Reasons for performing study: Time delays between collection of blood samples and biochemical analysis of equine blood are unavoidably common in equine practice. The effect that delays may have on the accuracy of results of blood biochemical analyses is not well established.
Hypothesis: Delays in processing of blood of up to 72h results in alterations in measured levels of common biochemical analytes that are of potential clinical relevance. Separation of serum prior to storage is protective against the effects of time delays.
Methods: Samples of clotted blood, separated serum and oxalate fluoride plasma from 20 horses were stored and analysed at 0, 24, 48 and 72 h. Graphical exploration of each analyte was undertaken. General linear models with fixed effects were fitted for the whole blood data. The mean bias and 95% limits of agreement were calculated, using bootstrapped data, to assess agreement between pairs of samples analysed at 0 h and other time points. Bland-Altman plots were used to explore general trends in the data. Paired t tests were used to compare the results from whole blood and separated serum.
Results: Delays in processing equine blood resulted in significant increases in measured concentrations of aspartate aminotransferase, creatine kinase, lactate dehydrogenase, total bile acids and magnesium. A significant decrease in concentration was identified for glucose (serum and oxalate fluoride preserved plasma). Separation of serum immediately following clot formation resulted in nonsignificant increases in accuracy for some analytes.
Conclusions and practical significance: Delays in processing of blood samples may result in biochemical changes of clinical relevance in individual cases; however, in the majority of cases, where delays are only a few days and a number of analytes are assessed concurrently, delays are unlikely to have an effect on the interpretation of results. Separation of serum following clot formation is of limited benefit. Clinical samples in which a delay in processing has occurred may be interpreted with reference to the data presented.