Introduction: Cohort studies typically bank biospecimens for many years prior to assay and the levels of analyte degradation is unknown. Long storage times could cause degradation and produce measurement errors that would bias estimation of the analyte effects on health outcomes. Such effects could also produce confounding with other factors under study.

Methods: We collected control samples from 22 non-study participants using the same enrollment criteria and specimen collection, processing, and storage protocols as in The Sister Study, a large cohort study conducted by the National Institute of Environmental Health Sciences. Samples were assayed for 21 analytes at collection and then six years later. For each sample, the difference between the result at baseline and at six years was calculated for each analyte. The difference, Yij = result6 year - resultbaseline, was then modeled using a mixed-effects model, with random effects for batch (bi) and a fixed effect for the overall mean difference: Yij = u + bi + eij , where , bi ~ N(0, s2batch) , i=1,2,3,5,6.

Results: Some of the analytes experienced a marked change in concentration after six years of frozen storage, compared to their baseline value. There were no significant changes in ten of the analytes. Two of the analytes, lactate dehydrogenase (LDH) and sex hormone binding globulin (SHBG), increased significantly in concentration over time. Two analytes, HDL cholesterol and luteinizing hormone (LH), decreased significantly but within the limits of the internal laboratory control variance. Significant reductions in excess of the laboratory control variance were found for 7 of the 21 analytes tested (aspartate transaminase (AST), total cholesterol, estradiol, glucose, protein, sodium, and triglycerides). Despite the evidence for systematic changes over long-term storage, correlations between baseline and later measures were high and there was little relation between the size of the correlation and the estimated mean difference across the two time points.

Conclusion: We conclude there are differences in assay results after long-term storage, although for most analytes the correlation coefficients were high and percent change small. Biobanks or cohort studies that bank samples should consider building in QC experiments designed to assess the impact of long-term storage on anticipated analytes of interest, when possible. By doing so, degradation could be corrected for to control bias due to the associated measurement error. Since not all cohorts were able to conduct these QC experiments, those that have carried out such assessment should share their results so others can use the data to evaluate the potential impact of measurement error on their study findings.

Citation Format: Cynthia Kleeberger, David Shore, Elaine Gunter, Dale P. Sandler, Sandra Deming-Halverson, Clarice R. Weinberg. The effects of long-term storage on commonly-measured serum analyte levels [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 4239. doi:10.1158/1538-7445.AM2017-4239