Mortality in Behçet's disease

Authors

  • D. Saadoun,

    Corresponding author
    1. Hôpital Pitié-Salpétrière, Assistance Publique-Hôpitaux de Paris, and Université Pierre et Marie Curie-Paris 6, Paris, France
    • Service de Médecine Interne, AP-HP, Hôpital Pitié-Salpêtrière, Service de Médecine Interne, 83 Boulevard de l'Hôpital, Paris F-75013, France
    Search for more papers by this author
  • B. Wechsler,

    1. Hôpital Pitié-Salpétrière, Assistance Publique-Hôpitaux de Paris, and Université Pierre et Marie Curie-Paris 6, Paris, France
    Search for more papers by this author
  • K. Desseaux,

    1. INSERM U717, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris, Paris, France
    Search for more papers by this author
  • D. Le Thi Huong,

    1. Hôpital Pitié-Salpétrière, Assistance Publique-Hôpitaux de Paris, and Université Pierre et Marie Curie-Paris 6, Paris, France
    Search for more papers by this author
  • Z. Amoura,

    1. Hôpital Pitié-Salpétrière, Assistance Publique-Hôpitaux de Paris, and Université Pierre et Marie Curie-Paris 6, Paris, France
    Search for more papers by this author
  • M. Resche-Rigon,

    1. INSERM U717, Hôpital Saint-Louis, Assistance Publique-Hôpitaux de Paris, Paris, France
    Search for more papers by this author
  • P. Cacoub

    1. Hôpital Pitié-Salpétrière, Assistance Publique-Hôpitaux de Paris, and Université Pierre et Marie Curie-Paris 6, Paris, France
    Search for more papers by this author
    • Dr. Cacoub has received consulting fees, speaking fees, and/or honoraria from AstraZeneca, Bristol-Myers Squibb, Sanofi-Aventis, Gilead Sciences, and Schering-Plough (less than $10,000 each) and from Roche and Servier (more than $10,000 each).


Abstract

Objective

To report the long-term mortality in patients with Behçet's disease (BD).

Methods

A cohort of 817 patients fulfilling the international criteria for BD from a single center in France were analyzed for causes of death, the standardized mortality ratio (SMR), and the factors associated with mortality.

Results

Among the 817 patients with BD, 41 (5%) died after a median followup of 7.7 years, of whom 95.1% were male. The mean ± SD age at death was 34.8 ± 11.9 years. Main causes of death included major vessel disease (mainly, arterial aneurysm and Budd-Chiari syndrome) (43.9%), cancer and malignant hemopathy (14.6%), central nervous system involvement (12.2%), and sepsis (12.2%). The mortality rate at 1 year and 5 years was 1.2% and 3.3%, respectively. There was an increased mortality among patients ages 15–24 years (SMR 2.99, 95% confidence interval [95% CI] 1.54–5.39) and those ages 25–34 years (SMR 2.90, 95% CI 1.80–4.49) as compared with age-and sex-matched healthy controls. The mortality decreased in patients older than age 35 years (SMR 1.23, 95% CI 0.75–1.92). In multivariate analyses, male sex (hazard ratio [HR] 4.94, 95% CI 1.53–16.43), arterial involvement (HR 2.51, 95% CI 1.07–5.90), and a high number of BD flares (HR 2.37, 95% CI 1.09–5.14) were independently associated with the risk of mortality.

Conclusion

The overall mortality in our BD cohort was 5% after a median followup of 7.7 years. Male sex, arterial involvement, and the number of flares were associated with mortality in BD.

Behçet's disease (BD) is a chronic, relapsing vasculitis of unknown etiology that is characterized by mucocutaneous, ocular, articular, vascular, gastrointestinal, and central nervous system manifestations (1, 2). BD significantly increases morbidity and mortality. The leading cause of morbidity in BD is eye involvement with the potential threat of visual loss. Few studies have addressed the mortality of Behçet's syndrome. Among 2,031 patients from Japan, 31.7% experienced clinical deterioration, and 0.9% died during the course of a single year of followup. In Turkey, 42 of 428 patients died, mainly due to major vessel disease and neurologic involvement. However, the remaining studies were mostly short term or had a small cohort, with the focus being on mortality related to a single organ system (i.e., central nervous system, major vessels). Furthermore, most of the reports on mortality in patients with BD come from Japan or Turkey. BD shows considerable geographic variation in clinical expression. For instance, patients with BD from Asia exhibit a higher frequency of gastrointestinal involvement as compared with those from the Mediterranean basin (3, 4).

The present study was undertaken to report the long-term mortality in BD. We analyzed the main causes of death and the standardized mortality ratio (SMR) in a cohort of 817 patients with BD from a single university center in France. Factors associated with mortality were assessed by multivariate analysis.

PATIENTS AND METHODS

Patients.

Clinical records of 817 consecutive patients fulfilling the international criteria for the classification of BD (5) were analyzed. All patients were referred to, and regularly followed up in, the internal medicine department of the Pitié-Salpétrière university hospital in Paris, France between 1974 and 2006. For each patient, the following data were collected: age at diagnosis of BD, sex, date of satisfaction of criteria for BD, geographic origin, and main features of BD, including mucocutaneous manifestations, ocular lesions, rheumatic disease manifestations (arthralgias, arthritis), neurologic involvement, and/or vascular involvement (venous and arterial lesions). The number of BD flares, treatment, outcome, and causes of death were recorded. Six hundred forty-two patients with BD were investigated for HLA–B5 typing.

Assessment of the SMR.

To compare the observed mortality with the expected mortality, the SMR was used. The SMR is the ratio of the observed mortality in patients to the mortality in the total French population of subjects with corresponding sex and year of birth. Only patients who were adults at diagnosis (age ≥15 years) were considered. Expected mortality rates were obtained from the French National Statistical Institute (http://www.ined.fr/fr/pop_chiffres/france/mortalite_causes_deces/deces_sexe_age). Patients were divided into 3 age groups (ages 15–24 years, 25–34 years, and ≥35 years).

Statistical analysis.

Data are summarized as frequencies and percentages for categorical variables. Quantitative variables are presented as the median and 25th–75th percentiles (interquartile range [IQR]) or mean ± SD. To take into account censored observations, the Kaplan-Meier method was used to estimate survival curves. The percentage of patients with a length of followup of <6 months, 1 year, and 2 years was 5%, 10%, and 17%, respectively. Factors associated with the occurrence of death were assessed using a Cox proportional hazards model. Hazard ratios (HRs) with their 95% confidence interval (95% CI) are presented as a measure of association. Occurrences of first oral ulcerations, genital ulcerations, eye involvement, articular involvement, venous involvement, and central nervous system involvement were considered as time-dependent covariates, as well as having ≥5 BD flares. All factors with a P value lower than 0.05 in the univariate analysis were included in a Cox proportional hazards multivariate model. All tests were 2-sided at the 0.05 significance level. Confidence intervals were calculated directly from the Poisson distribution (6, 7). Analyses were performed using the R statistical package.

RESULTS

Characteristics of the patients.

The main features of the 817 patients with BD are summarized in Table 1. The mean ± SD age at diagnosis was 31.5 ± 10.5 years, and 541 (66.2%) of the patients were male. Most of the patients originated from Europe (44.7%) and North Africa (41.1%). The HLA–B5 typing was positive in 40.3% of patients. The main clinical signs of BD included recurrent oral ulcerations (99.4%) and genital ulcerations (69.5%), skin lesions (mainly, pseudofolliculitis) (61.4%), eye involvement (62.9%), articular involvement (43.3%), venous involvement (36.8%), central nervous system involvement (26.9%), and arterial involvement (13.9%). The mean ± SD number of BD flares was 3.7 ± 2.6. In comparing the main characteristics between patients who died and those who remained alive, a higher frequency of male patients (92.7% versus 64.8%, respectively), a higher number of BD flares (mean ± SD 4.0 ± 2.7 versus 3.6 ± 2.7, respectively), a higher frequency of arterial involvement (36.6% versus 12.8%, respectively), a higher proportion of patients receiving immunosuppressant drugs (75.6% versus 51.6%, respectively), and a lower frequency of genital ulcerations (43.9% versus 70.8%, respectively) were observed among the patients who died (Table 1).

Table 1. Characteristics of the patients with Behçet's disease (BD) and comparison between patients who died and those who remained alive*
ParameterAll (n = 817)Died (n = 41)Alive (n = 776)
  • *

    Except where indicated otherwise values are the number (%) of patients. CNS = central nervous system.

Age at diagnosis, mean ± SD years31.5 ± 10.534.8 ± 11.931.4 ± 10.4
Male sex541 (66.2)38 (92.7)503 (64.8)
Ethnic origin   
 Europe365 (44.7)15 (36.6)350 (45.1)
 North Africa336 (41.1)19 (46.3)317 (40.8)
 Africa73 (8.9)4 (9.8)69 (8.9)
 Other43 (5.3)3 (7.1)40 (5.1)
HLA–B5 (n = 642)259 (40.3)11/31 (35.5)300/611 (49.1)
Oral ulceration812 (99.4)41 (100)771 (99.3)
Genital ulceration568 (69.5)18 (43.9)550 (70.8)
Ocular involvement514 (62.9)25 (61)489 (63.0)
CNS involvement220 (26.9)15 (36.6)205 (26.4)
Articular involvement354 (43.3)20 (48.8)334 (43.0)
Venous involvement301 (36.8)18 (43.9)283 (36.5)
Arterial involvement114 (13.9)15 (36.6)99 (12.8)
Number of BD flares, mean ± SD3.7 ± 2.64.0 ± 2.73.6 ± 2.7
BD flares ≥5179 (21.9)12 (29.3)167 (21.5)
Immunosuppressants431 (52.8)31 (75.6)400 (51.6)

When we compared the characteristics of the male patients with those of female patients, we found differences in terms of geographic origin, in that male patients originated from either Europe (39.1%) or northern Africa (47.7%), whereas female patients were mostly from Europe (66.2%). There was a higher number of BD flares (mean ± SD 3.7 ± 2.6 versus 3.5 ± 2.8, respectively) and a higher frequency of arterial involvement (90 [16.6%] of 540 versus 14 [5.1%] of 276, respectively) in male patients compared with female patients. However, there were no significant differences according to age at diagnosis, HLA–B5 status, frequency of main clinical features of BD, or immunosuppressant use. In the multivariate analysis, among male patients, arterial involvement (HR 2.5, 95% CI 1.2–5.6), genital ulceration (HR 0.5, 95% CI 0.3–1.1), and a high frequency (≥5) of BD flares (HR 2.5, 95% CI 1.2–5.6) were independently associated with mortality.

Causes of death.

Table 2 summarizes the main causes of death in the 41 patients with BD who died over followup. The mean ± SD age at death was 34.8 ± 11.9 years, with a mean time between diagnosis of BD and death of 62.8 ± 64.8 months. Main causes of death included arterial involvement (pulmonary arterial aneurysm [n = 3], thoracic aortic aneurysm [n = 3], myocardial infarction [n = 3], abdominal aortic aneurysm [n = 1], and cerebral aneurysm [n = 1]; 26.8% of patients who died), large venous involvement (Budd-Chiari syndrome [n = 4)] or pulmonary embolism [n = 3]; 17.1% of patients who died), cancer (lung, tongue, or bladder carcinoma [n = 1 each], malignant hemopathy [non-Hodgkin's lymphoma] [n = 2], and acute leukemia [n = 1]; 14.6% of patients who died), central nervous system involvement (n = 5; 12.2% of patients who died), and sepsis (pseudomonas septicemia [n = 2], candidosis septicemia [n = 1], tuberculosis [n = 1], and endocarditis [n = 1]; 12.2% of patients who died). Three patients who died from sepsis had an underlying disease, i.e., diabetes (n = 2) and bronchiectasis (n = 1). Other causes of death included pancreatitis (n = 1), AA amyloidosis (n = 1), thrombotic microangiopathy (n = 1), and unknown (n = 4).

Table 2. Causes of death in 41 patients with Behçet's disease (BD)
Patient/sex/age at death, yearsTime between diagnosis of BD and death, monthsNumber of BD flaresCause of death
1/M/5733Pseudomonas septicemia
2/M/4344Pulmonary artery aneurysm (massive hemoptysis)
3/M/3641Candidosis septicemia
4/M/4042Thoracic aortic aneurysm (death after surgical repair)
5/M/3243Cerebral aneurysm (sudden death by rupture)
6/M/3271Pulmonary arterial aneurysm (massive hemoptysis)
7/M/4071Ischemic heart disease
8/M/3093Endocarditis
9/M/4496Thoracic and abdominal aortic aneurysm (rupture of thoracic aneurysm)
10/M/21122Budd-Chiari syndrome
11/M/27143Pulmonary arterial aneurysm (Hughes Stovin syndrome)
12/M/18142Budd-Chiari syndrome
13/M/34202Budd-Chiari syndrome
14/M/22291Pulmonary embolism
15/M/24343Central nervous system disease with severe progressive course
16/M/223614Acute renal failure (thrombotic microangiopathy during cyclosporine therapy)
17/M/39364Ischemic heart disease
18/M/22377Abdominal aortic aneurysm (sudden death by rupture)
19/M/24373Unknown
20/M/57412Carcinoma (lung)
21/M/48473Carcinoma (tongue)
22/M/37507Tuberculosis
23/M/33502Ischemic heart disease
24/M/29516Pulmonary embolism (during thalidomide therapy)
25/M/48524Thoracic aortic aneurysm (death after surgical repair)
26/M/28542Unknown
27/M/30571Budd-Chiari syndrome (rupture of esophageal varice)
28/M/29749Pancreatitis
29/M/58773Acute leukemia
30/M/40774Central nervous system disease with severe progressive course
31/M/48826Central nervous system disease with severe progressive course
32/F/29823Central nervous system disease with severe progressive course
33/M/34921Unknown
34/F/22983Pseudomonas septicemia
35/M/271196AA amyloidosis (renal and digestive)
36/M/611364Non-Hodgkin's lymphoma, apsergillosis
37/M/331566Carcinoma (bladder)
38/M/271598Central nervous system disease with severe progressive course
39/M/171838Pulmonary embolism (during thalidomide therapy)
40/M/272288Unknown
41/M/492902Non-Hodgkin's lymphoma

Mortality rates and SMRs.

The cumulative rate of death was 1.2%, 2.1%, 3.3%, and 4.3% at 1, 3, 5, and 10 years, respectively (Figure 1). After a median followup of 7.7 years (IQR 3.4–13 years), 41 patients with BD (5%) died. The mortality as measured by the SMR (Figure 2) was divided into 3 age groups (15–24 years, 25–34 years, and ≥35 years) based on age at diagnosis of BD. In this analysis, 777 patients age 15 years or older were included. There was an increased mortality among the patients ages 15–24 years (SMR 2.99, 95% CI 1.54–5.39) and those ages 25–34 years (SMR 2.90, 95% CI 1.80–4.49) (Figure 2). The mortality decreased in patients older than age 35 years (SMR 1.23, 95% CI 0.75–1.92).

Figure 1.

Survival curve over followup of the 817 patients with Behçet's disease, among all patients (A) and according to sex (B) and ethnic origin (C).

Figure 2.

Standardized mortality ratio (SMR) with 95% confidence intervals for 777 patients with Behçet's disease. Only patients who were adults at diagnosis (age ≥15 years) were considered. Horizontal line indicates the cutoff.

Factors associated with mortality.

Results of the logistic regression analysis of factors associated with mortality are summarized in Table 3. In the univariate analysis, there was no significant association between mortality and age, HLA–B5 status, ethnic origin, oral ulceration, articular involvement, central nervous system involvement, or ocular and venous involvement. Among the patients who died, patients were predominantly male and had a significantly higher frequency of BD flares, a significantly higher frequency of corticosteroid and immunosuppressant use, a significantly higher frequency of arterial involvement, and a significantly lower frequency of genital ulceration. In the multivariate analysis, male sex (HR 4.94, 95% CI 1.53–16.43), arterial involvement (HR 2.51, 95% CI 1.07–5.90), genital ulceration (HR 0.49, 95% CI 0.25–0.98), and a high frequency (≥5) of BD flares (HR 2.37, 95% CI 1.09–5.14) were independently and significantly associated with mortality (Table 3).

Table 3. Factors associated with mortality in Behçet's disease (BD)*
ParameterUnivariate analysisMultivariate analysis
HR (95% CI)PHR (95% CI)P
  • *

    HR = hazard ratio; 95% CI = 95% confidence interval; CNS = central nervous system.

Age at diagnosis1.38 (0.6–3.3)0.49  
Male sex6.87 (2.1–22.3)0.0014.94 (1.5–16.4)0.007
Geographic origin    
 Europe1   
 North Africa2.48 (0.8–7.4)0.11  
 Africa1.71 (0.8–3.3)0.12  
 Other2.45 (0.7–8.5)0.16  
Number of BD flares2.47 (0.1–1.6)0.0182.37 (1.1–5.1)0.029
HLA–B50.51 (0.2–1.1)0.07  
Oral ulcerations0.54 (0.1–2.2)0.39  
Genital ulcerations0.41 (0.2–0.7)0.0060.49 (0.2–0.9)0.044
Ocular involvement0.78 (0.3–1.6)0.51  
CNS involvement1.42 (0.8–3.3)0.42  
Articular involvement1.02 (0.5–1.9)0.96  
Venous involvement1.06 (0.1–2.2)0.88  
Arterial involvement3.3 (1.4–7.8)0.0052.51 (1.1–5.9)0.034
Immunosuppressants2.39 (1.1–4.8)0.017  

We next tested the impact of being non-European versus being European on the risk of death (results not shown). Among European patients with BD (n = 365), there were 15 deaths (4.1%), while among non-European patients (n = 452), there were 26 deaths (5.7%). In the univariate analysis, the HR for the risk of mortality in non-European patients was 1.85 (95% CI 0.98–3.57) (P = 0.06). In the multivariate analysis, male sex (HR 4.6, 95% CI 1.4–15.2), arterial involvement (HR 2.4, 95% CI 1.1–5.3), genital ulceration (HR 0.5, 95% CI 0.2–0.9), and a high frequency (≥5) of BD flares (HR 2.4, 95% CI 1.1–5.3) were independently and significantly associated with mortality. The HR for the risk of mortality in non-European patients was 1.3 (95% CI 0.7–2.5) (P = 0.45).

DISCUSSION

In the present study, we assessed the long-term mortality in a cohort of 817 patients with BD. We analyzed the main causes of death and the SMR in our BD cohort. Factors associated with mortality were assessed by multivariate analysis. The most striking findings from this study are 1) the overall mortality of 5% among patients with BD after a median followup of 7.7 years, 2) the 3 times higher mortality in BD patients younger than age 35 years, 3) the independent association of mortality with male sex, arterial involvement, and a high number of BD flares, and 4) the specific causes of death in our patients with BD.

Among 817 patients with BD, 41 (5%) died after a median followup of 7.7 years. This was consistent with the results of previous studies from Morocco (10 deaths [3.2%] in 316 patients]) and from Turkey (6 deaths [3.9%] in 152 patients) (8, 9). However, the investigators in Turkey reported the long-term mortality in their cohort after surveying the data over 2 decades, which yielded a mortality rate of 9.8% (10). In the present study, most deaths occurred 5 years after the diagnosis of BD. The main causes of death included major vessel disease (i.e., arterial aneurysm and Budd-Chiari syndrome) and central nervous system involvement. Although not independently associated with mortality, central nervous system involvement accounted for 12% of deaths in our study. In large studies specifically addressing neurologic disease in BD, the mortality rate ranged between 5.5% and 20% (11, 12). The median period until death was 4 years after onset of neurologic signs (11).

The frequency of arterial involvement was 3 times higher among the patients who died. In the multivariate analysis, patients with BD who developed arterial involvement had a 2.5 higher likelihood of dying. Thoracic aorta and pulmonary arterial aneurysms were associated with the higher mortality. The rate of death in patients with aneurysm rupture has been found to be as high as 60% (13). We previously reported that the survival rate of patients with BD complicated by arterial lesions was 66% at 15 years (14). The most severe complication of pulmonary vasculitis is pulmonary arterial aneurysms (13, 15, 16). In a study by Hamuryudan et al involving 24 patients with BD who experienced pulmonary arterial aneurysm (16), 12 (50%) of the patients died 9.5 months after the onset of hemoptysis, despite receiving therapy. In a more recent report on patients with pulmonary arterial aneurysm, the same group observed an overall 5-year survival rate of 63%; patients diagnosed since 1992 had a 5-year survival rate of 80%. The improved prognosis was believed to be due to earlier diagnosis and rational use of immunosuppressive agents (17). Pulmonary arterial vasculitis almost exclusively affects male patients and has a strong association with a systemic pattern of vessel involvement at other sites (13, 16).

In addition to arterial involvement, large venous lesions were associated with an important increase in mortality, especially in those with Budd-Chiari syndrome. In the present study, Budd-Chiari syndrome accounted for 10% of deaths. Hepatic vein thrombosis was diagnosed in 12 patients, of whom 4 (33.3%) died. Behçet's syndrome is a common cause of Budd-Chiari syndrome. In a series of BD patients with large-vessel thrombosis (18), hepatic vein thrombosis was reported in 14 (26.4%) of 53 patients, and after a mean followup of 3 years, 10 (71.2%) of the patients died. In a study by Kural-Seyahi et al, among 42 patients with BD who died, Budd-Chiari syndrome was diagnosed in 3 patients and superior vena cava occlusion in 4 patients (10). In another study, involvement of the large vessels was found to be present in 15–35% of patients with BD (19). The prevalence of large vessel involvement, involving either the venous or arterial systems or both, in Turkish patients with BD was reported to be 49.2% in male patients and in 5.6% in female patients (10). All of the studies on BD confirm the predominance of vascular disease in male patients (13).

Male sex was the main factor associated with mortality in our cohort of 817 patients with BD. Thirty-eight (92.7%) of the 41 patients with BD who died were male. In multivariate analyses, the likelihood of death was increased almost 5 times when male sex was added as a covariate. Compared with female patients, the proportion of arterial involvement was 3 times higher in male patients. Male sex and a younger age at onset have been previously reported to markedly influence disease expression and the course of BD (3, 10, 20). Male patients tended to have more flares of BD compared with female patients. We found an independent association between a high number of BD flares and mortality. Patients who experienced ≥5 flares of BD had a 2 times increased likelihood of dying. This has never been previously reported. In a recent study of BD patients with neurologic involvement, the number of attacks (>2) and relapse during steroid dose tapering was associated with a poor prognosis (11). A high number of BD flares resulted in exposure of patients to repeated treatment with corticosteroids and immunosuppressants, which might account for the deaths due to sepsis, drug toxicity, or neoplasia. Alternatively, persistent inflammation may lead to complications such as AA amyloidosis.

Mortality, as measured by the SMR, was significantly higher in patients with BD as compared with age-and sex-matched healthy controls. The mortality rate was 3 times higher in BD patients younger than age 35 years as compared with healthy controls. In patients older than age 35 years, the mortality rate tended to reach that in healthy controls. Interestingly, previous studies have reported that the severity of this disease abates with time (3, 10, 20).

Thus, the overall mortality in our cohort of patients with BD was 5% after a median followup of 7.7 years. The mortality rate at 1 year and 5 years was 1.2% and 3.3%, respectively. Male sex, arterial involvement, and a high number of flares were independently associated with mortality in BD. When compared with age-and sex-matched healthy controls, the mortality rate was 3 times higher in patients with BD younger than age 35 years.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Saadoun had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Saadoun, Wechsler.

Acquisition of data. Saadoun, Wechsler, Le Thi Huong, Amoura, Cacoub.

Analysis and interpretation of data. Saadoun, Wechsler, Desseaux, Le Thi Huong, Amoura, Resche-Rigon, Cacoub.

Ancillary