Accurate blood-borne biomarkers are sought for diagnosis, prognosis and treatment stratification. Consistent handling of blood is essential for meaningful data interpretation, however, delays during processing are occasionally unavoidable. We investigated the effects of immediately placing blood samples on ice versus room temperature for 1 h (reference protocol), and holding samples on ice versus room temperature during a 3 h delay to processing. Using Luminex multi-plex assays to assess cytokines (n = 29) and diabetes-associated proteins (n = 15) in healthy subjects, we observed that placing blood samples immediately on ice decreased the serum levels of several cytokines, including PAI-1, MIP1-β, IL-9, RANTES and IL-8. During a delay to processing, some analytes, e.g. leptin and insulin, showed little change in serum or plasma values. However, for approximately half of the analytes studied, a delay, regardless of the holding temperature, altered the measured levels compared to the reference protocol. Effects differed between serum and plasma and for some analytes the direction of change in level varied across individuals. The optimal holding temperature for samples during a delay was analyte-specific. In conclusion, deviations from protocol can lead to significant changes in blood analyte levels. Where possible, protocols for blood handling should be pre-determined in an analyte-specific manner.