Capillary blood is not accurate in predicting blood ammonia values using an ammonia point-of-care test in dogs

To compare results for blood ammonia (BA) concentrations measured with a point-of-care (POC) device versus commercial diagnostic assay (CDA) for venous and capillary blood samples from dogs with normal BA and hyperammonemia.

Dogs were prospectively enrolled from January 2024 through July 2024 and grouped as being healthy (controls), having liver disease with normal BA, or having liver disease with hyperammonemia. All dogs had BA concentrations determined with a venous sample run on a CDA, a venous sample run on an ammonia POC device (POC venous [POC-V] method), and a capillary blood sample run on an ammonia POC device (POC capillary [POC-C] method). The results were compared across methods.

46 dogs were enrolled: 15 healthy dogs and 31 dogs with liver disease with normal BA (n = 16) or hyperammonemia (n = 15). The mean biases for BA concentration as measured with the POC-V and POC-C methods compared with the CDA method were -54.3 µg/dL (95% CI, -76.8 to 32.0) and 1.4 µg/dL (95% CI, -36.0 to 38.7), respectively. The mean bias of the POC-C method versus the POC-V method was 55.7 µg/dL (95% CI, 30.4 to 81.0). For the 31 dogs with CDA results for BA within reference limits, all were similarly classified with the POC-V method, whereas 25 of 31 (81%) were classified as normal with the POC-C method.

The BA in the POC-V and POC-C groups was, on average, underestimated when compared to the CDA. The BA in the POC-C group was consistently overestimated when compared to the POC-V group. Although both POC methods had good agreement in the classification of normal BA values, venous (vs capillary) samples yielded better results.

The use of a POC device to measure BA in venous blood, but not capillary blood, may be an alternative to CDAs in emergency settings.