we were able to define four main distinct phenotypes in more detail (Figure 3):

• characterized by the presence of GI symptoms, skin lesions, OU and GU, arthritis/arthralgias, and EN, also being mainly male Caucasians, with age at onset and at diagnosis greater than the median values;

• characterized by the co-presence of benign subset and pathergy;

• characterized by the presence of orchitis/epididymitis associated with neurological involvement and OM; and

• characterized by GI symptoms plus endoscopic lesions, LVI, and other involvement (fever, constitutional symptoms or serositis).

Figure 3

Discussion

The present work aimed at identifying and analyzing disease phenotypes in an Italian historical cohort of patients with BS. In agreement with other large BS cohort studies, patients included in the analysis were mostly women, the disease mainly occurred during early adulthood, and the muco-cutaneous involvement was the most prevalent (32–37).

Although there is large evidence from BS cohorts showing that severe manifestations more frequently occur in male patients (33–36, 38–40), we only observed a lower frequency of GU in male patients compared to female patients. Noteworthy, upon comparing this result with previous research involving European BS cohorts, we found that higher rates of uveitis, skin lesions, and venous thrombosis were observed in male patients (34, 35, 39) and that EN and arthritis were more common in female patients (34), with no significant difference in GU in both genders. However, gender did not significantly correlate with the benign or severe subset of disease in our analysis.

Moreover, older age at disease onset significantly correlated with skin lesions and other involvement, in contrast to previous studies that showed that patients with late disease onset frequently develop vascular, neurologic, and GI involvements compared to patients with early onset (41, 42). Last, we observed a negative correlation between HLA-B51 and pathergy, arthritis, and other involvement, while a meta-analysis of observational studies by Maldini et al. showed that HLA-B51 presence is mainly associated with decreased prevalence of GI involvement. Nevertheless, it is recognized that the contribution of HLA-B51 to the clinical presentation greatly varies with patient ethnicity (43).

To the best of our knowledge, this is the first study to assess the disease phenotypes in a BS cohort applying MCA. In the past years, multiple clinical phenotypes were identified performing cluster analysis and other association studies (16, 17, 44–47), but it has been argued that these methodologies may not be reliable for detecting existing disease phenotypes (23).

In this study, we were able to recognize four disease phenotypes. Specifically, phenotype A features a complete muco-cutaneous involvement associated with articular and GI manifestations in older male patients. As mentioned earlier, the association between skin lesions, mucosal ulcerations, and arthritis was widely described, as it is one of the first and most accepted disease phenotypes (17, 48, 49). Despite GI manifestations being conventionally severe, the absence of clear endoscopic findings suggests that this phenotype should be considered benign overall. Indeed, in clinical practice, it is not uncommon for patients with BS to refer non-specific mild abdominal symptoms that can be consistent with the disease. Phenotype B includes all the benign involvements of the disease. In contrast, patients of phenotype C show eye and neurologic involvement, in association with a less frequent manifestation like orchitis/epididymitis. Accordingly, Bitik et al. have previously reported a strong association between posterior uveitis and neurologic involvement (26), and other authors have recently observed a male predominance in ocular disease secondary to BS after performing a cluster analysis (44, 45).

Similarly, phenotype D is severe due to the presence of “complete” GI involvement (both symptoms and consistent endoscopic findings) and LVI, which consists in both venous and arterial lesions. Because of the very low prevalence of GI involvement, it was studied in few previous cluster analyses, and they showed that it rarely clusters with other organ involvement (45, 47). For what concerns LVI, strong association among peripheral venous and arterial lesions was found in previous research (50), as well as among peripheral vascular disease and cerebral venous sinus thrombosis (CVST) (51), leading to the definition of what is referred to as vascular or cardiovascular phenotype (3, 19). In addition, the vascular phenotype was mainly viewed as a separate phenotype in cluster studies (44, 47), and few associations with other clinical features were reported, such as a negative association with severe ocular lesions (20, 52). Of note, in phenotype D, vascular and GI involvement co-existed, in agreement with few other studies so far (53). Additionally, these two involvements suggest a highly inflammatory status; thus, unsurprisingly, auto-inflammatory symptoms like fever and serositis (referred to as other involvement in the current analysis) are also present in phenotype D.

This study provides valuable insights into specific BS clinical phenotypes, contributing to a more comprehensive understanding of BS heterogeneous clinical presentation. However, our study has some limitations, primarily due to the retrospective nature of the data. In addition, the monocentric design of our study might also have introduced potential selection biases. Another limitation is represented by the fact that these phenotypes are the result of artificial subgroups generated by the accumulation of clinical manifestations during the follow-up. Finally, considering that no significant associations were found in this regard, this type of analysis did not include ongoing therapies.

In conclusion, our analysis recognized four BS phenotypes within an Italian tertiary center. Our results add to prior evidence supporting the clinical perception that BS is a complex syndrome with different phenotypes consisting of different combinations of organ involvement. Phenotypes may have different pathogenetic mechanisms and different therapeutic implications. However, to date, pathogenetic studies in separate phenotypes of BS are still lacking, and comparative studies are needed to assess the response of tailored phenotype-based therapeutic approaches.

References

• 1

  EmmiGBettiolAHatemiGPriscoD. Behçet’s syndrome. Lancet. (2024) 403:1093–108. doi: 10.1016/S0140-6736(23)02629-6

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  BehcetH. Some observations on the clinical picture of the so-called triple symptom complex. Dermatology. (1940) 81:73–83. doi: 10.1159/000253787

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  BettiolAPriscoDEmmiG. Behçet: the syndrome. Rheumatol (Oxford). (2020) 59:iii101–7. doi: 10.1093/rheumatology/kez626

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  YaziciHSeyahiEHatemiGYaziciY. Behçet syndrome: a contemporary view. Nat Rev Rheumatol. (2018) 14:107–19. doi: 10.1038/nrrheum.2017.208

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  HatemiIHatemiGÇelikAF. Gastrointestinal involvement in behçet disease. Rheum Dis Clin North Am. (2018) 44:45–64. doi: 10.1016/j.rdc.2017.09.007

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  KalraSSilmanAAkman-DemirGBohlegaSBorhani-HaghighiAConstantinescuCSet al. Diagnosis and management of Neuro-Behçet’s disease: international consensus recommendations. J Neurol. (2014) 261(9):1662–76. doi: 10.1007/s00415-013-7209-3

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  TalaricoRPalaginiLd'AscanioAElefanteEFerrariCStagnaroCet al. Epidemiology and management of neuropsychiatric disorders in Behçet’s syndrome. CNS Drugs. (2015) 29(3):189–96. doi: 10.1007/s40263-015-0228-0

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  VerityDHMarrJEOhnoSWallaceGRStanfordMR. Behçet’s disease, the Silk Road and HLA-B51: historical and geographical perspectives: Behçet’s disease, the Silk Road and HLA-B51. Tissue Antigens. (1999) 54:213–20. doi: 10.1034/j.1399-0039.1999.540301.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  HatemiGSeyahiEFreskoITalaricoRUçarDHamuryudanV. Behçet’s syndrome: one year in review 2024. Clin Exp Rheumatol. (2024) 42(10):1999–2007. doi: 10.55563/clinexprheumatol/fqao16

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  BettiolAHatemiGVannozziLBarilaroAPriscoDEmmiG. Treating the different phenotypes of behçet’s syndrome. Front Immunol. (2019) 10:2830. doi: 10.3389/fimmu.2019.02830

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  YaziciHUgurluSSeyahiE. Behçet syndrome: is it one condition? Clinic Rev Allerg Immunol. (2012) 43(3):275–80. doi: 1007/s12016-012-8319-x

  • Google Scholar
  • View reference in article

• 12

  HatemiGSeyahiEFreskoITalaricoRUçarDHamuryudanV. Behçet’s syndrome: one year in review 2023. Clin Exp Rheumatol. (2023) 41(10):1945–54. doi: 10.55563/clinexprheumatol/7kdo9x

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  Di CianniFSulisAItalianoNMorettiMUrbanMLFerroFet al. The impact of environmental factors on aetiopathogenesis and clinical manifestations of Behçet’s syndrome. Clin Exp Rheumatol. (2024) 42(10):2008–13. doi: 10.55563/clinexprheumatol/bk6lc0

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  YaziciYHatemiGBodaghiBCheonJHSuzukiNAmbroseNet al. Behçet syndrome. Nat Rev Dis Primers. (2021) 7(1):67. doi: 10.1038/s41572-021-00301-1

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  TalaricoRItalianoNEmmiGPigaMCantariniLMattioliIet al. Efficacy and safety of infliximab or adalimumab in severe mucocutaneous Behçet’s syndrome refractory to traditional immunosuppressants: a 6-month, multicentre, randomised controlled, prospective, parallel group, single-blind trial. Ann Rheum Dis. (2024) 17:ard-2024-226113. doi: 10.1136/ard-2024-226113

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  TuncRKeymanEMelikogluMFreskoIYaziciH. Target organ associations in Turkish patients with Behçet’s disease: a cross sectional study by exploratory factor analysis. J Rheumatol. (2002) 29(11):2393–6.

  • Google Scholar
  • View reference in article

• 17

  KaracaMHatemiGSutNYaziciH. The papulopustular lesion/arthritis cluster of Behcet’s syndrome also clusters in families. Rheumatology. (2012) 51:1053–60. doi: 10.1093/rheumatology/ker423

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  YaziciHSeyahiE. Behçet syndrome: the vascular cluster. Turk J Med Sci. (2016) 46:1277–80. doi: 10.3906/sag-1605-28

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  SeyahiE. Phenotypes in behçet’s syndrome. Intern Emerg Med. (2019) 14:677–89. doi: 10.1007/s11739-019-02046-y

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  SuzukiTHoritaNTakeuchiMIshidoTMizukiYMizukiRet al. Clinical features of early-stage possible Behçet’s disease patients with a variant-type major organ involvement in Japan. Modern Rheumatol. (2019) 29(4):640–6. doi: 10.1080/14397595.2018.1494501

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  AridaAVaiopoulosGMarkomichelakisNKaklamanisPSfikakisPP. Are clusters of patients with distinct clinical expression present in Behçet’s disease? Clin Exp Rheumatol. (2009) 27(2 Suppl 53):S48–51.

  • Google Scholar
  • View reference in article

• 22

  DiriE. Papulopustular skin lesions are seen more frequently in patients with Behcet’s syndrome who have arthritis: a controlled and masked study. Ann Rheumatic Dis. (2001) 60:1074–6. doi: 10.1136/ard.60.11.1074

  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  IlgenU. Cluster analysis as a clinical and research tool in Behçet’s syndrome. Curr Opin Rheumatol. (2024) 36:3–8. doi: 10.1097/BOR.0000000000000980

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  SotaJRiganteDEmmiGLopalcoGOrlandoITosiGMet al. Behçet’s syndrome in Italy: a detailed retrospective analysis of 396 cases seen in 3 tertiary referral clinics. Intern Emerg Med. (2020) 15(6):1031–9. doi: 10.1007/s11739-019-02248-4

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  SuwaAHoritaNIshidoTTakeuchiMKawagoeTShibuyaEet al. The ocular involvement did not accompany with the genital ulcer or the gastrointestinal symptoms at the early stage of Behçet’s disease. Modern Rheumatol. (2019) 29(2):357–62. doi: 10.1080/14397595.2018.1457424

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  BitikBTufanASahinKSucullu KaradagYCan SandikciSMercanRet al. The association between the parenchymal neurological involvement and posterior uveitis in Behçet’s syndrome. Clin Exp Rheumatol. (2016) 34(6 Suppl 102):82–5.

  • Google Scholar
  • View reference in article

• 27

  MastroliaMVMarinelloDDi CianniFTalaricoRSimoniniG. Assessing quality of life in Behçet’s disease: a systematic review. Clin Exp Rheumatol. (2022) 40:1560–6. doi: 10.55563/clinexprheumatol/sian1b

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  LorenzoniVMarinelloDPallaIMoscaMTurchettiGTalaricoR. A cost-of-illness study of Behçet syndrome in Italy. Eur J Health Econ. (2024) 25(3):411–22. doi: 10.1007/s10198-023-01593-8

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  DavatchiFAssaad-KhalilSCalamiaKTCrookJESadeghi-AbdollahiBSchirmerMet al. The International Criteria for Behçet’s Disease (ICBD): a collaborative study of 27 countries on the sensitivity and specificity of the new criteria. J Eur Acad Dermatol Venereol. (2014) 28(3):338–47. doi: 10.1111/jdv.2014.28.issue-3

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  International Study Group for Behçet’s Disease. Criteria for diagnosis of Behçet’s disease. International Study Group for Behçet’s Disease. Lancet. (1990) 335(8697):1078–80.

  • Google Scholar
  • View reference in article

• 31

  NishisatoSGreenacreMJ. Correspondence analysis in practice Vol. 61. London: Academic Press (1996) p. 187–90.

  • Google Scholar

• 32

  HoumanMHNeffatiHBrahamAHarzallahOKhanfirMMiledMet al. Behçet’s disease in Tunisia. Demographic, clinical and genetic aspects in 260 patients. Clin Exp Rheumatol. (2007) 25(4 Suppl 45):S58–64.

  • Google Scholar
  • View reference in article

• 33

  Rodríguez-CarballeiraMAlbaMASolans-LaquéRCastilloMJRíos-FernándezRLarrañagaJRet al. Registry of the Spanish network of Behçet’s disease: a descriptive analysis of 496 patients. Clin Exp Rheumatol. (2014) 32(4 Suppl 84):S33–39.

  • Google Scholar
  • View reference in article

• 34

  LeccesePPadulaMLascaroNPadulaAD’AngeloS. Clinical phenotypes of Behçet’s syndrome in a large cohort of Italian patients: focus on gender differences. Scandinavian J Rheumatol. (2021) 50:475–8. doi: 10.1080/03009742.2021.1885735

  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  PipitoneNBoiardiLOlivieriICantiniFSalviFMalatestaRet al. Clinical manifestations of Behçet’s disease in 137 Italian patients: results of a multicenter study. Clin Exp Rheumatol. (2004) 22(6 Suppl 36):S46–51.

  • Google Scholar
  • View reference in article

• 36

  IshidoTHoritaNTakeuchiMKawagoeTShibuyaEYamaneTet al. Clinical manifestations of Behçet’s disease depending on sex and age: results from Japanese nationwide registration. Rheumatol (Oxford). (2017) 56(11):1918–27. doi: 10.1093/rheumatology/kex285

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  DavatchiFChams-DavatchiCShamsHNadjiAFaeziTAkhlaghiMet al. Adult Behcet’s disease in Iran: analysis of 6075 patients. Int J Rheum Dis. (2016) 19(1):95–103. doi: 10.1111/apl.2016.19.issue-1

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  TursenUGurlerABoyvatA. Evaluation of clinical findings according to sex in 2313 Turkish patients with Behçet’s disease. Int J Dermatol. (2003) 42:346–51. doi: 10.1046/j.1365-4362.2003.01741.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  BonitsisNGLuong NguyenLBLaValleyMPPapoutsisNAltenburgAKötterIet al. Gender-specific differences in Adamantiades-Behçet’s disease manifestations: an analysis of the German registry and meta-analysis of data from the literature. Rheumatol (Oxford). (2015) 54(1):121–33. doi: 10.1093/rheumatology/keu247

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  SotaJRagabGAlMaglouthILopalcoGTufanADireskeneliHet al. Influence of gender on Behçet’s disease phenotype and irreversible organ damage: Data from the International AIDA Network Behçet’s Disease Registry. Joint Bone Spine. (2024) 92(2):105819. doi: 10.1016/j.jbspin.2024.105819

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  RyuHJSeoMRChoiHJBaekHJ. Clinical phenotypes of Korean patients with Behcet disease according to gender, age at onset, and HLA-B51. Korean J Intern Med. (2018) 33:1025–31. doi: 10.3904/kjim.2016.202

  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  MastroliaMVBettiolAMarraniEMaccoraITaddeiEPagniniIet al. Behçet syndrome in children and adults: discovering similarities and differences by a comparative study. Rheumatol (Oxford). (2023) 62(SI2):SI189–95. doi: 10.1093/rheumatology/keac347

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  MaldiniCLavalleyMPCheminantMde MenthonMMahrA. Relationships of HLA-B51 or B5 genotype with Behcet’s disease clinical characteristics: systematic review and meta-analyses of observational studies. Rheumatol (Oxford). (2012) 51:887–900. doi: 10.1093/rheumatology/ker428

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  KılıçGKörüklüKFKumcuMGÇakırEKarkucakMKılıçE. Gender disparities in Behçet’s syndrome: identifying distinct phenotypes through cluster analysis. Immunol Res. (2024) 72:975–81. doi: 10.1007/s12026-024-09498-1

  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  SheCHHuDZouJBaoHFShenYCaiJFet al. Clinical phenotypes of adult-onset Behçet’s syndrome: a comprehensive cross-sectional study in China. Clin Rheumatol. (2024) 43:2261–71. doi: 10.1007/s10067-024-06995-4

  • CrossRef
  • Google Scholar
  • View reference in article

• 46

  OguzEBektaşM. Characteristics of patients with behçet disease from the van province, eastern Turkey: definition of disease clusters in a tertiary referral center. J Clin Rheumatol. (2023) 29:285–90. doi: 10.1097/RHU.0000000000001996

  • CrossRef
  • Google Scholar
  • View reference in article

• 47

  ZouJLuoJ-FShenYCaiJ-FGuanJ-L. Cluster analysis of phenotypes of patients with Behçet’s syndrome: a large cohort study from a referral center in China. Arthritis Res Ther. (2021) 23:45. doi: 10.1186/s13075-021-02429-7

  • CrossRef
  • Google Scholar
  • View reference in article

• 48

  KurosawaMTakenoYKirinoYSoejimaNMizukiM. Subgroup classification of Behçet’s disease using clinical information: analysis of a clinical database of patients receiving financial aid for treatment. In: ISBD Conf 2018. Rotterdam. (2018).

  • Google Scholar
  • View reference in article

• 49

  HatemiGFreskoITascilarKYaziciH. Increased enthesopathy among Behçet’s syndrome patients with acne and arthritis: An ultrasonography study. Arthritis Rheum. (2008) 58:1539–45. doi: 10.1002/art.23450

  • CrossRef
  • Google Scholar
  • View reference in article

• 50

  TascilarKMelikogluMUgurluSSutNCaglarEYaziciH. Vascular involvement in Behçet’s syndrome: a retrospective analysis of associations and the time course. Rheumatology. (2014) 53(11):2018–22. doi: 10.1093/rheumatology/keu233

  • CrossRef
  • Google Scholar
  • View reference in article

• 51

  TuncR. Cerebral venous thrombosis is associated with major vessel disease in Behcet’s syndrome. Ann Rheumatic Dis. (2004) 63:1693–4. doi: 10.1136/ard.2003.018515

  • CrossRef
  • Google Scholar
  • View reference in article

• 52

  HusseinMAEissaIMDahabAA. Vision-threatening behcet’s disease: severity of ocular involvement predictors. J Ophthalmol. (2018) 2018:1–6. doi: 10.1155/2018/9518065

  • CrossRef
  • Google Scholar
  • View reference in article

• 53

  SoejimaYKirinoYTakenoMKurosawaMTakeuchiMYoshimiRet al. Changes in the proportion of clinical clusters contribute to the phenotypic evolution of Behçet’s disease in Japan. Arthritis Res Ther. (2021) 23(1):49. doi: 10.1186/s13075-020-02406-6

  • CrossRef
  • Google Scholar
  • View reference in article