College of American Pathologists
Printable Version

  Glucose testing variability and the
  need for an expert oversight



cap today

May 2005
Feature Story

In 1979 a work group of the National Diabetes Data Group1 established the criteria, later endorsed by the World Health Organization Committee on Diabetes, that patients with a fasting or 2-h postprandial glucose concentration greater than 140 or 200 mg/dL, respectively, were to be considered diabetic.

In 1997, the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus was convened to reexamine the classification and diagnostic criteria for diabetes based on the 1979 publication of the National Diabetes Data Group. As a result of its deliberations, the committee recommended several changes to the diagnostic criteria for diabetes and for lesser degrees of impaired glucose regulation.2 The use of a fasting plasma glucose, or FPG, test for the diagnosis of diabetes was recommended, and the cut point separating diabetes from nondiabetes was lowered from a FPG > 140 mg/dL to > 126 mg/dL. This change was based on data that showed an increase in prevalence and incidence of diabetic retinopathy beginning approximately at a FPG of 126 mg/dL, as well as on the desire to reduce the discrepancy that existed in the number of cases detected by the FPG cut point of > 140 mg/dL and the 2-h value in the OGTT (2-h plasma glucose) of > 200 mg/dL.

In November 2003, the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus3 recommended that the criteria to diagnose diabetes remain as previously defined. However, the lower cut point defining impaired fasting glucose, or IFG, was reduced from > 110 mg/dL to > 100 mg/dL. Thus, "normal" would now be defined as a FPG of < 100 mg/dL. These diagnostic criteria for diabetes mellitus were restated in January 2005.4

Establishing a single number for glucose to determine whether a patient’s value is "normal" or "abnormal"—that is, a fasting plasma glucose of 100 mg/dL—is a way to help physicians remember when a patient’s fasting glucose is increased; however, it does not take into account the variability of the measurement itself.

There is a marked variability involved (both biologic and analytic) in glucose testing. Physicians are taught in medical school that this variability exists and that the biological variability is substantially greater than the analytic variability.5

The concentration of glucose in the serum of a fasting individual is not the same when measured on different days. The 2002 diabetes mellitus practice guideline of the National Academy of Clinical Biochemistry5 cites Ollerton, et al’s study6 that reports biological variabilities around 6.9 percent (coefficient of variation). Thus, even using a theoretically perfect measurement method for glucose, a patient with an average fasting glucose concentration of 126 mg/dL would have actual (reported) concentrations spanning from 109 to 143 mg/dL in 19 out of 20 different days, and even higher than 143 and lower than 109 mg/dL in the remaining one out of 20 days. This variability would be worse if the sample was obtained at different times of the day, since fasting plasma glucose is higher in the morning than in the afternoon.7

Even greater variability in the measurement of glucose is introduced by the manner in which the blood sample is collected and handled before its arrival in the laboratory. While plasma is recommended as the preferred specimen,4 some clinicians might be using results obtained on serum specimens (the liquid portion of blood after clotting). Serum intrinsically has glucose concentrations that are five percent higher than plasma.8

Furthermore, while glucose is stable for 72 hours in refrigerated serum or plasma that has been separated from blood cells by centrifugation, it changes rapidly if the cells are allowed to remain in contact with serum or plasma because the blood cells will continue to consume it. The rate of disappearance of glucose in the presence of blood cells has been reported to average 10 mg/dL per hour,9 but the rate increases with glucose concentration, temperature, white blood cell count, and other factors.10 This problem can be attenuated by the addition of fluoride to the blood collection tube (gray-top tube), but fluoride takes about one hour to effectively stop the consumption of glucose by blood cells.11,12 The rates of decline of glucose in the first hour after sample collection in tubes with and without fluoride are virtually identical.9

Since physician offices might separate plasma from cells at various times after collection or submit the uncentrifugated (whole blood) original fluoride tube for analysis, even a perfect test would underestimate to a variable extent the amount of glucose in the sample.

Some of the analytic variables involved in the testing of glucose include differences within the same lot of collection tubes, lot-to-lot variation in collection tubes, lot-to-lot variation of reagents, calibration, and instrument-to-instrument variation.

Due to the variability of measuring glucose, in order to make a diagnosis of diabetes, the American Diabetes Association, or ADA, recommended that a patient’s glucose value be confirmed on a subsequent day to make certain that the patient’s glucose value exceeded 126 mg/dL on more than one occasion.3,4

Instrument manufacturers establish their calibrator values for the glucose assay using the National Institute of Standards and Technology material, or material that has been standardized to NIST. Because of issues related to the lack of availability of the NIST standard in 2003 and in part of 2004, manufacturers apparently assigned glucose calibrator values that caused slightly elevated glucose results on patient samples. During this period of time, laboratories were well within the total allowable error as determined by both the CAP and the Clinical Laboratory Improvement Amendments of 1988. CLIA allows a total allowable error of ± 6 mg/dL or 10 percent for glucose, whichever value is greater. We believe the minimal positive shift caused by the slightly biased calibrator value assignment had little or no effect on patient results, especially when considered in the context of the significantly larger biological and preanalytical variability affecting plasma glucose measurements, particularly the negative bias due to the glycolytic activity of blood cells in the sample.

A review of results of CAP proficiency testing for glucose during this period revealed that all laboratories using automated glucose platforms/systems gave comparable results.

During this period, some physicians contacted laboratories saying they had a higher number of patients with abnormally high glucose values. We believe this was largely due to the stricter ADA classification for impaired fasting glucose imposed in 2004 and not due to any analytical issues.

Currently, the National Cholesterol Education Program and the Centers for Disease Control and Prevention jointly define the accuracy of laboratory tests for lipids and the proper protocol for sampling. As a result, there are standards against which laboratories are able to compare their lipid test results. There is no similar oversight committee for glucose assays, however, that defines appropriate preanalytical, analytical, and postanalytical requirements for testing.

We need to ensure at a minimum that NIST standard material is plentiful and is kept current. Physician awareness needs to be raised about the numerous variables affecting glucose measurements. Some clinicians should be reminded not to make the diagnosis of diabetes or impaired glucose tolerance based on a single measurement and to consider laboratory results in the clinical context.

We will need to work with the American Diabetes Association to assist in recommending guidelines for the most optimal glucose samples and encourage both the CAP and the ADA to consider forming an expert committee with representatives from laboratories, manufacturers, and practicing physicians to establish improved guidelines for glucose sample collection and processing. This expert committee will need to develop a national system to monitor and ensure the accuracy of glucose measurements and to address how to appropriately interpret borderline results and followup actions.


  1. National Diabetes Data Group. Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes. 1979;28:1039-1057.
  2. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the expert committee on the diagnosis and classification of diabetes mellitus. Diabetes Care. 1997;20:1183-1197.
  3. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Follow-up report on the diagnosis of diabetes mellitus. Diabetes Care. 2003;26:160-3167.
  4. American Diabetes Association. Standards of medical care in diabetes. Diabetes Care. 2005;28:S4-S36.
  5. Sacks DB, Bruns DE, Goldstein DE, Maclaren NK, McDonald JM, Parrott M. Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus. Clin Chem. 2002;48:436-472.
  6. Ollerton RL, Playle R, Ahmed K, Dunstan FD, Luzio SD, Owens DR. Day-to-day variability of fasting plasma glucose in newly diagnosed type 2 diabetic subjects. Diabetes Care. 1999;22:394-398.
  7. Troisi RJ, Cowie CC, Harris MI. Diurnal variation in fasting plasma glucose: Implications for diagnosis of diabetes in patients examined in the afternoon. JAMA. 2000;284:3157-3159.
  8. Ladenson JH, Tsai LM, Michael JM, Kessler G, Joist JH. Serum versus heparinized plasma for eighteen common chemistry tests: Is serum the appropriate specimen? Am J Clin Pathol. 1974;62:545-552.
  9. Chan AY, Swaminathan R, Cockram CS. Effectiveness of sodium fluoride as a preservative of glucose in blood. Clin Chem. 1989;35:315-317.
  10. Ladenson JH. Nonanalytical sources of variation in clinical chemistry results. In: Sonnenwirth A, Jarett L, eds. Clinical Laboratory Methods and Diagnosis. St. Louis, Mo.: CV Mosby; 1980:149-192.
  11. Le Roux CW, Wilkinson SD, Pavitt DV, Muller BR, Alaghband-Zadeh J. A new antiglycolytic agent. Ann Clin Biochem. 2004;41(Pt1):43-46.
  12. Foucher B, Pina G, Desjeux G, Cheminel V, Prevosto JM. Stability of glucose in blood collected with or without antiglycolytic agent. Ann Biol Clin (Paris). 2004;62(5):601-604.
Dr. Schwartz is vice president and chief laboratory officer; Dr. Reichberg is corporate pathologist, science and technology; and Dr. Gambino is chief medical officer emeritus—all at Quest Diagnostics, Corporate Medical Department, Lyndhurst, NJ.