Stability of 40 cytokines/chemokines in chronically ill patients under different storage conditions.

Numerous studies point to the utility of blood cytokine measurements in the diagnosis and treatment of human disease. Advances in detection allow robust multiplex analysis. However, cytokines are present at low levels and are produced and act in complex networks which can remain active in stored blood. A major barrier to the routine use of cytokines as clinical biomarkers is sample management prior to analysis. Studies on cytokine stability under storage frequently use 'spiked' normal control plasma or serum to generate detectable levels of the cytokines of interest. These conditions may oversimplify the reality of clinically complex samples and provide limited information regarding optimal management of whole blood samples prior to plasma separation. Cytokine stability has also been addressed previously using plasma from normal individuals under different conditions of anticoagulant use, storage time and temperature of storage. No studies have as yet been undertaken to address cytokine stability in critically ill patients which may differ from normal, healthy individuals due to underlying cofounders such as inflammation. To address these issues, we subjected samples from five patients exhibiting an inflammatory disease state to three storage extremes which might be encountered in a clinical setting, prior to analysis of 40 cytokines. Blood drawn into EDTA or ACD anticoagulant was immediately separated and plasma stored at -80 °C. Matched samples were stored as follows; whole blood overnight at room temperature, or whole blood or plasma stored 10 days at 4 °C. We used equivalence testing to determine the similarity of stored cytokine values to baseline values. In ACD plasma, Eotaxin, IL-6, IL-11, IL-15, IP10, MDC, MCP-1 met equivalence to baseline in all storage conditions while for EDTA plasma stored 10 days at 4 °C EGF, FGF2, Fractalkine, G-CSF, IL-1β, IL-5, IL-6, IL-7, IL-11, IP-10, TGFα and TNFα showed equivalence to baseline measurements. Intraclass Correlation Coefficients (ICC) were calculated and the following cytokines showed good to excellent agreement across all 4 storage conditions: Eotaxin, IL-5, IL-6, IL-11, IL-13, IP-10, MCP-1 and TNFα when collected in EDTA, and Eotaxin, IL-5, IL-6, IL-11, IL12-p40, IL-15, IP-10 and MCP-1 when collected in ACD. Five plasma cytokines were identified as being the least stable in both ACD and EDTA: IL-7, IL-9, IL12p70, RANTES, sCD40L, while IL-1β was identified as unstable stored in ACD plasma. This study identified several clinically important cytokines that are remarkably stable in blood and plasma, and some that stored poorly. To our knowledge, this is the first cytokine storage study to use medically unwell patient samples and equivalence testing to evaluate the stability of measured cytokine values after storage.