Is A1C the Best Measure of Glycemic Control?

For years, the A1C test has been considered the gold standard for assessing glycemic status in patients with diabetes. Numerous prospective and epidemiologic studies have featured ‘improved glucose control as measured by A1C’ as a primary end-point. From these studies, researchers have concluded that lowering A1C significantly reduces the risk and progression of diabetic complications.^1–3

For years, the A1C test has been considered the gold standard for assessing glycemic status in patients with diabetes. Numerous prospective and epidemiologic studies have featured ‘improved glucose control as measured by A1C’ as a primary end-point. From these studies, researchers have concluded that lowering A1C significantly reduces the risk and progression of diabetic complications.^1–3

In a study examining the relative contribution of postprandial glucose and fasting glucose of the A1C, Monnier et al.⁴ found that fasting glucose was a much greater contributor when A1C was 8.4% and higher whereas postprandial glucose was more influential when A1C was below 8.4%. From a practical perspective, this finding may provide some guidance for clinicians in their therapeutic decision-making. For example, it may be more effective to focus on therapies that target fasting glucose when the A1C is above 9%. Conversely, a patient with an A1C of 8% who is struggling to achieve their goal may benefit from therapies that target postprandial glucose (i.e. rapidacting insulin or secretagogues).

A1C measurement is a critical component of diabetes management; however, a key limitation of using A1C as a measure of glycemia is the lack of timeliness – it does not provide guidance to clinicians or patients regarding glucose excursions throughout the day as it is simply a crude measurement of mean glycemia. This is particularly relevant when considering new data that show that glycemic variability may also play an important role in the development of complications, independent of A1C levels. In turn, these data prompt reconsideration of the lessons learned from the Diabetes Control and Complications Trial (DCCT).

What the DCCT Really Showed
Few would dispute the conclusions drawn from the DCCT—that lower A1C levels are associated with reduced risk of complications. This result was seen in both the intensive and conventionally treated groups; however, when comparing patients who had identical A1C levels, there was a significant difference in risk for complications between the two groups. For example, the risk for retinopathy in conventionally treated patients with an A1C of 9% was more than 2.5 times greater than in intensively treated patients with the same A1C level (see Figure 1).⁵ The question of why this difference in risk exists now arises. One could certainly speculate that the difference may be due to reduced glycemic variability in the intensively treated group. The conventionally treated patients, who utilized twice-daily neutral protamine Hagedorn (NPH) insulin, had limited prandial coverage whereas the intensively treated group utilized basal-bolus therapy—NPH or ultralente insulin with regular insulin at meals—using either multiple daily injections (MDIs) or insulin pumps (with buffered regular insulin). Due to the fact that patients in the intensively treated group addressed both bolus and basal insulin needs, it is quite reasonable to assume that they experienced reduced frequency and severity of postprandial hyperglycemia, therefore reducing glycemic variability throughout the day. Although opponents of this hypothesis would argue that glycemic variability was not directly measured in the DCCT, few would disagree that the intensive therapy cohort received a more physiologic insulin regimen, which, as a goal, was to reduce both fasting and postprandial hyperglycemia.

It is important to consider that postprandial hyperglycemia should be regarded as a type of glycemic variability; the two are not necessarily synonymous. For example, a patient whose glucose ranges from 40mg/dL to 180mg/dL within a given day experiences the same glycemic variability as a patient whose glucose fluctuates between 150mg/dL and 290mg/dL. However, it is quite likely that the physiological consequences are much more significant in the second patient, given the persistent hyperglycemia. Despite the fact that postprandial hyperglycemia is not the only form of glycemic variability, it is the most common and, therefore, the most studied.

The Impact of Glycemic Variability
Over several years, numerous epidemiological studies such as the Diabetes Epidemiology: Collaborative Analysis of Diagnostic Criteria in Europe (DECODE) study, the Honolulu Heart study, and others have shown a strong association between postprandial and postchallenge glucose and risk for macrovascular disease. In addition to epidemiologic data, a number of recent experimental trials have demonstrated a clear connection between postprandial glucose variability and vascular tissue damage. Temelkova-Kurktschiev et al. demonstrated that post-challenge glucose and glycemic ‘spikes’ are more strongly associated with carotid intima-media thickening (IMT) than fasting glucose and A1C in patients at risk for or in the early stages of type 2 diabetes.⁶

Reducing postprandial glucose excursions is particularly important in light of new data that show a relationship between improved postprandial glycemic control and reduced atherosclerotic risk. A study by Esposito et al., looking at postprandial hyperglycemia as a possible risk factor for cardiovascular disease (CVD) in patients with type 2 diabetes, showed that reductions in postprandial hyperglycemia was associated with regression of carotid intima-media thickness—a known marker for atherosclerosis.⁷ Perhaps not surprisingly, proinflammatory cytokines were also reduced when postprandial hyperglycemia was targeted. Hanefeld et al.⁸ recently demonstrated that reductions in postprandial glucose levels through treatment with acarbose slowed the progression of carotid intima-media thickness in individuals with impaired glucose tolerance.⁸ Recent data from Ceriello et al. showed that postprandial hyperglycemia is accompanied by endothelial dysfunction in patients with type 2 diabetes and that rapid-acting insulin at mealtimes improved endothelial function.⁹
One study by Risso et al. examined apoptosis in human umbilical vein endothelial cells in cell culture.¹⁰ Cells were subjected to the following glycemic conditions:

• constant glucose at 5mmol concentration;
• constant glucose at 20mmol; and
• alternating glucose at 5mmol and 20mmol.

Measurements taken at seven and 14 days showed significantly greater apoptosis in cells subjected to alternating glucose concentrations (see Figure 2). The researchers concluded that variability in glycemic control may be more deleterious than a constant high concentration of glucose.

The Mechanism Behind Glycemic Variability and Endothelial Damage
Protein kinase C-beta (PKC-β) appears to be a critical step in the pathogenesis of vascular complications. Current opinion is that hyperglycemia feeds on advanced glycation end-products, DAG generation, and oxidative stress. All three of these will lead to PKC activation and vascular damage. It is therefore possible to use PKC-β as a downstream marker for oxidative stress, which in turn appears to be important as it relates to glycemic variability. A recent report by Ceriello et al. suggests that endothelial dysfunction is induced through oxidative stress caused by postprandial hyperglycemia and hypertriglyceridemia (hTG).¹¹

A study by Quagliaro et al. investigated PKC-β activity in human umbilical vein endothelial cells in cell culture using low-glucose, high-glucose, and alternating glucose concentrations.¹² Researchers observed the same relationship between steady-state glucose and alternating glucose as seen in the earlier study. PKC-β activity was significantly greater in cells exposed to alternating glucose concentrations compared with steady-state glucose concentrations (low or high). Alternating glucose levels has also been shown to be a key factor in the progression of nephropathy.¹³ Furthermore, a study by Piconi et al., which looked at the effect of alternating glucose levels on various adhesion molecules (E-selectin, intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule -1 (VCAM-1)), and interleukin-6 (IL-6) in umbilical vein human endothelial cells, showed the same increased expression when exposed to glycemic fluctuation.¹⁴

Current thinking proposes that hyperglycemia, free fatty acids, and insulin resistance feed into oxidative stress, PKC activation, and advanced glycated endproduct receptor (RAGE) activation, which in turn leads to vasoconstriction, inflammation, and thrombosis. Glycemic variability appears to be an important role in this mechanism.

Application in Clinical Practice
How can clinicians accurately and efficiently assess glycemic variability in patients? Use of selfmonitoring of blood glucose (SMBG) with datadownload technology is the first step. In a large epidemiological study of SMBG use in individuals with type 1 and type 2 diabetes, Karter et al. found that frequency of testing is linked to A1C—the greater the frequency of testing, the lower the A1C level.15 Although the data do not demonstrate a causal relationship between SMBG frequency and A1C improvement, it is generally well-accepted that SMBG provides the strong positive reinforcement for patient compliance and clearly provides the vital information necessary to make informed decisions about treatment regimens.
Accessing SMBG data has been greatly simplified through the use of advanced data management technology.¹⁶Virtually all current downloading software systems feature a standard deviation (SD) calculation, which is a good measure of glycemic variability. Many software programs provide standard deviation data at various times of the day.

Using the standard deviation to assess glycemic control requires significant data – five results at minimum, while 10 results are preferable. In addition, the average glucose value must be taken into consideration. For example, if a patient has an extremely high or low average glucose level, this alone demands attention regardless of the SD. After the average glucose has been effectively addressed, it makes sense to establish a target for SD. A simple target to strive for is:

SD x 2 < mean glucose For example, if a patient’s mean glucose is 140mg/dL, the target SD should be less than 70. The ideal target would be SD x 3 < mean glucose; however, this can be extremely difficult to achieve in patients with type 1 diabetes. A high SD can serve as an important indicator for a number of problems, including insulin deficiency (especially good with fasting blood glucose), poor matching of calories (especially carbohydrates) with insulin, gastroparesis, late administration of mealtime insulin (or missing injections completely), erratic snacking, or poor matching of basal insulin. Assuming that glycemic variability is found to be an independent risk for microvascular complications, it would be valuable to have an index that combines A1C with glycemic variability. One way of doing this would be to multiply the A1C by the standard deviation (identified through meter downloads) and divide by 100 (A1C x SD÷100); the higher the number, the greater the risk for complications. Clearly, this is only one of many options to consider, and it should be formally evaluated.The point is that a better number—one that combines the impact of A1C with glycemic variability—is needed to improve the ability to stratify patient risk. It is quite likely that the A1C and SD are not equivalent in their risks and a more complicated equation will be identified. However, there is growing scientific evidence that both of these factors are important in the pathogenesis of diabetes complications and the simplistic view of A1C alone needs updating. Conclusions
Measurement of A1C will continue to be an important tool for assessing glycemic control in patients with diabetes. However absent data is from randomized control trials, there is a growing body of evidence that strongly suggests that glycemic variability, independent of A1C, may also be a significant risk factor in the development of diabetic complications. With current SMBG systems combined with advanced data management technology, clinicians are able to calculate glycemic variability in their offices and then follow up with appropriate and effective therapeutic interventions. ■