March 2009
|
Type 2 Diabetes and Glycemic Control
For now, the approach to the type 2 diabetes epidemic should include attempts to prevent the disease and to treat it with the medications known to be safe and effective.
By David M. Nathan, MD
In a recent editorial in the Archives of Internal Medicine (2008;168:2064–2066), David M. Nathan, MD, discussed whether lower glycemia is better for the prevention of cardiovascular disease (CVD) in diabetes. The following is a summary of Dr. Nathan's comments.
In response to the enormity of the type 2 diabetes epidemic,1 many new medications have been developed for glycemic control. Six new classes of medications have been approved in the United States since 1995, in addition to new variations of older drugs. Both the promise of large profits in this category and the known benefits of glycemic control on long-term diabetes complications have fueled the development of new diabetes medications.
The DCCT (Diabetes Control and Complications Trial)2—conducted among type 1 diabetes patients—compared intensive therapy (goal = near normoglycemia, mean A1C 7.0% for >6 years) with conventional therapy (mean A1C approximately 9.0%). This trial revealed that intensive therapy reduced retinopathy, nephropathy, and neuropathy by up to 76%. The UKPDS (United Kingdom Prospective Diabetes Study) conducted among approximately 5,000 new-onset type 2 diabetes patients, demonstrated the benefits of intensive diabetes management with several antidiabetic medications.3,4 The active intervention group achieved a mean A1C of approximately 7.0% compared with 7.9% in the conventional treatment group and a reduction in microvascular complications proportional to that in the DCCT, taking into account the relative A1C values in the two studies.
Longer follow-up of patients in the DCCT found a 57% reduction in major CVD events among patients assigned to intensive treatment versus conventional therapy.5 CVD results in patients with type 2 diabetes remain less clear. In the UKPDS, metformin was the only agent found to significantly reduce CVD.6 The rationale for an A1C <7.0% in type 2 diabetes, therefore, is to reduce the risk of microvascular and neuropathic complications. To date, the CVD benefits of lowering glucose in type 2 diabetes are hypothetical. Editor's note: A 10-year follow-up of the UKPDS, published after Dr. Nathan's editorial, demonstrated a long-term benefit of all intensive therapy on CVD, with a reduction of approximately 15% with insulin or sulfonylurea and a 33% reduction with metformin.7
IS LOWER BETTER?
Further doubt was cast on the "lower-is-better" role of glycemia in CVD by last years' ACCORD (Action to Control Cardiovascular Risk in Diabetes)8 and ADVANCE (Action in Diabetes and Vascular Disease) study9 results. Neither trial showed a benefit on primary CVD outcomes using intensive intervention strategies (median A1C of 6.4% and 6.3%, respectively) versus standard treatment (A1C 7.5% and 7.0%, respectively). In fact, intensive therapy was associated with increased mortality in ACCORD, although the exact mechanisms are not known. It must be considered that antidiabetic drugs have the potential to cause harm or benefit through individual pharmacologic effects, independent of glycemia.
Several meta-analyses10-12 and editorials13-15 have considered the safety of current antidiabetic medications. The most controversial meta-analysis, by Nissen and Wolski10 suggested that the thiazolidinedione (TZD) rosiglitazone (Avandia, GlaxoSmithKline) might be associated with an increase in CVD. In a recent meta-analysis, Selvin et al16 examined the relative effects of older oral agents (metformin and second-generation sulfonylureas) compared with the newer TZDs and meglitinides on cardiovascular morbidity and mortality and total mortality. This meta-analyses was carefully constructed, however, it suffers like all meta-analyses from the bundling together of the results of large and small clinical trials.
That said, Selvin's report did shed some light on the effects of commonly used oral antidiabetic agents on CVD. Metformin appears to be associated with a reduced odds ratio (OR) of CVD mortality (OR, 0.74; 95% confidence interval [CI], 0.62–0.89) and a borderline effect on CVD morbidity (OR, 0.85; CI, 0.69–1.05) and total mortality (OR, 0.81; CI, 0.60–1.08), compared with any other treatment (diet, placebo, or other oral agent). No other single oral agent appears to be associated with CVD risk, however rosiglitazone (but not pioglitazone [Actos, Takeda]) was again clouded in suspicion with ORs >1.0 (albeit with large CIs).
ARE THESE DRUGS SAFE?
It appears that the answer to the simple question, "Is it safe?" is not simple. Consensus guidelines continue to emphasize rapid and aggressive management with a goal of achieving an A1C <7.0% (not <6.0%).17 It is important to note that higher glycemic goals should be considered for patients with a limited life expectancy or with severe comorbidities when the risks of intensive therapy may outweigh any benefits on long-term diabetic complications. Based on UKPDS and DCCT data regarding the beneficial effects of A1C 7.0% on microvascular and neuropathic outcomes2,4,5 and the cautionary note added by ACCORD with an A1C goal <6.0%,8 the consensus goal of <7% is still appropriate.
As far as the choice of medications, the consensus algorithm recommends metformin at the time of diagnosis, followed quickly by a sulfonylurea, insulin, or a TZD if A1C is not <7.0%.17 Editors note: The latest revision of the algorithm, published after this editorial, has relegated pioglitazone to a second-tier therapy and rosiglitazone is no longer recommended.18 The algorithm medications were chosen by their ability to reduce glycemia, clinical experience, other medication effects, tolerability, patient acceptance, expense, and safety.
Selvin et al's meta-analysis,16 which focuses on CVD safety, appears in agreement with the established algorithm: emphasizing metformin and the sulfonylureas and limiting TZDs because of cost and risk of edema and heart failure, and limiting rosiglitazone because of potential CVD risk. Unfortunately, insulin therapy was not included in the meta-analysis; however, prior randomized trials have not found an adverse effect of insulin therapy on CVD.4,18
PREVENTION, TREATMENT
Our approach to the type 2 diabetes epidemic should include attempts to prevent the disease19 and to treat it with the medications known to be safe and effective. Given the role of glycemic control in ameliorating the toll of microvascular and neuropathic complications, a goal of A1C <7.0% should be emphasized for patients likely to benefit. Furthermore, we should be aggressive in applying other interventions known to reduce CVD risk.
How we currently assess rare but clinically important adverse effects of diabetes management is unsatisfactory. Meta-analyses are unreliable, but it is also impractical to increase the size and duration of controlled clinical trials to provide adequate statistical power to detect relatively infrequent events. Therefore, new approaches are needed to ensure the safety of drugs without slowing their development. The phased introduction of new medications with uniform, standardized collection of adverse outcome data could potentially identify relatively rare complications before the drugs are used by millions of patients. Clinical databases may also provide an early alert to adverse outcomes. In the meantime, practitioners should continue to use aggressively well-established and safe treatments to improve the long-term health of type 2 diabetes patients.
David M. Nathan, MD, is Director of the Clinical Research Center and the Diabetes Center at Massachusetts General Hospital, and a Professor of Medicine at Harvard Medical School. Dr. Nathan is also Chairman of the Diabetes Prevention Program. He may be reached at dnathan@partners.org.
|
