The Diabetes Control and Complications Trial (DCCT)3 and the United Kingdom Prospective Diabetes Study (UKPDS)4,5 demonstrated the benefits of improving glycemic control in both type 1 and type 2 diabetes respectively. Both studies showed that improved glycemic control is associated with sustained decreased rates of microvascular and neuropathic complications.
The Diabetes Control and Complications Trial (DCCT)3 and the United Kingdom Prospective Diabetes Study (UKPDS)4,5 demonstrated the benefits of improving glycemic control in both type 1 and type 2 diabetes respectively. Both studies showed that improved glycemic control is associated with sustained decreased rates of microvascular and neuropathic complications. Current guidelines issued by the American Association of Clinical Endocrinologists (AACE) recommend that glycated hemoglobin (HbA1c) be adopted as the primary method of assessment of glycemic control, with an HbA1c target of ²6.5%,6 while the American Diabetes Association (ADA) recommends a target HbA1c of <7%.7 Improving glycemic control in individuals with type 2 diabetes has not been shown to be directly associated with macrovascular benefits in large-scale clinical intervention studies;8,9 indeed, there was a suggestion of harm associated with tight glycemic control in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) study. While the current HbA1c targets provide a generalized goal, it should be noted that the guidelines also stress that HbA1c targets should be individualized, taking into account factors such as the patient’s age, life expectancy, comorbid conditions, and preferences.8,9
The most recent consensus guidelines for the management of hyperglycemia in type 2 diabetes from the ADA and the European Association for the Study of Diabetes (EASD) recommend that initial management should always include lifestyle modifications.10 Due to the progressive chronic nature of type 2 diabetes, the initiation of pharmacotherapy is often necessary to reach or maintain the target HbA1c, with the biguanide metformin recommended as the initial drug. However, the majority of patients will require additional medications during the course of the disease, either to maintain or to intensify treatment. Several therapeutic options are now available for the treatment of type 2 diabetes, including metformin, alpha glucosidase inhibitors, thiazolidinediones, meglitinides, and insulin. More recently approved agents include dipeptidyl peptidase-4 (DPP-4) inhibitors, amylin analogs, and incretin mimetics. With the availability of multiple agents with differing pharmacological targets, the question then arises, in the absence of head-tohead clinical trials, of how best to compare the relative effectiveness of these newer treatments, especially against established agents, which have accumulated long-term efficacy and safety data.
Glycated Hemoglobin Levels—Change from Baseline Values
Obviously, HbA1c is clinically important: it is one of the fundamental benchmarks by which we assess glycemic control and serves as a key efficacy parameter in clinical trials. HbA1c levels reflect mean glycemic control over two to three months and thus provide a global assessment of glycemic control, in contrast to fasting glucose or post-prandial glucose, which are related more to specific instances. It is a validated surrogate for the short-term clinical consequences of hyperglycemia, as well as for the long-term microvascular complications of type 2 diabetes.
Observational analyses from the UKPDS trial showed that a 1% reduction in HbA1c in patients with type 2 diabetes was associated a reduction in long-term microvascular complications, regardless of the HbA1c threshold.11 However, the relationship between HbA1c and microvascular complications was curvilinear, which suggests that the relative benefits of improving HbA1c decrease as baseline HbA1c is reduced.
The primary end-point in the majority of clinical trials evaluating the effectiveness of drug therapy on glucose lowering is reduction in HbA1c from baseline levels. This end-point is also the basis for the final demonstration of efficacy for purposes of drug approval and labeling.12 In terms of clinical trial data, what do changes in HbA1c from mean baseline tell the physician in terms of clinical efficacy? Such changes provide an idea of the efficacy of a compound. However, it should be emphasized that if the study begins with a higher mean baseline HbA1c, the percentage change in HbA1c may be greater; conversely, smaller reductions in mean HbA1c levels may be observed in a study that starts with a lower mean baseline HbA1c. Thus, if compound A achieved a fall of 2% in HbA1c but fell from 9.5 to 7.5%, while compound B achieved a 1.5% fall but fell from 8 to 6.5%, the inference is that compound A has a greater efficacy; however, this may not be correct. Focusing solely on change from baseline could be misleading. Indeed, a recent analysis by Bloomgarden et al. that reviewed studies of five major classes of established glucose-lowering agents—including metformin, sulfonylureas, and thiazolidinediones—showed that the magnitude of HbA1c lowering is related to the baseline HbA1c level, irrespective of drug class.13 In patients with HbA1c baseline levels <8%, the reduction in HbA1c levels was only 0.1–0.2%. In comparison, in patients with HbA1c baseline levels ³10%, a 1.2% reduction was observed. A subsequent analysis on a per-patient basis supported this relationship between baseline bA1c and HbA1c lowering.14 These findings highlight the difficulty in interpreting the efficacy results of clinical trials of different agents if the trials have different baseline HbA1c levels.
Furthermore, it should also be noted that more recent clinical trials have included trial participants with lower baseline HbA1c levels (~8%) than previously studied in registration trials involving the more established agents, wherein baseline HbA1c was often in the 9–10% range. In studies evaluating the novel incretin therapies, the magnitude of the reduction in bA1c was dependent on the baseline HbA1c, with greater reductions seen in groups of participants with higher baseline HbA1c.15–17 This is another important factor that should be considered when comparing the relative efficacies of different agents from published trial data.
Percentage to Target Glycated Hemoglobin Goal
Although there has been an improvement in glycemic control in recent years, over 40% of diabetes patients fail to reach the target HbA1c goal of <7%.18 Furthermore, even with the introduction of additional oral agents, achieving and maintaining HbA1c <7%has proved difficult, particularly in patients with a longer duration of diabetes.19
In the controlled environment of a clinical trial, the percentage of patients achieving the recommended HbA1c target (<7% and/or <6.5%) end-point provides a good indication of efficacy and allows comparison of relative efficacy between agents in the absence of head-to-head clinical trials. Clearly, if the trial population includes an excess of individuals with difficult-to-treat diabetes and high baseline HbA1c levels, achieving a change from HbA1c baseline may be much easier than achieving a percentage of patients to goal. For example, if you have patients with a mean baseline HbA1c of >9%, the evidence suggests that it would be easier to achieve a reduction of 1.5% in a majority of that population. However, it is theoretically more difficult to achieve a significant percentage of people achieving <7% in the same population. While data on percentage to recommended HbA1c goal provide good guidance, achieving optimal target levels in clinical practice may be hindered by several factors, including clinical inertia, differences in prescription coverage among individual patients, and suboptimal patient adherence to lifestyle and pharmacological treatments. Moreover, the optimal targets of <7% or <6.5% are not applicable for all patients, especially older or frail patients and patients at increased risk for adverse complications from tight control, or those who have substantially reduced life expectancy due to comorbid conditions.
With the multitude of agents available for treating type 2 diabetes, and in the absence of head-to-head clinical trials, evaluating the relative effectiveness of newer treatments and assessing whether they have comparable efficacy to established agents may be confusing. In terms of clinical trial data, the change in HbA1c from mean baseline provides an idea of the efficacy of a compound, but absolute changes should be reviewed with consideration of the patient population in the trial. There can often be a great deal of variety in the characteristics of patients enrolled in various type 2 diabetes trials—particularly differences in baseline HbA1c levels—which can affect the analysis of study findings. The baseline HbA1c level correlates with the magnitude of HbA1c reduction that we can expect to see with drug therapy. The end-point of percentage of patients achieving recommended HbA1c targets also provides a good indication of a drug’s effectiveness, but again consideration needs to be made to achieving optimal treatment goals outside of the clinical trial environment. When reviewing clinical trial data for drugs, besides efficacy, safety is paramount. With a strict focus on HbA1c levels, there is a anger that complications such as hypoglycemia may be ignored, especially in patients with other underlying conditions and polypharmacotherapy. By incorporating the best evidence available on newer treatments while continuing to study these drugs, we will to be able to more accurately place them within the therapeutic armamentarium for diabetes. Furthermore, in order to maximize treatment, the achievement of treatment goals must be balanced with respect to the potential risks and benefits for individual patients.