Indirect estimation of serum enzymes reference intervals in adults using the reflimR and refineR algorithms

Area of Science:

• Clinical Chemistry
• Laboratory Medicine
• Biostatistics

Background:

• Clinical laboratories often use manufacturer-provided reference intervals (RIs) due to cost and logistical challenges.
• Indirect estimation methods provide a feasible alternative for deriving RIs from available laboratory data.
• Accurate RIs are crucial for interpreting patient results and guiding clinical decisions.

Purpose of the Study:

• To estimate RIs for eight serum enzymes using the R-based algorithm reflimR.
• To compare the performance of reflimR with refineR, manufacturer's Instructions for Use (IFU), and direct RI estimation methods.
• To evaluate the utility of reflimR for indirect RI estimation in a clinical laboratory setting.

Main Methods:

• Retrospective analysis of adult outpatient data for eight serum enzymes (ALP, ALT, amylase, AST, CK, GGT, LDH, lipase) from January 2021 to May 2022.
• Estimation of RIs using the reflimR and refineR algorithms.
• Comparison of estimated RIs with manufacturer's IFU and directly derived RIs.
• Statistical analysis including Shapiro-Wilk test for data distribution, Mann-Whitney U test for group comparisons, and Spearman's correlation.

Main Results:

• Sex-specific RIs were necessary for ALP, ALT, AST, CK, and GGT.
• ReflimR generally produced wider intervals compared to refineR.
• Agreement between reflimR and refineR was 88.5%, with parametric RIs at 72.7% and IFU-based RIs at 62.5%.
• The lowest agreement was observed with non-parametric RIs (55.0%).

Conclusions:

• ReflimR is a practical tool for indirect RI estimation using routine laboratory data.
• Its performance is comparable to refineR and parametric methods, supporting its use for verifying or updating local RIs.
• This study is the first to apply reflimR to the Turkish population and compare it with refineR and IFUs, highlighting its potential in resource-limited settings.