Background: The prevalence of hemolysis, icterus, and lipemia (HIL) was determined for residual whole blood specimens analyzed for clinical chemistry parameters on blood gas analyzers. The frequency and potential impact of exogenous interference from iodide, salicylate, and thiocyanate (metabolite of sodium nitroprusside) on analysis of whole blood chloride was also assessed.
Methods: Over an approximately two month period at an academic medical center, indices for HIL were determined on Roche cobas c502 analyzers for 1,986 residual whole blood specimens that had been previously analyzed for clinical chemistry parameters on Radiometer ABL90 FLEX blood gas analyzers. To examine exogenous interferences, retrospective analysis was performed over multiple years to ascertain whether patient samples analyzed for whole blood chloride were potentially affected by interference from iodide, salicylate, or thiocyanate.
Results: Some degree of hemolysis (defined as hemolysis index of greater than 60) was present in 9.7% of the whole blood specimens. Increasing rates of hemolysis were associated with higher whole blood potassium concentrations. Nearly 60% of specimens with potassium concentrations between 6.0 and 6.9 mEq/L had hemolysis indices of 100 or greater, and 75% of specimens with a potassium concentration of 7.0 mEq/L or greater were severely hemolyzed (hemolysis index of 300 or greater). In contrast to the hemolysis results, icterus and lipemia were determined to have minimal impact on patient results. For the exogenous interferences, we did not identify any patient samples where elevated salicylate levels or pharmaceutical iodide administration overlapped with whole blood chloride analysis (out of 75,887 and 169,229 total chloride measurements, respectively). We did, however, find that for patients receiving nitroprusside therapy in the inpatient setting, whole blood chloride concentrations were significantly higher during nitroprusside therapy [106.7 +/- 6.2 mEq/L (mean, SD)] compared to before or after nitroprusside therapy (103.1 +/- 7.7 mEq/L).
Conclusions: In this analysis of whole blood specimens, hemolysis is a common interference and likely to introduce meaningful biases, as illustrated with potassium analysis. Icterus, lipemia, salicylate, and iodide appear unlikely to cause clinically significant bias. Nitroprusside therapy introduced a slight rise in whole blood chloride concentrations that probably has minimal clinical significance.