AbstractBackground: Serum, as a standard material for the determination of numerous analytes, has disadvantages. The coagulation process leads to an artificial increase of the potassium concentration of approximately 0.3 mmol/L in serum samples compared to plasma. Consequently, plasma reflects the in vivo situation more accurately. The aim of the present analyses was to establish reference intervals (RIs) for potassium using data from a population-based study for serum (PS) and plasma (PP).Methods: Serum was used from 2897 subjects aged 20–79 years, participating in the 5-year follow-up of the Study of Health in Pomerania (SHIP 1), a population-based study in northeast Germany. In addition, 2483 samples (serum and plasma) from a population of blood donors (DONOR-SHIP) were used. Finally, calculated RIs were reevaluated in 202,350 potassium values from hospitalized patients. All measurements were performed on a Siemens Dimension RxL Max HM with ion-selective electrodes. Using the sample pairs from DONOR-SHIP, a regression formula for the transformation of PS to PP was calculated. This formula was applied to the serum data from SHIP 1 to calculate corresponding plasma values. RIs (2.5th and 97.5th percentile) were defined with quantile regression and bootstrap method in SHIP 1.Results: RIs for PS and PP were 3.7–5.1 mmol/L and 3.5–4.6 mmol/L, respectively. Clinically relevant age- or sex-specific tendencies were not detected. The difference between PS and PP is dependent on platelet count and potassium concentration.Conclusions: The study permitted the establishment of RIs for PS and PP on a population-based study. For serum, the influence of platelet count and absolute potassium concentration on the results should be taken into account.
Preventing misdiagnosis of diabetes in the elderly: age-dependent HbA1c reference intervals derived from two population-based study cohorts