Assessment of Point-of-Care Measurement of Urinary Creatinine and Electrolytes in the Intensive Care Unit

Comparison of point-of-care NOVA CCX blood gas analyzer to laboratory analyzer in a population of healthy adult cats

OBJECTIVE

To compare the level of agreement of measurement of analytes (sodium, chloride, potassium, urea nitrogen [UN], creatinine, glucose) in a population of healthy adult cats between the point-of-care (POC) analyzer and laboratory analyzer. To establish reference intervals for the POC analyzer in healthy adult cats.

DESIGN

Prospective observational study.

SETTING

University teaching hospital.

ANIMALS

Fifty-five cats were screened. Seven cats were excluded due to aggression that prohibited phlebotomy, and 1 cat was excluded due to prolonged restraint; 47 cats were enrolled.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

In this patient population, reference intervals for the POC analyzer were calculated: sodium 145-157 mmol/L; chloride 116-124 mmol/L; potassium 3.4-5.5 mmol/L; UN 5.71-13.9 mmol/L (16-39 mg/dl); creatinine 74.3-189.2 mmol/L (0.84-2.14 mg/dl); and glucose 4-11.8 mmol/L (72-213 mg/dl). Comparison between the POC analyzer and laboratory analyzer using the Bland-Altman method was performed. The bias for each analyte is as follows: sodium 1.55 mmol/L; chloride 0.99 mmol/L; potassium 0.21 mmol/L; UN -0.25 mmol/L (-0.7 mg/dl); creatinine 9.73 mmol/L (0.11 mg/dl); and glucose 0.5 mmol/L (9.79 mg/dl).

CONCLUSIONS

Reference intervals for each analyte were similar to other chemistry analyzers. There was no significant difference between the POC and laboratory analyzers in analysis of UN, with a statistically significant difference observed with sodium, potassium, chloride, creatinine, and glucose. However, the values are likely not sufficiently different to alter initial clinical decisions regarding patient care.

Pitfalls in the management of severe hyponatraemia

The correction of severe hyponatraemia is complicated by practical and theoretical limitations of formulae used to predict response to treatment, and random errors in the measurement of sodium concentration. We report the case of an elderly woman with a symptomatic serum sodium concentration of 93 mmol/l. Correction of serum sodium unintentionally exceeded conventional targets but, as in the majority of such cases, there were no neurological sequelae.