Highly active antiretroviral therapy (HAART) can cause mitochondrial toxicity. The concentration of mitochondrial DNA (mtDNA) in peripheral blood cells has been reported to be a marker of this toxicity. However, these observations are controversial and were drawn from small series. Thus, we analysed the value of blood mtDNA as a marker of mitochondrial toxicity in a large cohort of human immunodeficiency virus (HIV)-infected out-patients during routine clinical evaluations. Real-time quantitative PCR was used to determine the mtDNA to nuclear DNA (nDNA) ratio in peripheral blood mononuclear cells from 157 consecutive HIV-1-infected patients (13 naive, 144 receiving HAART) and 30 HIV-1-uninfected patients. The mtDNA to nDNA ratio was significantly lower in both groups of HIV-infected patients than in the control group. No significant difference was observed between treated and naive HIV-infected patients. Lactataemia was significantly lower in controls than in the group of HIV-treated patients. None of the treated patients had lactataemia >5 mmol/l or bicarbonates <20 mmol/l. Triglyceride levels were significantly higher in the HAART-treated patients than in the nontreated patients. Clinical symptoms of lipodystrophy were observed in 62 HAART-treated patients. These symptoms were not associated with an abnormal mtDNA to nDNA ratio or plasma triglyceride concentration. The mtDNA to nDNA ratio was lower in DDI/D4T-treated patients than in AZT/3TC-treated patients. In conclusion, there are no obvious links between the mtDNA to nDNA ratio in peripheral mononuclear cells and any clinical symptoms or lactate level. Thus, the mtDNA to nDNA ratio in leukocytes does not seem to be an accurate marker of mild and/or long-term mitochondrial toxicity.