Evaluation of point-of-care blood glucose measurements in patients with diabetic ketoacidosis or hyperglycemic hyperosmolar syndrome admitted to a critical care unit.

BACKGROUND: Point-of-care (POC) blood glucose (BG) measurement is currently not recommended in the treatment of patients presenting with diabetic ketoacidosis (DKA) or hyperglycemic hyperosmolar syndrome (HHS). METHODS: We prospectively evaluated and compared capillary and venous POC BG values with laboratory venous glucose in patients with DKA or HHS admitted to one critical care unit over 8 months. RESULTS: Venous laboratory glucose was strongly correlated with venous (r = 0.98) and capillary (r = 0.96) POC glucose values, though POC glucose values were higher than venous laboratory values (venous POC 21 ± 3 mg/dl, capillary POC 30 ± 4 mg/dl; both p < .001). Increased plasma osmolality had no effect on glucose meter error, while acidemia (pH < 7.3) was associated with greater glucose meter error (p = .04) independent of glucose levels. Comparing hypothetical insulin infusion rates based on laboratory venous glucose to actual infusion rates based on POC glucose values showed that 33/61 insulin infusion rates would have been unchanged, while 28 out of 61 rates were on average 7% ± 2% higher. There were no instances of hypoglycemia in any of the patients. CONCLUSIONS: Overall, both venous and capillary POC BG values were safe for the purpose of titrating insulin infusions in patients with severe hyperglycemia. Acidemia, but not hyperosmolality, increased POC BG value errors.

The impact of inpatient point-of-care blood glucose quality control testing.

Analyze the effectiveness of mandated point-of-care (POC) blood glucose (BG) meter quality control (QC) testing. All POC BG QC tests were analyzed to evaluate operator and strip/meter error rates and institutional cost. POC BG QC test failure (17/103,580 over 24 months) was low and no meters failed subsequent linearity testing. Examining individual QC measures shows that operator error occurs frequently and total error rate is related to QC familiarity (>50 QC tests/month, 2.4%; <50 QC tests/month, 3.8%, p < .001). Even among the most competent operators, strip/meter error (1.2 ± 0.3%) accounted for 50% of total error. Compared with manufacturer-recommended QC testing, Joint Commission mandated POC BG QC testing during 2008/2009 incurred excess costs of approximately US$127,000. POC BG meter failure within current guidelines is rare and does not justify the cost of daily QC testing. Frequent QC testing can identify operators needing retraining in POC testing. Strip/meter QC errors are common, are not prevented by current QC testing standards, and may contribute to clinical errors.