Background: Previous reports show that parathyroid hormone (PTH) concentrations may vary widely depending on the assay used to assess PTH. In this cross-sectional study, we aim to determine the usefulness of standardizing blood handling for optimal interpretation of PTH in patients with chronic kidney disease.
Study design: Diagnostic test study.
Setting & participants: Predialysis blood was sampled in 34 long-term hemodialysis patients at a single academic medical center.
Index test: PTH was measured by using 6 different automated second-generation assays (Elecsys, Advia Centaur, LIAISON, Immulite, Architect, and Access assays), 3 blood specimen types (serum, EDTA plasma, and citrate plasma), and 2 consecutive days of measurement (after thawing and 18 hours later with samples having been let at room temperature).
Reference test: None.
Results: A mixed statistical analysis model showed that the nature of the assay (P < 0.001) and nature of the blood sample (P < 0.001) significantly influenced variability in PTH concentrations, whereas day of measurement (day 1 or 2) did not (P = 0.5). Most PTH variability was caused by observations (96.8%), then manufacturer's kit (2.5%), and last, specimen type (0.7%). PTH concentrations measured in citrate plasma were lower with every assay method used than those observed in serum or EDTA plasma. The interaction between manufacturer and specimen type was of moderate statistical significance (P = 0.04). To evaluate the potential clinical consequence of PTH measure variability, we classified patients according to Kidney Disease Outcomes Quality Initiative cutoff values (PTH < 150 pg/mL; PTH, 150 to 300 pg/mL; and PTH > 300 pg/mL). Overall, statistical classification agreement was moderate to high for comparison between assays and high to very high between different blood samples and between days of measurement. However, we found that up to 11 of 34 patients were classified in different categories with some assays (LIAISON versus Architect) and up to 7 of 34 in different categories with different blood specimen type (citrate plasma versus serum [corrected] in LIAISON assay).
Limitations: This is a cross-sectional study that used single lots of reagents. There currently is no reference method for the measurement of PTH and no recombinant PTH standard for PTH assay.
Conclusion: PTH variability caused by the nature of the assay and/or blood specimen type is large enough to potentially influence clinical decision making. A specified collection method therefore should be used for PTH measurements. In routine practice, we recommend serum PTH over EDTA or citrate plasma.