We compared the quality of reference measurements for serum potassium in four reference laboratories from three different European countries, using a panel of 60 native patients' samples. The reference methods were based on either ion chromatography (one laboratory) or flame atomic emission spectrometry (three laboratories). Performance specifications for serum potassium measurements were defined as a maximum overall coefficient of variation (CV) of 1.5%, a maximum bias of 0.65% and a maximum total error of 3.0%. The overall imprecision for all laboratories was in the range of 0.7 to 1.3%, and was thus below the proposed specification of 1.5%. However, two laboratories reported 12 and 13 quadruplicates with CVs exceeding this limit. The mean bias (expressed as deviation from the overall mean of all laboratories) for all reference laboratories was < 0.65%. In the lower concentration range, however, one laboratory exceeded this limit. No laboratory measured samples with a total error above 3.0%. From these results, it can be concluded that the reference measurements, and, thus, also the reference methodologies, based on ion chromatography and flame atomic emission spectrometry were equivalent, and able to satisfy current analytical specifications for serum potassium measurements.