Type 2 diabetes (T2D) is a progressive multisystem heterogeneous disease. Apart from its well known microvascular and ischemic macrovascular complications, recent evidence has accumulated that patients with this disease are also at an increased risk of congestive heart failure, fractures, cancer, pancreatitis, and serious infections such as pneumonia. Existing expert guidelines recommend tailoring therapy to the individual patient. However tailoring therapy remains a challenge for clinicians because evidence of risks and benefit on patient-oriented outcomes, as well as the underlying treatment preferences of patients for these outcomes, remains incomplete.
Risks and Benefits: A Deficit of Evidence
Efficacy of the 10-plus classes of oral drugs, injectables, and versions of short- and long-acting insulins on glycated hemoglobin is well established from short-term, placebo controlled trials of moderate size. However the effectiveness of these agents in the real world remains uncertain. Although the linkages between glycated hemoglobin reduction and improved microvascular outcomes appears to be reasonable, we still lack prospective clinical trial evidence demonstrating, for example, that glycemic control mediated via incretin mimetics such as exenatide or DPP-IV inhibitors, or via the thiazolidinediones, leads to a reduction in the macrovascular complications of T2D. Moreover, the applicability of evidence generated from highly restrictive clinical trials to the more comorbid and older general populations remains unclear. For these reasons, our knowledge of the benefits of individual drugs and drug classes remains limited.
Equally lacking is sufficient evidence about risks, which, along with efficacy data, is essential if we are to quantify and evaluate information on risk along with benefit in order to tailor therapy. For example, there is insufficient high quality evidence from controlled studies for clinicians on the potential harms, if any, of the newly approved therapies for T2D. Initially exenatide and now sitagliptin carry regulatory warnings about pancreatitis, based largely on biological plausibility and spontaneous reports. Clinical trials have been underpowered to detect the risk of pancreatitis as this is a rare event. Two epidemiologic studies of incretin mimetics have shown no evidence of increased risk. The case of the thiazolidinediones linked to heart failure, fractures, and now serious pneumonia and for rosiglitazone to myocardial infarction should serve as a reminder to not equate absence of information on safety as proof of safety.
Lack of Data on Patient Preferences
In addition to a lack of risk/benefit information, there is a paucity of data on patient preferences in T2D. A simple search for patient preferences and type 2 diabetes, yielded only eight citations in February 2011 in contrast to the voluminous research on T2D. Without this data, tailored therapy is unrealistic. It is illogical to expect clinicians to balance the probabilities of various beneficial outcomes such as glycemic control or various harms such as hypoglycemia because probabilities of events in themselves have no direct clinical relevance without information on underlying patient preferences. This is because, in a shared decision-making model, underlying patient preferences for various outcomes are an inextricable part of the ultimate risk and benefit of any given treatment. Thus, it would not be unreasonable for a post-menopausal female with comorbid congestive heart failure and osteoporosis to forego thiazolidinedione therapy, however effective it might have been found to be in terms of glycemic control, for alternative options if her treatment preferences are weighted to avoid fractures and congestive heart failure. In some cases preferences for patients with T2D are weighted to provide relief of symptoms of depression and pain, which may collide with clinician’s priorities on laboratory targets.
Thus, carefully eliciting patient preferences for various outcomes is needed to tailor therapy for the individual patient with type 2 diabetes.