Abstract
In clinical practice, glycemic control is generally assessed by measuring and interpreting glycated hemoglobin levels, however, this test should be run under standardized conditions. We focus here on the crucial steps to ensure IFCC standardized HbA1c results, pointing out several residual weak points, mostly relating to the laboratory end-user (calibration, quality control materials, and EQAS). We also review the use of HbA1c for diagnosing diabetes and the various indicators useful for assessing glucose variability because in some cases they seem to represent a patient’s glucose profile more accurately than one-off HbA1c assays. Finally, the potential utility of glycated albumin and the glycation gap, the costs involved and the laboratory management issues are briefly discussed.
Introduction
Diabetes is considered a health problem of epidemic proportions. The IDF Diabetes estimated that 285 million people around the world had diabetes in 2010, representing about 7% of the world’s adult population [1]. The ADA recommends screening for type 2 diabetes and assessing the risk of future diabetes in asymptomatic adults of any age who are overweight or obese and have one or more additional diabetes risk factors, and also in asymptomatic adults with no risk factors ≥45 years of age [2]. In the light of the findings of an International Expert Committee, the ADA recently recommended using the HbA1c test to diagnose diabetes, adopting a threshold of 6.5% (48 mmol/mol) [3]. These recommendations have also been adopted by the American Association of Clinical Endocrinologists/American College of Endocrinology [4] and, to some degree, by the IDF [5].
This document reviews some of the topics, in the field of laboratory medicine, relevant to the diagnosis and monitoring of diabetes, paying particular attention to the use of HbA1c, glucose variability indexes and glycated albumin.
Critical issues for ensuring standardized HbA1c measurements
Many papers have been written about the various analytical aspects relevant to a correct HbA1c measurement, making it difficult to add anything new. It is worth mentioning some excellent reviews on the topic [6, 7], and also a recent contribution [8] in which we have tried to group together the analytical interferences that could be overcome by appropriate sample handling or by choosing the most appropriate method for measuring HbA1c, from those in whom some particular physiological or pathological conditions may be present, usually not known a priori (see table 1 in [8]).
Essential steps in order to provide IFCC standardized HbA1c test results.
| Actions | Tools |
|---|---|
| Choice of method | Assess the IFCC certificate (ask the manufacturer) |
| Calibration | Enter the IFCC target values provided by the manufacturer (ask the manufacturer for traceability to the IFCC reference system) |
| Reporting the HbA1c result | Use the mmol/mol units (possibly converting them afterwards into % units) Report decisional limits (not reference intervals) |
| Monitoring long-term imprecision | Internal Quality Control with materials at normal and high HbA1c content Calculate the CVs monthly (or over a longer time interval) |
| Evaluating trueness | Regular participation in EQAS exercises (commutable materials, IFCC target values assigned by the IFCC reference measurement procedures) |
Something that warrants further attention, in our opinion, is the standardization of the methods at the end-user level. Some recent papers emphasize that this issue needs to be considered carefully. For instance, Elder et al. pointed out that further assay standardization would be advisable in the UK, either because internal data are obtained by different analytical systems, or because of an unacceptable variability in the data collected through the UKNEQAS program, in which the results were quite often excessively variable for clinical purposes [9]. A similar warning was published by Lenters et al. in a more recent work, in which they showed an unacceptable level of imprecision for one in five laboratories involved in a Dutch/Belgian external quality assessment scheme (EQAS) with 220 laboratories participating [10]. Analytical problems also seem to arise if certain point-of-care testing (POCT) systems are used to measure HbA1c instead of using a conventional laboratory technique [11, 12]. Finally, recent data from a German EQAS exercise again seem to indicate that inter-laboratory variation is not entirely acceptable, and that the overall mean still differs significantly from the one obtained by the IFCC reference measurement procedure, or by a newly-developed LC-ID-MS procedure [13].
To shed some light on this issue, we would like to retrace the metrological chain for the traceability of laboratory activities, and to repeat that achieving a global standardization involves duties and responsibilities at various levels, from the scientific societies and metrological institutes to the diagnostics manufacturers, and finally to the end-user clinical laboratories [14, 15]. In the specific case of HbA1c, the whole chain has been defined [16], primary materials have been prepared and characterized [17], an IFCC-approved reference method has been developed [18], and a network of reference laboratories has been created [19]. Close cooperation between the manufacturers and the IFCC Network has been in place for several years with a dedicated EQAS, the IFCC Monitoring Program [20]. Master equations were developed and published for converting the results obtained using methods calibrated with the IFCC reference system into those obtained by using NGSP-aligned methods [21], and an international consensus on the implementation of the IFCC reference system was signed [22], and renewed 3 years later [23]. An agreement with the manufacturers on the implementation of the reference system has also finally been reached [24]. Analytical goals for imprecision, bias and total error have been defined, based either on clinical needs [25] or on biological variations [26], and it has ultimately been demonstrated that these goals are different when the HbA1c results being reported are expressed in different units [27].
Coming back to the laboratory operators, we feel that these professionals should essentially accomplish all the steps outlined in Table 1 in order to be IFCC standardized.
First of all, a valid method should be chosen, and this entails conducting a careful analysis of the IFCC certificate. This document (see the facsimile in Figure 1) is obtainable directly from the manufacturer and certifies to the method’s alignment with the IFCC reference system thanks to the manufacturer’s participation in the above-mentioned monitoring program [28]. In the example shown in Figure 1, the trueness of the measurements is reported for three HbA1c levels (30, 60 and 90 mmol/mol): on the whole, this performance is very good, showing only a slight underestimation at low HbA1c values, balanced by a slight overestimation at the highest HbA1c concentrations. A bias of ±1 mmol/mol (approx. equivalent to ±0.1%, expressed in NGSP units) is perfectly acceptable. As for the reproducibility, the reported value of 2.4% is also acceptable, the goal being within 2.8% [26]. The report also shows a figure for linearity, and this should theoretically always be ≥0.999.
![Figure 1
Example of a certificate released to a manufacturer by the coordinator of the IFCC network, reporting the results of the Monitoring Program (reprinted with the Editor’s permission from [28]) modified.](/document/doi/10.1515/cclm-2012-0594/asset/graphic/j_cclm-2012-0594_fig_001.jpg)
Example of a certificate released to a manufacturer by the coordinator of the IFCC network, reporting the results of the Monitoring Program (reprinted with the Editor’s permission from [28]) modified.
A second point concerns the method’s calibration: this should be done by inserting the IFCC target values provided by the manufacturer directly in the apparatus, but in our experience, this is not always what happens in the field. In some cases, calibration is done in percentage units, and then reporting is handled in mmol/mol units using the ‘master equation’, which is not really right!
Then there is a third point, reporting. According to the consensus statement on the implementation of the IFCC reference system [22, 23], this has to be done in mmol/mol, and possibly also in a second unit. Various national societies have taken a position on this aspect, and some European countries have the IFCC units already in place, while others have double reporting systems, and an IFCC integrated project is now dedicated to this topic at global level [29]. It is not our intention to discuss the above-mentioned consensus statement in the present document, but simply to make the point that it becomes complicated, in the long run, to retain both units because:
1) the analytical goals differ for mmol/mol and percentage units [27]; 2) only IFCC units should be used, given the relatively poor specificity of the NGSP reference system, as illustrated by the following considerations. Let us consider the famous DCCT study, and particularly the relationship between mean plasma glucose and HbA1c [30]. As we all remember, a lesson learned from that study was that a worsening plasma glucose control is associated with a higher risk of microvascular complications, and possibly also with a greater risk of macrovascular and neurological complications. It was demonstrated, moreover, that a 1% (10 mmol/mol) increase in HbA1c correlated with an average increase in mean plasma glucose of approximately 36 mg/dL, as confirmed by a later re-assessment of the DCCT data (see figure 1 in [31]). Figure 2 shows the relationship between mean plasma glucose and HbA1c emerging from the DCCT study (Figure 2A), and we have plotted both axes starting from the origin for both glucose and HbA1c. We have also plotted a confidence interval of ±15% glucose values, as proposed at the time for the ADAG study [32]. There is a clearly evident negative intercept of the regression line, due essentially to the limited specificity of the NGSP method. In Figure 2B, we have plotted the regression line of the same data in the DCCT study processed by Rohlfing et al. [31], this time by converting the NGSP-HbA1c values into IFCC-HbA1c values using the master equation published by Hoelzel et al. [21], adapted so as to convert the percentage units into mmol/mol units [i.e., NGSP-HbA1c=0.0915 (IFCC-HbA1c)+2.15%]. Clearly, the regression line now passes through the origin of both axes, as one might logically expect based on practical biochemical considerations (no glucose, no HbA1c).
![Figure 2
Relationship between mean plasma glucose calculated from seven-point blood glucose profiles and glycated hemoglobin in the Diabetes Control and Complications Trial (DCCT).
(A) Original data, reproduced from [30]. (B) Same data, with HbA1c expressed in SI units.](/document/doi/10.1515/cclm-2012-0594/asset/graphic/j_cclm-2012-0594_fig_002.jpg)
Relationship between mean plasma glucose calculated from seven-point blood glucose profiles and glycated hemoglobin in the Diabetes Control and Complications Trial (DCCT).
(A) Original data, reproduced from [30]. (B) Same data, with HbA1c expressed in SI units.
Still on the subject of reporting, most clinical laboratories report reference intervals and we feel that this custom should be abandoned, for at least a couple of reasons. First of all, Braga et al. demonstrated that between-subject variability is much greater than within-subject variability [26], with a mean CVi (2.5%) lower than the figure (3.4%) reported by Ricòs et al. [33], and a slightly higher mean CVG (7.1%) compared with the one published in the database (5.1%). We do not want, to this regard, to discuss which number is more correct. However, it is important to remember that the biological variability is an incompressible component of the total variability, due to a set of physiological and biochemical events not easily distinguishable separately from each other. In particular, previous works have established that there is some heterogeneity in red cell life span as well in non-diabetic subjects recruited from the general population, as in type 1 and type 2 diabetic subjects [34]. Mean RBC ages were found between 38 and 60 days in non-diabetic controls and between 39 and 56 days in diabetic subjects, and the impact of red cell age on HbA1cwas such that three patients with similar glucose control could have very different HbA1c levels, i.e., between 7.5% and 9.9% (58–85 mmol/mol, respectively), if they would have mean RBC ages between 38 and 60 days. In other words, variation of red cell life span could result in variation of ±12% (CV) in the HbA1c values, at the same extent of glucose control [34]. Another potential source of between subjects variability in HbA1c could be related also to differences in glucose transport/gradient across the red cell membrane, as clearly proven by an elegant study with radioactive glucose analogues [35]. In this investigation, the authors speculated that differences between two different individuals showing various gradients in glucose concentrations between outside and inside the red cells could result in a 25% difference in the level of hemoglobin glycation.
Therefore, in our opinion, in the same way as for other analytes, it is preferable to use subject-based reference intervals in such cases [36] or, if such values are unavailable, then clinical decisional limits (diagnostic cut-offs) should be adopted. ADA also seems to recommend this [37]. It is worth bearing in mind that, given the small value of intra-individual variation, only two specimens are needed to determine an individual’s homeostatic HbA1c set point to within 5% of the true mean value [26]. There is evidence of HbA1c values being different in different ethnicities, and some data indicate an effect of age (see later on). Indeed, the latest version of the AACC guidelines for laboratory analyses in the diagnosis and treatment of diabetes mellitus [38] also contain a rather generic comment on the reference intervals (For NGSP-certified assay methods, reference intervals should not deviate substantially (e.g., >0.5%) from 4% to 6% (20–42 mmol/mol). For these various reasons, we feel that using cut-off values would be preferable to either lower limits [HbA1c levels below 4.3% (23 mmol/mol) are very rare in our experience] or higher limits [HbA1c≥6.5% (48 mmol/mol)]. In Table 2, we propose some of the most useful decisional limits for orienting the interpretation of HbA1c test results. In particular, as concerns pregnancy, O’Connor et al. recently defined trimester-specific reference ranges for HbA1c using a HPLC method well-aligned with the IFCC [40]. As for the reference change values, which are useful for assessing the significance of changes in results obtained by analyzing serial samples from the same individual, an average CD of 10% can be used, as reported elsewhere [26].
Possible decisional limits for HbA1c.
| HbA1c (%) | HbA1c (mmol/mol) | Utility | Reference |
|---|---|---|---|
| 4.0 | 20 | Lower limit of healthy subjects | [35]a |
| 5.4 | 36 | Upper limit in pregnancy (1st and 2nd trimester) | [36] |
| 5.7 | 39 | Lower limit for prediabetes | [39] |
| 6.4 | 46 | Upper limit for prediabetes | [39] |
| 6.5 | 48 | Diagnosis of diabetes | [3–5] |
| 7.0 | 53 | ADA primary goal of therapy | [2] |
aIn our experience we very rarely find healthy subjects with HbA1c<4.5% (<26 mmol/mol).
Finally, particular attention should be paid to monitoring imprecision and trueness. This can be accomplished by a regular use of the IQC procedures to confirm and verify the manufacturer’s declared performance of commercial systems, and by participating in the EQAS, using commutable control materials with target HbA1c values assigned by the reference measurement procedure, in order to be able to define the uncertainty of laboratory measurements [15]. Regarding the IQC, we would like to draw attention to the proper use of commercial control materials. In the package insert that comes with the materials, the assigned values are usually reported together with confidence intervals or ranges (depending on the manufacturer). In several cases we have found that these intervals are really too wide, and this can be very misleading in daily practice. For instance, just to quote two major diagnostics manufacturers, the leaflet for the Lyphochek Diabetes Control (Bio-Rad Laboratories) mentions a range from 8.8% (73 mmol/mol) to 11.2% (99 mmol/mol) for Level 2 control, and for the PreciControl HbA1c material (Roche) we find a range from 7.3% (56 mmol/mol) to 10.5% (91 mmol/mol). What if these were whole blood samples instead of control materials? How can we accept that a given patient could have an HbA1c of 7.3% (56 mmol/mol) according to a measurement obtained at one clinical laboratory and 10.5% (91 mmol/mol) according to another? We strongly urge the manufacturers to reduce the ranges of their control materials, at least enough to make them compatible with the proposed total error limit.
Trueness is best assessed by participating routinely in an EQAS program, providing that commutable materials are used and the HbA1c titer is assigned by the IFCC reference measurement procedure, or by reference laboratories. A list of the laboratories capable of providing such a service is available in the JCTLM database [41]. Unfortunately, these conditions are never met by lyophilized materials, which always suffer from non-commutability problems, as already demonstrated [39, 42].
HbA1c for diagnosis and monitoring
Background
Numerous studies have recently been published on this issue and some of them warrant consideration. In an unselected emergency department population, Menchine et al. found a very high proportion of adults with undiagnosed diabetes [22% of patients with a BMI≥30 kg/m2 had an HbA1c≥6.5% (48 mmol/mol)] and a high prevalence of suboptimally controlled diabetes (almost 45% of those with a history of diabetes). The authors consequently supported the use of simple diagnostic criteria, such as HbA1c tests, in emergency departments for screening purposes [43]. Other authors have supported the use of HbA1c measurements in acute care settings, since their value is uninfluenced by recent eating or acute illness, and the cut-offs used to screen for pre-diabetes and diabetes are similar to those used in outpatient settings [44]. A further analysis on the 12,485 participants in the Atherosclerosis Risk in Communities (ARIC) study, and on a subpopulation of 691 participants in the NHANES III, identified a better performance of HbA1c assay [≥6.5% (≥48 mmol/mol)] in diagnosing prevalent diabetes and predicting incident diabetes than the use of single or repeated standard glucose measurements [45]. In a Dutch population, on the other hand, Riet et al. analyzed 2753 participants in the New Hoorn Study and found that considering HbA1c levels ≥6.0% (42 mmol/mol) as diagnostic would mean missing almost half of the people with diabetic glucose levels. The cut-off used to screen for diabetes was an HbA1c level of 5.8% (40 mmol/mol), which achieved a sensitivity of 72% and a specificity of 91%: this cut-off would detect 72% of patients with diabetes and 30% of cases at high risk of developing diabetes. Finally, 44% of patients with newly-diagnosed diabetes had HbA1c levels <6.0% (42 mmol/mol) [46] when tested using a standardized DCCT method (reverse-phase exchange chromatography).
Adding a glucose-based method to the process for diagnosing diabetes improves the performance of HbA1c, as confirmed by other analyses [47, 48]. In 2332 oral glucose tolerance test (OGTT)-diagnosed diabetic subjects from a Chinese population aged 35–75 years, FCG (fasting capillary blood) measurements performed better than HbA1c as a screening tool for newly-diagnosed diabetes and pre-diabetes [49]. Using HbA1c as a single diagnostic tool for detecting type 2 diabetes may result in an overestimation of the cases by comparison with the OGTT-based method, as reported by Mostafa et al. in a study involving a multi-ethnic cohort [50].
Another analysis on 6890 adults without previously-diagnosed diabetes from the 1999–2006 NHANES study showed a reasonable consistency between HbA1c and fasting glucose measurements for the purposes of diagnosing diabetes (k=0.60; 95% CI 0.55, 0.64), although discrepancies were reported for non-Hispanic black patients [51].
Ethnicity
The question of whether HbA1c and glucose homeostasis differ among various ethnic groups has recently been debated. In particular, Chapp-Jumbo et al. conducted an in-depth analysis and found significantly higher HbA1c levels in healthy black adult offspring of parents with type 2 diabetes than in their white counterparts [5.68±0.033% vs. 5.45±0.028% (39±2 vs. 37±2 mmol/mol)], after adjusting for age, hemoglobin, hematocrit, BMI, fasting plasma glucose, and glucose AUC [52]. In an Italian cohort of 1019 Caucasian individuals, there was again a moderate consistency between HbA1c levels and FPG criteria in diagnosing type 2 diabetes (k coefficient=0.522), but 35.9% of the individuals in this study with diabetes according to the FPG criteria alone had HbA1c levels <6.5% (48 mmol/mol), confirming that different diagnostic criteria identify different diabetic populations [53].
Analyzing the Leicester Ethnic Atherosclerosis and Diabetes Risk (LEADER) cohort showed that using HbA1c as a diagnostic test significantly increased the frequency of cases of diabetes, especially among the southern Asian population, while approximately one-third of the people previously identified as diabetic by OGTT were found non-diabetic using HbA1c assay [50]. In a Chinese population (2332 subjects aged 35–74 years), using the diagnostic cut-off of 6.5% (48 mmol/mol) for HbA1c generated a sensitivity of <30% in both men and women, i.e., lower than for plasma glucose. The optimal cut-off for detecting diabetes in this population was an HbA1c level of 5.6% (38 mmol/mol), i.e., significantly lower than in the new ADA recommendations [49], but this cut-off proved unsuitable for the South African population, for which Zemlin et al. identified a level of 6.1% (43 mmol/mol) as the optimal cut-off [54].
Finally, an analysis on six studies conducted from 1999 to 2009 in Denmark, the UK, Australia, Greenland, Kenya and India tested the probability of HbA1c levels being ≥6.5% (48 mmol/mol) among cases of diabetes diagnosed by means of OGTT. The prevalence of diabetes was lower in four of the six studies when the HbA1c-based criteria were used. Indeed, the probability of HbA1c≥6.5% (48 mmol/mol) among the OGTT-diagnosed cases ranged from 17% to 78%, with significant differences between different ethnic groups. Although methodological differences may have influenced these results, the impact of ethnicity on the different screening procedures needs to be clarified [55]. In this setting, an extensive analysis of the NHANES survey covering the years from 1988 to 2006 has shown that the prevalence of diagnosed diabetes has increased since 1988, while the percentage of undiagnosed diabetes has decreased and so has the prevalence of subjects at high risk of developing diabetes, though the latter two conditions remain common in the USA, and elderly and minority groups are disproportionately affected. The fact that HbA1c detects a much lower prevalence of diabetes than plasma glucose levels should therefore be borne in mind [56].
Age effect
When it comes to elderly patients, the use of HbA1c to diagnose diabetes is somewhat controversial. A cross-sectional analysis on data from the Health, Aging and Body Composition study at year 4 (2000–2001) compared the sensitivity and specificity of HbA1c-based diagnoses with those based on FPG in the elderly. The proportions of undiagnosed diabetes were 3.1% and 2.7%, respectively, using FPG≥126 mg/dL and HbA1c≥6.5% (48 mmol/mol) [57]. Another study conducted on Asian patients using ROC curve analysis found that an HbA1c test result of 6.2% (44 mmol/mol) was the best cut-off for diagnosing diabetes, but the sensitivity and specificity of the area under the ROC curve deteriorated as the age groups increased [58]. In addition, there are several practical implications to consider when HbA1c is adopted as the preferred tool for diagnosing diabetes, as already explained [8].
Judging from the above-mentioned studies, it seems that using HbA1c alone to diagnose diabetes would lead to a number of diagnoses being overlooked, while the sensitivity and specificity of the assay could be improved by combining it with fasting blood glucose or the OGTT.
HbA1c and risk of diabetic complications
The preference for the use of HbA1c in the diagnosis of diabetes has also been supported on the strength of its supposed prediction of cardiovascular risk but, here again, published studies provide contradictory evidence. In a Spanish cross-sectional epidemiological survey, individuals without known diabetes underwent an OGTT and HbA1c assay, and several cardiovascular risk factors were assessed. The authors demonstrated that subjects newly diagnosed as diabetic according to the new criteria recommended by the ADA had a worse cardiovascular risk profile [a higher BMI and waist circumference, lower high-density lipoprotein (HDL) cholesterol concentrations, higher HOMA-IR and fibrinogen concentrations] than patients diagnosed using the glucose-based criteria [59]. In a study including approximately 1300 diabetic patients prospectively followed up for cardiovascular disease, Paynter et al. identified significant improvements in the ability to predict CVD risk using models that included HbA1c levels by comparison with classifications of diabetes as a cardiovascular disease equivalent. HbA1c has sometimes been shown to identify individuals at risk of diabetes in much the same way as fasting blood glucose, but is better able to predict cardiovascular events (multivariable-adjusted hazard ratio of 1.95 for HbA1c≥6.5% [48 mmol/mol]), supporting the use of HbA1c assay as a test for diagnosing diabetes [60]. However, in a study conducted in 1015 patients with coronary artery disease (CAD) and no previous diagnosis of diabetes, the results of HbA1c measurements were compared with those of an OGTT for the purposes of early diabetes diagnosis. The analysis demonstrated that OGTT identifies more patients with glucose impairments who undergo coronary angiography than the HbA1c test using a cut-off of ≥6.5% (48 mmol/mol). In this set of patients, therefore, measuring HbA1c alone seems to miss a substantial proportion of cases with hitherto-unknown diabetes. OGTT measurements were also found to correlate significantly with the severity of CAD (ρ=0.09, p=0.01), unlike the HbA1c assay findings [61]. Moreover, as demonstrated in a study by Sturm et al., HbA1c values were sometimes inversely associated with all-cause mortality in diabetic dialysis patients, in addition to presenting no significant associations with CVD events [62].
Long-term glycemic control has proven to be one of the most important parameters for preventing chronic complications, and microvascular disorders, in particular. Identifying indicators better able to predict daily glycemic profiles has consequently been an important field of clinical research in recent years. HbA1c represents the combination of both fasting and postprandial glucose variations over a 3-month period [63], proving very useful in providing background information to orient effective patient treatment decisions [64]. The importance of HbA1c as a ‘gold standard’ for monitoring blood glucose in patients with diabetes was confirmed by the DCCT and UKPDS mega-trials [30, 65], which demonstrated the close relationship between HbA1c values and the onset of chronic complications of diabetes in the long-term. In the follow-up of these trials, despite no further differences in the treatment administered, the advantage of intensive diabetes treatment persisted and even extended to a reduction of patients’ micro- and macrovascular events, regardless of their HbA1c values [66, 67]. This phenomenon is called ‘metabolic memory’ and has to do with an as yet unexplained HbA1c-independent mechanism that contributes to the onset of chronic complications [68]. One of the most important determinants in this mechanism concerns the formation of advanced glycation endproducts (AGEs) and their accumulation in diabetic subjects, reflecting cumulative metabolic and oxidative stress, as reported elsewhere [69].
More recently, the A1c-derived average glucose (ADAG) study described the relationship between HbA1c and average blood glucose. This important work by Nathan et al. showed that a simple linear relationship exists between these two parameters, quite well-described by a simple regression equation. The conclusion was therefore that HbA1c levels could be expressed as estimated average glucose (eAG, since it is a calculated average) in both DM1 and DM2 [32]. It should be noted, however, that some ethnic groups were under-represented in this trial (especially the Asian and African populations), and its findings are only applicable to populations in stable glycemic control with no suggestion of any erythrocyte disorders. A large inter-individual variability was also noticeable. As a matter of fact, the adoption of the eAG was not endorsed in the consensus statement on HbA1c standardization [22].
In an effort to arrive at a comprehensive assessment of glycemic control, the term ‘glucose triad’ has also recently been introduced, meaning fasting glucose, postprandial blood glucose and HbA1c [70]. A number of studies have investigated the contribution of these three components to overall 24-h hyperglycemia, often with contradictory results.
Monnier et al. [71] showed that a gradually worsening glycemic control follows a ‘three-step’ temporal sequence: the earliest changes occur in the three postprandial periods (after breakfast, lunch and dinner); at a later stage, hyperglycemia is more severe in the early morning (the so-called ‘dawn phenomenon’); the last step corresponds to nocturnal hyperglycemia. Postprandial glucose levels thus contribute the most (70%) to the lower quintile (HbA1c <7.3% [56 mmol/mol]), whereas fasting hyperglycemia appears to be the main contributor to the overall diurnal hyperglycemia in patients with poorly-controlled disease [HbA1c >9.3% (78 mmol/mol)] [64]. According to the above-mentioned authors, postprandial hyperglycemia would be an early causal event in chronic glycemic damage, even in patients with well- or moderately-controlled diabetes. However, Borg et al. (again based on data from the ADAG study) showed more recently that, among all the SMBG measurements, preprandial glucose levels had a greater effect on HbA1c than postprandial levels, although FBG levels alone failed to provide a clear indication of overall hyperglycemia [70].
To sum up, no single blood glucose measurement obtainable during the day accurately predicts HbA1c [72], and that is why the contribution of pre- and postprandial glucose concentrations to overall glycemic control (HbA1c), and the role of postprandial glucose targets in disease management are still being debated. As suggested in several reports, many patients do not reach the HbA1c targets set by the published guidelines [73]. However, the goal is to cure the patient’s diabetes, not their hyperglycemia, so an optimal glycemic control, as Ceriello suggested, ‘equates to HbA1c (at target)+fasting plasma glucose (to target)+postprandial glucose (to target) without hypoglycemia and weight gain’ [74]. The low sensitivity of HbA1c levels in describing glycemic profiles implies a need to diversify the use of indicators for monitoring diabetic patients in the long-term. The indicators of fasting glucose, postprandial glucose and glucose variability each have their own pathophysiological value with a view to providing increasingly individualized treatment options.
Glucose variability
The acute glucose fluctuations that occur in diabetes have recently been considered an additional factor in the onset of chronic complications of the disease [75, 76]. In fact, a study by Monnier et al. [77] demonstrated that the urinary excretion of 8-isoprostane (a marker of oxidative stress) correlated positively with glycemic variability, as assessed by calculating the mean amplitude of glycemic excursion (MAGE). Ceriello et al. [78] also emphasized the role of acute hyperglycemic spikes in the physiopathology of micro- and macrovascular complications of DM, showing in particular that more limited hyperglycemic excursions coincide with lower levels of some oxidative stress markers. These data were not confirmed by Kilpatrick et al. [79], however, when they retrospectively analyzed the findings of the DCCT study: while mean blood glucose was predictive of microvascular complications, the within-day glucose variability (calculated as the standard deviation around the mean of a seven-point glycemic profile) was unrelated to chronic complications.
Finally, glucose variability has been seen as a plausible candidate for predicting severe hypoglycemia, because this condition is preceded by blood glucose disruptions [80]. Several studies have reported a decline in the occurrence of hypoglycemia episodes coinciding with a lower glucose variability [81].
Numerous glucose variability indicators [82] have been developed, but none of them have so far come to be considered the ‘gold standard’ for assessing glycemic variability in diabetic individuals (Table 3). There is also still no consensus as to which indicator is preferable for assessing glucose variability and for how long glycemic variability measurements should be continued.
Glucose variability indicators.
| Indicator | Meaning | Advantages | Disadvantages |
|---|---|---|---|
| Mean glucose (MG): | Mean of all glucose values | Simple and straight forward to interpret; it provides the best correlation with HbA1c values | It does not attribute more importance to hypo- or hyperglycemia |
| Standard deviation (SD): | SD of all glucose values | Simple to determine, the most classical and well-known method used by physicians | It is affected by extremely low and high glucose values, but fails to capture the rapidity and magnitude of glycemic spikes (not applicable to asymmetrical non-Gaussian distributions); it is affected by mean glucose |
| Interquartile range (IQR): | 75th percentile – 25th percentile of glucose values | ||
| J index | Combining information from mean glucose and SD→0.001×(mean+SD)2 | Combining information from the two most used CGM indicators; strongly correlated with time spent within target range | Almost unknown and not used in clinical practice; sensitive to hyperglycemia but relatively insensitive to hypoglycemia |
| Coefficient of variation (% CV): | (100×SD)/MG | Simple, unaffected by mean glucose | |
| Percentage within target range | Defined as time spent within a given glucose range (the most often used ranges are: 80–200 mg/dL, 70–180 mg/dL, 70–140 mg/dL) | Simple and straightforward to interpret, effective (sensitive to clinical intervention) | The choice of an arbitrary target range may not be optimal |
| Percentage below target range | Time spent below a given cut-off (the most often used cut-offs are: <80 mg/dL and <50 mg/dL) | Simple and straightforward to interpret, sensitive in detecting hypoglycemic events during the day | It does not attribute more weight to extremely low values |
| Percentage above target range | Time spent above a given cut-off (the most often used cut-offs are: >250 mg/dL, >200 mg/dL, >180 mg/dL, >140 mg/dL) | Simple and straightforward to interpret, sensitive in detecting hyperglycemic events during the day | It does not attribute more weight to extremely high values |
| Mean amplitude of glycemic excursion (MAGE) | Calculated as the arithmetic mean of the differences between consecutive glycemic peaks and nadirs, only considering changes in glycemic values of more than 1 SD | Describes major fluctuations; directly proportional and strongly correlated with the SD of all glucose values | Rarely used in clinical practice; statistically less efficient than SD |
| Mean of daily differences (MODD) | Mean of the absolute differences between glycemic gaps observed during the same time interval on 2 consecutive days, as an expression of ‘between-days’ glucose variability | The only indicator to describe between-days variability | Originally defined only for 2 consecutive days, assuming similar meals, physical activity and therapy on both days |
| Continuous overlapping net glycemic action (CONGAn) | The SD of the glycemic differences recorded between a specific point on the CGM profile and a point n hours previously (where n=1,2,3,4,…) | Sensitive in capturing ‘short-term’ glucose variability; opportunity to compare variability during fasting (CONGA night) and in postabsorptive period (CONGA day); strongly correlated with SD for certain n values (when n~2.5) | Difficult to interpret for physicians who do not use CGM systems very often |
| High blood glucose index (HBGI) | 27.695×(loge [glucose]1.084–5.381), for glucose values >112.5 mg/dL | Greater weight attributed to hyperglycemic values | Does not predict hypoglycemic events or low glucose trends; obscure mathematical derivation |
| Low blood glucose index (LBGI) | 7.695×[loge (Glucose)1.084–5.381], for glucose values <112.5 mg/dL | Greater weight attributed to hypoglycemic values | Does not predict hyperglycemic events or high glucose trends; obscure mathematical derivation |
| Average daily risk range (ADRR) | Average sum of |HBGI for maximum glucose|+|LBGI for minimum glucose| for each day | Combines information from HBGI and LBGI; moderate correlation with time spent within target range, determined by self-monitoring of blood glucose (SMBG) | Calculated using extreme values (min and max); more appropriate for SMBG than for CGM; for SMBG it takes 30 days, obtaining at least 4 values a day |
| Area under curve (AUC): | very similar to the percentage below or above a given target range. It can be used to describe the total glycemic exposure in a given glucose range or at a particular time of day, e.g., the postprandial AUC is calculated from the preprandial glucose levels to the highest peak over a 2-h period after a meal. |
In a study of ours, glucose variability indicators were recorded over at least 48 h using a continuous glucose monitoring (CGM) system in patients with type 1 diabetes, patients with type 2 diabetes taking multiple daily injections of insulin therapy, and type 2 patients treated with OHA and/or basal insulin. Baseline and postmonitoring HbA1c levels were also measured. HbA1c correlated significantly only with average glucose (AG) levels and the high blood glucose index (HBGI), and only in type 1 diabetic patients. There was greater glucose variability in patients with longstanding type 1 diabetes, irrespective of their HbA1c levels. An insulin regimen with MDI correlated strongly with HbA1c, but was unassociated with glucose variability. HbA1c values therefore identify states of sustained hyperglycemia and seem to be unaffected by hypoglycemic episodes or marked glucose variability, thus revealing shortcomings as a ‘gold standard’ indicator of metabolic control. Glucose variability indicators adequately represent the glucose profile of type 1 patients and identify any worsening glycemic control (typical of longstanding diabetes) more accurately than one-off HbA1c tests [83].
Judging from the above considerations, it is clear that what we know about the relationships between glycemic variability and the onset of diabetic complications is very limited, so prospective randomized trials specifically addressing this topic are needed to further strengthen the evidence in the literature.
Glycated albumin
The use of other proteins, such as glycated albumin, and the measurement of advanced glycation endproducts as predictors of the risk of chronic diabetic complications have not been mentioned in the guidelines, not even the most recent, and this is essentially because of the lack of sufficiently strong evidence to support their application [63]. We are nonetheless of the opinion that such markers should be used more. We would particularly like to stress the importance of measuring glycated albumin when interpreting HbA1c measurements could be hampered by certain pre-analytical conditions (e.g., in cases of chronic hemolytic anemia or hemoglobin variants). It is also worth bearing in mind that glycated albumin can be measured with a specific, robust and reliable enzymatic method that overcomes the limitations of the previous fructosamine test [84]. It could prove extremely interesting to apply this new test in re-assessing the data used to calculate the glycation gap (GG), proposed for the first time by Cohen et al. [85], whose work proved that GG is reproducible. In a cohort of 40 type 1 diabetic patients with a history of the disease spanning more than 15 years, the authors found that a 1% increase in their GG was associated with a three-fold increase in the cases of nephropathy, and this parameter correlated better with the severity of the nephropathies than either HbA1c or fructosamine. Cohen et al. subsequently demonstrated that GG is at least 70% inherited, as are HbA1c levels (60% inherited) [86]. Apparently, a third of the heritability of HbA1c is shared with that of GG, while the remainder is HbA1c-specific [87]. It has also been suggested that an index derived from HbA1c and glycated albumin [CLD-HbA1c=0.5 (HbA1c+AG/3)] could be more representative of mean blood glucose [88].
Diabetes and laboratory management
The most efficient medical management of diabetes mellitus, especially as regards the long-term degenerative complications, represents a key challenge not only for patients’ well-being and quality of life, but also for economic reasons. The costs of screening, diagnosing diabetes, and ensuring optimal disease control have to be considered in view of the costs of the disease’s treatment and the global social consequences for patients with severe degenerative complications, which greatly affect their independence and quality of life. Diabetes mellitus also has a substantial influence per se on the clinical outcomes and/or costs of managing patients with other pathologies, such as elderly patients with heart disease [89, 90].
It has been clearly demonstrated that the costs of improving glycemic control, and thereby delaying degenerative complications, in type 2 diabetic patients are lower than the costs generated by such complications [91]. Diabetic patients’ participation in disease management programs is also associated with lower medical costs [92]. A recent paper by Driskell et al. highlighted that glycated hemoglobin assays were inappropriately prescribed in 50% of cases [93], and this justifies the correct use of markers of glycemic control such as HbA1c for predicting all types of cost-generating complications. This applies to both short- and long-term costs [94]. The use of such markers is governed by various clinical practice guidelines, which should thoroughly consider the needs and characteristics of all sub-populations to ensure a better use of healthcare resources [95]. In the case of gestational diabetes mellitus, a recent meta-analysis reviewing relevant publications from 1975 to 2009 has shown that HbA1c could become a credible alternative to OGTT, reducing the burden and cost of identifying women at risk of associated adverse outcomes [96].
However, our current knowledge of the pathophysiology of diabetes may be influenced by new concepts with an impact on screening strategies. To give an example, the identification of a particular amino acid profile could be a promising predictor of the onset diabetes mellitus [97]. Such new strategies might enable the earlier implementation of corrective measures before the disease becomes irreversible.
Conclusions
Much remains to be done in order to optimize the use of laboratory tests for assessing glycemic control in diabetic patients. To our mind, what we still have to do can be summarized as follows.
Some manufacturers need to improve their methods because there is an inacceptable bias and/or imprecision with respect to the desired goals.
Laboratory professionals should carefully analyze the IFCC certificates before choosing a method for measuring HbA1c.
Several manufacturers should improve the information they provide together with their control materials. It is absolutely unacceptable to use control materials with assigned ranges that are wider than the total error limits for the quantity being measured.
EQAS providers need to demonstrate the commutability of the materials used in the inter-laboratory exercises; and the HbA1c targets must be assigned by means of reference measurement procedures or by reference laboratory services, as stated by the JCTLM.
Further studies are needed to ascertain whether other glycated proteins, such as glycated albumin, are clinically useful for routinely monitoring patients’ glycemic status.
The quality of glucose measurements obtained by POCT and CGM systems needs to be improved, to make most of the glucose variability indexes more robust and traceable to the laboratory data.
It would be advisable to be able to measure red cell age by a simple, automated procedure, in order to reduce a significant component of the between-subjects biological variability and to normalize the HbA1c concentration with respect to cell age.
This work was partially supported by a grant from the MIUR (PRIN 2008, recipients AM and AL).
Conflict of interest statement
Authors’ conflict of interest disclosure: The authors stated that there are no conflicts of interest regarding the publication of this article.
Research funding: None declared.
Employment or leadership: None declared.
Honorarium: None declared.
References
1. Shaw JE, Sicree RA, Zimmet PZ. Global estimates of the prevalence of diabetes for 2010 and 2030. Diabetes Res Clin Pract 2010;87:4–14.10.1016/j.diabres.2009.10.007Search in Google Scholar PubMed
2. American Diabetes Association. Standards of medical care in diabetes 2010. Diabetes Care 2010;33(Suppl 1):62–9.10.2337/dc10-S062Search in Google Scholar PubMed PubMed Central
3. The International Expert Committee. International Expert Committee report on the role of the A1c assay in the diagnosis of diabetes. Diabetes Care 2009;33:1327–34.10.2337/dc09-9033Search in Google Scholar PubMed PubMed Central
4. Rodbard HW, Blonde L, Braithwaite SS, Brett EM, Cobin RH, Handelsman Y, et al. AACE Diabetes Mellitus Clinical Practice Guidelines Task Force. American Association of Clinical Endocrinologists medical guidelines for clinical practice for the management of diabetes mellitus. Endocr Pract 2007;13 (Suppl 1):1–68.10.4158/EP.13.S1.1Search in Google Scholar PubMed
5. International Diabetes Federation. Global Guideline for type 2 diabetes. Brussels: International Diabetes Federation, 2005.Search in Google Scholar
6. John WG. Haemoglobin A1c: analysis and standardization. Clin Chem Lab Med 2003;41:1199–212.10.1515/CCLM.2003.184Search in Google Scholar PubMed
7. Little RR, Rohlfing CL, Sacks DB. National Glycohemoglobin Standardization Program (NGSP) Steering Committee. Status of hemoglobin A1c measurement and goals for improvement: from chaos to order for improving diabetes care. Clin Chem 2011;57:205–14.10.1373/clinchem.2010.148841Search in Google Scholar PubMed
8. Lapolla A, Mosca A, Fedele D. The general use of glycated haemoglobin for the diagnosis of diabetes and other categories of glucose intolerance: still a long way to go. Nutr Metab Cardiovasc Dis 2011;21:467–75.10.1016/j.numecd.2011.02.006Search in Google Scholar PubMed
9. Elder C, Hinchliffe C, Wright N. HbA1c league tables: might assay variability encourage foul play? J Clin Pathol 2010;63:660.10.1136/jcp.2009.071340Search in Google Scholar PubMed
10. Lenters-Westra E, Weykamp C, Schindhelm RK, Siebelder C, Bilo HJ, Slingerland RJ. One in five laboratories using various hemoglobin A1c methods do not meet the criteria for optimal diabetes care management. Diabetes Technol Therap 2011;13:1–4.10.1089/dia.2010.0148Search in Google Scholar PubMed
11. Lenters-Westra E, Slingerland RJ. Six of eight hemoglobin A1c point-of-care instruments do not meet the general accepted performance criteria. Clin Chem 2010;56:44–52.10.1373/clinchem.2009.130641Search in Google Scholar PubMed
12. Matteucci E, Giampietro O. Point-of-care testing in diabetes care. Mini-reviews in Medicinal Chemistry 2011;11:1–7.10.2174/138955711794519474Search in Google Scholar PubMed
13. Kaiser P, Reinauer H. Diabetes melllitus: the long way of standardization of HbA1c to the level of highest metrological order. Ger Med Sci 2011;9:1–4.Search in Google Scholar
14. Panteghini M. Traceability as a unique tool to improve standardization in laboratory medicine. Clin Biochem 2009;42:236–40.10.1016/j.clinbiochem.2008.09.098Search in Google Scholar PubMed
15. Panteghini M. Application of traceability concepts to analytical quality control may reconcile total error with uncertainty of measurement. Clin Chem Lab Med 2010;48:7–10.10.1515/CCLM.2010.020Search in Google Scholar PubMed
16. Mosca A, Goodall I, Hoshino T, Jeppsson JO, John WG, Little RR, et al. Global standardization of glycated hemoglobin measurement: the position of the IFCC working group. Clin Chem Lab Med 2007;45:1077–80.10.1515/CCLM.2007.246Search in Google Scholar PubMed
17. Finke A., Kobold U, Hoelzel W, Weykamp C, Jeppsson JO, Miedema K. Preparation of a candidate primary reference material for the international standardisation of HbA1c determinations. Clin Chem Lab Med 1998;36:299–308.10.1515/CCLM.1998.051Search in Google Scholar PubMed
18. Jeppsson JO, Kobold U, Barr J, Finke A, Hoelzel W, Hoshino T, et al. Approved IFCC reference method for the measurement of HbA1c in human blood. Clin Chem Lab Med 2002;40:78–89.10.1515/CCLM.2002.016Search in Google Scholar PubMed
19. Weykamp C, John WG, Mosca A, Hoshino T, Little R, Jeppsson JO, et al. The IFCC reference measurement system for HbA1c: a 6-year progress report. Clin Chem 2008;54:240–8.10.1373/clinchem.2007.097402Search in Google Scholar PubMed
20. IFCC. Available from: http://www.ifcchba1c.com/. Accessed on 24 November, 2012.Search in Google Scholar
21. Hoelzel A, Weykamp C, Jeppsson JO, Miedema K, Barr JR, Goodall I, et al. IFCC reference system for measurement of hemoglobin A1c in human blood and the National Standardization Schemes in the United States, Japan, and Sweden: a method-comparison study. Clin Chem 2004;50:166–74.10.1373/clinchem.2003.024802Search in Google Scholar PubMed
22. Consensus statement on the worldwide standardization of the hemoglobin A1C measurement: the American Diabetes Association, European Association for the Study of Diabetes, International Federation of Clinical Chemistry and Laboratory Medicine, and the International Diabetes Federation. Diabetes Care 2007;30:2399–400.10.2337/dc07-1752Search in Google Scholar PubMed
23. Hanas R, John G. on behalf of the International HbA1c Consensus Committee. 2010 consensus statement on the worldwide standardization of the hemoglobin A1c measurement. Clin Chem 2010;56:1362–4.10.1373/clinchem.2010.150540Search in Google Scholar PubMed
24. Panteghini M, John WG, Mosca A, Weykamp C. IFCC Report. Implementation of standardization of HbA1c measurement. Clin Chem Lab Med 2008;46:573–4.Search in Google Scholar
25. Mosca A, Branca MT, Carta M, Genna ML, Giorda CB, Ghidelli R, et al. Recommendations for the implementation of international standardization of glycated hemoglobin in Italy. Clin Chem Lab Med 2010;48:623–6.10.1515/CCLM.2010.140Search in Google Scholar PubMed
26. Braga F, Dolci A, Montagnana M, Pagani F, Paleari R, Guidi GC, et al. Revaluation of biological variation of glycated hemoglobin (HbA1c) using an accurately designed protocol and an assay traceable to the IFCC reference system. Clin Chim Acta 2011;412:1412–6.10.1016/j.cca.2011.04.014Search in Google Scholar PubMed
27. Weykamp CW, Mosca A, Gillery P, Panteghini M. The analytical goals for hemoglobin A1c measurement in IFCC units and National Glycohemoglobin Standardization Program units are different. Clin Chem 2011;57:1204–6.10.1373/clinchem.2011.162719Search in Google Scholar PubMed
28. Mosca A. Some practical advice on how to implement the international standardization of glycated hemoglobin measurement in Italy [Italian]. Biochimica Clinica 2011;35:36–41.Search in Google Scholar
29. IFCC. Available from: http://www.ifcc.org/ifcc-executive-board-and-council/eb-task-forces/task-force-hba1c/. Access verified on 24 November, 2012.Search in Google Scholar
30. The Diabetes Control and Complications Trial Research Group. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 1993;329:977–86.10.1056/NEJM199309303291401Search in Google Scholar PubMed
31. Rohlfing CL, Wiedmeyer HM, Little RR, England JD, Tennill A, Goldstein DE. Defining the relationship between plasma glucose and HbA1c. Diabetes Care 2002;25:275–8.10.2337/diacare.25.2.275Search in Google Scholar PubMed
32. Nathan DM, Kuenen J, Borg R, Zheng H, Schoenfeld D, Heine RJ, for the A1c-derived average glucose (AFAG) study group. Translating the A1c assay into estimated average glucose values. Diabetes Care 2008;31:1–6.10.2337/dc08-0545Search in Google Scholar PubMed PubMed Central
33. Westgard QC. Available from: http://www.westgard.com/biodatabase1.htm. Accessed on 24 November, 2012.Search in Google Scholar
34. Cohen RM, Franco RS, Khera PK, Smith EP, Lindseell CJ, Ciraolo PJ, et al. Red cell life span heterogeneity in hematologically normal people is sufficient to alter HbA1c. Blood 2008;112:4284–90.10.1182/blood-2008-04-154112Search in Google Scholar PubMed PubMed Central
35. Khera PK, Joiner CH, Carruthers A, Lindsell CJ, Smith EP, Franco RS, et al. Evidence for interindividual hetereogeneity in the glucose gradient across the human red blood cell membrane and its relationship to hemoglobin glycation. Diabetes 2008;57:2445–52.10.2337/db07-1820Search in Google Scholar PubMed PubMed Central
36. Burtis CA, Ashwood ER, Bruns DE. Tietz textbook of clinical chemistry and molecular diagnostics. 5th ed. St. Louis, MO: Elsevier Saunders, 2012.Search in Google Scholar
37. American Diabetes Association. Standards of medical care in diabetes – 2010. Diabetes Care 2010;33(Suppl 1):S11–61.10.2337/dc10-S011Search in Google Scholar PubMed PubMed Central
38. Sacks DB, Arnold M, Bakris GL, Bruns DE, Horvath AR, Kirkman S, et al. Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus. Clin Chem 2011;57:e1–47.10.1373/clinchem.2010.161596Search in Google Scholar PubMed
39. Mosca A, Paleari R, Madè A, Ferrero C, Locatelli M, Ceriotti F. Commutability of control materials in glycohemoglobin determinations. Clin Chem 1998;44:632–8.10.1093/clinchem/44.3.632Search in Google Scholar
40. O’Connor C, O’Shea PM, Owens LA, Carmody L, Avalos G, Nestor L, et al. Trimester-specific reference intervals for haemoglobin A1c (HbA1c) in pregnancy. Clin Chem Lab Med 2012;50:905–9.10.1515/cclm.2011.397Search in Google Scholar PubMed
41. BIPM. Available from: http://www.bipm.org/jctlm. Accessed on 24 November, 2012.Search in Google Scholar
42. Ponzo P, Dolci A, Scapellato L, Milano M, Paleari R, Mosca A, et al. Evaluation of commutability for Cobas Integra analytical system of the EQA materials for hemoglobin A1c. Biochimica Clinica 2008;32:44–7.Search in Google Scholar
43. Menchine MD, Arora S, Camargo CA, Ginde AA. Prevalence of undiagnosed and suboptimally controlled diabetes by point-of-care HbA1c in unselected emergency department patients. Acad Emerg Med 2011;18:1–4.10.1111/j.1553-2712.2011.01014.xSearch in Google Scholar PubMed PubMed Central
44. Silverman RA, Thakker U, Ellman T, Wong I, Smith K, Ito K, et al. Hemoglobin A1c as a screen for previously undiagnosed prediabetes and diabetes in an acute-care setting. Diabetes Care 2011;34:1908–12.10.2337/dc10-0996Search in Google Scholar PubMed PubMed Central
45. Selvin E, Steffes MW, Gregg E, Brancati FL, Coresh J. Performance of A1C for the classification and prediction of diabetes. Diabetes Care 2011;34:84–9.10.2337/dc10-1235Search in Google Scholar PubMed PubMed Central
46. Van’t Riet E, Alssema M, Rijkelijkhizen JM, Kostense PJ, Nijpels G, Dekker JM. Relationship between A1c and glucose levels in the general Dutch population. Diabetes Care 2010;33:61–6.10.2337/dc09-0677Search in Google Scholar PubMed PubMed Central
47. Cavagnolli G, Comerlato J, Comerlato C, Renz PB, Gross JL, Camargo JL. HbA1c measurement for the diagnosis of diabetes: is it enough? Diab Med 2011;28:31–55.10.1111/j.1464-5491.2010.03159.xSearch in Google Scholar PubMed
48. Hu Y, Liu W, Chen Y, Zhang M, Wang L, Zhou H, et al. Combined use of fasting plasma glucose and glycated hemoglobin A1c in the screening of diabetes and impaired glucose tolerance. Acta Diabetol 2010;47:231–6.10.1007/s00592-009-0143-2Search in Google Scholar PubMed
49. Zhou X, Pang Z, Gao W, Wang S, Zhang L, Ning F, et al. Performance of HbA1c and fasting capillary blood glucose test for screening newly diagnosed diabetes and pre-diabetes defined by OGTT in Qingdao China. Diabetes Care 2010;33:545–50.10.2337/dc09-1410Search in Google Scholar PubMed PubMed Central
50. Mostafa SA, Davies MJ, Webb D, Gray LJ, Srinivasan BT, Jarvis J, et al. The potential impact of using glycated haemoglobin as the preferred diagnostic tool for detecting type 2 diabetes mellitus. Diabetes Med 2010;27:762–9.10.1111/j.1464-5491.2010.03015.xSearch in Google Scholar PubMed
51. Carson AP, Reynolds K, Fonseca V, Muntner P. Comparison of A1C and fasting glucose criteria to diagnose diabetes among US adults. Diabetes Care 2010;33:95–7.10.2337/dc09-1227Search in Google Scholar PubMed PubMed Central
52. Chapp-Jumbo E, Edeoga C, Wang J, Dagogo-Jack S. Ethnic disparity in hemoglobin A1c levels among normoglycemic offspring of parents with type 2 diabetes. Endocr Pract 2011;2:1–18.Search in Google Scholar
53. Marini MA, Succurro E, Arturi F, Ruffo MF, Andreozzi F, Sciacqua A, et al. Comparison of A1c, fasting and 2-h post-load plasma glucose criteria to diagnose diabetes in Italian Caucasians. Nutr Metab Cardiovasc Dis 2012;22:561–6.10.1016/j.numecd.2011.04.009Search in Google Scholar PubMed
54. Zemlin AE, Matsha TE, Hassan MS, Erasmus RT. HbA1c of 6.5% to diagnose diabetes mellitus – does it work for us? The Bellville South Africa Study. PLoS One 2011;6:e22558.10.1371/journal.pone.0022558Search in Google Scholar PubMed PubMed Central
55. Christensen DL, Witte DR, Kaduka L, Jorgensen ME, Borch-Johnsen K, Mohan V, et al. Moving to an HbA1c-based diagnosis of diabetes has a different impact on prevalence in different ethnic groups. Diabetes Care 2010;33:580–2.10.2337/dc09-1843Search in Google Scholar PubMed PubMed Central
56. Cowie C, Rust K, Byrd-Holt DD, Gregg EW, Ford ES, Geiss LS, et al. Prevalence of diabetes and risk for diabetes using hemoglobin A1c criteria in the US population in 1998–2006. Diabetes Care 2010;33:562–8.10.2337/dc09-1524Search in Google Scholar PubMed PubMed Central
57. Lipska KJ, De Rekeneire N, Van Ness PH, Johnson KC, Kanaya A, Koster A, et al. Identifying dysglycemic states in older adults: implications of the emerging use of hemoglobin A1c. J Clin Endocrin Metab 2010;95:5289–95.10.1210/jc.2010-1171Search in Google Scholar PubMed PubMed Central
58. Tay TL, Foo JP, Tan E, Chen R, Khoo J, Soh SB, et al. HbA1c may not be a sensitive determinant of diabetic status in the elderly. Diab Res Clin Pract 2011;92:e31–3.10.1016/j.diabres.2011.01.003Search in Google Scholar PubMed
59. Boronat M, Saavedra P, López-Rìos L, Riaño M, Wägner AM, Nóvoa FJ. Differences in cardiovascular risk profile of diabetic subjects discordantly classified by diagnostic criteria based on glycated hemoglobin and oral glucose tolerance test. Diabetes Care 2010;33:2671–3.10.2337/dc10-0529Search in Google Scholar PubMed PubMed Central
60. Selvin E, Steffes MW, Zhu H, Matsuhira K, Wagenknecht L, Pankow J, et al. Glycated hemoglobin, diabetes, and cardiovascular risk in non-diabetic adults. N Engl J Med 2010;362:800–11.10.1056/NEJMoa0908359Search in Google Scholar PubMed PubMed Central
61. Doerr R, Hoffmann U, Otter W, Heinemann L, Hunger-Battefeld W, Kulzer B, et al. Oral glucose tolerance test and HbA1c for diagnosis of diabetes in patients undergoing coronary angiography; the Silent Diabetes Study. Diabetologia 2011;54:2923–30.10.1007/s00125-011-2253-ySearch in Google Scholar PubMed
62. Sturm G, Lamina C, Zitt E, Lhotta K, Haider F, Neyer U, et al. Association of HbA1c values with mortality and cardiovascular events in diabetic dialysis patients. The INVOR studies and review of the literature. PLoS One 2011;6:e20093.10.1371/journal.pone.0020093Search in Google Scholar
63. 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 (Position Statement). Diabetes Care 2002;25:750–86.10.2337/diacare.25.4.750Search in Google Scholar
64. Monnier L, Lapinski H, Colette C. Contribution of fasting and postprandial glucose increments to the overall diurnal hyperglycemia of type 2 diabetic patients: variations with increasing levels of HbA1c. Diabetes Care 2003;26:881–5.10.2337/diacare.26.3.881Search in Google Scholar
65. UK Prospective Diabetes Study group: intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patient with type 2 diabetes. Lancet 1998;352:837–53.10.1016/S0140-6736(98)07019-6Search in Google Scholar
66. Lachin JM, Genuth S, Nathan DM, Zinman B, Rutledge BN. For the DCCT/EDIC Research Group: effect of glycemic exposure on the risk of microvascular complications in the Diabetes Control and Complications Trial – Revisited. Diabetes 2008;57:995–1001.10.2337/db07-1618Search in Google Scholar PubMed
67. Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA. 10-year follow-up of intensive glucose control in type 2 diabetes. N Engl J Med 2008;359:1577–89.10.1056/NEJMoa0806470Search in Google Scholar PubMed
68. Wax J. Metabolic memory: a vascular perspective. Cardiovasc Diabetol 2010;9:51.10.1186/1475-2840-9-51Search in Google Scholar PubMed PubMed Central
69. Meerwaldt R, Links T, Zeebregts C, Tio R, Hillebrands JL, Smit A. The clinical relevance of assessing advanced glycation endproducts accumulation in diabetes. Cardiovasc Diabetol 2008;7:29.10.1186/1475-2840-7-29Search in Google Scholar PubMed PubMed Central
70. Monnier L, Colette C. Target for glycemic control. Diabetes Care 2009;2:S199–204.10.2337/dc09-S310Search in Google Scholar PubMed PubMed Central
71. Monnier L, Colette C, Dunseath GJ, Owens DR. The loss of postprandial glycemic control precedes stepwise deterioration of fasting with worsening diabetes. Diabetes Care 2007;30:263–9.10.2337/dc06-1612Search in Google Scholar PubMed
72. Borg R, Kuenen JC, Carstensen B, Zheng H, Nathan D, Heine RJ, et al. Associations between features of glucose exposure and HbA1c. The A1c-Derived Average Glucose (ADAG) study. Diabetes 2010;59:1585–90.10.2337/db09-1774Search in Google Scholar PubMed PubMed Central
73. Skyler JS, Bergenstal R, Bonow RO, Buse J, Deedwania P, Gale EA, et al. Intensive glycemic control and the prevention of cardiovascular events: implication of the ACCORD, ADVANCE and VA diabetes trials. Circulation 2009;119:1–7.10.1161/CIRCULATIONAHA.108.191305Search in Google Scholar PubMed
74. Ceriello A. The glucose triad and its role in comprehensive glycaemic control: current status, future management. Int J Clin Pract 2010;64:1705–11.10.1111/j.1742-1241.2010.02517.xSearch in Google Scholar PubMed PubMed Central
75. Hirsch IB, Brownlee M. Should minimal blood glucose variability become the gold standard of glycemic control? J Diabetes Complications 2005;19:178–81.10.1016/j.jdiacomp.2004.10.001Search in Google Scholar PubMed
76. Brownlee M, Hirsch IB. Glycemic variability: a haemoglobin A1c-independent risk factor for diabetic complications. J Am Med Assoc 2006;295:1707–8.10.1001/jama.295.14.1707Search in Google Scholar PubMed
77. Monnier L, Mas E, Ginet C, Michel F, Villon L, Cristol JP, et al. Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. J Am Med Assoc 2006;295:1681–7.10.1001/jama.295.14.1681Search in Google Scholar PubMed
78. Ceriello A, Esposito K, Piconi L, Ihnat MA, Thorpe JE, Testa R, et al. Oscillating glucose is more deleterious to endothelial function and oxidative stress than mean glucose in normal and type 2 diabetic patients. Diabetes 2008;57:1349–54.10.2337/db08-0063Search in Google Scholar PubMed
79. Kilpatrick ES, Rigby AS, Atkin SL. Mean blood glucose compared with HbA1c in the prediction of cardiovascular disease in patient with type 1 diabetes. Diabetologia 2008;51:365–71.10.1007/s00125-007-0883-xSearch in Google Scholar PubMed
80. Kovatchev BP, Cox DJ, Farhy LS, Straume M, Gonder-Frederick L, Clarke WL. Episodes of severe hypoglycemia in type 1 diabetes are preceded and followed within 48 hours by measurable disturbances in blood glucose. J Clin Endocrinol Metab 2000;85:4287–92.Search in Google Scholar
81. Kudva YC, Basu A, Jenkins GD, Pons GM, Vogelsang DA, Rizza RA, et al. Glycemic variation and hypoglycemia in patients with well-controlled type 1 diabetes on a multiple daily insulin injection program with use of glargine and ultralente as basal insulin. Endocr Pract 2007;13:244–50.10.4158/EP.13.3.244Search in Google Scholar PubMed
82. Klonoff DC. Continuous glucose monitoring: roadmap for 21st century diabetes therapy. Diabetes Care 2005;28:1231–9.10.2337/diacare.28.5.1231Search in Google Scholar PubMed
83. Sartore G, Chilelli NC, Burlina S, Di Stefano P, Piarulli F, Fedele D, et al. The importance of HbA1c and glucose variability in patients with type 1 and type 2 diabetes: outcome of continuous glucose monitoring (CGM). Acta Diabetol 2012. DOI: 10.1007/s00592-012-0391-4.10.1007/s00592-012-0391-4Search in Google Scholar PubMed
84. Paroni R, Ceriotti F, Galanello R, Leoni GB, Panico A, Scurati E, et al. Performance characteristics and clinical utility of an enzymatic method for the measurement of glycated albumin in plasma. Clin Biochem 2007;40:1398–405.10.1016/j.clinbiochem.2007.08.001Search in Google Scholar PubMed
85. Cohen RM, Chener TC, Holmes YR, Joiner CH. Discordance between HbA1c and fructosamine. Diabetes Care 2003;26:163–7.10.2337/diacare.26.1.163Search in Google Scholar PubMed
86. Cohen RM, Snieder H, Lindsell CJ, Beyan H, Harva MI, Blinko S, et al. Evidence for independent heritability of the glycation gap (glycosylation gap) fraction of HbA1c in nondiabetic twins. Diabetes Care 2006;29:1739–43.10.2337/dc06-0286Search in Google Scholar PubMed
87. Leslie RD, Cohen RM. Biologic variability in plasma glucose, hemoglobin A1c, and advanced glycation end products associated with diabetes complications. J Diabetes Sci Technol 2009;3:635–43.10.1177/193229680900300403Search in Google Scholar PubMed PubMed Central
88. Koga M, Kasayama S, Kanehara H, Bando Y. CLD (chronic liver disease)-HbA1c as a suitable indicator for estimation of mean plasma glucose in patients with chronic liver disease. Diabet Res Clin Pract 2008;81:258–62.10.1016/j.diabres.2008.04.012Search in Google Scholar PubMed
89. Bogner HR, Miller SD, De Vries HF, Chhatre S, Jayadevappa R. Assessment of cost and health resource utilization for elderly patients with heart failure and diabetes mellitus. J Card Fail 2010;16:454–60.10.1016/j.cardfail.2010.01.007Search in Google Scholar PubMed PubMed Central
90. Sun Y, Toh MP. Impact of diabetes mellitus (DM) on the health-care utilization and clinical outcomes of patients with stroke in Singapore. Value Health 2009;12:S101–5.10.1111/j.1524-4733.2009.00639.xSearch in Google Scholar PubMed
91. Nuckols TK, McGlynn EA, Adams J, Lai J, Go MH, Keesey J, et al. Cost implications to health care payers of improving glucose management among adults with type 2 diabetes. Health Serv Res 2011;46:1158–79.10.1111/j.1475-6773.2011.01257.xSearch in Google Scholar PubMed PubMed Central
92. Dall TM, Roary M, Yang W, Zhang S, Chen YJ, Arday DR, et al. Health care use and costs for participants in a diabetes disease management program, United States, 2007–2008. Prev Chronic Dis 2011;8:A53.Search in Google Scholar
93. Driskell OJ, Holland D, Hanna FW, Jones PW, Pemberton RJ, Tran M, et al. Inappropriate requesting of glycated hemoglobin (HbA1c) is widespread: assessment of prevalence, impact of national guidance, and practice-to-practice variability. Clin Chem 2012;58:906–15.10.1373/clinchem.2011.176487Search in Google Scholar PubMed
94. Aagren M, Luo W. Association between glycemic control and short-term healthcare costs among commercially insured diabetes patients in the United States. J Med Econ 2011;14:108–14.10.3111/13696998.2010.548432Search in Google Scholar PubMed
95. Leach MJ, Segal L. Are clinical practical guidelines (CPGs) useful for health services and health work force planning? A critique of diabetes CPGs. Diabet Med 2010;27:570–7.10.1111/j.1464-5491.2010.02981.xSearch in Google Scholar PubMed
96. Katon J, Williams MA, Reiber G, Miller E. Antepartum A1c, maternal diabetes outcomes, and selected offspring outcomes: an epidemiological review. Paediatr Perinat Epidemiol 2011;25:265–76.10.1111/j.1365-3016.2011.01195.xSearch in Google Scholar PubMed PubMed Central
97. Wang TJ, Larson MG, Vasan RS, Cheng S, Rhee EP, McCabe E, et al. Metabolite profiles and the risk of developing diabetes. Nat Med 2011;17:448–53.10.1038/nm.2307Search in Google Scholar PubMed PubMed Central
©2013 by Walter de Gruyter Berlin Boston
