Hemolysis in samples for blood gas analysis

Hemolysis causes clinically significant bias in four of twelve parameters generated during blood gas analysis. That is the headline finding of a recently published study that is, according to the authors, the first ever to investigate the effect of hemolysis on blood gas analytes. Venous blood was sampled from nine healthy volunteers and transferred to heparinized tubes. 

Half of each sample was submitted for blood gas analysis without further treatment, and the other half was mechanically hemolyzed by aspiration through a very narrow bore needle prior to blood gas analysis. The following twelve parameters were generated during blood gas analysis of both non-hemolyzed and hemolyzed samples: pH, pO₂, pCO₂, HCO3^-, p50, sO₂ , base excess, ctHb, COHb, MetHb, Ca²⁺ and K⁺. 

The blood that remained from the 18 samples after blood gas analysis was centrifuged and the recovered plasma samples were submitted for estimation of cell-free hemoglobin concentration (hemolysis index). The nine plasma samples recovered from pretreated blood returned a mean hemoglobin concentration of 8.9 ± 1.5 g/L. 

The equivalent concentration for plasma from all untreated blood samples was <0.5 g/L, thereby confirming a marked degree of hemolysis in the pretreated blood samples and absence of hemolysis in the untreated blood samples. Analysis of blood gas results revealed no statistically significant difference between hemolyzed and non-hemolyzed samples for the following four parameters: ctHb, p50, base excess and MetHb. 

For the remaining parameters there was a statistically significant bias. In the case of pH, pO₂, sO₂, COHb and Ca²⁺, hemolyzed samples returned values lower than non-hemolyzed samples, and in the case of pCO₂, HCO₃^- and K⁺, hemolyzed samples returned values greater than non-hemolyzed samples.

The observed biases was demonstrated to be clinically significant for only four parameters, pO₂ (mean bias: –4.9 %, range: –9.6 % to –0.2 %), pCO₂ (mean bias: +4.1 %, range: +1.7 % to +6.6 %), Ca²⁺ (mean bias: –7.0 %, range: –11.3 % to –2.8 %) and K⁺ (mean bias: +152 %, range: +150 % to +155 %). 

The marked effect of hemolysis on ionized calcium (Ca²⁺) and potassium (K⁺) measurement is very well documented and therefore not unexpected, but the effect on partial pressure of oxygen (pO₂) and partial pressure of carbon dioxide (pCO₂) is a novel observation. 

In discussion of their findings the authors concede that the degree of hemolysis tested in this study was ”remarkably high” and that a lesser degree of hemolysis may result in clinically insignificant bias for pCO₂ and pO₂.