Evolution in the American Diabetes Association Standards of Care

Glycemic treatment goals have been discussed twice annually for at least the past 3 years that I have served on the Professional Practice Committee of the American Diabetes Association (ADA). These were first established over a decade ago without the benefit of any of the landmark outcomes studies that are available for review today. The overall recommendations have been subsequently well supported by the Stockholm Diabetes Intervention Study,1 the Diabetes Control and Complications Trial,2 the Kumamoto study,3 and, most recently, the U.K. Prospective Diabetes Study (UKPDS).4,5 In each of these trials, a more intensively treated group achieved an average hemoglobin A_1c (A1C) of ∼7%, and there was an associated reduction in microvascular endpoints and trend toward improvement in macrovascular endpoints. In each, the more stringently controlled group had a higher incidence of hypoglycemia, greater weight gain, and increased costs, although the net benefit with respect to endpoint reduction was felt to adequately compensate for these adverse effects.

However, the trend towards reduction in cardiovascular events in these trials is in part counterbalanced by the Veterans Administration Cooperative Study,6 a feasibility study of modest size and short duration, in which there was a nonstatistically significant increase in cardiovascular events in the intensive treatment group. Furthermore, interpretation of the UKPDS results is complicated by the fact that there was a statistically significant reduction in cardiovascular disease (CVD) events in the obese group treated with metformin (Glucophage), but a lesser effect in those treated with insulin and sulfonylurea despite their having a modestly greater reduction in A1C. Thus, it remains uncertain whether more intensive glycemic control is associated with cardiovascular benefit. At least three large-scale multicenter trials are underway to examine the effect of glycemic control on CVD events, and they will hopefully provide a definitive answer to this question within the decade.

Given that the interventional studies above generally aimed for normalization of glucose levels and that epidemiological analyses do not demonstrate a threshold of A1C below which reductions in complications are not observed,7 it has been argued that we should target an A1C even within the normal range. However, the ADA continues to take a conservative view of what we know about the risks and benefits of diabetes treatment. The available outcomes studies only achieved an A1C of 7%, in part because of adverse events such as hypoglycemia despite extreme levels of commitment and expertise among both staff and subjects. Because of this, one can argue that we do not robustly understand the true risk of hypoglycemia and the true benefit in the form of microvascular and macrovascular risk reduction at A1C levels <7% in standard clinical practice.

Clearly in the range of <7%, the rate of severe hypoglycemia climbs steeply and the absolute risk of microvascular complications is small, providing only modest opportunity for further risk reduction. Because there is no proof that lowering A1C independent of technique will produce a lowering of CVD events, it is hard to argue that lower A1C targets are justified to reduce the burden of CVD events in our society. We could also contemplate whether there would be adverse effects related to long-term exposure to poorly studied combinations of medications, greater expense, life disruption caused by greater attention and effort to achieve lower glycemic targets, or the potential that extreme efforts expended in achieving very stringent glycemic goals will result in less attention paid to other health risks by patients or providers.

With the advent of improvements in oral antidiabetic agents, insulin analogs, and glucose monitoring technology, achieving lower glycemic targets is easier in the 21st century than in the previous decade, and, thus, these advances could mitigate some of the risks. That said, we do not know the risks or the benefits of these technologies in long-term studies, and it remains possible that there are predictable (excessive weight gain with thiazolidinediones or altered rates of retinopathy related to insulin-like growth factor-1–mediated effects of insulin analogs) or unpredictable adverse events that have just not been evaluated in long-term studies.

At least in 2000, with almost 40% of Americans with diabetes having A1C results >9.5%8 and even excellent health care systems reporting average A1C results of ∼8% in both type 1 and type 2 diabetic patients,9 the relative argument of whether the A1C target should be <7% or <6.5% seems relatively irrelevant. The former target, established by the ADA, is well supported by long-term randomized prospective studies. The latter target, proposed by the American College of Endocrinology (ACE)10 and the European Diabetes Policy Group,11 is a hopeful attempt to extend observational (i.e., epidemiological) studies to further reduce risks for patients with diabetes.

The recently published results of the Women’s Health Initiative study12 of conjugated equine estrogens and medroxyprogesterone acetate remind us that the conclusions of observational studies do not necessarily inform what clinical practice should be and require confirmation through interventional studies. Fortunately, the Action to Control Cardiovascular Risk in Diabetes study13 and others will explore the relative risks and benefits of a treatment program targeting an A1C of <6% versus a program aiming for A1C results between 7.0 and 7.9%.

In this issue of Clinical Diabetes (p. 27), we reprint an abridged version of the recently updated ADA position statement “Standards of Medical Care for Patients With Diabetes Mellitus,” which constitutes the highlights of this important document. Review of the ADA standards of care is an excellent reminder of our state of knowledge in diabetes management and how much is not yet definitively known.

In previous years, the standards of care document included a table on glycemic control showing, among other things, an A1C goal of <7%, but a level at which “additional action” was suggested of <8%. The explanatory text noted that the “action” level was not acceptable as a goal and was instead indicative of the need for significant changes in treatment. Nonetheless, the “additional action” column was often misunderstood as justification for no intervention at levels of A1C of ≤7.9%. It has therefore been dropped in the 2003 revision.

The standards of care historically have not included guidelines for postprandial glucose targets because there are no clinical trials in either type 1 or type 2 diabetes to demonstrate the benefits and risks of monitoring and targeting particular postprandial glucose levels. The report of an ADA consensus conference on postprandial blood glucose14 suggested that there are effective A1C-lowering agents that primarily target postprandial glucose levels (e.g., rapid-acting insulin analogs, short-acting secretagogues, and α-glucosidase inhibitors) and that monitoring postprandial glucose levels may allow for dosage adjustment of these agents. Furthermore, the report recognized that there are patients with diabetes who have average fasting glucose levels within targets but whose A1C results are elevated. In such patients, monitoring and specifically treating postprandial elevations may bring about improvements in A1C, perhaps at a lower risk of hypoglycemia and weight gain than further lowering fasting and premeal glucose levels.14 Since 2001, ACE has recommended targeting a 2-hour postprandial glucose <140 mg/dl in an effort to achieve near-normal glycemia.10 Consistent postprandial glucose levels of <140 mg/dl could be predicted to produce average A1C levels <6%,15 but the risks, benefits, and costs of such an approach are completely unknown. In the 2003 update of the ADA standards of care, a postprandial target of <180 mg/dl is included, not because there are adequate studies to suggest a target postprandial glucose level, but instead to provide a level of postprandial glucose unlikely to be seen in normal individuals16 and one that would be associated with an A1C of modestly <7%.15 

Finally, and perhaps most significantly, at least symbolically, the ADA 2003 standards of care document has changed from including separate tables that deal with glycemic, blood pressure, and lipid goals to a single table titled “Summary of Recommendations for Adults With Diabetes Mellitus.” This suggests the equivalent importance of each of these targets in avoiding the complications of diabetes.

There are dozens of other changes in this year’s version of the ADA standards of care. Most are modest in their impact on diabetes care, and some represent modifications of opinion, generally in the absence of randomized clinical trial data, aimed at bringing the ADA standards in line with the standards put forth by other organizations.

It is important to keep in mind as we work with individual patients that these guidelines are largely derived from incomplete data sets in groups of people. Though the treatment targets pursued in individual patients need to generally conform to these recommendations, they must be individualized, and responses to interventions must be carefully monitored.

To reach these targets, patients will often require a dozen or more pharmaceutical agents on top of a program of lifestyle intervention, with substantial potential for therapeutic misadventure related to drug side effects, drug-drug interactions, and the possibility of overwhelming patients because of the complexity and expense of the regimens required. Though the goals of therapy seem pretty cut-and-dried, there is still substantial art in the science of pursuing them. Controversy regarding the particular targets recommended by various groups is greater than the consensus that many patients are inadequately controlled by any standards and that substantial barriers to care exist in clinical practice.