Zuping Tang

Effects of Different Hematocrit Levels on Glucose Measurements With Handheld Meters for Point-of-Care Testing

OBJECTIVES To determine the effects of low, normal, and high hematocrit levels on glucose meter measurements and to assess the clinical risks of hematocrit errors. DESIGN Changes in glucose measurements between low and high hematocrit levels were calculated to determine hematocrit effects. The differences between glucose measured with meters and with a plasma glucose method (YSI 2300) also were compared. SETTING Six hand-held glucose meters were assessed in vitro at low (19.1%), normal (38.5%), and high (58.3%) hematocrit levels, and at 6 glucose concentrations ranging from 2.06 mmol/L (37.1 mg/dL) to 30.24 mmol/L (544.7 mg/dL). RESULTS Most systems, regardless of the reference to which they were calibrated, demonstrated positive bias at lower hematocrit levels and negative bias at higher hematocrit levels. Low, normal, and high hematocrit levels progressively lowered Precision G and Precision QID glucose measurements. Hematocrit effects on the other systems were more dependent on the glucose concentration. Overall, Accu-Chek Comfort Curve showed the least sensitivity to hematocrit changes, except at the lowest glucose concentration. CONCLUSIONS We strongly recommend that clinical professionals choose glucose systems carefully and interpret glucose measurements with extreme caution when the patients hematocrit value changes, particularly if there is a simultaneous change in glucose level.

Point-of-Care Glucose Testing Effects of Critical Care Variables, Influence of Reference Instruments, and a Modular Glucose Meter Design

OBJECTIVE To assess the clinical performance of glucose meter systems when used with critically ill patients. DESIGN Two glucose meter systems (SureStepPro and Precision G) and a modular adaptation (Immediate Response Mobile Analysis-SureStepPro) were assessed clinically using arterial samples from critically ill patients. A biosensor-based analyzer (YSI 2700) and a hospital chemistry analyzer (Synchron CX-7) were the primary and secondary reference instruments, respectively. PATIENTS AND SETTING Two hundred forty-seven critical care patients at the University of California, Davis, Medical Center participated in this study. OUTCOME MEASURES Error tolerances of +/-15 mg/dL for glucose levels </=100 mg/dL and +/-15% for glucose levels >100 mg/dL were used to evaluate glucose meter performance; 95% of glucose meter measurements should fall within these tolerances. RESULTS Compared to the primary reference method, 98% to 100% of SureStepPro and 91% to 95% of Precision G measurements fell within the error tolerances. Paired differences of glucose measurements versus critical care variables (Po(2), pH, Pco(2), and hematocrit) were analyzed to determine the effects of these variables on meter measurements. Po(2) and Pco(2) decreased Precision G and SureStepPro measurements, respectively, but not enough to be clinically significant based on the error tolerance criteria. Hematocrit levels affected glucose measurements on both meter systems. Modular adaptation did not affect test strip performance. CONCLUSIONS Glucose meter measurements correlated best with primary reference instrument measurements. Overall, both glucose meter systems showed acceptable performance for point-of-care testing. However, the effects of some critical care variables, especially low and high hematocrit values, could cause overestimated or underestimated glucose measurements.

Effects of Drugs on Glucose Measurements With Handheld Glucose Meters and a Portable Glucose Analyzer

Thirty drugs used primarily in critical care and hospital settings were tested in vitro to observe interference on glucose measurements with 6 hand-held glucose meters and a portable glucose analyzer. Paired differences of glucose measurements between drug-spiked samples and unspiked control samples were calculated to determine bias. A criterion of +/- 6 mg/dL was used as the cutoff for interference. Ascorbic acid interfered with the measurements on all glucose devices evaluated. Acetaminophen, dopamine, and mannitol interfered with glucose measurements on some devices. Dose-response relationships help assessment of drug interference in clinical use. High dosages of these drugs may be given to critically ill patients or self-administered by patients without medical supervision. Package inserts for the glucose devices may not provide adequate warning information. Hence, we recommend that clinicians choose glucose devices carefully and interpret results cautiously when glucose measurements are performed during or after drug interventions.

Oxygen effects on glucose meter measurements with glucose dehydrogenase- and oxidase-based test strips for point-of-care testing

ObjectivesTo determine the effects of different oxygen tensions (Po2) on glucose measurements with glucose dehydrogenase (GD)-based and glucose oxidase (GO)-based test strips, to quantitate changes in glucose measurements observed with different Po2 levels, and to discuss the potential risks of oxygen-derived glucose errors in critical care. DesignVenous blood from healthy volunteers was tonometered to create different oxygen tensions simulating patient arterial Po2 levels. Venous blood from diabetic patients was exposed to air to alter oxygen tensions simulating changes in Po2 during sample handling. Whole-blood glucose measurements obtained from these samples with six glucose meters were compared with reference analyzer plasma glucose measurements. Glucose differences were plotted vs. different Po2 levels to identify error trends. Error tolerances were as follows: a) within ±15 mg/dL of the reference measurement for glucose levels ≤100 mg/dL; and b) within ±15% of the reference measurement for glucose levels >100 mg/dL. Setting and Subjects Five healthy volunteers in the bench study and 11 diabetic patients in the clinical study. ResultsIn the bench study, increases in Po2 levels decreased glucose measured with GO-based amperometric test strips, mainly at Po2 levels >100 torr. At nearly constant glucose concentrations, glucose meter systems showed large variations at low (39 torr) vs. high (396 torr) Po2 levels. Glucose measured with GD-based amperometric and GO-based photometric test strips generally were within error tolerances. In the clinical study, 31.6% (Precision PCx), 20.2% (Precision QID), and 23.0% (Glucometer Elite) of glucose measurements with GO-based amperometric test strips, 14.3% (SureStep) of glucose measurements with GO-based photometric test strips, and 4.6% (Accu-Chek Advantage H) and 5.9% (Accu-Chek Comfort Curve) of glucose measurements with GD-based amperometric test strips were out of the error tolerances. ConclusionsDifferent oxygen tensions do not significantly affect glucose measured with the GD-based amperometric test strips, and have minimal effect on GO-based photometric test strips. Increases in oxygen tension lowered glucose measured with GO-based amperometric test strips. We recommend that the effects of different oxygen tensions in blood samples on glucose measurements be minimized by using oxygen-independent test strips for point-of-care glucose testing in critically ill and other patients with high or unpredictable blood Po2 levels.

Multiplex polymerase chain reaction detection enhancement of bacteremia and fungemia.

Objective:To test a multiplex real-time polymerase chain reaction (PCR) method for simultaneous detection of multiple organisms in bloodstream infections. Methods:Prospective observational study at the University of California Davis Medical Center (Sacramento, CA). Two hundred adult (>18 yrs) patients from the emergency room, intensive care units, and general medicine wards at risk of a bloodstream infection and who manifested signs of systemic inflammatory response syndrome (SIRS). Whole blood samples for PCR testing were collected at the same time as blood culture (BC). PCR results were compared to blood and other culture results. Results:PCR detected potentially significant bacteria and fungi in 45 cases compared to 37 by BC. PCR detected the methicillin resistance (mecA) gene in all three culture-confirmed methicillin-resistant Staphylococcus aureus cases. More than 68% of PCR results were confirmed by blood, urine, and catheter culture. Independent clinical arbitrators could not rule out the potential clinical significance of organism(s) detected by PCR, but not by BC. PCR did not detect Enterococcus faecalis in five BC-confirmed cases. On average, seven patient samples could be tested simultaneously with the PCR method in 6.54 ± .27 hrs. Conclusions:Multiplex PCR detected potentially significant bacteria and fungi that were not found by BC. BC found organisms that were not detected by PCR. Despite limitations of both BC and PCR methods, PCR could serve as an adjunct to current culture methods to facilitate early detection of bloodstream infections. Early detection of microorganisms has the potential to facilitate evidence-based treatment decisions, antimicrobial selection, and adequacy of antimicrobial therapy.

Effects of pH on glucose measurements with handheld glucose meters and a portable glucose analyzer for point-of-care testing.

OBJECTIVES To determine pH effects on glucose measurements obtained with the latest generation of glucose devices, to quantitate changes in glucose measurements obtained over a wide pH range, and to assess the potential clinical risks of pH effects with use of point-of-care glucose testing. DESIGN Paired differences of glucose measurements between pH-altered and parallel control samples with target pH 7.40 were calculated. SETTING A pH range of 6.94 to 7.84 was used to evaluate pH effects on glucose measurements in vitro with 6 handheld glucose meters and a portable glucose analyzer at both normal, 4.81 mmol/L (86.6 mg/dL), and high, 11.16 mmol/L (201 mg/dL), glucose levels. MAIN OUTCOME MEASURES Glucose measurements obtained from test samples and control samples were compared by calculating paired differences, which were plotted against pH to show pH effects on glucose meter measurements. RESULTS At the normal glucose level, different pH levels did not interfere significantly with glucose measurements. At the high glucose level, a trend whereby low pH decreased and high pH increased glucose measurements was observed on the Precision G and the Precision QID glucose meters. CONCLUSION Because of potential risk in diabetic patients with ketoacidosis and in other patients with acid-base disorders, we recommend that clinicians choose glucose devices carefully and interpret the measurements cautiously when point-of-care glucose testing is performed in critically ill patients with acidemia, alkalemia, or changing acid-base status.

Oxygen Effects on Glucose Measurements with a Reference Analyzer and Three Handheld Meters

Oxygen may affect glucose meter and reference analyzer measurements. We evaluated the effects of changes in blood oxygen tension (Po2) on Accu-Chek Comfort Curve (Roche Diagnostics, Indianapolis, IN), Precision G, (Abbott Laboratories, Bedford, MA) and One Touch II (Lifescan, Milpitas, CA) glucose meter measurements, and on Yellow Springs Instruments (YSI) (Yellow Springs, OH) reference analyzer measurements. Venous blood drawn from healthy volunteers was adjusted to three glucose levels of 80, 200, and 400 mg/dL, each tonometered with six different Po2 levels (40, 80, 160, 240, 320, and 400 torr). To quantitate oxygen effects on reference analyzer measurements, glucose differences between test sample (Po2 changed) and control (Po2 80 torr) were calculated (YSItest-YSIcontrol). The threshold for determination of oxygen effects was +/-2 SD, where 2 SD was fro

Whole-blood glucose and lactate: Trilayer biosensors, drug interference, metabolism, and practice guidelines

OBJECTIVE To assess the effects of 30 of the most commonly used critical care drugs on measurements obtained with trilayer electrochemical biosensors on a reference analyzer (ABL625-GL), to determine metabolic changes in glucose and lactate in vitro, and to formulate guidelines for whole-blood analysis of these 2 analytes. DESIGN Serial measurements were taken of changes in glucose and lactate levels caused by metabolism in whole blood in vitro over time. A parallel control study of drug interference with measurements of glucose and lactate in whole blood and of dose-response relationships in whole-blood samples and in plasma samples also was conducted. RESULTS At room temperature, whole-blood metabolism decreased glucose levels -2.3% at 15 minutes, -4.6% at 30 minutes, and -6.4% at 45 minutes. Metabolism increased lactate levels 11.4% at 15 minutes, 20.6% at 30 minutes, and 26.7% at 45 minutes in vitro. Paired differences between drug-spiked and control samples were calculated to determine interference (corrected for metabolism). The threshold for determination of interference was +/-2 SD from within-day precision, equal to +/-0.18 and +/-0.10 mmol/L for glucose and lactate, respectively. Only mannitol (C(6)H(14)O(6)) interfered with glucose and lactate measurements. At a concentration of 24 mg/mL, mannitol decreased whole-blood glucose levels by an average of 0.711 mmol/L (12.8 mg/dL) and whole-blood lactate levels by 0.16 mmol/L (1.4 mg/dL). Mannitol interference with measurements may have resulted from suppression of hydrogen peroxide formation in the enzymatic reactions in the biosensors, repartitioning of water between erythrocytes and plasma, or from other mechanisms. CONCLUSIONS Most critical care drugs had no significant effects on the trilayer electrochemical biosensors. Whole-blood analysis should be performed within 15 minutes for lactate and within 30 minutes for glucose because of metabolism in vitro. Mannitol effects on glucose measurements may be clinically significant in mannitol-induced acute renal failure and therefore should be considered for appropriate diagnosis and treatment of critically ill patients.

Point-of-Care Testing: Millennium Technology for Critical Care

Point-of-care testing (POCT) is an important diagnostic tool used in various locations in the hospital, especially in critical care settings such as the intensive care unit (ICU), the operating room (OR), and the emergency department (ED).

Validation of oxygen saturation measurements in a canine model of hemoglobin-based oxygen carrier infusion.

This study was designed to validate oxygen saturation measurements from the NOVA CO-Oximeter (NOVA Biomedical Corporation, Waltham, MA), the i-STAT System (Sensor Devices, Waukesha, WI), and the Corning 170 blood gas analyzer (Bayer Corporation, East Walpole, MA) under conditions similar to the clinical application of a hemoglobin-based oxygen carrier (HBOC, hemoglobin glutamer–200 [bovine]; Oxyglobin, Biopure Corporation, Cambridge, MA).A canine model was used for both in vitro and in vivo experiments. In vivo experiments were conducted in a canine laboratory, and in vitro experiments were conducted in a tonometry laboratory. Study subjects were six mixed-breed dogs, each weighing approximately 30 kg. In the first set of experiments, the target blood po2 levels were reached by tonometry. In the second set of experiments, quantitative measurements of total oxygen content with the LEXO2CON-K (HOSPEX Fiberoptics, Chestnut Hill, MA) were performed, immediately followed by measurements with the NOVA CO-Oximeter and the i-STAT system. HBOC was added in concentrations of 16.2, 32.5, 65, and 97.5 g/L. To analyze the clinical significance of the differences in the results obtained with the each investigated instrument, blood samples from dogs treated with HBOC after acute hemorrhagic shock were used. Oxygen saturation, oxygen content, and po2 were measured. There was a strong correlation between the oxygen saturation values measured with the investigated instruments in samples after tonometry and known po2. The total calculated oxygen content varied by 5% based on results generated by calculations using the investigated instruments. The results did not change with different oxygenation of the sample. The differences among methods were not significant when the HBOC concentration was 16.2 g/L. Higher concentrations of HBOC increased the difference between calculated and measured oxygen content; the i-STAT system demonstrated a greater deviation compared with the results of the other two instruments. Systemic oxygen uptake using the investigated instruments showed a high correlation with values based on LEXO2CON-K measurements (R = 0.97 for CO-Oximeter, R = 0.96 for Corning 170 blood gas analyzer, and R = 0.79 for i-STAT system). Systemic oxygen uptake values based on CO-Oximeter and Corning 170 blood gas analyzer data showed 75% accuracy; i-STAT system accuracy was 63% for control samples and 50% for samples after HBOC infusion.