Measurement of circulating glucose concentrations: the time is now for consistency among methods and types of samples

Abstract

Diabetes mellitus was first described in the Egyptian papyrus of Ebers in 1552 BC by people who noted the attraction of insects to the “sweet urine” of those who produced abnormal amounts and later by tasters of urine who could fashion a diagnosis by recognizing the “sweet taste”. The name “diabetes”, initially used in the 1st century AD, comes from the Greek word meaning “to pass through”. “Mellitus”, a Latin word meaning “sweet as honey”, was added in the 18th century. The first chemical tests to measure sugar in urine were developed in the early 19th century. Early in the last century, whole-blood colorimetric analyses became available to clinicians, despite complexities in the preparation and analysis of samples. The term “blood glucose” remains a prominent feature to describe the state of glycemia in patients with diabetes, whether as a fasting value, during a glucose tolerance test, or in monitoring glucose values to determine the insulin dose. Measurements of glucose in venous whole blood, capillary whole blood, or plasma still define diabetes mellitus by WHO criteria (1), with the American Diabetes Association (ADA) referring only to concentrations in plasma (2). Modern applications of glucose assays practically diverge in the diagnosis vs monitoring of treatment(s) of diabetes. Thus, diagnosis is best made in a certified laboratory with the measurement of glucose in plasma after a fast of at least 8 h (3). The benefits of relating closely monitored glucose concentrations to treatment decisions have become feasible and practical with the availability of small devices capable of measuring with precision and accuracy glucose from low to high concentrations in very small quantities of blood or extracellular fluid. The fact that most devices sample whole blood presents a dilemma when expressing values in plasma vs whole blood. In this issue, D’Orazio et al. (4) have admirably and directly addressed these conundrums with a strong recommendation to report all measured glucose concentrations (whether assayed in plasma, whole venous blood, whole capillary blood, or serum) as if they were assayed in plasma. Their commentary (and importantly, Fig. 1 in their report) suggests a constant to convert values in whole blood to comparable units in plasma. They aver that uniform application of their recommendations will rationalize the use of glucose measurements for diagnosis and monitoring of treatment. Certainly the consistent reporting of all values in plasma equivalents will reduce the 11% bias when values are reported in whole-blood equivalents but interpreted as equating to values in plasma. In addition, they recommend expressing glucose values in SI units. During the usual course of managing their disease, patients with diabetes mellitus will experience glucose measured on different samples by different techniques. For example, they may encounter the healthcare system in the emergency department or otherwise in a hospital or clinic and have venous blood collected and processed for measurement of glucose in plasma. Alternatively, if treated with insulin, they will quite frequently determine glucose concentrations with some kind of hand-held device, which most likely measures glucose in whole blood obtained from capillaries (5). Harmonization of the values reported by these two sets of instruments remains an imperative for clinical medicine in general and the management of diabetes in particular. Thus, the clinician may (and usually does) review the results with the expectation that the glucose values derived from measurement of whole blood or plasma are equivalent. However, that may not always be the case. D’Orazio et al. (4) have proposed a sensible solution with values reported exclusively as equivalents in plasma. This is not the first time these recommendations have been placed before the community of laboratory professionals. In 2001, an IFCC working group in a report (6) very similar to that of D’Orazio et al. (4) (and with several authors in common) recommended that glucose meters be harmonized to the concentration of glucose in plasma, regardless of the technology or type of sample. Two of the authors presented the same recommendation in a still earlier article (7). Similarly, the Clinical and Laboratory Standards Institute (formerly NCCLS) guidelines for point-of-care blood glucose testing published in 2002 recommend that institutions reporting plasma glucose results from instruments in the clinical laboratory should use point-of-care devices that report plasma-equivalent results (8). The practical applications of these recommendations remain somewhat uncertain. D’Orazio et al. (4) emphasize that the WHO guidelines for the diagnosis of diabetes mellitus quote glucose concentrations in plasma, venous whole blood, and capillary whole blood. If one were to use the contrasting concentrations to discriminate “normal” vs impaired glycemia vs diabetes mellitus according to the WHO guidelines, manufacturers o