[en] OBJECTIVES: In migraine, an interictal reduction of mitochondrial energy metabolism and a preventive effect of high-dose riboflavin were reported. To explore the relation between the two, we tested if the therapeutic response to riboflavin is associated with specific mitochondrial DNA (mtDNA) haplogroups. We focused our attention on haplogroup H, which is known to differ from others in terms of energy metabolism. METHODS: Sixty-four migraineurs completed a 4-month open trial with riboflavin (400 mg QD) and were genotyped blindly for mtDNA haplogroups. RESULTS: Forty patients responded to riboflavin treatment and 24 were nonresponders. The mtDNA haplogroup H was found in 29 subjects (20 migraine without aura, 9 migraine with aura). Riboflavin responders were more numerous in the non-H group (67.5%). Conversely, nonresponders were mostly H (66.7%). The difference between the two groups was significant (chi(2) = 7.07; p = 0.01). The presence of aura had no influence on riboflavin's effectiveness (chi(2) = 0.113; p = 0.74) and was not associated with a particular haplogroup (chi(2) = 0.55; p = 0.46). CONCLUSIONS: In this pharmacogenetic study, riboflavin appears to be more effective in patients with migraine with non-H mitochondrial DNA haplotypes. The underlying mechanisms are unknown, but could be related to the association of haplogroup H with increased activity in complex I, which is a major target for riboflavin. Our results may have ethnic implications, since haplogroup H is chiefly found in the European population.