Journal Logo


The Long-Term Mortality and Morbidity of Behçet Syndrome

A 2-Decade Outcome Survey of 387 Patients Followed at a Dedicated Center

Kural-Seyahi, Emire MD; Fresko, Izzet MD; Seyahi, Nurhan MD; Ozyazgan, Yilmaz MD; Mat, Cem MD; Hamuryudan, Vedat MD; Yurdakul, Sebahattin MD; Yazici, Hasan MD

Author Information
  • Free



A number of studies have addressed the mortality and morbidity of Behçet syndrome (1,2,4,6–8,10–12,14,18,22,24,25,27–29,31–38,40,41,43,46,47). Of these, only 1 was a true prospective cohort study (40); the authors reported that, among 2,031 patients from Japan, 31.7% of the patients clinically deteriorated, and 0.9% died during the course of a single year’s follow-up. In a retrospective cohort study (41), we reported an overall mortality of 3.9% among a group of 152 patients at 10 years. The remaining studies were mainly short-term and retrospective, aiming to assess either mortality or the outcome of a single organ system like the eye, central nervous system, or the pathology in major vessels.

In the current survey, we address both mortality and overall morbidity in a sizeable cohort of patients with Behçet syndrome over 20 years. To our knowledge, this is the first study to look at the outcome of Behçet syndrome over such a long time.

The main associates of mortality in Behçet syndrome are major vessel pathology and neurologic involvement. In an earlier retrospective cohort survey (22), we observed 50.0% mortality among 24 male patients with pulmonary artery aneurysms in 2 years after the onset of this complication. With a similar study design, Akman-Demir at al (1) reported a 20.0% mortality rate at 7 years among 15 patients with neurologic involvement. In contrast, the leading cause of morbidity in Behçet syndrome is eye involvement with the potential threat of visual loss. The reported rate of this loss has usually been quite high. In 1 study from Israel, 75.0% of patients had lost useful vision 6–10 years after the onset of symptoms, despite treatment (8).


Between 1999 and 2000, we evaluated the outcome of 428 consecutive patients whose first clinical visits were between July 1977 and December 1983 (inclusive) in the multidisciplinary Behçet syndrome outpatient clinic at Cerrahpasa Medical Faculty of the University of Istanbul, Istanbul, Turkey. This dedicated clinic was founded in 1977 and had been the first of its kind both locally and internationally. The main facility, the Cerrahpasa Hospital, is a 2,500-bed teaching hospital, and the referrals to our center, located in the rheumatology division of the department of medicine, are either from the various departments of the hospital or from outside physicians. The clinic has been held once a week since its inception. All patients are initially evaluated by a rheumatologist, dermatologist, and ophthalmologist. The patients are referred to the departments of neurology, pulmonary disease, gastroenterology, and vascular surgery when necessary, and all management problems are discussed in a joint meeting of representatives of the various disciplines during the same day the patients are seen. Access to the clinic is without charge if the patient does not have any financial coverage. The patients have to pay for their own medication (∼30% of our patients) if they do not have health insurance. Currently there are about 5,000 Behçet syndrome patients registered at the clinic, where many controlled drug trials in Behçet syndrome have been conducted (19,20,21,23,42,45).

We attempted to contact every patient in our cohort by telephone calls, home visits, or mail. All thus contacted were called back to the outpatient clinic for a formal evaluation. For patients who could not come to the hospital for reevaluation, a detailed telephone interview was conducted. A standard questionnaire was used for the patients who were evaluated in the clinic or by telephone calls. The questionnaire sought information on the clinical symptoms related to Behçet syndrome with special emphasis on eye, vascular, and neurologic disease and neoplasms. A formal physical examination was performed in patients who came to the hospital. Ophthalmology, dermatology, neurology, and vascular surgery consultations were also obtained in those instances where the patient had complaints or established pathology related to the corresponding subspecialty. All patients had fulfilled O’Duffy criteria for diagnosis (32) at their initial visit.

In 41 (9.6%) patients, no outcome information was available at the end of 2 decades. All outcome analyses were calculated only among the 387 patients in whom outcome information had been available at the end of 2 decades. This included the 42 patients who had died.

Assessment of mortality

Mortality was statistically analyzed only in males by calculating age standardized mortality ratios (SMR) (9). SMRs were calculated using information from the Turkish Mortality Table 1980–1990 (13). Patients were divided into 3 age-groups (14–24 yr, 25–34 yr, and 35–50 yr) based on age at initial visit, to assess the impact of age on SMRs. Two male patients who were older than 50 years at their first visits were excluded from the analyses. The SMR was determined according to the formula: Observed deaths in the Behçet population/Expected deaths in the age- and sex-matched Turkish population.

Ninety-five percent confidence intervals were calculated for the upper and lower boundaries. For male patients, 7 yearly SMRs were calculated alongside the SMR for the entire 2 decades to look for trends in change in mortality over time. No formal death statistics were made for females because of small numbers.

Assessment of visual outcome

Male and female patients were analyzed separately. To assess visual outcome, patients were stratified according to the presence or absence of eye disease at their initial visits to the clinic. Group I patients represented those without, and Group II those with, initial eye disease. Subgroup designations of A, B, and C stood for increasing degrees of severity for eye disease. Group II C patients (those with no useful vision at their initial presentation at the clinic; see Results) were not included in the outcome analyses.

All visual acuity scores were measured on a 10/10 Snellen scale. The most severe degrees of visual impairment like finger counting, light perception, and phthisis bulbi were counted as a vision acuity less than 1/10. A reporting of the loss of useful vision by the patient on the telephone interview was also taken as a visual acuity less than 1/10.

Kaplan-Meier analyses were performed to evaluate the time-dependent loss of useful vision. The main outcome in these analyses was the emergence of visual acuity less than 1/10 in both eyes (bilateral loss of useful vision).

The last recorded ophthalmologic evaluations from the charts were used to analyze the visual outcomes of the 21 patients with eye disease among the 42 patients who had died.

Outcome information about those patients in whom the last visual acuity could be obtained by a formal ophthalmologic evaluation and information about those in whom an assessment could be made only in a telephone interview were initially assessed separately. They were also pooled in the same Kaplan-Meier curves when it was observed that there were no significant differences in trend between the 2 assessments (see Results).

The emergence of new eye disease during the disease course was also evaluated. Only 2 patients (1 male and 1 female) reported such new eye disease during the telephone interviews. The remaining information concerning new eye disease was obtained by ophthalmologic examinations.

Assessment of mucocutaneous, articular, vascular, and neurologic manifestations

The details of the mucocutaneous, articular, vascular, and neurologic manifestations at the beginning of the study were determined from the charts. A formal inquiry into the current presence of these manifestations was made in 245 patients (153 male, 92 female) by outpatient clinic evaluation and in 100 patients (70 male, 30 female) by telephone interviews. For the final standings of the mucocutaneous manifestations and arthritis, only information obtained by the outpatient clinic reevaluation or by the telephone—and not those available on the charts—were considered in the analyses, and the data from physical examinations and telephone interviews were evaluated both separately and after being pooled. A manifestation was considered to be currently present if it occurred at least once during the previous year of the final clinic visit or the telephone interview.

For the vascular and neurologic lesions, however, information from the patient charts was included in the analyses, similar to what was done for the assessment of visual acuity.

We used different methods of data collection for mucocutaneous lesions and arthritis and for eye, central nervous system, and major vessel disease because of our contention that the usually less severe manifestations of mucocutaneous disease and arthritis would potentially have a lower chance of being recorded in the charts during follow-up.

Unemployment and dependence

Similar to what was done in mucocutaneous disease and arthritis, employment and dependence information was obtained from 345 patients (223 males, 122 females) at the end of 2 decades. The household labor force survey for the year 2002 of the Institute of Statistics of Turkey (5) was used to compare the observed unemployment rates with those given for the general population.

Dependence (needing another person’s assistance in his/ her daily activities) was defined as formerly outlined in a neurologic outcome study (2). Those patients who had lost their useful vision were included among the dependent.


Information about the drugs used before the first visit and during the follow-up was also assessed. Colchicine, corticosteroids, azathioprine, cyclosporine, and cyclophosphamide use were tabulated from the patient charts. Of the systemic medications, levamisole and sulfasalazine use were not formally assessed. They had been used in a small number of patients. Similarly, nonsteroidal antiinflammatory drug, acetazolamide, and topical medication use were not tabulated.

Statistical analysis

Quantitative variables were expressed as mean ± standard deviation (SD). Median and interquartile range (IQR) were used when the distribution was skewed. Differences between the Kaplan-Meier curves were evaluated by the log-rank tests. Chi-square tests were used to analyze nonparametric data.


Figure 1 is a flow chart detailing the evaluation process, while Table 1 shows the initial demographic and clinical characteristics of the study cohort. Male patients had significantly more papulopustular lesions, eye involvement, vascular disease, and arthritis, while females had more erythema nodosum. No outcome information was available in only 41 (9.6%) patients (see Figure 1).

Initial demographic and clinical characteristics of the cohort
The evaluation process.

Table 2 compares the initial clinical characteristics of the patients who could not be contacted at the end of 2 decades with the rest of the cohort. It is to be noted that those patients for whom no outcome information was available had similar clinical and demographic characteristics when they were first seen in the clinic.

Initial demographic and clinical characteristics of patients who were lost to follow-up compared with the rest of the cohort

Seventy-six percent (31/41) of the patients who could not be reached had lost contact with the outpatient clinic during the first 3 years after their initial visit. Before we started our survey, only 48 (11.2%) of our cohort had made a formal clinic visit within the previous 12 months. In the remaining 339/428 (79.2%) patients (excluding those patients who could not be reached for the purposes of this survey), the median time of outpatient clinic attendance after the first visit was 4.0 years (IQR: 1.0–12.0 yr).

By the end of 2 decades, 245 (153 male, 92 female) (57.2%) patients were reevaluated in the clinic while 100 (70 male, 30 female) (23.4%) patients were interviewed by telephone. A factor that seemed to be important in determining whether a patient came back to the hospital for reevaluation was the distance the patient lived from the center. There were significantly more patients living in Istanbul among those who could be reevaluated in the clinic (187/245, 76.3%) compared with those who were contacted by telephone (38/100, 38.0%) (χ2 = 4.31; p = 0.0001).


During a median of 19 years of follow-up (IQR: 17–21 yr), 39 of 262 (14.9%) male and 3 of 125 (2.4%) female patients died. No postmortem examinations were available. Mortality information for 14/42 patients (13 male, 1 female) (33.3%) was based on medical records (all died in the hospital), while for the remaining 28 patients (26 male, 2 female), it was obtained from immediate family members (first-degree relatives or spouses, 23/42 [54.8%]) or from other sources (more distant relatives or neighbors, 5/42 [11.9%]). Therefore, we label all cause of death information as “probable.”

Table 3 compares the clinical characteristics of patients who died with characteristics of those who remained alive. Patients who died had initially higher frequencies of eye, vascular, and nervous system disease. Table 4 shows that major vascular involvement and especially pulmonary arterial aneurysm were the main associates of mortality, followed by central nervous system involvement. It is also noted in this table that the information on cause of death was similar for the major organ systems, regardless of the information source.

Initial demographic and clinical characteristics of patients who died and patients still alive
Probable cause of death and source of information

The increased mortality among male patients was also apparent by the SMRs (Figure 2). For the 14–24 year old age-group, this reached 10 times that observed in the general population for the whole follow-up period. There was a trend for mortality to taper down, or at least stabilize, as a function of the time of follow-up (see Figure 2). This was true for each age-group.

Standardized mortality ratios (SMRS with 95% confidence interval) of a cohort of 260 male patients with Behçet syndrome. (Two patients were excluded because they were over age 50 yr at their initial visits; see text.) †Number of patients who had died.

Visual outcome

Figure 3 shows the mode of evaluation of eye disease and defines the patient subsets according to the presence or absence of eye involvement and its degree of severity at the time of the initial clinic visits. Of the total 221 patients who had eye disease, either present at the initial visit or developed during the course of this survey, outcome information on visual acuity was obtained by a formal ophthalmologic evaluation in 178/221 (80.6%).

Assessment of eye disease. Group I: Patients without eye disease at initial visit. Group II: Patients with eye disease at initial visit. Group II A: Patients with visual acuity of at least 5/10 in both eyes at initial visit. Group II B: Patients with eye disease with visual acuity of less than 5/10 in at least 1 eye but still maintaining some useful vision (visual acuity of 1/10 or more in at least 1 eye). Group II C: Patients with no useful vision (visual acuity less than 1/10) in both eyes at initial visit.

Male patients

Figure 4 shows that among 125/169 (74.0%) male patients, eye involvement had its onset within 2 years of the fulfillment of the O’Duffy criteria.

Emergence of eye involvement relative to the fulfillment of O’Duffy criteria in 169 male and 52 female patients. 0 = time the patient fulfilled O’Duffy criteria.

Among the 30 patients of Group II A (the group with less severe eye disease initially) who were evaluated by ophthalmologic evaluation, 6 (20.0%) had lost useful vision, as had 2/17 (11.8%) patients who were interviewed over the telephone (χ2 = 0.101; p =0.750). This was similar in Group II B, where 23/55 (41.8%) of the patients with ophthalmologic examinations had lost useful vision while 8/19 (42.1%) patients who had telephone interviews had lost useful vision (χ2= 0.0001, p =1.0). With no significant differences in outcome between the 2 methods of evaluation of visual outcome, we pooled the data from the 2 methods of data collection and also thus compared the visual acuities for Groups II A and II B (Figure 5a).

Loss of useful vision (<1/10) in both eyes over time. A. male patients;B. female patients. Group II A: Patients with visual acuity of at least 5/10 at both eyes at initial visit. Group II B: Patients with eye disease with visual acuity of less than 5/10 in at least 1 eye but still maintaining some useful vision (visual acuity of 1/10 or more in at least 1 eye).

In Group II A, 8/47 (17.0%) patients lost useful vision in both eyes during the follow-up. In 4 (8.5%) patients of the same group, the visual acuity in only 1 eye became less than 1/10 and remained at that level for 9, 10, 14, and 15 years, respectively.

In contrast to results in Group II A, among the Group II B patients, visual acuity in both eyes became less than 1/10 by the end of 2 decades in 31/74 (41.9%) (log rank p = 0.0098). In 8 patients (10.8%), visual acuity in only 1 eye became less than 1/10 and remained at that level for a median of 10 years (IQR: 3.6–16.3 yr). Finally, in 21 (51.2%) of the 41 patients of group II B with visual acuity less than 1/10 in 1 eye at the beginning, visual acuity remained greater than 1/10 in the better eye for a median of 19 years (IQR: 17.0–20.5 yr), without progressing to bilateral loss of useful vision.

Female patients

Among 32/52 (61.5%) female patients, the eye involvement had its onset within 2 years of the fulfillment of the O’Duffy criteria (see Figure 4). Fourteen patients in Group II A had ophthalmologic examinations while the remaining 3 had telephone interviews. No patient in either group had lost useful vision at the end of 20 years. In Group II B, 10 patients had ophthalmologic examination and 7 had telephone interviews. Two in each group had lost useful vision (χ2 = 0.0001; p = 1.00). The data were pooled as for male patients because there were no differences in trend in the data obtained by either method.

In 2/17 (11.8%) patients of Group II A the visual acuity in only 1 eye became less than 1/10 (Figure 5b) and remained at that level for 3 and 18.3 years, respectively, during follow-up. Thus neither of the patients progressed to have bilateral loss of useful vision.

Among 17 patients of group II B, the visual acuity in both eyes became less than 1/10 in 4 (23.5%) (see Figure 5b). In 4 (23.5%) patients in the same group the visual acuity in only 1 eye became less than 1/10 and remained at that level for 13, 17, 17, and 20 years, respectively. In 3 patients of group II B the visual acuity in either eye was already less than 1/10 at the first visit. In 1 of them the visual acuity remained less than 1/10 in only 1 eye for 20 years without progressing to bilateral loss of useful vision.

Eye disease during follow-up

The visual acuity in both eyes became less than 1/10 in 3 patients (13.0%) among the 23 males who had eye involvement for the first time during the follow-up. This outcome was similar to that observed among the Group II A patients, the group with less severe eye disease at their initial visit. In 1 patient (4.4%), the visual acuity in only 1 eye became less than 1/10 and remained at that level for 8 years.

Among the 14 female patients who developed new eye disease during follow-up, the visual acuity in only 1 eye became less than 1/10 and remained at that level for 12 and 10 years, respectively, in 2 patients (14.3%). None of the female patients in this category lost useful vision in both eyes. Thus the prognosis of eye disease that appeared late during the follow-up was more favorable in both sexes.

Bilateral versus unilateral eye disease among male and female patients

Eye involvement was bilateral in 134/169 (79.3%) male patients and 33/52 (63.5%) female patients at their first visits (χ2 = 4.57; p = 0.032). Similarly, at the end of 2 decades, 147 males (87.0%) and 37 females (71.2%) had bilateral eye disease (χ2 = 6.06; p = 0.01).

Mucocutaneous and articular manifestations

There was a statistically significant decrease in all mucocutaneous disease manifestations and in arthritis in patients of both sexes at the end of the study, regardless of the method of data collection (Table 5). Only the decrease in the number of patients with oral ulcerations was more pronounced among those interviewed over the telephone compared with those reevaluated in the hospital (176/245 [71.8%] versus 44/100 [44.0%], respectively; χ2 = 22.63; p = 0.0001).

Number of patients with mucocutaneous and articular manifestations at the time of initial and final visits*

It is also noteworthy that 94 of 345 (27%) patients (pooled for hospital reevaluations and telephone interviews) would have fulfilled O’Duffy criteria for diagnosis at the end of the follow-up, if one considered the disease manifestations that had been present only within the previous year (data not shown).

Vascular involvement

The bulk of information related to vascular pathology was based on medical records or hospital reevaluations. Only 3 males and 1 female interviewed by telephone reported new development of deep vein thrombosis of the lower extremities. In 2 of these patients the presence of deep vein thrombosis of the legs had been diagnosed by other physicians. The other 2 patients described chronic swelling of the legs with slow healing ulcerations. We interpreted these complaints as evidence for chronic thrombophlebitis.

The frequency of vascular involvement among male patients was considerably more than that observed among females (129/262 [49.2%] versus 7/125 [5.6%], respectively; χ2 = 69.00; p = 0.0001). Among 21 patients with arterial disease (pulmonary or peripheral arteries), there was only 1 female patient. This patient developed a Takayasu-like arterial occlusive disease at age 38 years, after 14 years of fulfilling criteria for diagnosis of Behçet syndrome.

A high proportion of patients with arterial disease also had venous disease. Seven of 10 patients with pulmonary artery aneurysms and 8 of 10 patients with abdominal and peripheral arterial disease had deep vein thrombosis of the lower extremities. On the other hand, occlusion of vena cava, either superior or inferior, was not seen in association with arterial disease, either pulmonary or peripheral. However, a significant proportion (33/36, 91.7%) of patients with vena caval disease had thrombosis of the deep veins of the extremities.

We noted that all patients who had been diagnosed with pulmonary artery aneurysms had hemoptysis as their initial symptom. Only 1 of 10 patients with pulmonary arterial aneurysms was alive by the end of 2 decades. He was symptom free and had a totally normal chest computed tomography scan after being treated with cyclophosphamide for 2 years.

In contrast, 8 of 10 patients with abdominal and peripheral arterial disease were symptom free within a median of 12.0 years (IQR: 6.1–17.5 yr) after surgical correction. One patient with an aneurysm of the abdominal aorta died immediately after the operation, and another with the same manifestation had a recurrence 5 years after a successful resection.

Medians of 7 and 5 years, respectively, were required for the onsets of arterial and vena caval disease after the fulfillment of the O’Duffy criteria (Table 6). This is unlike what we observed in eye disease, where about 83% of patients of either sex developed eye involvement within 5 years of fulfilling diagnostic criteria (see Figure 4).

Vascular Involvement

Neurologic involvement

All information on neurologic disease was based on charts and hospital reevaluations; none of the patients interviewed by telephone complained of new onset of neurologic involvement. The frequency of neurologic disease was significantly more among males compared with females (34/262 [13.0%] versus 7/125 [5.6%], respectively; χ2 = 4.12; p = 0.042).

Parenchymal involvement was the most common type of neurologic disease (Table 7). No patients had peripheral nervous system disease. Only 1 male patient had concomitant parenchymal disease and dural sinus thrombosis; no female patient had dural sinus thrombi. The median time of onset of neurologic involvement was, as for vascular disease, relatively late (for males: median, 5.5 yr, IQR: 1.3–11.0 yr; for females: median, 8.0 yr, IQR: 4.5–19.0 yr) during the disease course.

Neurologic involvement

Unemployment and dependence

Of the 223 male patients reached at the end of 2 decades (mean age: 50.9 ± 7.7 SD yr when interviewed), 103 (46.2%) were unemployed, whereas among the 122 females contacted (mean age: 50.9 ± 8.4 SD yr), 111 (91.0%) were not working. These percentages were considerably higher than the ∼10%–15% rate of unemployment reported for the general population of Turkey, for either gender (5).

Fifty-seven males among the patients reached at the end of 2 decades (25.6%), were dependent. In 44/57 (77.2%) patients, the loss of useful vision seemed to be the main reason behind this dependence, and in 3 (5.3%) patients, severe central nervous disease seemed to be the cause. However, 8/57 (14.0%) of the dependent males had combined severe eye and central nervous system disease, so the relative contribution of either organ system involvement to the total disability these patients experienced was difficult to ascertain. In addition, 2/57 (3.5%) male patients were dependent due to erosive osteoarthritis of the knees, probably secondary to episodes of inflammatory arthritis due to Behçet syndrome.

Ten females (8.2%) among the 122 interviewed were dependent at the end of 2 decades. Of these 7 had lost useful vision and 2 had severe central nervous system disease. None of these patients had combined eye and central nervous system disease. One final patient was dependent due to primary osteoarthritis, unrelated to Behçet syndrome.

It is noteworthy that none of the patients with vascular disease complained of being dependent at the end of 20 years.


Among 262 male and 125 female patients who were surveyed, 8 male patients had cancer (Table 8). Four had died. Among those, 2 had a history of cytotoxic drug use. One had used cyclophosphamide for 4 years before his first visit. The other had used chlorambucil for an unknown period of time before his initial visit and azathioprine for almost 2 years during follow-up.

Characteristics of patients who developed malignancies (all males)


Drug use during follow-up is shown in Table 9. Patients with ocular involvement (Group II) were treated mainly with corticosteroids and azathioprine. As expected, patients who initially had eye involvement (Group II) were prescribed more systemic drugs compared with patients who did not have eye involvement initially.

Systemic drug use*


We surveyed the 20-year outcome of a cohort of patients with Behçet syndrome. The O’Duffy criteria (32) had been used for diagnosis in this cohort, because the International Study Group Criteria (26) had yet to be formulated at the initial clinic visits of these patients. Nevertheless, at the initial visit all members of our cohort would also have fulfilled the International Study Group criteria in current use (data not shown).

Table 10 summarizes the results of the previous outcome studies in Behçet syndrome and compares the main results of this study with those previously reported. Our results are, in general, in conformity with those reported by others and by our group in the past. The information about more severe disease among male patients (7,25,27,31,35,37,41,43), the main causes of morbidity (1,2,4,6–8,10–12,14,18,22,24,25,27–29,31–38,41,43,46,47), the tendency for the disease to abate with the passage of time (27,29,34,43), the late onset of neurologic (2,36,38) and vascular disease (24,28,37) compared with mucocutaneous and eye disease, as well as the association of pulmonary artery aneurysms and deep vein thrombosis of the lower legs (14,22), have all been mentioned before. On the other hand, there is also considerable new information at hand.

Outcome studies in Behçet syndrome
(continued) Outcome studies in Behçet syndrome
(continued) Outcome studies in Behçet syndrome


We report here the highest mortality thus far in Behçet syndrome, to our knowledge. This probably is the direct result of the duration of follow-up and the diligent attempt to account for the outcome of every member of a large cohort. For example a study from Greece (27) reported no mortality among a smaller number of patients followed for shorter and varying periods of time. Similarly, in the past we had also said that mortality was not increased in Behçet syndrome (43).

A novel observation of the current study is that mortality, as measured by SMR, tended to decrease with the passage of time in all age-groups. This is in parallel with our data showing that some disease manifestations also decreased with the passage of time (see Table 5) and is different from the situation in some other chronic diseases, such as rheumatoid arthritis and systemic lupus, where increasing rates of mortality with respect to disease duration have been reported (30,39). This observation suggests that the disease-initiating insult(s), and the subsequent pathology, at least for eye, mucocutaneous disease, and arthritis, might be more short lasting in Behçet syndrome compared with that observed in these 2 diseases. This has obvious pathogenetic and management implications.

Eye disease

This survey formally confirmed the earlier impressions (11,25,31,33,37) that the onset of eye disease in Behçet syndrome is most of the time within years, if not at disease onset. This observation also might be tied to what we have just said about the “initiating insult.”

The more sinister prognosis of eye disease among male patients is well recognized (31,35,43), and was apparent in the current survey. Until recently, the prognosis of eye disease was thought to be quite poor (8,25,29,31,37,47). More recently Ando and colleagues (4) reported an improved prognosis. They suggested, as we point out here, that the visual acuity at the time of presentation was important in determining outcome. Furthermore the improved prognosis has been attributed to the more recent use of cyclosporine for eye disease (4).

Our data suggest that late-onset eye disease had a better prognosis. It can be said that this is due to, by definition, a shorter follow-up (median, 15 yr, IQR: 11.0–17.0 yr) in these patients compared with the rest of the cohort. However, useful loss of vision mostly occurs in the first 7 years of the disease, and the follow-up period in our patients was probably sufficient for the development of this complication (see Figures 5a and 5b).

Among those patients with eye disease from the start, bilateral loss of useful vision was 64/146 (43.8%) among male patients and 8/38 (21.1%) among females at the end of 2 decades. On the other hand, 25/146 (17.1%) of this same group of males and 4/38 (10.5%) of the females had already lost useful vision at the inception. Thus there were relatively few patients who had marked deterioration of visual acuity once proper treatment was initiated. As shown in Figures 5a and 5b, the initial vision seems to be clearly important in determining visual outcome. These figures also support our former suggestion that treatment started early at the beginning of eye involvement gives better outcome results (21). Even in those patients with initial unilateral loss of useful vision in 1 eye and severely restricted vision in the other, however, some visual acuity could be preserved in the better eye for many years. This observation, we believe, is another important outcome of this survey and contradicts the general pessimistic belief that severe eye disease almost always implies an eventual loss of useful vision in Behçet syndrome (29).

Finally, very few of our patients with eye disease had received cyclosporine at the time of their active disease, and it is most probable that the current outcome of eye disease is even better than what we report here.

Mucocutaneous lesions and arthritis

We are unaware of published reports about the outcome of mucocutaneous lesions in Behçet syndrome. We have observed that the frequency of all mucocutaneous lesions decreased in frequency with the passage of time. The same was true for arthritis. These findings strengthen the contention that the severity of the disease in Behçet abates with time.

Vascular disease

Of our male patients, 49.2% (129/262) had vascular involvement (arterial and venous). This figure is considerably higher than that previously reported in a large series (28). The relatively short follow-up times of the previous cross-sectional studies and the late occurrence of arterial and venous large-vessel disease (a median of 7 years for arterial disease and a median of 5 years for venous large-vessel disease) that we observed in our cohort, probably accounted for this difference.

It is instructive that only 1 of our female patients had arterial disease; a finding also noted by Hamza et al (24). Hemoptysis, a symptom that should be considered an emergency in young male patients, especially in areas where Behçet syndrome is endemic, was the most frequent manifestation of pulmonary aneurysms in our series. We should also point out that the grim outcome for pulmonary artery aneurysms noted in this survey, only 1/10 alive at the end of 20 years, has somewhat improved within the last decade, probably due to earlier recognition and prompt treatment (18).

Neurologic disease

The frequency of neurologic involvement in our cohort was 13.0% in males and 5.6% in females by the end of 2 decades. This frequency, as is the case with the frequency of the vascular disease, differs from the usually quoted frequency of neurologic involvement of about 5%–6% in Behçet syndrome (36,38). About two-thirds of the neurologic events emerged during follow-up and were not evident at the first visit. This also would not have come out in cross-sectional studies. The nature of the central nervous system manifestations observed in the current study was similar to what has been reported in 2 recent publications, again from Turkey (2,38).

Unemployment and dependence

We are aware of no previous studies on unemployment in Behçet syndrome, and, to our knowledge, there have been only 2 previous reports (2,38) on disability and dependence, both among neurologic cases. Siva et al (38), reporting on a later cohort from our clinic, observed a similar disability among patients with neurologic disease at the end of a mean follow-up of 2.9 ± 3.2 years. Data on unemployment and dependence in other diseases have been scarce in Turkey, so the information about dependence and unemployment in this series is difficult to interpret. On the other hand the unemployment rate of 46% among our male and 91% among our female patients at the end of 20 years is strikingly high by any comparison, and clearly merits further detailed and controlled data collection.


In our survey, we observed 8 patients (all male) with cancer (4 deaths due to cancer) among our analyzed cohort of 387 patients in 2 decades. This gives an annual incidence rate of 103/105 for both sexes and 153/105 among males. Firat and Celik (17) reported a crude yearly cancer incidence of 90/105 among the general population of Turkey in 1995. Also, recently, Fidaner et al (16) reported a 158/105 crude yearly incidence of cancer among the males in the Izmir region in Turkey, a figure similar to that found in our survey. This suggests that the frequency of cancer among male Behçet patients is not much different from that observed in the general population. However, the real cancer frequency in Behçet may be somewhat higher than what we have observed due to the proportion of patients in our study who could not be reached and those who died early due to causes directly related to Behçet syndrome.

Our relatively small sample size might account for our observation of no neoplasms among female patients.


Our survey indicated that we had managed most of our patients without the use of systemic drugs. Although we are unaware of formal reports, we suspect that our experience may be somewhat different than that of some other centers interested in Behçet syndrome. Our early (3) and recently confirmed (45) experience about the ineffectiveness of colchicine in some clinical settings may be a reason for this pattern of drug use. On the other hand, since our survey for drug use was entirely based on patient records, it is likely that the figures we give in Table 9 underestimate the drug use of these patients during the entire 2 decades. As we indicate in the Results section, many of our patients for whom we have outcome information had not been faithful attendees of the clinic. It is thus conceivable that they used medications prescribed by other physicians or simply had not sought medical care during the times they were away from our clinic. In brief, the drug use patterns as presented here should not be taken as true guidelines for patient management.

Limitations and conclusions

The 3 main methodologic concerns in our survey are 1) its retrospective nature, 2) the duality of data collection, and 3) the reliability of information related to mortality. We were able to obtain outcome information on 90.4% of our patients at the end of 20 years. We reason that this rather high retrieval rate lessens the impact of the retrospective nature of our study. It was also gratifying to note that the initial disease severity among the patients who could not be contacted at the end of 20 years was not significantly different from that of those who could be contacted (see Table 2). In fact, female patients, the sex group with less severe disease, were somewhat overrepresented among the patients lost to follow-up. Furthermore, the number of patients with oral ulceration was not as high (see Table 5) among those who were interviewed by telephone compared with those who had a formal physical examination at the end of 20 years. These suggest that severe disease and the associated disability were not the main reasons that determined whether a patient came for evaluation at the hospital or was interviewed by telephone. One significant reason for a telephone interview was the distance the patient lived from the hospital (see Results).

The information on visual acuity, the most important measure of outcome of eye disease, could be obtained by telephone interviews in only 43/221 (19.5%) of the patients with eye involvement. However, we used a rather conservative measure of visual acuity, loss of useful vision, as an outcome measure in all our eye assessments. This makes us think that the visual outcomes we report are also close to the true state of affairs. The similar trends observed in our data about eye disease, mucocutaneous manifestations, and arthritis between those patients who had telephone interviews compared with those who had hospital reevaluations, and the small number of patients (n = 4) who reported new vascular episodes in telephone interviews suggest that the duality of data collection was not a significant source of bias.

We obtained information about whether a patient was dead or alive at the end 20 years from a source other than the hospital record or a first-degree relative in only 5/42 patients (11.9%). Thus, we think that the data about mortality by itself were quite reliable. However, information about the specific cause of death was less solid in that none of our patients had autopsies. Permission to perform postmortem examinations is rather difficult to get in Turkey, presumably due to religious concerns and cultural idiosyncrasies. Thus, we chose to call all information related to the cause of death “probable.” The similarity, on the other hand, between the probable causes of death based on hospital records and on information obtained from first-degree relatives supports the reliability of the cause of death information at hand.

Finally, our survey reflects the clinical experience of 1 center following and treating almost exclusively patients of Turkish origin. Behçet syndrome shows considerable geographic variation in clinical expression. For example, the patients from the Far East have a higher (∼50%) frequency of intestinal disease (37) compared with those from the Mediterranean basin (46). Furthermore, many patients from the West have less complete forms of the disease, and it has been suggested that the term Behçet syndrome should be reserved for those sporadic patients, reserving the designation Behçet disease for those patients only from the areas where the condition is endemic (15). Thus, caution is in order in generalizing our observations made in an area where Behçet syndrome has the highest reported frequency (44).


1. Akman-Demir G, Baykan-Kurt B, Serdaroglu P, Gurvit H, Yurdakul S, Yazici H, Bahar S, Aktin E. Seven-year follow-up of neurologic involvement in Behcet syndrome. Arch Neurol 53: 691–4, 1996.
2. Akman-Demir G, Serdaroglu P, Tasci B. Clinical patterns of neurological involvement in Behcet’s disease: Evaluation of 200 patients. The Neuro-Behcet Study Group. Brain 122: 2171–82, 1999.
3. Aktulga E, Altac M, Muftuoglu A, Ozyazgan Y, Pazarli H, Tuzun Y, Yalcin B, Yazici H, Yurdakul S. A double blind study of colchicine in Behcet’s disease. Haematologica 65: 399–402, 1980.
4. Ando K, Fujino Y, Hijikata K, Izawa Y, Masuda K. Epidemiological features and visual prognosis of Behcet’s disease. Jpn J Ophthalmol 43: 312–7, 1999.
5. Basic indicators of household labor force survey, second quarter of 2002. Republic of Turkey, Prime Ministry State Institute of Statistics (SIS) web page. URL
6. Bayraktar Y, Balkanci F, Bayraktar M, Calguneri M. Budd-Chiari syndrome: A common complication of Behcet’s disease. Am J Gastroenterol 92: 858–62, 1997.
7. Benamour S, Zeroual B, Bennis R, Amraoui A, Bettal S. [Behcet’s disease. 316 cases] Presse Med 19: 1485–9, 1990.
8. Benezra D, Cohen E. Treatment and visual prognosis in Behcet’s disease. Br J Ophthalmol 70: 589–92, 1986.
9. Brennan P, Jones M. A guide to measuring mortality using standardized mortality ratios and survival analysis. Rheumatology in Europe 25: 5–8, 1996.
10. Chajek T, Fainaru M. Behcet’s disease. Report of 41 cases and a review of the literature. Medicine (Baltimore) 54: 179–96, 1975.
11. Cochereau-Massin I, Wechsler B, Le Hoang P, Le Thi Huong D, Girard B, Rousselie F, Godeau P. Pronostic oculaire de la maladie de Behcet. J Fr Ophtalmol 15: 343–7, 1992.
12. Colvard JM, Robertson DM, O’Duffy D. The ocular manifestations of Behcet’s disease. Arch Ophthalmol 95: 1813–7, 1977.
13. Duransoy L. 1980–1990 Turkish mortality tables. Unpublished PhD Thesis. Istanbul: Mimar Sinan University, 1993.
14. Efthimiou J, Johnston C, Spiro SG, Turner-Warwick M. Pulmonary disease in Behcet’s syndrome. Q J Med 58: 259–80, 1986.
15. Ehrlich GE. Behcet disease and the emergence of thalidomide. Ann Intern Med 128: 494–5, 1998.
16. Fidaner C, Eser SY, Parkin DM. Incidence in Izmir in 1993–1994: First results from Izmir Cancer Registry. Eur J Cancer 37: 83–92, 2001.
17. Firat D, Celik I. Cancer statistics in Turkey and in the world, Apa Ofset, 1998.
18. Hamuryudan V, Er T, Tanverdi M, Fresko I, Yurdakul S, Yazici H. Prognosis of the pulmonary arterial aneurysms in Behcet’s disease: An update. Ann Rheum Dis 60(Suppl): 55 (OP 0045), 2001.
19. Hamuryudan V, Mat C, Saip S, Ozyazgan Y, Siva A, Yurdakul S, Zwingenberger K, Yazici H. Thalidomide in the treatment of the mucocutaneous lesions of the Behcet syndrome. A randomized, double blind, placebo-controlled trial. Ann Intern Med 128: 443–50, 1998.
20. Hamuryudan V, Moral F, Yurdakul S, Mat C, Tuzun Y, Ozyazgan Y, Direskeneli H, Akoglu T, Yazici H. Systemic interferon alpha 2b treatment in Behcet’s syndrome. J Rheumatol 21: 1098–100, 1994.
21. Hamuryudan V, Ozyazgan Y, Hizli N, Mat C, Yurdakul S, Tuzun Y, Senocak M, Yazici H. Azathioprine in Behcet’s syndrome: Effects on long-term prognosis. Arthritis Rheum 40: 769–74, 1997.
22. Hamuryudan V, Yurdakul S, Moral F, Numan F, Tuzun H, Tuzuner N, Mat C, Tuzun Y, Ozyazgan Y, Yazici H. Pulmonary arterial aneurysms in Behcet’s syndrome: A report of 24 cases. Br J Rheumatol 33: 48–51, 1994.
23. Hamuryudan V, Yurdakul S, Rosenkaimer F, Yazici H. Inefficacy of topical alpha interferon in the treatment of oral ulcers of Behcet’s syndrome: A randomized, double blind trial. Br J Rheumatol 30: 395–6, 1991.
24. Hamza M. Large artery involvement in Behcet’s disease. J Rheumatol 14: 554–9, 1987.
25. Imai Y. Studies on prognosis and symptoms of Behcet’s disease in long-term observation. Jpn J Clin Ophthalmol 25: 665–94, 1971.
26. International Study Group for Behcet’s Disease. Criteria for diagnosis of Behcet’s disease. Lancet 335: 1078–80, 1990.
27. Kaklamani VG, Vaiopoulos G, Kaklamanis PG. Behcet’s Disease. Semin Arthritis Rheum 27: 197–217, 1998.
28. Koc Y, Gullu I, Akpek G, Akpolat T, Kansu E, Kiraz S, Batman F, Kansu T, Balkanci F, Akkaya S, Telatar H, Zileli T. Vascular involvement in Behcet’s disease. J Rheumatol 19: 402–10, 1992.
29. Mamo JG, Baghdassarian A. Behcet’s disease. Arch Ophthalmol 71: 38–48, 1964.
30. Minaur NJ, Jacoby RK, Cosh JA, Rasker JJ. Prospective study of rheumatoid arthritis: Function, disease activity and mortality after 40 years. Arthritis Rheum 43(Suppl): 385, 2000.
31. Mishima S, Masuda K, Izawa Y, Mochizuke M, Namba K. Behcet’s disease in Japan: Ophthalmological aspects. Trans Am Ophthalmol Soc 77: 225–79, 1979.
32. O’Duffy JD, Goldstein NP. Neurologic involvement in seven patients with Behcet’s disease. Am J Med 61: 170–8, 1976.
33. Pazarli H, Ozyazgan Y, Bahcecioglu H, Yazici H, Hattat N, Yurdakul S, Muftuoglu A. Ocular involvement in Behcet’s syndrome in Turkey. In: Lehner T, Barnes CG, eds. Recent advances in Behcet’s disease. International Congress and Symposium Series No. 103. London: Royal Society of Medicine Services, pp 267–8, 1986.
34. Pivetti Pezzi P, Gasparri V, De Liso P, Catarinelli G. Prognosis in Behcet’s disease. Ann Ophthalmol 17: 20–5, 1985.
35. Sakamoto M, Akazawa K, Nishioka Y, Sanui H, Inomata H, Nose Y. Prognostic factors of vision in patients with Behcet disease. Ophthalmology 102: 317–21, 1995.
36. Serdaroglu P. Behcet’s disease and the nervous system. J Neurol 245: 197–205, 1998.
37. Shimizu T, Ehrlich GE, Inaba G, Hayashi K. Behcet disease (Behcet syndrome). Semin Arthritis Rheum 8: 223–60, 1979.
38. Siva A, Kantarci OH, Saip S, Altintas A, Hamuryudan V, Islak C, Kocer N, Yazici H. Behcet’s disease: Diagnostic and prognostic aspects of neurological involvement. J Neurol 248: 95–103, 2001.
39. Urowitz MB, Bookman AA, Koehler BE, Gordon DA, Smythe HA, Ogryzlo MA. The bimodal mortality pattern of systemic lupus erythematosus. Am J Med 60: 221–5, 1976.
40. Yamamoto S, Toyokawa H, Matsubara J, Yanai H, Inaba Y, Nakae K, Ono M. A nation-wide survey of Behcet’s disease in Japan. 1. Epidemiological survey. Jpn J Ophthalmol 18: 282–90, 1974.
41. Yazici H, Basaran G, Hamuryudan V, Hizli N, Yurdakul S, Mat C, Tuzun Y, Ozyazgan Y, Dimitriyadis I. The ten-year mortality in Behcet’s syndrome. Br J Rheumatol 35: 139–41, 1996.
42. Yazici H, Pazarlı H, Barnes CG, Tuzun Y, Ozyazgan Y, Silman A, Serdaroglu S, Oguz V, Yurdakul S, Lovatt GE, Yazici B, Somani S, Muftuoglu AU. A controlled trial of azathioprine in Behcet’s syndrome. N Engl J Med 322: 281–5, 1990.
43. Yazici H, Yurdakul S, Tuzun Y, Pazarli H, Altac M, Ozbayrak H, Ozyazgan Y, Ulku B, Ozdogan H, Muftuoglu A. Influence of age of onset and patient’s sex on the prevalence and severity of manifestations of Behcet’s syndrome. Ann Rheum Dis 43: 783–9, 1984.
44. Yurdakul S, Gunaydin I, Tuzun Y, Tankurt N, Pazarli H, Ozyazgan Y, Yazici H. The prevalence of Behcet’s syndrome in a rural area in northern Turkey. J Rheumatol 15: 820–2, 1988.
45. Yurdakul S, Mat C, Tuzun Y, Ozyazgan Y, Hamuryudan V, Uysal O, Senocak M, Yazici H. A double-blind trial of colchicine in Behcet’s syndrome. Arthritis Rheum 44: 2686–92, 2001.
46. Yurdakul S, Tuzuner N, Yurdakul I, Hamuryudan V, Yazici H. Gastrointestinal involvement in Behcet’s syndrome: A controlled study. Ann Rheum Dis 55: 208–10, 1996.
47. Zierhut M, Sal J, Pleyer U, Kotter I, Durk H, Fierlbeck G. Behcet’s disease epidemiology and eye manifestations in German and Mediterranean patients. German J Ophthalmol 4: 246–51, 1995.
© 2003 Lippincott Williams & Wilkins, Inc.