Abstract
An audit of troponin measurement protocols in use in Wales showed significant variations in practice with no consensus on analytical methods, or in the selection of a decision limit for the diagnosis of myocardial damage. Peer review data on assay imprecision at concentrations approaching the analytical limit of detection are lacking. The objective of the study was to establish clinically relevant precision profiles for the troponin methods used throughout Wales, which could be used to develop a standardized approach to the selection of a decision limit. A series of five pools of human serum spiked with troponin I-T complex were prepared and stored at -7°C until despatch. Five sets of each pool were dispatched to all Welsh laboratories and stored at -20°C until analysis. The analysis protocol consisted of two replicates of each pool per batch, and two batches per day for five days (n=20). All the laboratories performed all the measurements in the same week. The lowest concentration providing a 10% coefficient of variation (CV) was 0.02 µg/L for the Roche method and 0.11 µg/L for the Beckman AccuTnI. The lowest concentrations that could be distinguishable from the 99th percentile reference limit were 0.02 µg/L for the Roche method and 0.07 µg/L for the Beckman AccuTnI. Current methods do not achieve the 10% CV at the 99th percentile reference limit proposed by the European Society of Cardiology. Use of the lowest concentration that can be reliably distinguished from the 99th percentile reference limit offers a novel alternative decision limit, which provides a slightly lower concentration than at the 10% CV but maintaining confidence in the assay that false-positive rates will be minimized.