A-093 Assessing Alkaline Phosphatase Reference Intervals Using Gamma-Glutamyl Transferase as a Surrogate for Normal Liver Function
H Brown,N C Spies,C W Farnsworth
DOI: https://doi.org/10.1093/clinchem/hvae106.092
IF: 12.114
2024-10-27
Clinical Chemistry
Abstract:BackgroundThere are limited published studies assessing reference intervals (RIs) for hepatic enzymes, particularly for alkaline phosphatase (ALP) and gamma-glutamyl transferase (GGT). In the US, ALP and GGT RIs were established on Roche assays using a European cohort of ≤ 450 patients that may not reflect population variance across practice settings. Further, age- and sex-specific variation in ALP and GGT may warrant tiered RIs, but evidence to justify this approach is limited. We aimed to evaluate the manufacturer's RIs by comparing them to indirectly estimated RIs (IRIs) from routine clinical data.Methods217,333 AST, ALT, ALP, and GGT results (Roche c702) performed between 1/1/2019-12/31/2022 on outpatients ≥21 years were extracted from the laboratory information system. The first ALP for each patient was identified, and the earliest AST, ALT, and GGT drawn within 7 days was included. This approach was repeated for GGT. Concordance and correlation were calculated using Weighted Cohen's Kappa and Spearman's rank correlation coefficient respectively. {refineR}, an algorithm for calculating IRIs, was applied to patients with "normal" liver function (defined as normal AST, ALT for all IRIs, plus normal GGT for ALP IRIs, and ALP for GGT IRIs). IRIs were estimated for all males, all females, females ≤50, and females >50 years and compared to the method specific manufacturer RI (ALP: 40-129 Units/L for males and 35-104 Units/L for females; GGT: 8-61 Units/L for males and 5-36 Units/L for females).ResultsOf the 71,783 outpatients, 54.9% were female, 74% self-identified as white and 19% as black. The median age was 58 (IQR: 43-68 years), the median ALP concentration was 77 (62-97 Units/L), and the median GGT concentration was 28 (17-57 Units/L). A moderate agreement between ALP and GGT was observed (Kappa = 0.48, 95% confidence interval: 0.46-0.51). A positive correlation between ALP and GGT was observed r = 0.51 (p 50 years, respectively. For GGT, no significant differences were observed for the lower limits. The upper limit for all patients with normal AST/ALT/ALP (n = 2,446) was 62 [55-70 Units/L]. For males, all females, females ≤50 and >50 years was 58 [46-69 Units/L], 40 [39-51 Units/L], 31 [23-39 Units/L], and 56 [41-71 Units/L].ConclusionsFor the majority of patient subsets, the manufacturer RI agreed well with the IRIs. However, significant differences in IRIs were observed for the upper limit of ALP IRIs of all females with normal hepatobilliary function and in GGT for all females and females >50 years. Further investigation of the impact of preselecting patient populations on IRI estimation and a direct RI study on select patient subsets is worth investigation.
medical laboratory technology