Quality of Glucose Measurement with Blood Glucose Meters at the Point-of-Care: Relevance of Interfering Factors

Introduction

S elf-monitoring of blood glucose (SMBG) is a central part of modern diabetes therapy and key to achieve and maintain good metabolic control in individuals with diabetes. Despite some recent discussions about the optimal target level for metabolic control, it still appears to be logical to keep metabolic control as close as possible to the physiologic level as long as this is in accordance with the individual therapeutic target and adverse effects (e.g., hypoglycemic events) can be avoided. Optimal coverage of prandial insulin needs relies on knowledge of the preprandial glycemic level by means of SMBG, estimation of the carbohydrate content of the meal, and calculation of the insulin dose. Patients are trained for this complex task by participation in teaching and training programs.

In line with the need to take all aspects of SMBG into account to achieve a blood glucose (BG) measurement with a high quality, it is not sufficient to check for (potential) interferences that impair the accuracy of a given BG meter (see below), as this is only one component in the overall “total system performance” of the meter in the hands of patients. Thus, also safeguards must be provided as well as consistent, easy-to-understand labeling. ¹ In addition, quality assurance standards should be set and achieved or exceeded. Support and education must be provided at all times.

BG meters are also used regularly nowadays in clinical settings, e.g., on the ward for diabetes treatment and in the so-called point-of-care-testing. The main reason for doing is that the handling time for glucose measurement in the central laboratory is too long for rapid treatment adjustments (e.g., tight glycemic control within the intensive care unit). The requirements the BG meters have to fulfill (also with respect to accuracy and precision) are different between the patient and the clinical setting: for the individual patient it is not the absolute value which is key (precision), but the relative difference (accuracy). In contrast, the clinical setting tends to apply the quality criteria used for clinical chemistry laboratory analyzers to the handheld BG meters.

Modern BG meters allow rapid glucose measurement in a small blood volume with a high reliability under daily life conditions and in the hospital setting. It is of note that the systems per se have a measurement quality that comes close to that of laboratory analyzers. However, like with each and every (medical) technology, certain factors can impair measurement quality. BG meters have to fulfill certain quality requirements stated in International Organization for Standardization ISO NORM 15197, 2004(E). It appears as if not all BG meters that are on the market stick to these requirements (and that required for a Conformité Européenne [CE] mark; see below) in practice, especially some of the low-cost providers. ² If a given BG meter with a given lot of test strips fulfills the requirements, the questions remains if all devices and lots do so, especially under daily life conditions. Unfortunately, no independent academic institution exists that runs systematic evaluation of BG meter performance regularly. Recently a discussion was initiated by the Food and Drug Administration (FDA) to improve the accuracy of BG meters in the point-of-care testing setting by asking for fulfillment of higher requirements.

Clearly it is of relevance to know to which extent endogenous/exogenous substances in blood influence the reliability of SMBG measurements. Interfering substances can either originate from drugs prescribed to or taken by the patient under daily life or from special therapies initiated by the medical staff in the hospital setting, e.g., peritoneal dialysis. In fact, in recent years several serious and even a few fatal clinical events have occurred based on inaccurate BG measurements due to interfering substances. ³

The aim of this (mini-)review is to provide an overview of the (published) knowledge about the impact of interfering substances and other factors on the measurement quality of (modern) BG meters. One word of caution is required right away: publications that are <5 years old probably do not adequately reflect the performance of the most recent generation of BG meters. In view of the very rapid progress of the technology used for BG measurement, results obtained with a previous generation of a given BG meter can differ massively from that obtained with the recent one. Additionally, it was checked if certain “performance parameters” of six different BG meters, which use different enzymatic measurement technologies, have an impact on the measurement quality. It is well known that all enzymatic reactions interfere with certain substances; however, they differ from each other significantly with respect to which substance and to which degree they interfere. Historically glucose oxidase (GOD) was used predominately for BG measurement; however, test strips based on this enzyme have a significant oxygen dependency. To overcome this limitation—especially in the clinical setting where blood samples of different origin (arterial, venous, capillary) are used—the enzyme glucose dehydrogenase (GDH) was been used to develop test strips without a significant oxygen dependency.

Substances Known to Impair Glucose Measurement

When a drop of blood is applied to the BG test strip, plasma-like filtrate passes into the reagent layer of the test strip, which contains enzymes, coenzymes, mediators, and indicators. Table 1 gives an overview of the enzymes and coenzymes used in the test strips of six different BG meters. The enzymes GDH and GOD confer specificity for glucose. However, neither GDH nor GOD is completely specific for glucose; depending on the enzyme–coenzyme combination used (GDH-flavin adenine dinucleotide, pyrroloquinolinequinone [PQQ]-dependent GDH [GDH-PQQ], or GDH-nicotinamide adenine dinucleotide) different substances can interfere with the principal reaction (Table 2). Such interferences can affect the reliability of BG measurements (for a detailed analysis of the published literature, see below).

Two substances are discussed in more detail: (1) acetaminophen and (2) maltose. The latter substance gained attention recently by the regulatory agencies as this interference can have deleterious effects under special clinical conditions. ^1,3

Acetaminophen (paracetamol), as the most widely used analgesic and antipyretic agent, is used in various formulations by about 200 million consumers worldwide. Many acetaminophen-containing drugs are freely available over-the-counter, and some patients use them in high doses. Above certain concentration levels acetaminophen starts to affect BG measurements, causing inaccurately high BG results. Because of different drug metabolization rates the concentration level of acetaminophen at which BG results are affected varies from patient to patient. Unfortunately, the risk of possible interference with the reliability of SMBG measurements is most often not mentioned in the labeling of acetaminophen-containing drugs. However, also in the labeling of BG test strips the risk of interference with acetaminophen is not described unanimously, i.e., the amount of information provided by the different manufacturer differs widely (Table 2).

Maltose is a disaccharide consisting of two glucose molecules linked by an α-1,4-glycosidic bond. Physiologically, maltose is an intermediate product in the degradation of polysaccharides (e.g., starch) that are an important part of our diet. Such carbohydrates are enzymatically digested in the intestine and completely hydrolyzed by different enzymes to the monosaccharides glucose, galactose, and fructose. Therefore, maltose is not detectable in blood under normal conditions, i.e., after oral uptake. However, maltose is used in certain drugs as an excipient, i.e., as a stabilizing agent or an osmolality regulator. A comprehensive list of maltose-containing or maltose-generating drugs is given in Table 3. In fact, only parenterally administered maltose can lead to detectable maltose levels in blood. For example, maltose is contained in solutions used for peritoneal dialysis or in immunoglobulin solutions that are intravenously administered. From the listed drugs—assuming that they are administered at the recommended dose and infusion time—only three of the seven drugs (Extraneal^® [Baxter Healthcare Corp., Deerfield, IL], Octagam^® [Octapharma Pharmazeutika, Lachen, Switzerland], and Adept^® [Baxter Healthcare Corp.]) lead to concentrations of maltose in blood that are likely to have a clinically significant interference with BG measurement.

The manufacturers of the drugs, listed in Table 3, as well as the BG meter manufacturers, have implemented a variety of measures to mitigate the risk of maltose interference with SMBG:

Appropriate drug/medical device labeling

A recent review of the existing literature about the pharmacokinetics of maltose concluded that the circulating levels of intravenously administered maltose—like with most other substances/drugs—depends on the patient's health status (e.g., healthy vs. renally impaired), the dose, infusion rate, and total infusion time. ¹⁴ Based on the doses of maltose applied, the patient-dependent metabolization of maltose, and the detection limit for maltose for a given BG meter (see below), a time period of at least 24 h is needed after maltose was applied before BG measurement provides reliable data again.

Only test strips for BG measurement based on the GDH-PQQ enzyme exhibit an interference with maltose. ¹ Any concentration of maltose in blood causes inaccurately high BG measurement results, because of the nonspecificity of the GDH-PQQ enzyme for the monosaccharide glucose alone but also because it reacts with the disaccharide maltose. Clearly the manufacturers of BG meters have evaluated the impact of maltose on the reliability of the BG measurement with their systems. For example, Ghys et al. ¹⁵ investigated the influence of maltose on two point-of-care BG meter systems (Precision PCx^® [Abbott Laboratories, Abbott Park, IL] and Accu-Chek^® Inform^® [Roche Diagnostics, Indianapolis, IN]). Maltose interfered with the BG meter based on GDH-PQQ (Accu-Chek Inform), but not with the meter based on GDH-NAD (Precision PCx). The manufacturers of BG meters have tried to improve the glucose specificity of the GDH-PQQ enzyme to reduce the cross-reactivity with maltose to <2% of the wild-type. ^16,17 Use of such a genetically modified GDH-PQQ enzyme should minimize the risk of maltose interference.

It is also worth acknowledging that the number of patients with diabetes treated with intravenous maltose (glucose measurements might also be performed in patients without diabetes in an intensive care unit!) appears to be extremely small, as such therapies are limited to highly specific conditions in a hospital setting or in specialized outpatient centers: among the 124 million patients with diabetes worldwide (estimation of the International Diabetes Federation for 2007) it is assumed that 352,000 (0.28%) have a disease that is potentially treatable with a drug containing maltose. ¹⁸ More specifically, from the 13.6 million patients diagnosed with diabetes in the United States, this number is 39,700 (0.29%), and about 173,000 (0.14%) patients with diabetes worldwide (19,700 [0.15%] in the United States) are actually treated with a parenterally administered drug containing maltose (drug list and estimated plasma concentrations, 2008, from IMS Health, Inc., Norwalk, CT).

Other Factors Influencing Measurement Quality of BG Meters

Beside substance interferences, BG meters differ in their specific performance parameters and specifications when it comes to BG measurement (Table 4); for example, the hematocrit in a given blood sample can have an impact on the measurement result. The hematocrit normally ranges from 37% to 47% for women and from 42% to 52% for men. A low hematocrit (as a measure for the concentration of blood cells in a blood sample) can result in inaccurately high BG measurement results. Anemia, certain types of cancer, chronic and end-stage renal disease, malnutrition or specific diet deficiencies, rheumatoid arthritis, and other conditions (e.g., during intensive care unit treatments of neonatal as well as adult patients) may lead to a low hematocrit. Anemia occurs when there is a shortage of red blood cells or when red blood cells are not correctly formed. In patients on dialysis the hematocrit normally stabilizes between 20% and 25%, i.e., it is lower than in healthy subjects. Therefore, BG measurements performed in such patients may be inaccurate if a BG meter is used that is sensitive to the hematocrit.

Knowledge About Interfering Substances and Other Factors

A literature search of the database PubMed for relevant, English publications from January 1, 1980 to August 10, 2009 was performed—one overall search and three specific searches with the following search terms: (1) overall search = (“interference” OR “performance” OR “limitations”) AND (“SMBG” OR “self-monitoring of blood glucose” OR “self monitoring of blood glucose” OR “blood glucose self-monitoring” OR “blood glucose self monitoring” OR “blood glucose monitoring” OR “blood glucose measurement”), (2) maltose-specific search = “maltose” AND (“SMBG” OR “self-monitoring of blood glucose” OR “self monitoring of blood glucose” OR “blood glucose self-monitoring” OR “blood glucose self monitoring” OR “blood glucose monitoring” OR “blood glucose measurement”), (3) “acetaminophen-specific search” = (acetaminophen OR paracetamol) AND (“SMBG” OR “self-monitoring of blood glucose” OR “self monitoring of blood glucose” OR “blood glucose self-monitoring” OR “blood glucose self monitoring” OR “blood glucose monitoring” OR “blood glucose measurement”), and (4) hematocrit-specific search = (“hematocrit” OR “haematocrit”) AND (“SMBG” OR “self-monitoring of blood glucose” OR “self monitoring of blood glucose” OR “blood glucose self-monitoring” OR “blood glucose self monitoring” OR “blood glucose monitoring” OR “blood glucose measurement”).

The “overall search” resulted in 275 hits (30 selected), the “maltose specific search” in four hits (three selected), the “acetaminophen specific search” in 10 hits (nine selected), and the “hematocrit specific search” in 42 hits (16 selected). Some of the hits from the different searches were identical. The abstracts of the hits were screened for relevance ( = selected). Relevant publications were screened for additional, pertinent citations. Finally, 29 publications were included in our literature survey (Table 5).

Analysis of the Published Literature

The phenomenon of substance/drug interference and other factors influencing the quality of BG measurement has been known for a long time. In many studies several different factors were studied at the same time; therefore it was decided to present the studies in historical order and not separate them into those dealing with interfering substances and those presenting data about other factors.

Already in 1985 Rice and Galt ⁴⁰ described the in vitro interference of ascorbic acid, acetaminophen, or salicylic acid with SMBG systems. They spiked blood from patients without diabetes—with BG levels in the normal range—with different therapeutic concentrations of the above-mentioned drugs. They observed a drug concentration-dependent interference with the three SMBG systems assessed. As stated above, the recommendation made in their publication (“BG values measured by SMBG systems should be interpreted with caution for patients taking ascorbic acid, acetaminophen, or salicylic acid”) must be interpreted with caution as the technology of BG test strips has evolved rapidly in the last 25 years. In the year 1988 electrochemical BG sensors were introduced. Updike et al. ⁴² in 1988 described the first enzyme-electrode BG sensor designed specifically for pocket-portable SMBG systems. They reported that the new sensor type has the following advantages:

It is reusable for at least 30 days.

They found that the sensor is inherently linear, independent of hematocrit, and shows no interferences with blood plasma constituents, heparin, or acetaminophen.

Giordano et al. ²⁷ in 1989 described the influence of altitude on seven SMBG systems. They observed that six of the seven SMBG systems underestimated BG levels at “high” altitude (2,073 m).

Devreese et al. ²² in 1993 compared five SMBG systems with respect to hematocrit interference. In the high BG glucose range there was a decrease in BG values with increasing hematocrit for all meters, but the error was smaller for Accutrend^® (Roche Diagnostics) and Glucocard^® (Arkray, Kyoto, Japan). In the normoglycemic range the results obtained with Accutrend and Glucocard were not influenced by even extreme hematocrit values.

Moatti-Sirat et al. ³⁵ in 1994 presented an evaluation of a glucose sensor coated by sulfurized polytetrafluorethylene. They observed a clear reduction of the sensitivity of the glucose sensor to acetaminophen in rats and humans.

Wiener ⁴⁴ in 1995 compared a point-of-care glucose testing (POCGT) system based on the hexokinase reaction with a laboratory reference method with respect to hematocrit and ambient temperature. The author stated that if this system is used as intended, there is no clinically significant influence of these two parameters.

Tang et al. ⁴¹ in 2000 performed a comprehensive study assessing possible drug interferences in six SMBG systems and one POCGT system in vitro. They included 30 drugs used primarily in critical care and hospital settings and assessed possible drug interferences at different drug dosage levels relevant in such settings, at both low (80 mg/dL) and high (200 mg/dL) BG levels. Paired differences of BG measurement results between drug-spiked samples and unspiked control samples were calculated. In this study ascorbic acid interfered with BG measurements of all SMBG/POCGT systems evaluated; acetaminophen, dopamine, and mannitol interfered with some of the systems. The clinical relevance of the drug interferences were determined by comparing dose–response curves (i.e., drug concentration vs. glucose difference) with the therapeutic levels and the reported clinical concentrations of the drug. The authors recommended that clinicians choose SMBG systems carefully and interpret results cautiously when BG measurements are performed during or after drug interventions. In view of the widespread usage of these drugs this statement is far-reaching and indicates an issue that does not exist in most cases with modern BG meters.

Hussain et al. ³⁰ in 2000 studied the influence of elevated hematocrit values on BG measurements in 180 newborns. They measured BG with test strips in both capillary and venous blood samples, performed a BG measurement with a laboratory reference method (hexokinase), and determined the hematocrit value. Their study suggested that elevated hematocrit values in newborns affect the BG measurement results in capillary blood samples less than in venous blood samples.

Louie et al. ³² in 2000 investigated the clinical effects of critical care variables, including pressure of O₂ (pO₂), pH, pressure of CO₂, and hematocrit, on two BG measurements in 247 critical care patients. Interestingly, only the hematocrit affected both meters in a clinically relevant manner: low hematocrit resulted in a positive bias, whereas high hematocrit resulted in a negative bias.

Pecchio et al. ³⁹ in 2000 tested the effects of altitude (sea level vs. 3,000 m) with two BG meters in a small field test with six healthy subjects. At sea level, both meters tended to underestimate BG levels; at 3,000 m one tended to overestimate, and the other tended to underestimate, BG levels. However, the bias was not clinically meaningful in both cases, and the authors therefore concluded that both meters can be used safely at moderately high altitude.

Vote et al. ⁴³ in 2001 studied the accuracy of two POCGT systems under hyperbaric oxygen conditions, i.e., with a high pO₂. Whereas one system was almost unaffected by high pO₂, the other system showed considerable bias and imprecision under hyperbaric oxygen conditions. However, the practical relevance of this evaluation is unclear.

Nawawi et al. ³⁸ in 2001 measured the BG levels in 114 venous blood samples with a BG meter and a reference method at temperatures from 21–22°C to 33–34°C. They found no effect of the ambient temperature on the measurement quality of this meter.

Fink et al. ²⁴ in 2002 studied the influence of altitude, temperature, and relative humidity on the precision and accuracy of BG measurement with seven meters. They performed a thorough field analysis during climbing on a mountain (change in altitude from 0 m to 3,749 m). The precision of the BG measurement was affected by altitude as follows: BG levels were underestimated by approximately 1–2% for each 300 m gain in altitude. However, this effect was less significant after appropriate adjustment for temperature and relative humidity. Beside altitude, temperature and relative humidity also affected the performance of the BG meters studied. That is why in modern BG meters an efficient temperature compensation algorithm is implemented.

Moberly et al. ³⁶ in 2002 studied the absorption, plasma kinetics, and elimination of icodextrin ( = maltose) and its metabolites in 13 patients treated with peritoneal dialysis. The metabolism of absorbed icodextrin and the resulting rise in plasma level of small glucose polymers did not result in hyperglycemia or hyperinsulinemia. ²⁵

Gutman et al. ²⁸ in 2004 analyzed both mandatory and voluntary reports on in vitro diagnostic devices to the FDA from 2000 to 2002. The vast majority of in vitro diagnostic device reports (84%, n = 18,959) were on BG meters; only 333 in vitro diagnostic device reports were voluntary reports from users. Among the user reports, the most common problems are false high or low BG values and erratic BG values.

Kannan et al. ³¹ in 2004 reported on two cases of patients with diabetes receiving an intravenous immunoglobulin preparation containing maltose (Intragam^®, CSL Ltd., Broadmeadows, VIC, Australia) who experienced hypoglycemia because of inappropriate antidiabetes therapy adaptations due to falsely elevated BG readings.

Erythrocytes are a major source of response variation in biosensor electrodes measuring glucose in whole blood. This blood-to-plasma difference is called the “hematocrit effect.” Typically, sensors for BG exhibit a decreasing response to BG in the presence of increasing hematocrit levels. In an effort to minimize the hematocrit effect, Forrow and Bayliff ²⁶ in 2005 presented a commercial whole BG biosensor with a low sensitivity to hematocrit based on an impregnated porous carbon electrode that excludes erythrocytes and is consequently able to measure glucose with an acceptable quality in venous, capillary, arterial, and neonatal blood over a wide hematocrit range of 20–70%.

Michel et al. ³⁴ in 2005 evaluated the usability of the POCGT device Glucometer Elite XL (Bayer, Tarrytown, NY) for screening for neonatal hypoglycemia in comparison with a laboratory reference method. In blood samples from 869 newborn infants they found a negligible dependence of BG on hematocrit with both the POCGT device and the laboratory reference method. However, they advised (1) shifting the cutoff for hypoglycemia detection of the POCGT device from 2.6 mmol/L to 3.2 mmol/L to compensate deviating results and (2) confirming hypoglycemia by additional measurements in the clinical laboratory.

Hoftman ²⁹ in 2005 studied the interference between Extraneal peritoneal dialysis solution and BG measurements performed with the GDH-based Accu-Chek BG meter. As described above, an overestimation of BG was observed because of “cross-reactivity” of the measurement of this meter. Hoftman referred to a case report by Mehmet et al. ³³ from 2001 on three patients with insulin-treated diabetes and end-stage renal disease treated with peritoneal dialysis. They experienced symptomatic hypoglycemia in spite of high BG readings, which is fully understandable in hindsight knowing the maltose-induced measurement error.

Bamberg et al. ¹⁹ in 2005 studied the effect of adverse storage conditions on the performance of BG test strips for SMBG. Test strips were stored in open or closed vials, at the conditions recommended by the manufacturer (+2–25°C) or at adverse conditions, i.e., elevated temperature (+37°C), direct light, increased humidity, or refrigeration (+4–8°C) during a period of 50 days. Not surprisingly, BG test strips stored in closed vials performed better than BG test strips stored in open vials. For open BG test strip vials in adverse storage conditions, the refrigerated environment offered the best stability (35–50 days); direct light and increased humidity resulted in a reduced stability (3–14 days).

Floré and Delanghe ²⁵ in 2006 also reported on the interference of icodextrin (maltose) used in continuous ambulatory peritoneal dialysis with BG meters based on the enzyme GDH.

Bilen et al. ²⁰ in 2007 compared the performance of BG test strips based on GDH versus GOD with respect to altitude. They observed that the BG values measured at about 2,000 m altitude with BG test strips based on GDH were higher than those measured in the reference laboratory, but not that measured with BG test strips based on GOD. They suggested usage of GOD-based BG test strips in elevated regions to detect hypoglycemia.

As described above, Ghys et al. ¹⁵ in 2007 investigated the influence of hematocrit and maltose on two POCGT systems (Precision PCx and Accu-Chek Inform). Both systems showed a negative bias primarily at high hematocrit levels, but differed with respect to maltose.

The influence of the perfusion index on the quality of BG measurements in critically ill patients was investigated by Desachy et al. ²¹ in 2008. They compared BG values determined by a meter versus a laboratory reference method and observed that a low perfusion index (reflecting peripheral hypoperfusion) was associated with poor measurements quality, i.e., disagreement between meter and laboratory reference by more than 20%. This study highlights that under such conditions in hospitals BG measurements have to be interpreted with care.

Eastham et al. ²³ in 2009 studied the prevalence of interfering substances with POCGT systems based on GDH-PQQ BG test strips in a California community hospital. Over a 12-month period they identified all inpatients with serum uric acid concentrations >10 mg/dL, hematocrit <20% or >55%, serum total bilirubin concentrations >20 mg/dL, serum acetaminophen concentrations >8 mg/dL, and serum triglyceride concentrations >5,000 mg/dL. From 6,885 hospital admissions, 84 patients (1.2%) were identified as having interfering substances. From these patients, 30 (0.4%) had an active order for an insulin product during the interference time interval. As a consequence, in this specific hospital a screening and notification procedure was installed that identifies and “flags” patients who are potentially affected by substance interference: their BG values are not determined with POCGT but in the hospital core laboratory.

In a recent overview on SMBG Montagnana et al. ³⁷ in 2009 specifically commented on interferences of BG meter with hematocrit, maltose, and acetaminophen. The authors highlighted that the operators of such devices should be aware of such factors before taking measurement results for granted. ¹

Requirements that BG Meters Have to Fulfill for a CE Label

BG meters used for SMBG belong to the group of in vitro diagnostic medical devices. In Europe, such systems have to comply with the directive 98/79/EC on in vitro diagnostic medical devices, which specifies general requirements to ensure overall safety and quality. ⁴⁵ Additionally, the standard DIN EN ISO 15197:2003 specifies requirements for SMBG systems, e.g., with regard to system performance, accuracy, and precision. ⁴⁶ Manufacturers of BG meters have to provide evidence of conformity with these norms in order to get the CE label for their products. They submit the accuracy data required by the standard to a notified body of the regulatory authority. The CE label is granted based on the manufacturer's accuracy data as well as an assessment of the manufacturer's quality management system. In contrast, in order to get an FDA approval the manufacturers have to provide detailed preclinical and clinical data.

Total System Performance

Improving the analytical performance of BG meters ( = accuracy) and avoiding interferences are important; however, there is a chain of additional aspects that determine the quality with which individuals with diabetes measure their capillary BG levels in daily life. So, if one is interested in the question with which quality people with diabetes measure their glycemia, and not that highly trained technicians under optimally environmental conditions obtain, these factors have to be taken into account in order to improve the total system performance of a BG meter: system limitations, system safeguards, labeling, quality assurance, and support and education. One can assume that higher levels of patient health and safety can be achieved with the adoption of additional industry standards for these aspects; however, systematic studies investigating these are missing widely.

Current standards mandate that only one lot of test strip lot is used for testing. In view of the fact that—very much depending on the manufacturing process—considerable differences between lots exist, this should be increased to three lots to demonstrate good lot-to-lot performance. Today, clinical evaluations of test strips most often are conducted at the manufacturer's site. To mitigate the risk of biased results, such evaluations should be performed at external sites, ideally by an independent research organization.

Meters and testing supplies can be affected by environmental (e.g., temperature, humidity, and air pressure) and physiological variables (hematocrit). Currently, there is no standardization how to evaluate the impact of such parameters on the quality of BG measurement. Consumers should have confidence that whatever product they choose has been rigorously tested under different circumstances and proven effective.

A system safeguard means that individuals with diabetes should be warned of factors that can cause an incorrect reading that are otherwise not obvious to them. Whereas this is (or at least should be) obvious for everybody on the diabetes team, for people with diabetes it is quite often not clear that an incorrect BG reading can happen because the test strip expired, that the test strip is damaged by extreme temperature conditions, etc. Thus, it is not trivial to warn individuals with diabetes that they should not leave a test strip vial open, etc. Safeguards can help to prevent common user mistakes by simple labels on the test strip vial, but also technology can be of help, if, for example, underdosing (not enough blood on the test strip) is automatically detected. Also, test strip expiration, exposure to abnormal temperature, and low meter battery can be detected by “smart” test strips/BG meters.

Interestingly, the product labels on diagnostic material must be submitted to regulatory authorities, like the FDA, for clearance; however, there are no standards governing their format. Therefore, each diabetes product comes with a unique label ( = instructions for use). Depending on the amount of information provided—and how it is provided—this can be difficult to understand for individuals with diabetes. User-friendly standards regarding product labels—which are standard in other areas—most probably are of help: they should be clear, concise, and easily understandable by people with diabetes and clinicians. This might sound trivial; however, in reality the information provided by the labels is not easy for patients to understand and does not meet the needs of a diverse audience. Labeling should also provide information about system accuracy, interferences, limitations of use, and product safeguards. A standardizing labeling would also allow individuals with diabetes to make a truly informed choice about what products are best in managing their care.

Unfortunately, no specific guidelines exist for the handling of complaints about medical devices, analysis of trends, and the continued performance of each device. All manufacturers should comply with regulatory requirements for medical device reporting, including international medical device reporting reporting. Also, a specific guidance for complaint handling and reporting of events associated with SMBG should be established that allows detecting issues that show up with a certain lot of test strips or BG meter. Such practices should help to ensure patient safety. Offering support regarding the proper use and administration of diabetes products is too important not to be a requirement for all makers of test strips, meters, and other diabetes testing supplies.

Technical support should not be limited to being provided by each manufacturer just from 9 to 5, but 24 h a day, 7 days a week. Technical product support by a well-trained call center (with around the clock service!) is also quite helpful in this sense. People with diabetes quite often have the feeling that they are left alone with many practical questions once they have bought the BG meter. Ongoing product education (including appropriate materials) would also be of help for individuals with diabetes. Comprehensive technical support and education help to ensure that people with diabetes receive appropriate assistance to be safe and live a healthy life.

Conclusions

In daily life, people with diabetes (and physicians) tend to believe the measurement results of their BG meter without questioning the number provided. In fact, this is a valid assumption in most cases. However, in the framework of an adequate patient instruction for the handling of a given BG meter, individuals with diabetes should be instructed about factors that can impair the measurement quality of BG meters (e.g., the impact of temperature and altitude) to enable a reliable performance of SMBG. They should also be informed about the differences between BG meters in their sensitivity to interfering substances, e.g., acetaminophen. ¹ Ideally a given patient should stick to a given BG meter to avoid differences in the results depending on the enzyme technology used in different meters. For people with diabetes it is of paramount interest that the BG measurement works reliable not only at room temperature and at normal altitude, but also in direct sunlight on the beach or while skiing in the winter.

Under clinical conditions the influencing factors are different, whereas environmental factors are of limited relevance, and interfering substances/drugs become of increasing relevance. The BG meters used differ in their susceptibility for such substances, i.e., the currently used generation of BG test strips that uses the GDH-PPQ enzyme shows a relevant interference with maltose. However, it has to be acknowledged that the respective patient group—patients who obtain highly specific, parenterally administered therapies—is quite small. This enzymatic technology also has profound advantages when it comes to other factors impairing BG monitoring, for example, the oxygen dependency is minimal. Usage of this type of test strips in combination with parenterally administered maltose-containing drugs is inadequate (off-label).

Only through a comprehensive approach to total system performance can individuals with diabetes be assured that their testing supplies and products are safe, accurate, and effective. As people with diabetes and physicians/hospital teams usually do not read the product labeling/package inserts of BG test strips/potentially interfering medication, adequate training within the practices/clinical quality management system is mandatory. Thus, we should not focus our view on the analytical performance of BG meters only, but also have the diagnostic performance of the “total system performance” in mind. As this is a novel approach, it will take some time until acceptance for a more systematic approach is achieved.