In this case-control study, we observed a statistically significant association between low peripheral blood mtDNA copy number and increased risk of RCC. Our finding replicates that of an earlier hospital-based study that found individuals with mtDNA copy number below the median among controls to have a 50% increased risk of RCC, with a dose-response relationship between lower mtDNA copy number and increasing risk 
. In our study, the association with low mtDNA copy number appeared to be stronger for men than women, and for White than for African American participants, although tests of interaction were not statistically significant. Xing et al. did not report any differences in this association between men and women in their study, which consisted entirely of White participants. They did note findings suggestive of effect modification by smoking, which we did not observe in our study.
The association with blood mtDNA copy number observed in these two case-control studies is consistent with evidence that mtDNA copy number is decreased in RCC cells. Reduced mtDNA content in RCC tissue relative to adjacent normal tissue was detected by Meiehofer et al. for 34 of 37 cases and by Selvanayagram et al. for 8 of 13 cases 
. Selvanayagram et al. also observed mitochondrial depletion in five of six RCC cell lines 
. The findings by Meierhofer et al. were independent of stage, metastasis, ploidy or proliferative activity, suggesting that the mechanisms associated with mtDNA depletion are early carcinogenic events which do not change with tumor progression. There are several possible mechanisms associated with low mtDNA copy number that could potentially contribute to RCC development, including resistance to apoptosis 
, up-regulation of nuclear factor-kappa B signaling 
, and increased production of reactive oxygen species 
. Further research will be necessary to fully elucidate the underlying pro-neoplastic mechanisms.
Strengths of our study include its large sample size, the inclusion of African Americans, and its population-based design. A limitation intrinsic to our study design, and that of Xing et al. 
, is the use of blood collected after diagnosis. As a consequence, we cannot determine whether the association with low peripheral blood mtDNA copy number reflects an etiologic mechanism or an effect of disease. Our finding for low mtDNA remained when we restricted the analysis to cases with localized disease and those treated by surgery only, and mtDNA copy number was not associated with time from diagnosis to blood collection. While these results argue against the presence of bias due to tumor- or treatment-induced effects, we nonetheless cannot rule out reverse causation as an explanation for our findings. Prospective studies of mtDNA copy number and RCC involving pre-diagnostic blood specimens are needed to resolve this question.
In conclusion, our case-control study offers confirmatory evidence that low peripheral blood mtDNA copy number is associated with RCC risk. These findings support a role for mitochondrial effects in RCC development. More research is needed to understand the biologic mechanisms underlying this association, including investigations involving peripheral blood, tumor and adjacent normal kidney tissue, and to assess whether the association is replicable in prospective studies.