BACKGROUND: Genetic predisposition to dideoxynucleoside-induced mitochondrial dysfunction might be related to mitochondrial DNA (mtDNA) polymorphisms. Severe hyperlactataemia is probably the best model to assess such a predisposition. METHODS: For this exploratory study in White European and Black African HIV-infected adults, hypervariable region 1 of mtDNA samples from peripheral blood mononuclear cells or buccal smears of patients who have developed confirmed severe hyperlactataemia was sequenced. Additionally, 21 single nucleotide polymorphisms and a 9 bp deletion were genotyped to assign mtDNA haplogroups. Finally, entire mtDNA sequencing was performed in a subset of European samples. Samples were obtained from Black African cases and controls recruited from a single centre in Johannesburg, South Africa and from white European cases from Amsterdam, London and Zurich. RESULTS: A total of 40 cases and 38 controls from Johannesburg were included. All of the cases and 33 controls were receiving stavudine-based therapy at the time of the index date (P=0.024). The distribution of mtDNA haplotypes was not different between cases and controls (P=0.137), and neither were the predicted haplogroups (P=0.751). In total, 11 of the 12 European cases were on stavudine and/or didanosine at the time of the event. No hypervariable region 1 haplotype was consistently found in the European cases. Sequencing of the entire mtDNA from three of these cases supported the absence of any shared mutations other than major alleles frequently seen in the mtDNA database. CONCLUSIONS: We did not find an association between homoplasmic inherited mtDNA polymorphisms and severe hyperlactataemia. Our data do not support the existence of non-synonymous mtDNA mutations that explain an increased predisposition to dideoxynucleoside-induced mitochondrial dysfunction.