As a physician, I appreciate the convenience of blood glucose point-of-care testing (POCT), which allows almost immediate assessment of the critically ill patient. It is also a valuable resource for many of the more than 20 million Americans diagnosed with diabetes—one that empowers patients to take an active role in their own health.
In my training and practice, I have seen the results of glucose POCT lead to the reversal of life-threatening conditions like diabetic ketoacidosis and hypoglycemic seizures. Despite the clear benefits, however, glucose POCT is not without flaws. Its convenience is sometimes tied to inaccurate results that can cause errors in medical management.
History of glucose POCT
Ames produced the first commercially available glucose POCT in 1970, which measured blood glucose levels using the glucose oxidase (GO) reaction. Weighing in at five pounds and at a cost of $500, it was primarily available in the emergency room setting. The device revolutionized care by streamlining medical management and eliminated wait times of up to an hour for clinical laboratories to provide time-sensitive results.
Over the ensuing five decades, glucose POCT has evolved, securing a foothold in everyday clinical practice. Researchers eventually developed a prototype for non-invasive glucose POCT that measured the concentrations in the aqueous humor of the eye and continuous glucose monitoring (CGM) provides feedback as often as every five minutes for patients managed on insulin therapy. While innovative, these technologies do not yet rival the prevalence of standard glucose POCT, which has migrated into patients’ homes as a critical tool in diabetes management.
Etiology of inaccuracies
Accurate glucose POCT depends on a host of factors, from manufacturing technique to operator training and patient factors. Errors within any of these areas can compound, leading to worsening inaccuracy. While no true gold standard exists for glucose testing, POCT results are compared to laboratory plasma glucose levels when surprising or inconsistent. The following is a list of common sources of error in glucose POCT:
- Whole blood measurements
POCT glucometers conveniently use whole blood measurements at the patient’s bedside, then convert the results to plasma glucose concentrations, the standardized form used in clinical practice. While the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) recommends a conversion factor of 1.11, the equation assumes standardized parameters for hematocrit, plasma, and red blood cell water concentrations that may not be accurate in individual patients.
- Enzymatic tests
POCT glucometers use GO or glucose-1-dehydrogenase (GDH) reactions. GO is affected by hydration status and blood oxygen levels. While variation in these parameters is less common in ambulatory settings, they can cause significant errors in dehydrated or hypoxic acutely ill patients. The GDH reaction (of which there are three subtypes) is not specific to glucose and can cause overestimation of glucose levels due to the presence of other blood sugars, such as galactose. In comparison, clinical laboratories use the hexokinase reaction, which is the most accurate but does not have the suitable shelf life required for distribution and storage necessary for POCT glucometers.
- Manufacturer errors
All manufacturers do not necessarily conform to the IFCC conversion factor and may use alternative methods to calibrate devices. In addition, there is significant variability in the quality of glucose test strips, which can be damaged by heat, moisture, or altitude.
- Patient factors
Clinicians should be aware that POCT glucometers tend to be less accurate in patients who are hypoglycemic (as opposed to hyperglycemic), hemodynamically unstable, or have poor peripheral circulation. Additionally, elevated blood concentrations of bilirubin, lipids, and uric acid can interfere with accuracy. Newer glucose POCT allows sampling from areas other than the fingertip, which may differ from traditional fingerstick results.
- Operator training
Both patients and healthcare providers using glucometers must be trained adequately in the use of the devices. Operator errors include insufficient device cleaning, test strip handling, lack of handwashing, and errors in the timing of sampling (i.e., sampling after eating when testing should be before meals).
Some patients, notably those with Type 1 Diabetes, use CGM to provide near-continuous assessment of glucose levels. Generally, patients calibrate CGMs to routine glucose POCT. However, a recent device is factory calibrated and does not require routine fingerstick testing. Regardless, CGMs use the same enzymatic reactions—most commonly GO— to monitor glucose levels via the interstitial fluid. As such, they can suffer from similar errors as routine POCT.
Patients can easily upload both glucose POCT and CGMs readouts to software that allows seamless integration of the number, timing, and results of patient glucose levels for physician review.
Guidelines for accuracy
Both the International Organization for Standardization (ISO) and Clinical and Laboratory Standards Institute (CLSI) have recently published guidelines regarding accuracy in glucose POCT. CLSI’s guidelines are more rigid as they include POCT use in hospital settings:
- ISO: 95 percent of results should be within 15 mg/dL when glucose levels are less than 100 mg/dL and within 15 percent when levels are greater than 100 mg/dL.
- CLSI: 95 percent of results should be within 12 mg/dL when glucose levels are less than 100 mg/dL and within 12.5 percent when levels are greater than 100 mg/dL. In addition, when the cutoff level is 75 mg/dL, 98 percent of results should be within 15 mg/dL and 20 percent for results below and above the cutoff, respectively.
Accuracy of commonly available glucometers
A recent evaluation of commercially available glucometers revealed that more than half do not meet guideline recommendations, leaving patients at risk of at least one in 20 results being incorrect. Importantly, glucose meters are increasingly inaccurate in hypoglycemic patients, which is concerning due to the high risk of hypoglycemia with certain diabetic medications. Table 1 indicates the overall accuracy of selected glucometers.
Table 1: Accuracy of several commercially available glucometers
The future of non-invasive glucose testing
Although it has been more than three decades since the first noninvasive glucose POCT was developed, there are still no widely available commercial devices in use. Despite the numerous technologies, which include infrared light spectroscopy, radio wave impedance, and analysis of exhaled metabolites, significant errors in accuracy remain, with some devices being pulled from the market due to significant error in glucose POCT results. Hurdles remain to accurately detect glucose levels through the skin, due to significant variations in water concentration in body fluids and impedance from circulating proteins.
The future remains bright, with a plethora of innovations seeking to free patients from finger sticks. I eagerly await what will come next to improve the care of diabetic patients.