The CTLA-4 gene encodes a T-cell receptor that is involved in the regulation of T-cell activation. Recent studies have demonstrated an association of a microsatellite polymorphism [variant lengths of a dinucleotide (AT)n repeat] lying in exon 3 of the CTLA-4 gene, specifically a 106-bp allele, with autoimmune disorders, such as Graves' disease, autoimmune Addison's disease and autoimmune hypothyroidism. The aim of the present study was to determine whether the same polymorphism of the CTLA-4 gene was associated with vitiligo, a disorder that may have an autoimmune origin and can be present in patients who have one or more autoimmune diseases. CTLA-4 gene microsatellite polymorphisms were determined for 74 vitiligo patients (53 without any autoimmune disorder; 21 with one or more autoimmune disease) and 173 healthy controls, who had no clinical evidence of either vitiligo or any other autoimmune disorder, by polymerase chain reaction amplification of genomic DNA and resolution of the products on polyacrylamide sequencing gels. The frequency of the 106-bp allele was significantly increased (chi2 = 5. 2; P = 0.02) in vitiligo patients as a whole, in comparison with control subjects. However, when the patients were classified according to the absence or presence of an autoimmune disorder, the frequency of the 106-bp allele was significantly increased (chi2 = 14.8; P = 0.0001) only in the group of vitiligo patients who also had an autoimmune disease. However, the fact that 17 of 21 patients also had either autoimmune thyroiditis or autoimmune Addison's disease probably accounts for the apparent association of vitiligo and the 106-bp allele in this patient group. However, it was found that the relative risk of 3.2, conferred by the 106-bp allele in this group of vitiligo patients, was greater than that found for patients with only either Graves' disease, autoimmune hypothyroidism or autoimmune Addison's disease, with values of 2.1, 2.2 and 2.2, respectively. For the group of patients without an autoimmune disorder, there was no significant difference (chi2 = 0.2; P = 0.64) in the frequency of the 106-bp allele when compared with control subjects. These results indicate that there is no association between the 106-bp allele and vitiligo, at least when the disorder occurs in the absence of an autoimmune disease.