HM07 - Agreement Between Prospective and Indirect Hematology Reference Intervals and Influence of Age in Cats

Abstract:

Introduction: Small sample sizes limit the ability of traditional direct reference intervals (RIs) to represent the diversity of the patient population. Indirect RIs using newer statistical methods to create RIs from real-world laboratory data may overcome some of these limitations.

Objective: Compare direct prospective hematology RIs to indirect RIs from retrospective data and to assess the effect of age. Animals: 125 healthy cats (1–10 years) were enrolled for the prospective direct RI study; 49,527 feline CBC submissions were evaluated for the indirect RI study.

Methods: Direct RIs were estimated per CLSI guidelines. Indirect RIs were estimated using the refineR algorithm on CBC results from the same labs and time period. Age-restricted (1-10 years) indirect RIs were also estimated. Two-One-Sided-Tests (TOST, α = 0.05) for equivalence of reference limits determined whether differences fell within ASVCP total allowable error limit. P-values were calculated from 1000 bootstrap resamples. Jaccard Index was used to evaluate similarity of age-restricted sets.

Results: Upper and lower reference limits were statistically equivalent for RBC, hematocrit, hemoglobin, MCHC, MCV, and eosinophils. The RIs estimated from age-restricted CBC submissions were closer to those of the direct study for 12 of 15 measurands, particularly RBCs, neutrophils, monocytes, lymphocytes.

Conclusions: Reference limits for many RBC indices in cats were equivalent between the methodologies. Age contributed to differences between estimated RIs. refineR provides RIs that better represent the age range of the patient population and can be used to augment methods of establishing reference intervals.