Non-linearity in lipase assays: A multicentric comparison on different analysers

Area of Science:

• Clinical Chemistry
• Biomedical Engineering
• Analytical Chemistry

Background:

• Non-linearity in lipase assays has been observed on Roche analysers, leading to gaps in results distribution.
• This phenomenon, however, has not been extensively studied across different analytical platforms.

Purpose of the Study:

• To investigate and quantify non-linearity in lipase assays across multiple automated clinical chemistry analysers.
• To evaluate the impact of non-linearity on lipase activity measurements and identify potential gaps in results distribution.

Main Methods:

• Preparation of eighteen lithium-heparinized plasma pools with decreasing lipase activities for multicentric evaluation.
• Modeling non-linearity using polynomial and linear regression relative to dilutions and manufacturer's limit of linearity (MLL).
• Graphical identification of distribution gaps via histograms and calculation of upper gap limits.

Main Results:

• Manufacturer's Limit of Linearity (MLL) and lipase activity at MLL varied significantly across analysers (e.g., Abbott Architect C16000: 1200 U/L, Roche Cobas c503: 300 U/L).
• Biases at MLL were generally acceptable (within ±14.2%), except for the Roche Cobas c503 (-17.3%) and Randox Lipase on Abbott Alinity (-18.6%).
• Specific reagent-analyser combinations showed varying performance, with Sentinel Lipase on Abbott systems demonstrating low biases.

Conclusions:

• Non-linearity in lipase assays can introduce significant biases, particularly at the manufacturer's limit of linearity.
• The Roche Cobas c503 and Randox method on Abbott Alinity exhibited unacceptable biases, indicating potential issues with linearity or reagent performance.
• Further investigation into assay linearity and performance across different platforms is crucial for accurate lipase activity determination.