Most true cases cluster along the ancient Silk Route: from Japan, where the disease is the leading cause of acquired blindness [3], to the Mediterranean littoral and areas of the former Ottoman empire [4]. Points in between (Iraq, Iran, China, and Korea) report additional thousands of cases (the Japanese and Turkish languages share Altaic roots, suggesting remote common ancestry of the people who speak these languages and perhaps explaining the genetic seeding along the Silk Route that connects the disparate areas of Asia and the Mediterranean coast) [5]. Western medical centers, even those with a special interest in the disease, accumulate only a handful of patients with Behcet disease; even then, some of these patients claim ancestors from Silk Route countries [6].

One way to distinguish Behcet disease from the Behcet syndrome (which has a presentation similar to that of Behcet disease but can be linked to some other named underlying disease) may be to limit the diagnosis of Behcet disease to patients who have ancestors from an area along the Silk Route (Turkey, the Balkan States, southern Italy, northern Africa, the Middle East, Iran, China, Korea, and Japan). This distinction is probably legitimate because genetic predisposing factors obviously play an important role in the disease; however, one must be wary of the possibility of circular reasoning because the distinction precludes a wider definition of the disorder. Nevertheless, patients from the identified populations at risk, although differing slightly in the variability of presentation depending on the areas they come from, have a more unfavorable prognosis than the patients sporadically described in the western world because of more severe uveitis, vascular changes, and demyelinating disease. In applying diagnostic criteria, moreover, one must remember that these criteria are intended for classification in studies, not necessarily for diagnosis of the condition of an individual patient in the physician's office.

Although arthritis presents as a minor manifestation of Behcet disease, rheumatologists have taken the lead in studying and treating the disease, generally in concert with ophthalmologists, dermatologists, neurologists, and other specialists [2]. No single diagnostic laboratory feature identifies the disease, but immunologists have focused on genetic features: the region of HLA-B51 in particular and heat shock protein genes [7, 8]. Genetics alone cannot be the entire story because in the large populations of Japanese persons and persons of Japanese descent in Hawaii and Brazil, Behcet disease is exceedingly rare; in contrast, the prevalence of the disease is 13.6 per 100 000 persons in Japan itself (range, >30 per 100 000 in Hokkaido to 9.9 per 100 000 in Kyushu) [9]. Studies of heat shock proteins have implicated an infectious precipitation, perhaps by streptococci or herpes simplex viruses [10].

The recurrent aphthae of the mouth, vulva and cervix, or scrotum can be extremely recalcitrant and are echoed in the cecum and right colon [11]. Because of demonstrable hyperreactivity of neutrophils [12] and the possible role of these cells in producing localized vasculitis responsible for the lesions [13, 14], colchicine, cytotoxic drugs, and corticosteroids have been prescribed, with indifferent success for the ulcerative lesions. As early as 1979, however, thalidomide curtailed attacks in reported series [15]. This drug, banned globally because of its teratogenic effects, still survived under highly controlled conditions as a result of its effectiveness in patients with lepromatous leprosy. It has recently emerged as a potential treatment for the ulcerative lesions and wasting of AIDS [16, 17]. In addition, as shown in a controlled study reported in this issue [18], the new availability of thalidomide has revived the drug's use in Behcet disease.

Unfortunately, thalidomide may not be the hoped-for cure because lesions recur after cessation of therapy. Studies must be done to determine whether a longer course of therapy or restarting therapy would lead to better results or whether combining thalidomide with another currently favored cytotoxic agent would enhance the success rate. There are hints that uveitis, epididymitis, and arthritis might also respond to thalidomide in time; erythema nodosum worsens, however, and the effects on caval thrombosis, arterial occlusions and aneurysms, and demyelinating disease are unknown [19]. Any further trials, especially of combination therapy, should be conducted only in areas where the prevalence of the disease is high, not in Europe and the Americas, where cases occur only sporadically. Yazici and colleagues, authors of the report published in this issue and longtime leaders in Behcet disease research, should be encouraged to continue these studies.

In the hands of knowledgeable and experienced Turkish, Middle Eastern, and Japanese investigators, thalidomide can be a valuable weapon against this disease. But what of U.S. and other western physicians who see only isolated cases, at most a handful? Enthusiasm for and rash adoption of thalidomide could lead to unnecessary treatment of patients who do not need such an approach and could create sequelae that would once again deprive us of this drug (not just because of teratogenicity but also because of neuropathy and hypersedation). Thalidomide should be given only when the diagnosis is firm (preferably confirmed by a second expert opinion) and under rigidly controlled protocols, perhaps with the advice of an ethics board. Meanwhile, the search for thalidomide analogues and mutants continues; this search includes attempts to separate the chiral components in the racemate in the hopes of creating safer but still effective drugs.

Grand-round cases divert us from the humdrum of everyday practice, but diagnostic eurekas must be tempered by therapeutic sobriety. Let us not kill a potentially useful drug, given a second chance, by injudicious prescribing.