To the Editor:

We read with interest the article by Amoura et al on the association of extracranial deep vein thrombosis with R92Q tumor necrosis factor receptor superfamily 1A (TNFRSF1A) mutation in Behçet's disease (BD) (1). In discussing the results, the authors refer to the in vitro findings that under inflammatory conditions tumor necrosis factor α (TNFα) may have an antithrombotic effect by stimulating inducible nitric oxide (NO) synthase, leading to a generation of NO (2). They suggest that the R92Q mutation may interfere with the local antithrombotic effect of TNFα. Although this is an attractive hypothesis, there are some problems with using it to explain vascular thrombosis associated with BD.

The authors analyze cerebral thrombophlebitis and extracranial thrombosis separately. Neuro-BD has 2 subtypes of clinical involvement (parenchymal and vascular), and vascular neuro-BD is accepted as being clinically similar to extracranial thrombosis. There seems to be no explanation for the difference in the observed frequencies of the R92Q mutation in cerebral thrombosis (0 of 10 patients) and noncerebral thrombosis (6 of 20 patients). A second point is that if TNFα has a role in vascular BD, other polymorphisms affecting the levels of TNFα secretion are also expected to have an association with thrombosis. TNF promoter allele TNF-1031C is associated with BD in Caucasian patients, with a relative risk of 2.3 (3). However, no association with any disease subtype is reported in that study (3). Similarly, no association of TNFα promoter region polymorphisms at positions −308 and −376 with vascular involvement is observed in patients with BD from Turkey (4).

However, the study by Amoura et al points to another issue, i.e., the role of background genetic load of the susceptible population and disease associations. Investigators in our group observed increased MEFV gene mutations (associated with familial Mediterranean fever, another autoinflammatory disorder) in patients with BD from Turkey compared with healthy controls (26% versus 9%), especially in patients with vascular involvement (55%) (5). Similarly, since the R92Q mutation is present mainly in Caucasian populations, an association with another inflammatory disorder such as BD can be expected; however, this may be difficult to confirm in populations with a higher prevalence of BD, but with no significant presence of the R92Q mutation.


  1. Top of page

Haner Direskeneli MD*, Guher Saruhan-Direskeneli MD†, * Marmara University Medical School, † Istanbul University, Istanbul, Turkey