Insufficient mixing of thawed serum samples leading to erroneous results - experience from a field study and use of a correction procedure.

Incorrect analysis results that are close to expected might not be recognized in scientific studies or routine patient care. In two field studies we obtained unexpected results in a large number of samples. The present study aimed to identify the source of error in the samples from these studies and to validate a method to obtain correct results. Pre-analytical procedures were scrutinized, giving no indications of inappropriate pre-analytical sample handling in the field or during transport in a tropical climate. Using a new set of samples from volunteers in simulation experiments, we observed the known concentration gradient of analytes sampled in gel as well as plain tubes after freezer storage and thawing. Experiments demonstrated that mixing of samples by vortexing alone was not sufficient to disrupt the gradient formed by freezing and thawing, which appeared to cause the problem encountered when we in field studies analyzed and biobanked large sample sets by robot pipetting. A correction procedure was introduced, in which the obtained value of an analyte was multiplied by a correction factor calculated for each sample using the expected sodium level (140 mmol/L) divided by the measured sodium value. When it was validated on results from the simulation experiments, we repeatedly found that the correction lead to results very close to true values for analytes of different size and charge. Usefulness of the procedure was demonstrated when applied to a large set of field study results.