LAUSR

We note with interest the technical note from Ilardo et al. [1] regarding the correction of potassium results for the effect of collection haemolysis. We are however concerned that their approach has some significant deficiencies related to a failure to appreciate the uncertainties involved in the correction process [2]. A major issue lies with their method for the haemolysis interference study. This involved pooling serum samples at varying potassium levels and spiking with a single preparation of haemolysate. This approach does not account for variations in the ratio of potassium to haemoglobin between individuals. Dimeski et al. [3] showed that the red cell potassium/haemoglobin ratio was variable between individuals with a range between 0.0029 and 0.0053mmol/10mg (i.e. potassium in mmol/L, haemoglobin, measured using the Roche H index, in mg/dL). Hawkins also showed a range of 0.002–0.0035mmol/10mg for hospital patients [4,5], as opposed to the fixed figure of 0.004mmol/10mg Ilardo obtained [1]. On top of this is the possible effect of the method of lysing the cells, with the addition of water and freeze/thawing giving an average of 0.0028 [4,5] and physical lysis 0.004 [3], both of which are lower than seen in the comparison between real haemolysed and non-haemolysed samples where a value of 0.0051 was found [6]. Additional factors that may add to the uncertainty are the white cell count, the precision of haemoglobin measurement (likely to be small), and any factors associated with the collection that may be more common with haemolysed samples, such as tissue damage due to a difficult collection. While a corrected result may be better than an uncorrected result, awareness of the uncertainties will provide important information on the limitations of any correction [2]. Disclosure statement