Clinical records of 769 consecutive patients fulfilling the international criteria of classification for BD 5 were analyzed. All patients were referred to and regularly followed in the Internal medicine department of the Pitié-Salpétrière university hospital, Paris, France between 1974 and 2010. For each patient, the following data were collected: age at diagnosis of BD, gender, date of criteria for BD, geographic origin, main features of BD including mucocutaneous manifestations, ocular lesions, rheumatologic manifestations (athralgia, arthritis), neurologic involvement and/or cardiovascular involvement (venous, arterial and cardiac lesions). The number of BD flare, treatment, outcome and causes of death were recorded. 568 patients with BD were investigated for HLA B5 typing. The 3 largest ethnic groups originated from Europe (n = 369), North Africa (n = 350) and sub Saharan Africa (n = 50). Ethnicity was defined as the country of origin of the patient’s parents and grandparents and subjects were classified into 1 of the following defined ethnic groups: European, North African (Moroccan, Algerian, Tunisian, and Egyptian) and sub Saharan African. We considered only the European, North African and sub Saharan African patients, given the relatively small number of patients in the other racial groups. The diagnosis of neuro-BD was based on objective neurological symptoms not explained by any other known disease or therapy associated with neuroimaging findings suggestive of BD related central nervous system (CNS) involvement and/or cerebrospinal fluid (CSF) findings showing aseptic inflammation. Patients with non-parenchymal CNS involvement, or who did not show any abnormality on neurological examination or those without evidence of objective neurological involvement (i.e. only headache or dizziness) were excluded. The diagnosis of cardiovascular involvement included venous, arterial and cardiac lesions related to BD. Diagnosis of venous thrombosis was based on imaging data: venous doppler sonography, phlebography, cavography, computed tomography angiography and/or angio-magnetic resonance imaging (MRI). Diagnosis of arterial manifestations was based on imaging data: doppler sonography, arteriography, computed tomography angiography and/or angio-MRI. Cardiac involvement (pericarditis, myocardiopathy, myocardial infarction, endomyocardial fibrosis, thrombosis, valve insufficiency and/or stenosis) was considered to be related to BD if it was contemporaneous with BD flares and if other known causes of cardiac disease had been excluded.

We systematically screened the medical literature via PubMed using the following keywords: “Behcet’s disease”, “Behcet’s syndrome”, “ethnology”, “ethnicity”, and “epidemiology”. We only analyzed cases series published after 1988 in English or in French. Studies focusing on ethnicity related phenotype or prognosis differences in adults BD patients were included. However, it should be acknowledged that these different studies have different approaches in case findings. Some studies were based on epidemiological estimates while others describe patients from tertiary referral centres and the latter were based on questionnaire or population samples. The methodological differences can be substantial and might affect the outcome of different cohorts.

Data are summarized as frequencies and percentages for categorical variables. Quantitative variables are presented as medians and 25th and 75th percentiles. Patients characteristics were compared using χ2 tests (or χ2 tests with Yates’s correction or Fisher’s exact test when required) for categorical data and Kruskal-Wallis rank sum tests for quantitative data. Survival was estimated using the Kaplan-Meier method. Patients were censored at the date of their last visit. Factors associated with the occurrence of death were assessed using a Cox proportional hazard model. Proportional hazard assumptions were checked. Hazard-ratios (HR) with their ninety-five percent confidence intervals (95% CI) are presented as a measure of association. All factors with P-value lower than 0.05 in the univariable analysis were included in a multiple Cox proportional hazard model. A model selection based on p-value was then performed. All tests were two-sided at the 0.05 significance level. Analyses were performed using R 2.15.1 statistical package. This study has been carried out in compliance with the Helsinki declaration.

The main features of the 769 BD patients are summarized in Table 1. The median (IQR) age at diagnosis was 30.9 (24.9-37.2) years with 535 (69.6%) male. Patients originated from Europe (47.9%), North Africa (45.5%) and sub Saharan Africa (6.5%). The HLA B5 typing was positive in 51.4% of cases. Main clinical signs of BD included eye involvement (65.9%), genital ulcerations (60.4%), articular involvement (48.6%), cardiovascular involvement (42.6%) and central nervous system involvement (31.7%). The median number of BD’s flare was of 3 (2–5).

*Except where indicated otherwise values are the median, IQR or n, percentage.

CNS, central nervous system; BD, Behçet’s disease; CV, cardiovascular.

The comparison between main characteristics according to ethnic origin of BD patients showed a higher frequency of male in sub Saharan and North African patients compared to those from Europe (78% and 78% vs 52.3%, respectively). The male/female ratio in the sub-Saharan and North African groups was of 3.5. There was a higher frequency of CNS involvement (48% vs 32.3% vs 29.5%) (p = 0.035), a higher rate of death (12% vs 6% vs 3.5%) (p = 0.029) and a lower frequency of HLA B51 allele (29.4% vs 49.2% vs 55.8%) (p = 0.009) in patients from sub Saharan Africa compared to those from North Africa and Europe, respectively. The time between first symptoms and diagnosis of BD was not significantly different according to ethnic groups (Table 1). There were no significant differences regarding the 3 separated periods (i.e. before 1990, between 1990 and 2000, and after 2000) of management of BD according to the 3 ethnics groups.

The 5-, 10- and 15 year survival rate was of 90%, 87% and 81% in sub Saharan African BD patients as compared to 96%, 93% and 91% and 99%, 96% and 94% in those from North Africa and Europe (p = 0.015), respectively (Figure 1).

Comparison between alive and deceased patients showed a higher proportion of men and of cardiovascular involvement and a lower frequency of HLA B5 genotype in the deceased group (Table 2). The analysis of factors associated with mortality is summarized in Table 3. In univariate analysis, there was no significant association between mortality and oral ulceration, articular involvement, central nervous system involvement, and eyes involvement. Patients who died were younger (p = 0.002), more frequently of male gender (p = 0.001), had a higher frequency of cardiovascular involvement (p = 0.009), had more frequently glucocorticosteroids (p = 0.013) and immunosuppressants use (p = 0.020), more frequently originated from sub Saharan Africa (p = 0.015) and had a lower frequency of genital ulceration (p = 0.006). In multivariate analysis (Table 3), male gender (OR: 5.01, CI: 1.51-16.65), cardiovascular involvement (OR: 2.24, CI: 1.15-4.36) and African sub-Saharan origin (OR: 2.62, CI: 0.98-6.97) were independently associated with mortality.

CNS, central nervous system, BD, Behçet's disease.

*HLAB5 was available for 538 out of 729 patients.

Except where indicated otherwise values are the median, IQR.

CNS, central nervous system, BD, Behçet's disease.

We found 11 manuscripts, with a total of 910 patients that assessed ethnicity related phenotypic or prognosis differences in BD 6 7 8 9 10 11 12 13 14 15 16 . Three of them 6 9 12 were excluded because two compared their patients’ characteristics to data reported in the literature, and one was using the same cohort of BD patients previously reported. We analysed 8 studies (783 patients), over a period of 1988 to 2012, that have addressed ethnicity related differences according to phenotype and outcome of BD 7 8 10 11 13 14 15 16 . The Table 4 summarized the main characteristics and conclusions of studies analysing ethnicity related differences in BD.

*Difference with respect to phenotype or outcome of BD. CNS, central nervous system.