Carboxyhemoglobin as a diagnostic and prognostic biomarker of hemolytic anemias in dogs.

Abstract

BACKGROUND: Endogenous production of carbon monoxide during hemoglobin metabolism leads to the formation of carboxyhemoglobin. Carboxyhemoglobin concentration is abnormally high in humans with hemolytic anemia (HA). HYPOTHESIS: Measurement of carboxyhemoglobin concentration can discriminate HA from other forms of anemia. ANIMALS: Twenty-seven dogs with HA (immune-mediated HA, n&#xa0;=&#xa0;22; microangiopathic HA, n&#xa0;=&#xa0;5), 27 dogs with non-HA (kidney disease, n&#xa0;=&#xa0;14; immune-mediated thrombocytopenia, [n&#xa0;=&#xa0;6]; miscellaneous, n&#xa0;=&#xa0;7) and 24 nonanemic control dogs. METHODS: Prospective cohort study. Carboxyhemoglobin quantification, a CBC and biochemistry profile were performed upon admission, and survival to hospital discharge and at 30&#x2009;days were the measured outcomes. Groups were compared by the Mann-Whitney and Kruskal-Wallis tests. Receiver-operator characteristic (ROC) analyses were used to examine the predictive utility of carboxyhemoglobin for the diagnosis of HA in anemic dogs. RESULTS: Carboxyhemoglobin (median [interquartile range]) differed between dogs with HA (7.7% [2.5%]) and non-HA (3.6% [1.05]; P&#x2009;<&#x2009;.001) and dogs with HA and nonanemic dogs (3.5% [0.65%]; P&#x2009;<&#x2009;.001). No difference was detected between nonHA and nonanemic dogs. The area under the ROC curve for carboxyhemoglobin as predictor of HA in anemic dogs was 0.997 (95% CI, 0.99-1.00). Three optimal cut-off points were identified, including 5.05%, 4.55% and 4.85%, with corresponding sensitivity/specificity of 92.6%/100%, 100%/92.6% and 96.3%/96.3%, respectively. Neither carboxyhemoglobin nor any of the CBC or chemistry analytes were associated with survival. CONCLUSIONS AND CLINICAL IMPORTANCE: Carboxyhemoglobin proved an excellent predictor of HA in dogs and might constitute a useful, ancillary tool for diagnosing and monitoring hemolytic anemias.