Evidence based evaluation of the diagnostic performance of conventional reticulocyte parameters in dogs with regenerative anemia

Background: Reticulocyte counts (Ret#), reticulocyte percentage (Ret%), and reticulocyte production index (RPI) are used to estimate bone marrow regeneration. No information on the actual diagnostic relevance of these parameters for canine regenerative anemia is available. Objectives: To assess the diagnostic relevance of reticulocyte parameters for canine regenerative anemia through a retrospective analysis of our caseload and a systematic review of literature. Methods: Data recorded at our Institution since 1993 to 2009 were screened to select 148 cases classified as: (1) non-anemic (n = 57); (2) regenerative anemia (RA; n = 37); (3) non-regenerative anemia (NRA; n = 27); (4) ‘‘pre-regenerative’’ anemia (PRA; n = 23). Values of Ret#, Ret%, and RPI were used to calculate sensitivity, specificity, positive and negative predictive values (PPV and NPV respectively), and positive likelihood ratios (LR1). ROC curves were designed to assess how reticulocyte parameters identify dogs with RA and dogs with RA or PRA. A systematic review of the literature was also performed followed by the calculation of sensitivity, specificity, PPV, NPV, and LR1 on published data. Results: In our caseload, dogs with RA had significantly higher Ret#, Ret%, and RPI than other groups. The area under the ROC curve to diagnose RA, as well as sensitivity, specificity, PPV, NPV, and LR1, were significantly higher for Ret% and RPI than for Ret#. When dogs with RA and PRA are merged in a single group reticulocyte parameters provided unsatisfactory diagnostic performances. The systematic review of literature confirms that Ret% and RPI identify RA better than Ret#. Conclusions: Ret% and RPI are more powerful than Ret# to identify RA in dogs. Reticulocytes parameters are not able to differentiate RA from PRA