Yes – it is important: Louis Monnier, MD
No – it is not important: Eric S Kilpatrick, MD
Dr Monnier began his talk by asking whether or not glycemic variability should be considered a component of dysglycemia. He defined dysglycemia as elevations in fasting plasma glucose (FPG), elevations in postprandial plasma glucose (PPG), and acute glucose fluctuations from peak to nadir. He discussed how both the duration and magnitude of glycemic excursions contribute to sustained hyperglycemia. Chronic hyperglycemia, in turn, contributes to an excess of glycation and activation of oxidative stress. Urinary isoprostanes, as measured by the F2 isoprostane 8-iso prostaglandin F2-alpha (8-iso PGF2-alpha), provide a reliable measure of the activation of oxidative stress over a 24-hour period. In people with type 2 diabetes, isoprostanes are increased 2-fold compared with healthy individuals. Both FPG and PPG elevations contribute to the activation of oxidative stress. Using continuous glucose monitoring (CGMS) technology, it is now possible to measure mean amplitude of glycemic excursions (MAGE) to characterize glycemic fluctuations. In his work, Dr Monnier demonstrated that MAGE assessment of glucose fluctuation was correlated with isoprostanes, suggesting that glycemic fluctuations also activate oxidative stress. Based on this research, Dr Monnier concluded that glycemic fluctuations probably should be considered a component of dysglycemia.
Dr Monnier went on to discuss whether or not we have the appropriate tools for measuring glycemic variability. To address this question, he compared 2 approaches: standard deviation (SD) around mean glucose values and MAGE. He suggested that although the former may be more appropriate from a statistical point of view, the latter may be more appropriate from a pathophysiological point of view. To support this view, he demonstrated data showing that among patients with 1 large glucose variation, SD may not be the most appropriate measurement strategy. Additionally, SD derived from a 7-point glycemic profile may not capture all data. Based on available data, Dr Monnier concluded that more research was needed to determine the appropriateness of the tools currently available for assessing glycemic variability.
The third question that Dr Monnier addressed was whether glycemic variability was a risk factor for diabetic complications. He began by discussing data showing the importance of PPG excursions to cardiovascular risk. FPG (defined as good, borderline, or poor control) was not a significant predictor of myocardial infarction (MI) incidence, but postprandial control has been shown to be predictive. He also discussed research demonstrating the effect of acarbose on MI incidence among patients with impaired glucose tolerance. He concluded that there is a small amount of evidence to support the role of PPG in diabetic complications but insufficient evidence to support the role of glycemic variability at this time.
The last question Dr Monnier addressed was whether or not measurements of glycemic variability open a new era for clinical trials to test in the future? He felt that the answer to this question was definitely yes. In studies of incretin mimetics and dipeptidyl peptidase-4 inhibitors, some have used multiple point glucose profiles and postprandial excursions in response to a test meal, but none have assessed glycemic variability using CGMS technology. Dr Monnier ended his talk by stating that dysglycemia and its consequences, the risk for complications, are greater than the sum of their parts. It is important to take into account FPG, PPG, and acute glucose fluctuation.
Dr Kilpatrick next presented the opposing view that glycemic variability is not important. He began with what he termed “the search for endpoint evidence.” He began by discussing the Diabetes Control and Complications Trial (DCCT) and the view that mean blood glucose is the important factor in predicting diabetic complications, regardless of how that mean was arrived at. Dr Kilpatrick reexamined DCCT data to examine whether or not glucose variability independently increases the risk of microvascular complications. In univariate analyses glucose variability (as measured by SD) was higher in the conventionally-treated group. However, in multivariate analyses, glucose variability was no longer associated with risk of complications, and neither PPG nor FPG preferentially contributed to risk of diabetic microvascular complications. However, mean blood glucose value remained highly predictive of diabetic complications. The same data were reanalyzed using MAGE, and comparable results were achieved. Dr Kilpatrick concluded that in the DCCT, mean blood glucose determines risk of diabetic microvascular complications, regardless of how that mean was arrived at.
Dr Kilpatrick next addressed whether glucose variability affects risk for macrovascular complications. In the DCCT there were relatively few macrovascular events, which is why the effect of tight glycemic control on risk for cardiovascular disease (CVD) may not have been observed. However, even though A1C did not predict CVD risk in the DCCT, mean blood glucose was predictive of CVD risk. Again, the results of reanalyzing data from the DCCT and Epidemiology of Diabetes Intervention and Complications (EDIC) demonstrated that only mean blood glucose values predicted CVD risk. Dr Kilpatrick concluded that mean blood glucose determines risk of diabetic macrovascular complications, regardless of how that mean was arrived at.
When looking at glycemic variability in type 2 diabetes, Dr Kilpatrick turned to the UKPDS. In this study, no glucose profiles were obtained. Working off the assumption that patients on insulin have more glycemic variability than patients treated with oral agents, he examined the effect of insulin, chlorpropamide, and glibenclamide, and found no difference in endpoints. Based on the available data, Dr Kilpatrick concluded that mean blood glucose has a much greater effect than glycemic variability, noting that he felt like he was “slaying a beautiful hypothesis by an ugly fact.” However, he acknowledged that he would still treat glucose variability because in order to reduce mean blood glucose, you must reduce glycemic excursions. He returned to Dr Monnier’s work showing that at lower A1C values, PPG is the main contributor.
