Effects of Long-Term Storage Time and Original Sampling Month on Biobank Plasma Protein Concentrations

Technical stability study to look for significant covariate effects of sampling strategy and sample storage on serum protein levels. The focus of this paper is not on the specific cohorts used or markers assayed, but rather to establishbroader principles for the use of biobank samples. The authors state the importance of understanding the potential covariates involved for longitudinal studies, or retrospective studies with samples having different storage methods and/or different storage times. Looked at the effects on serum protein levels of sample storage time, chronological age of subjects at the time of sampling, and the season/month of the year at sampling time. Long sample storage (2-28 years) affected the levels of 18 out of 122 proteins and was estimated to account for ~5-35% of the observed variance. Subject age at sampling (after adjustment for storage time) affected 70 proteins, accounting for 1-33% of the variance. After adjusting for storage time and age, season and month of sampling time affected 15 and 36 proteins, respectively (unclear whether sampling time affects protein levels due to UV levels, pollen levels or both). They conclude that sample handling history should be treated with equal importance to covariates such as age or gender when using biobank samples in proteomics-based epidemiological studies.