Analysis of molecular markers in biological fluids has been proposed as a tool for early detection and monitoring of cancer. Circulating plasma DNA concentrations have been found to be higher in cancer patients than in cancer-free control subjects, but little is known about the effect of specimen storage on plasma DNA concentrations. Here we investigated the impact of long-term storage of both plasma samples and purified plasma DNA on the reproducibility of plasma DNA quantification as determined using real-time polymerase chain reaction analysis. The analysis was performed on samples from a subset of 34 lung cancer patients and 28 matched control subjects selected from 200 subjects in our previously published case-control study and from 117 cancer-free smokers enrolled in a lung cancer screening program. Two samples of plasma and isolated DNA were assessed for each patient, with a median of 41 months between the first and second assessments for participants in the case-control study and 9 months for participants in the screening study. DNA levels declined substantially between the two assessments at an average rate of approximately 30% per year. These data provide valuable information for the rational planning of retrospective studies of banked series of biological samples, particularly if collected over a long period of time, as can occur in large clinical trials.