Michael Gibbons1, Sharon Klim2, Arthur Mantzaris3, Oliver Dillon1, Anne-Maree Kelly2,4. 1. Western Health, Melbourne, Victoria, Australia. 2. Joseph Epstein Centre for Emergency Medicine Research, Western Health, Melbourne, Victoria, Australia. 3. Emergency Department, Western Health, Melbourne, Victoria, Australia. 4. School of Medicine-Western Clinical School, The University of Melbourne, Melbourne, Victoria, Australia.
Abstract
OBJECTIVE: The aims of this study were to establish the bias (mean difference) and 95% limits of agreement (LoA) between electrolyte values (sodium and potassium) and haemoglobin between whole blood analysed by the ED resuscitation room blood gas analyser and specimens analysed using standard techniques in the central hospital laboratory and to determine the proportion of analyses falling outside defined clinically acceptable LoA and pathology expert defined standards. METHODS: Prospective cohort study. Paired blood gas analyser and laboratory samples taken no more than 10 min apart were included. The primary outcome of interest was bias and 95% LoA by Bland-Altman analysis. Subgroup analyses for values outside the normal range were also conducted. RESULTS: Three hundred and fifty-two sample pairs were included in the analysis. For sodium concentration the bias was 0.6 mmol/L (95% LoA -3.3 to 4.6 mmol/L). For potassium concentration the bias was 0.21 mmol/L (95% LoA -0.36 to 0.79 mmol/L). For haemoglobin concentration the bias was -1.6 g/dL (95% LoA -10.2 to 6.9 g/dL). For sodium and haemoglobin concentrations, >95% of results fell within the defined clinically acceptable limits. For potassium concentration, >90% of results fell within the defined clinically acceptable limits. In general, serum sodium and potassium concentrations were slightly higher than blood gas levels and for haemoglobin serum levels were slightly lower. CONCLUSION: Agreement between blood gas analysis and laboratory analysis for sodium, potassium and haemoglobin concentrations shows acceptable agreement for use in time critical clinical decision-making in ED.
OBJECTIVE: The aims of this study were to establish the bias (mean difference) and 95% limits of agreement (LoA) between electrolyte values (sodium and potassium) and haemoglobin between whole blood analysed by the ED resuscitation room blood gas analyser and specimens analysed using standard techniques in the central hospital laboratory and to determine the proportion of analyses falling outside defined clinically acceptable LoA and pathology expert defined standards. METHODS: Prospective cohort study. Paired blood gas analyser and laboratory samples taken no more than 10 min apart were included. The primary outcome of interest was bias and 95% LoA by Bland-Altman analysis. Subgroup analyses for values outside the normal range were also conducted. RESULTS: Three hundred and fifty-two sample pairs were included in the analysis. For sodium concentration the bias was 0.6 mmol/L (95% LoA -3.3 to 4.6 mmol/L). For potassium concentration the bias was 0.21 mmol/L (95% LoA -0.36 to 0.79 mmol/L). For haemoglobin concentration the bias was -1.6 g/dL (95% LoA -10.2 to 6.9 g/dL). For sodium and haemoglobin concentrations, >95% of results fell within the defined clinically acceptable limits. For potassium concentration, >90% of results fell within the defined clinically acceptable limits. In general, serum sodium and potassium concentrations were slightly higher than blood gas levels and for haemoglobin serum levels were slightly lower. CONCLUSION: Agreement between blood gas analysis and laboratory analysis for sodium, potassium and haemoglobin concentrations shows acceptable agreement for use in time critical clinical decision-making in ED.