Accuracy of measurements of hemoglobin and potassium in blood samples from peripheral catheters

OBJECTIVE: To compare measurements of hemoglobin and potassium in blood samples obtained from peripheral venous catheters with measurements in samples obtained via venipuncture. SUBJECTS: 53 adult volunteer patients admitted to a medical ward. METHODS: For each patient, a 10-mL blood sample was obtained from a 20-gauge or larger peripheral venous catheter after a 3-mL volume of fluid was discarded. Immediately afterward, additional samples were obtained by venipuncture, and all samples were sent to the laboratory for measurement of hemoglobin and potassium levels. Separate venipuncture samples were obtained from both extremities in another 10 subjects to assess intratechnique variation. Results were analyzed by using limits-of-agreement analysis. RESULTS: Blood samples were successfully obtained from a peripheral catheter in 32 (60%) of the 53 patients. The success rate was 78% with 18-gauge catheters and 42% with 20-gauge catheters (P = .021). The 95% agreement interval in the set of differences between values for samples obtained via catheter and samples obtained via venipuncture was +/- 0.37 mmol/L for potassium and +/- 7.6 g/L for hemoglobin. The 95% agreement interval for the venipuncture control group was +/- 0.21 mmol/L for potassium and +/- 7.3 g/L for hemoglobin. The preset acceptable interval for potassium measurements was +/- 0.3 mmol/L, and that for hemoglobin measurements was +/- 10 g/L. CONCLUSIONS: Obtaining blood samples from peripheral catheters is an acceptable alternative to venipuncture for hemoglobin measurements but not for potassium measurements. Larger gauge catheters are better than small-gauge catheters for obtaining blood samples.