Date: Friday 30th of July 2010
Laboratory News Directory is not responsible for the content of external internet sites
Point of care meters are widely used for measuring glucose in hospitalised patients. Here we ask, just how reliable are the results of current hospital glucose meters?
There is increasing awareness that the clinical accuracy and reliability of commonly used meters can be adversely affected by components and substances found in the specimen matrix of hospitalised patients. Managing and maintaining glucose levels in critically ill hospitalised patients plays a role in improving clinical outcomes by reducing mortality and morbidity rates as well as length of stay1-4. In recognition of this glycaemic control protocols are now widely used in critical care settings and blood glucose testing has become a high volume testing parameter in hospitals.
Rapid and frequent testing is required in order to manage safe and effective glycaemic control and because of this blood glucose testing in hospital patients has devolved from central laboratory testing to the use of near patient blood gas analysers or more commonly used POC glucose meters5,6.
POC Glucose meters were originally developed for self-monitoring of glucose levels in adult diabetics in a home environment. In this environment, patients are generally in good health and the meters are able to perform an important function helping diabetics to self-test and safely self-administer Insulin. However many of these meters have migrated into use in hospital settings with limited validation of their use in critical patient settings requiring glycemic control7-9. In addition developments in meter technology have been focused on improving ease of use, establishing IT connectivity, and testing smaller blood volumes more rapidly and not aimed at overcoming some of the patient and specimen barriers that can adversely affect clinical accuracy.
As hospital glucose meter testing has become more widespread questions have surfaced about the accuracy and reliability of results and it is now recognised that the clinical accuracy of nearly all of the glucose meters commonly used in hospitals are affected by components or substances often present in the blood matrix of critical care patients7-9. This raises concern about the impact of inaccurate glucose results on insulin therapy decision-making particularly as insulin has been labelled as one of the top five high risk medications in the in patient setting10.
One of the key components affecting glucose meter accuracy is the red blood cell concentration (haematocrit level) present in the whole blood specimen, which can adversely affect glucose meter enzyme reaction rates. Many different types of patients within a hospital setting will have abnormal haematocrit levels including, ICU, NICU, trauma, oncology, dialysis and surgical patients. In an adult ICU settings between 30-70% of patients can have abnormally low haematocrit levels which can decline by 1.5 - 2% per day spent in ICU. In NICU patients abnormally high haematocrit levels can often occur11-13.
Recent studies have shown that abnormal haematocrit levels can have a significant adverse affect on the accuracy of current glucose meters14-16. A number of different reasons have been proposed to explain the haematocrit effect on commonly used glucose strips. An increase in the number of red blood cells in the whole blood may mechanically impede diffusion of plasma into the reagent reaction region of the strip by blocking the pores in the mesh membranes or decreasing the plasma volume available to diffuse to the reaction surface. Haematocrit changes may alter blood viscosity, therefore, decreasing the fluid permeability into the reagent reaction layer. In addition, the increased viscosity results in a slower rate of diffusion that leads to measurement errors14,16.
High haematocrit levels can give falsely low glucose results and conversely low haematocrit levels can give rise to falsely high glucose results. Many of these studies show an adverse affect on clinical accuracy within the hematocrit range claimed by manufacturers as having no effect. In adult ICU patients inaccurate results giving falsely elevated glucose results increases the risk of insulin dosing errors and also increases the risk of hypoglycaemia17.
In addition to haematocrit a number of other drugs (eg acetaminophen, ascorbic acid), hormones, metabolites and additives such as maltose have been shown to significantly affect the clinical accuracy of the most commonly used glucose meters8,16, 18-19. As a result it is now recognised that the poor accuracy and reliability of these commonly used meters can increase the risk of an adverse incident as demonstrated in a number of recent regulatory alerts in critical care patients20-23.
A recent assessment of commonly used glucose meters by simulation modelling indicates that a total glucose error rate of 15% or more compared to the true glucose value increases the likelihood of a two-step insulin dosing error and the likelihood of an adverse incident24. A recent evaluation of four commonly used hospital glucose meters in a critical care setting demonstrated average error rates of >16%, with some results showing >35% bias compared to the reference value25. The meters fail to meet the acceptance criteria of the international quality standard for glucose testing, ISO 15197, which in itself has criteria limits set higher at +20%26. The consequences of using inaccurate meters increases the risk of insulin dosing errors, which in turn can result in a failure to maintain safe and effective glycaemic control and as a result increase the risk of hypoglycaemia.
Increasing awareness of the inadequate accuracy of many commonly used glucose meters has meant that some critical care units have been cautious about using them, preferring instead to use blood gas analysers. This limits testing flexibility and is not patient friendly in an NICU setting because of the larger specimen volumes used. The advent of more accurate meters will improve testing reliability, flexibility and ease of use throughout the hospital environment and avoid the need for tiered testing approaches.
The poor accuracy of many commonly used glucose meters is related to design limitations that do not correct for interfering substances. In response to the need for improved glucose meter accuracy a recently available glucose meter StatStrip Glucose (Nova Biomedical) has been developed specifically for hospital use and designed to eliminate specimen interference. The patented multi-well and multilayer technology incorporates control wells that measure and correct for hematocrit interference and also corrects for other common interfering substances. The electrochemistry technology is layered onto a gold platform providing a stable and robust surface for the electrochemical reaction kinetics. The gold layer also provides a platform for batch-to-batch manufacturing consistency and reproducibility that eliminates the need for batch-to-batch calibration.
The design of the meter has been challenged in several recent analytical studies and the accuracy has been shown to be unaffected by abnormal haematocrit levels and other common interferences27-32. A meta-analysis performed on the results of 1703 samples obtained from 35 US evaluation sites showed close concordance to the reference method (figure 1) with results meeting ISO 15197 acceptance criteria33.
One of the published studies reported that: "The StatStrip glucose meter correlated best with a plasma hexokinase reference method and over a wide range of glucose concentrations and was least significantly impacted by specimen hematocrit and other interfering substances. This should allow for better management of critically ill patients on tight glycaemic control protocols."
To challenge the design of StatStrip Glucose in a clinical setting, studies have been performed in hospitals throughout Europe including ICU patients (Belgium and Italy), NICU patients (UK and Holland) and in dialysis patients (UK)30-32,34-35. In these studies the accuracy of results was shown to be unaffected by a wide range of abnormal hematocrit levels present in the patient populations tested. These studies demonstrated that StatStrip met the acceptance criteria of ISO 15197 and the TNO (Netherlands Organisation for Applied Scientific Research) approved protocol for glucose meter measurements, which were not achieved by the other commonly used glucose meters tested. In an ICU study on 500 patients the performance and accuracy of StatStrip glucose was shown to be equivalent to the BGA method routinely used35.
These studies substantiate that the new generation design concept of glucose meters - like the StatStrip Glucose - eliminates common interferences and in doing so sets a new standard in clinical accuracy for POC glucose meters used to manage glucose levels in hospitalised patients.
References: