Secondary Logo

Journal Logo

Reducing Potential Medical Errors in Point-of-Care Testing: Clinical Effects of Heparin and EDTA on Handheld Glucose Meter Results

Kost, Gerald J. MD, PhD; Louie, Richard F. BS; Tang, Zuping MD; Lee, Judith H. MT (ASCP); Somsanith, Keith J.; Tran, Nam K.

Point of Care: The Journal of Near-Patient Testing & Technology: March 2002 - Volume 1 - Issue 1 - p 2–8
Article
Free

The objective of this study was to determine if heparin or ethylenediaminetetraacetic acid (EDTA) affected whole-blood glucose measurements obtained with test strips and common handheld meter systems. Lithium heparin (15 USP units/mL), potassium (K2) EDTA (1.8 mg/dL), and no-additive aliquots of venous blood from fifty-five diabetic patients were tested using four handheld glucose meter systems: the SureStep Pro and the One Touch II (Lifescan, Milpitas, CA), the Precision PCx (Abbott Laboratories, MediSense Products, Bedford, MA), and the Accu-Chek Comfort Curve (Roche Diagnostics, Indianapolis, IN), and a the YSI 2700 reference analyzer (Yellow Springs Instruments, Yellow Springs, OH). Coefficients of variation for within-day and between-day precision ranged from 1.4% to 6.8% for the glucose meter systems. Heparin and EDTA produce system-specific effects. Anticoagulant versus no-additive test strip comparisons showed that the SureStep Pro, with heparin and EDTA, and the Precision PCx, with EDTA, produced statistically significant negative mean paired differences. Heparin minimized the frequency of errors in glucose meter results. Heparinized samples produced the lowest frequency of discrepant meter results outside clinical error tolerances (± 15 mg/dL, glucose ≤ 100 mg/dL; ± 15%, glucose > 100 mg/dL) when compared with YSI 2700 reference results for paired samples. Specimen restrictions arise from anticoagulant selection, hematocrit range, physiological factors, and processing time. To reduce the potential for medical errors, the relative effects of different specimen anticoagulants should be considered carefully when conducting comparison studies, planning clinical applications, and interpreting glucose results.

From the Point-of-Care Testing Center for Teaching and Research (POCT·CTR), School of Medicine, University of California, Davis.

Address correspondence and reprint requests to Gerald J. Kost, MD, PhD, 3453 Tupper Hall, Medical Pathology, School of Medicine, University of California, Davis, Davis, CA 95616 (e-mail: gjkost@ucdavis.edu).

Back to Top | Article Outline

Introduction

Recent Institute of Medicine reports,1,2 and ongoing professional debate,3–8 highlight controversy and uncertainty regarding the frequency and significance9 of medical errors. Errors in laboratory testing can contribute to inappropriate care or modification of therapy.10,11 Although exact preanalytical, analytical, and postanalytical error rates are not known, pointof-care diagnostic testing may contribute to serious medical errors.12–14

As more testing shifts to the bedside,15,16 and is performed by increasing numbers of physicians, nurses, and nurse practitioners, adequate safeguards must be in place to prevent medical errors and reduce risk. Therefore, the goal of this paper is to increase awareness of the potential for anticoagulants to contribute to analytical errors in handheld glucose testing.

Back to Top | Article Outline

Methods

Glucose meter systems

The four glucose meter systems17 included: the SureStep Pro and the One Touch II [Hospital] (Life-Scan, Milpitas, CA), the Precision PCx (Abbott Laboratories, MediSense Products, Bedford, MA), and the Accu-Chek Comfort Curve (Roche Diagnostics, Indianapolis, IN).

Back to Top | Article Outline

Reference instrument

A YSI 2700 (Yellow Springs Instruments, Yellow Springs, OH) dual-channel analyzer served as the reference instrument. The YSI 2700 measures glucose using an oxidase-based method. Plasma was assayed for all systems except the One Touch II, for which whole blood was assayed, in accordance with the vendor-specified type of sample for reference measurements.

Back to Top | Article Outline

Subject protocol

Fifty-five diabetic adult patients from the Diabetes Clinic at the University of California, Davis Health System were included in the study. The experimental protocol was approved by the Human Subjects Committee. Venous blood was collected in a 10 mL syringe and then aliquoted into 3 mL and 4 mL blood collection devices (Vacutainer, Becton Dickinson and Company, Franklin Lakes, NJ) that contained either lithium heparin (15 USP units/mL) or K2 EDTA (1.8 mg/dL), respectively. These tubes were filled completely with blood. The blood remaining in the syringe served as the no-additive sample. Hematocrit was measured manually by centrifugation.

Back to Top | Article Outline

Glucose measurements

The application of the different sample types to the glucose meter test strips, and the order of glucose meter systems evaluated were randomized. All wholeblood glucose measurements were completed within 15 minutes. Plasma measurements requiring centrifugation were completed within 20 minutes from the time of venous blood collection to minimize the effects of glycolysis.

Glucose meter measurements were obtained using two lots of test strips per glucose meter system. Whole-blood and plasma glucose measurements were obtained with the YSI 2700 reference analyzer. Glucose results are reported in mg/dL. To convert to System International units: mmol/L = 0.05551 x mg/dL.

Back to Top | Article Outline

Effects of anticoagulants

For anticoagulant (A) analysis, glucose differences (ΔA) were calculated between the meter measurements obtained with heparin or EDTA samples and with no-additive samples (ΔA = MeterHeparin-MeterNo-Additive or MeterEDTA-MeterNo-Additive). These paired differences assessed anticoagulant effects on glucose test strip results. Student t test for paired differences was used to determine statistical significance. A value of P < 0.05 was considered statistically significant.

For error tolerance (ET) analysis, glucose paired differences (y axis) were calculated between meter and YSI 2700 reference measurements for each sample type (ΔET = MeterHeparin-YSIHeparin, MeterEDTA-YSIEDTA, or MeterNo-Additive-YSIHeparin) and compared graphically to the average (x axis) of the two YSI 2700 results for the samples containing anticoagulants. Use of the average on the x axis aligned individual patient results vertically in the y direction for facile comparison of the paired differences.

Back to Top | Article Outline

Error tolerances

The number and percentage of glucose paired differences (ΔET) outside error tolerances of ± 15 mg/dL for glucose, ≤ 100 mg/dL, or ± 15% for glucose > 100 mg/dL were determined to assess the clinical accuracy of glucose meter results relative to the YSI 2700 reference analyzer results.18 Fewer than 5% of the points (ΔET) should lie outside the error tolerances.

Back to Top | Article Outline

Precision evaluation

Two or three levels of aqueous quality control solutions provided by the manufacturers were tested 20 consecutive times to assess within-day precision for each of the four glucose meter systems. Evaluation of between-day precision was conducted over 20 experimental days. The within-day precision and between-day precision of the YSI 2700 reference analyzer were evaluated with three glucose levels (50, 200, and 400 mg/dL) of NERL 1343-Standard Glucose Solutions (New England Reagent Laboratory, East Providence, RI). Precision data are reported as the mean, standard deviation (SD), and coefficient of variation (CV) (CV = [Mean/SD] × 100%) of the within-day and between-day trials.

Back to Top | Article Outline

Results

Glucose and hematocrit ranges

Blood glucose levels ranged from 46 to 554 mg/dL. Comparisons totaled 55 for all systems except the SureStep Pro, for which glucose levels in two samples exceeded the meter measurement limit of 500 mg/dL and were excluded from the results. Hematocrit ranged from 32% to 55%. Only results for one lot of glucose test strips are presented below. The second lot produced similar results.

Back to Top | Article Outline

Effects of anticoagulants

Figure 1 displays the effects of heparin and EDTA on glucose meter measurements. Table 1 summarizes the quantitative results. With the SureStep Pro, both heparin and EDTA generated statistically significant mean paired differences (ΔA). With the Precision PCx, only EDTA generated statistically significant differences. Other anticoagulant effects were not statistically significant.

Fig 1

Fig 1

Table 1

Table 1

Back to Top | Article Outline

Error tolerance analysis

Figure 2 shows error tolerance analysis plots for each of the four glucose meter systems for the three sample types. The Comfort Curve glucose meter was the only system for which all points (ΔET) fell within the error tolerances.

Fig 2

Fig 2

Table 2

Table 2

Lithium heparin samples generated minimal numbers and percentages of points violating the error tolerance criteria (Table 2). Ethylenediaminetetraacetic acid tended to decrease glucose meter results obtained with most of the systems.

The negative bias of the Precision PCx system for samples containing EDTA resulted in a high frequency of values falling outside the lower error tolerance. Note that vendor specifications for the One Touch II state that using EDTA may cause inaccurately low results if the sample hematocrit exceeds 55%.

Back to Top | Article Outline

Precision

Table 3 summarizes within-day (3A) and betweenday (3B) precision results for the different levels of aqueous control tested and the four glucose meter and YSI systems. The SureStep Pro had the lowest CV (1.4%, middle level), and the Precision PCx had the highest CV (6.8%, low level) observed for within-day precision trials for glucose meters.

Table 3

Table 3

Back to Top | Article Outline

Limitations

Table 4 presents the characteristics of the four glucose meter systems, including the test strip chemistry and the analytical principle of each. Venous blood samples were processed rapidly, and as much as was feasible, anaerobically, but may have been subject to changes in pO2, which can affect glucose test strips depending on the principle of glucose measurement and other factors.18–21

Table 4

Table 4

For uniformity of comparison and interpretation of results, the effects of both heparin and EDTA were evaluated on all four systems, even though test strip package inserts may not rule out the use of EDTA explicitly. However, lithium heparin was recommended as an acceptable anticoagulant in the package inserts of all four glucose meter systems.

Back to Top | Article Outline

Discussion

Ethylenediaminetetraacetic acid chelates the cations, Ca++, and to a lessor extent, Mg.++ K3 EDTA produces a shrinkage of erythrocytes that results in an approximately 2% decrease in the packed cell volume.22–24 In contrast, the EDTA salt used in this study, K2 EDTA, is preferred for the determination of the packed cell volume because of a smaller change in ionic strength and less effect on erythrocytes. The collection device manufacturer stated that K3 EDTA is being phased out in favor of K2 EDTA. Nonetheless, K2 EDTA affected glucose meter measurements, possibly in part through changes in erythrocyte volume, flow, or stacking on test strips and other factors.

The mechanisms of EDTA-related effects could include subtle osmotic changes25 or multiple influences, such as surface charge disruption, viscosity differences, altered morphology,26 platelet phenomena,27–30 or unknown interactions of EDTA with plasma flow boundary layers and test strip glucose reactions. One manufacturer noted that glucose results may decrease when EDTA use is accompanied by high hematocrit. The collection device manufacturer stated that prompt sample processing limits cell shrinkage caused by EDTA. Heparin may affect glucose results through a similar set of complex actions including changes in sample pH.31

Few studies address anticoagulant-induced errors in glucose testing. Lui et al.32 documented a 4% decrease with EDTA-anticoagulated samples when glucose was measured by an amperometric point-of-care method. In the same study, fluoride-oxalate and heparin changed glucose results, 22% and -1.9%, respectively. Manufacturers commonly exclude fluoride, oxalate, or iodoacetate additives. Published data and professional experience suggest that heparin affects glucose meter measurements the least and can be recommended for use when direct whole-blood measurement without anticoagulation is not possible.

The results of this study show that minimal analytical errors were produced when lithium heparin was used as the sample anticoagulant. A recent study showed that sodium citrate, used to anticoagulate indwelling arterial catheters, interferes with reliable measurements of glucose and other analytes, even after discarding 9 mL of blood.33 Hence, samples taken from such lines may produce errors. With whatever additive is used, or without any additive, whole-blood and plasma specimens should be processed promptly to avoid preanalytical changes due to the metabolism of glucose.34

The authors recommend that users of glucose meter systems: a) consult the package insert for anticoagulant specifications, b) contact the manufacturer for current information, and if necessary, c) perform an appropriate investigation when unusual additives (e.g., thrombin inhibitors35,36) are necessary. The type of anticoagulant in the collection device, if used, should be clarified. To reduce the potential for medical errors in point-of-care glucose testing,37 the relative effects of different specimen anticoagulants should be considered and communicated carefully when conducting comparison studies, planning clinical applications, and interpreting glucose results.

Back to Top | Article Outline

Acknowledgments

The authors thank the companies who contribute generously to the POCT·CTR and enabled this research to be performed. This research was presented by Richard Louie, recipient of a Young Investigator Award, at the annual meeting of the Academy of Clinical Laboratory Physicians and Scientists in Seattle in June, and by Judith Lee at the annual meeting of the American Association for Clinical Chemistry in Chicago in July, 2001.

Back to Top | Article Outline

References

1. Kohn LT, Corrigan JM, Donaldson MS, eds. Committee on Quality of Health Care in America, Institute of Medicine. To Err is Human: Building a Safer Health System. Washington, DC: National Academy Press; 2000.
2. Richardson WL, Berwick DM, Bisgard JC, et al. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academy Press; 2001.
3. Leape LL. Institute of Medicine medical error figures are not exaggerated. JAMA 2000;284:95-97.
4. McDonald CJ, Weiner M, Hui SL. Deaths due to medical errors are exaggerated in Institute of Medicine report. JAMA 2000;284:93-95.
5. Sox HC, Woloshin S. How many deaths are due to medical error? Getting the number right. Effective Clin Pract 2000;3:277-283.
6. Bell CM, Redelmeier DA. Mortality among patients admitted to hospitals on weekends as compared with weekdays. N Engl J Med 2001;345:663-668.
7. Halm EA, Chassin MR. Why do hospital death rates vary? N Engl J Med 2001;345:692-694.
8. Hayward RS, Hofer TP. Estimating hospital deaths due to medical errors: preventability is in the eye of the reviewer. JAMA 2001;286:415-420.
9. Sirota RL. The Institute of Medicine's report on medical error: implications for pathology. Arch Pathol Lab Med 2000;124:1674-1678.
10. Plebani M, Carraro P. Mistakes in a stat laboratory: types and frequency. Clin Chem 1997;43:1348-1351.
11. Witte DL, VanNess SA, Angstadt DS, Pennell BJ. Errors, mistakes, blunders, outliers, or unacceptable results: how many? Clin Chem 1997;43:1352-1356.
12. Kost GJ. Guidelines for point-of-care testing: improving patient outcomes. Am J Clin Pathol 1995;104(Suppl 1):S111-S127.
13. Kost GJ, Ehrmeyer SS, Chernow B, et al. The laboratory-clinical interface: point-of-care testing. Chest 1999:115:1140-1154.
14. Kilgore ML, Steindel SJ, Smith JA. Continuous quality improvement for point-of-care testing using background monitoring of duplicate specimens. Arch Path Lab Med 1999;123:824-828.
15. Kost GJ. The hybrid laboratory: shifting the focus to the point of care. Med Lab Observer 1992;24(9S):17-28.
16. Kost GJ. Point-of-Care Testing ⇒ The Hybrid Laboratory ⇒ Knowledge Optimization. In: Kost GJ, ed. Handbook of Clinical Automation, Robotics, and Optimization. New York: John Wiley and Sons; 1996:757-838.
17. Tang Z, Louie RF, Kost GJ. Principles and performance of point-of-care instruments. In: Kost GJ, ed. Principles and Practice of Point-of-Care Testing. Philadelphia: Lippincott Williams and Wilkins; 2002.
18. Kost GJ, Vu HT, Lee J. Multicenter study of oxygen insensitive handheld glucose point-of-care testing in critical care/hospital/ambulatory patients in the United States and Canada. Crit Care Med 1998;26:581-590.
19. Louie RF, Tang Z, Sutton DV, Lee JH, Kost GJ. Point-of-care glucose testing: effects of critical care variables, influence of reference instruments, and a modular glucose meter design. Arch Pathol Lab Med 2000;124:257-266.
20. Tang Z, Louie RF, Payes M, Chang KC, Kost GJ. Oxygen effects on glucose measurements with a reference analyzer and three handheld meters. Diabetes Technol Ther 2000;2:349-362.
21. Tang Z, Louie RF, Lee JH, Lee DM, Miller EE, Kost GJ. Oxygen effects on glucose meter measurements with glucose dehydrogenase- and oxidase-based test strips for point-of-care testing. Crit Care Med 2001;29:1062-1070.
22. Koepke JA, van Assendelft OW, Bull BS, Richardson-Jones A. Standardization of EDTA anticoagulation for blood counting procedures. Labmedica 1988/89;5:15-17.
23. International Council for Standardization in Haematology: Expert Panel on Cytometry. Recommendations of the International Council for Standardization in Haematology for ethylenediaminetetraacetic acid anticoagulation of blood for blood cell counting and sizing. Am J Clin Path 1993;100:371-372.
24. Bull BS, Koepke JA, Simson E, van Assendelft OW. Procedure for determining packed cell volume by the microhematocrit method. Vol 20. No 18. Wayne, PA: National Committee for Clinical Laboratory Standards; 2000: Document H7-A3;4.
25. Nevius DB. Osmotic error in electronic determination of red cell volume. Am J Clin Pathol 1963;39:38-41.
26. Pinteric L, Manery JF, Chaudry IH, Madapallimattam G. The effect of EDTA, cations, and various buffers on the morphology of erythrocyte membranes: an electron-microscope study. Blood 1975;45:709-724.
27. Pegels JG, Bruynes ECE, Engelfriet CP, von dem Borne AEGKr. Pseudothrombocytopenia: An immunological study on platelet antibodies dependent on ethylene diamine tetra-acetate. Blood 1982;59:157-161.
28. Lombarts AJPF, de Kieviet W. Recognition and prevention of pesudothrombocytopenia and concomitant pseudoleukocytosis. Am J Clin Pathol 1988;89:634-639.
29. Bizzaro N, Brandalise M. EDTA-dependent pseudothrombocytopenia: association with antiplatelet and antiphospholipid antibodies. Am J Clin Pathol 1995;103:103-107.
30. Schrezenmeier H, Muller H, Gunsilius E, Heimpel H, Seifried E. Antocoagulant-induced pseudothrombocytopenia and pseudoleucocytosis. Thromb Haemost 1995;73:506-613.
31. Tang Z, Du X, Louie RF, Kost GJ. Effects of pH on glucose measurements with handheld glucose meters and a portable glucose analyzer for point-of-care testing. Arch Pathol Lab Med 2000;124:577-582.
32. Liu KF, Ng WY, Thai AC. An amperometric measurement—the ExacTech pen meter. Ann Acad Med Singapore 1990;19:473-476.
33. Cardinal P, Allan J, Mmath BP, Hindmarsh T, Jones G, Delisle S. The effect of sodium citrate in arterial catheters on acid-base and electrolyte measurements. Crit Care Med 2000;28:1388-1392.
34. Kost GJ, Nguyen TH, Tang Z. Whole-blood glucose and lactate: trilayer biosensors, drug interference, metabolism, and practice guidelines. Arch Pathol Lab Med 2000;124:1128-1134.
35. Lyon ME, Fine JS, Henderson PJ, Lyon AW. D-phenyalanyl-L-propyl-L-arginine chloromethyl ketone (PPACK): alternative anticoagulant to heparin salts for blood gas and electrolyte specimens. Clin Chem 1995;41:1038-1041.
36. Lyon AW, Harding SR, Drobot D, Lyon ME. Use of thrombin inhibitors ex vivo allows critical care clinical chemistry and hematology testing on common specimens. Clin Biochem 1997;30:121-127.
37. Kost GJ. Preventing medical errors in point-of-care testing: security, validation, performance, safeguards, and connectivity. Arch Pathol Lab Med 2001;125:1307-1315.
Keywords:

Bedside testing; Error tolerance; Glucose test strip; Hematocrit; Institute of Medicine; K2EDTA; Osmotic effects; Medical errors; YSI reference method.

© 2002 Lippincott Williams & Wilkins, Inc.