Impact of reference interval source on result classification in pediatric coagulation testing

Interpretation of pediatric coagulation results depends on the reference interval (RI) applied, yet the extent to which RI source affects result classification is rarely quantified. This retrospective single-center study determined how RI source affects out-of-RI flagging in pediatric coagulation testing. Analyzer-specific RIs for prothrombin time (PT), activated partial thromboplastin time (aPTT), and fibrinogen in children aged 1 to <18 years were derived from 37,167 routine results measured on Roche Cobas t511/t711 analyzers (PT, n = 17,000; aPTT, n = 17,001; fibrinogen, n = 3166). Indirect RIs were estimated using refineR, continuous RIs using generalized additive models for location, scale and shape (GAMLSS), and direct a posteriori RIs from CLSI EP28-A3c healthy subsets. Out-of-RI flagging rates were compared across manufacturer, locally derived indirect, continuous, direct and published Turkish pediatric RIs, and transferability of published Turkish PT and aPTT intervals was assessed using a bootstrap analogue of the CLSI EP28-A3c 20-sample procedure. Indirect, continuous, and direct RIs were broadly concordant. Manufacturer aPTT limits flagged 32.3% of results outside reference limits (95% CI, 31.6-33.0), compared with 10.7% (10.2-11.2) using continuous RIs. Published same-platform Turkish PT and aPTT intervals did not meet transferability criteria in any single-year age group. Local pediatric coagulation RIs markedly reduced excess abnormal flagging, particularly for aPTT. RI source can materially alter pediatric coagulation result classification, supporting local verification and analyzer-specific implementation before adopting external or manufacturer-supplied limits.