Application of Comparative Effectiveness Research in Evaluation of Treatments for Type 2 Diabetes Mellitus

Supplements and Featured Publications, Evaluating the Role of Incretin-Based Therapies in the Management of Type 2 Diabetes, Volume 17, Issue 2 Suppl


reviews available CER among the incretin-based agents and for older T2DM therapies. CER is expected to open new avenues for research to clarify best practices in the treatment of T2DM, as well as possibly reduce treatment costs and improve the overall quality of public health.

Comparative effectiveness research (CER) is a relatively new strategy in drug development and healthcare designed to aid consumers, clinicians, purchasers, and policy makers in determining the best treatment options for individual patients. While defined in scope of clinical research, CER has the potential to be applied to other outcomes including economic evaluation of treatments and their potential impact on overall healthcare utilization by using all available data from clinical trials, systematic reviews, meta-analyses, and observational studies. Whereas clinical trials have mainly compared new agents with placebo (either alone or added to background therapy), CER compares new treatments head-to-head with active comparators to determine a patient’s best available options. Because of the emphasis on the patient, and its potential to identify the best treatment options and produce substantial cost savings, CER is likely to be an integral part of American healthcare reform. The Affordable Care Act’s emphasis on CER should benefit patients with type 2 diabetes (T2DM), a major chronic disease that affects 26 million Americans. The most recent innovations in marketed T2DM therapies include 2 classes of drugs focused on the incretin system: glucagon-like peptide-1 (GLP-1) receptor agonists and dipeptidyl peptidase-4 (DPP-4) inhibitors. This article

(Am J Manag Care. 2011;17:S41-S51)

Comparative Effectiveness Research in Evidence-Based Medicine

The Institute of Medicine (IOM) defines comparative effectiveness research (CER) as “the generation and synthesis of evidence that compares the benefits and harms of alternative methods to prevent, diagnose, treat, and monitor a clinical condition or to improve the delivery of care.” CER serves “to assist consumers, clinicians, purchasers, and policy makers to make informed decisions that will improve healthcare at both the individual and population levels."1 Half of the total healthcare spending in the United States is attributed to hospital care and physician services (31% and 21%, respectively), with prescription drugs accounting for only 10% of costs. However, technology and prescription drugs are thought to be major drivers of healthcare spending (the number of prescriptions purchased increased 39% in the United States between 1999 and 2009), along with chronic diseases, the aging of the population, and administrative costs.2,3

CER may include new data, new analyses of old data, and systematic reviews of previous research.4 It can inform decision-making by providing an objective scientific base of evidence that directly compares one diagnostic or therapeutic modality with another. CER establishes an evidence base by analyzing and consolidating different types of data from:

  • Traditional randomized controlled clinical trials, especially head-to-head trials
  • Systematic reviews and meta-analyses of data from previous studies, especially those combining and comparing results from placebo-controlled trials of multiple single agents
  • Prospective and retrospective observational studies of realworld practice using registries, claims data, medical records review, and other databases

CER differs from research traditionally undertaken for new drugs and devices.5 The US healthcare system has historically given precedence to studies that weighed the safety and efficacy of a single product against a placebo control (with or without background therapy), which has also been a standard for US Food and Drug Administration (FDA) approval. Until recently, corporations had little incentive to make direct comparisons to treatments other than their own products. The basis of CER is the comparison of active agents, reaching conclusions on safety and efficacy for heterogeneous patient populations and specific subpopulations,5 including patients ordinarily excluded from randomized controlled trials, to provide more generalizable results which better reflect real-world situations faced by practitioners and patients.5

Well-designed CER identifies patient characteristics that can make meaningful differences in outcomes,4 which can be clinical or economic, and demonstrates the utility of different treatments for different populations. Applying CER to decision-making processes by government, insurers, managed care, and clinicians should improve quality of care while containing costs. Therefore, this new way of evaluating scientific evidence may soon become essential to healthcare delivery in the United States.

One example that illustrates the ideal use of CER involves implantable cardiac defibrillators (ICDs). ICDs were evaluated as a viable treatment compared with drug therapy in the prevention of cardiac death among patients at risk of ventricular tachycardia or fibrillation.5 At that time, ICD use was considered controversial in terms of efficacy, clinical use, and cost. Results from MADIT I, which randomized patients having ventricular arrhythmias unresponsive to medication and who had had a myocardial infarction to either ICD or usual care, showed a considerable survival benefit for patients given an ICD. MADIT I was followed by MADIT II, which expanded the patient population and also showed results that favored ICD use compared with usual medical therapy (albeit not as strongly as MADIT I). These results helped to guide the initial clinical indications of ICDs as well as the expansion of those indications to patient subpopulations. Economic data were also examined and analyses found that although ICDs were costly, they were associated with a reasonable return on investment, particularly for patients at greatest risk, and thus provided evidence to include ICDs in health insurance benefit packages.

Similar to the evaluation of ICDs, the evaluation of the value of drug therapies, especially new ones, is a challenge. With increasing costs of pharmaceuticals clearly visible, acquisition costs of new agents have been a key focal point in differentiating therapies. Considering that drug utilization is a small percentage of total healthcare costs, has the ability to influence medical utilization (eg, emergency care, hospitalization), and has been shown in some instances to reduce medical costs over and above acquisition costs,6 CER is an opportunity for new (and existing) agents to demonstrate their value in impacting outcomes outside the pharmacy silo.

The Government Perspective

Medicare and Medicaid coverages have often deferred to provider decisions that vary widely and may lack an evidence base. Standards set by these government insurers are often followed by private-sector payers. Meanwhile, costs have increased rapidly while national outcomes remain the same as or sometimes worse than those of other developed countries.

CER has become an integral part of healthcare reform. The 2009 American Recovery and Reinvestment Act allocated $1.1 billion for CER.5,7 The 2010 Affordable Care Act established the Patient-Centered Outcomes Research Institute to set CER priorities, create and implement a research agenda, and distribute findings to decision makers.8 The Agency for Healthcare Research and Quality (AHRQ) sponsors Evidence-based Practice Centers that conduct and regularly update Comparative Effectiveness Reviews.9 These reviews consolidate and analyze the evidence on pharmaceuticals, medical devices, and other interventions, define strengths and limitations of that evidence, and include strategies for optimizing organization, management and delivery of medications, devices, and services.9 Through initiatives like these, CER can help determine metrics of healthcare quality and provide an evidence base for national health programs and decisions, including decisions about Medicare and Medicaid payments.

The Physician Perspective

Some physicians have viewed CER as a threat to autonomy or the doctor-patient relationship and a harbinger of treatment rationing.5 However, CER can identify and fill knowledge gaps that cause uncertainties in clinical practice.4 CER also opens new research opportunities for physicians and scientists.5 Whereas corporate funding has sometimes limited research, recent legislation ensures greater government funding for head-to-head drug trials, meta-analyses, and other studies that broaden the clinical knowledge base.

CER can spur innovation and create disincentives for “metoo” drugs and devices that do not surpass existing modalities but complicate clinical decisions.5 Additionally, CER can protect against “blind” cost containment by providing evidence to underlie organizational financial incentives and payer decisions.5 Ultimately, CER should lead to important medical decisions being guided by science and not by the marketplace.5

On the clinical level, CER can potentially improve quality of care by providing a scientific evidence base. Discussing the research underlying clinician recommendations can reinforce patient understanding and enhance the doctor-patient relationship. When included in the medical record, the evidence base for decisions can potentially inhibit litigation and reduce wasteful “defensive medicine.” Thus, contrary to fears, CER can work to the benefit of physicians and patients. However, it is important to note that physicians will still need to make additional interpretations with regard to therapeutic convenience, likelihood of adherence, and costs.

The Managed Care Perspective

Cost containment often uses imprecise “blunt instruments” of financial incentives, new programs, and organizational changes.5 However, CER can “surgically” tailor changes in healthcare practice and management that ultimately benefit public health. Industry-funded research offers little incentive to prove the superiority of older, simpler, and cheaper diagnostic and treatment modalities. CER can reveal insights that favor these modalities; for example, hypertension studies that demonstrate the superiority of diuretics versus ACE inhibitors.5 CER can discourage the research and development of equivalent, expensive drugs and devices and instead promote the development of comparably effective, but less expensive, modalities.5 Moreover, CER may reveal that specific therapies are associated with improved outcomes, such as fewer emergency department visits and hospitalizations—the key drivers in healthcare costs.

CER can provide the evidence base to justify new efficiencies and cost-saving measures, including stratification of treatment options. CER findings are likely to be endorsed and disseminated by such recognized authorities as specialty medical societies and the IOM, potentially reducing conflicts between managed care and practitioners. With this starting base of agreement, stakeholders can then work together on improving access to approved treatments. As CER findings are increasingly accepted by the medical community, practitioner choice of preferred treatments should increase and costly, labor-intensive processing of claims and appeals for non-preferred treatments should decrease.

Other innovative value-based programs (eg, patient-centered medical homes, pay for performance) can incorporate the best practices established by CER to further improve care and reduce costs.

CER in Diabetes: Oral Antidiabetic Drugs and Insulin

According to the Centers for Disease Control and Prevention, currently 25.6 million Americans age 20 and older have diabetes and about 215,000 people younger than age 20 have diabetes. An estimated 79 million Americans age 20 and older have prediabetes.10 If current trends continue, 1 in 3 US adults are predicted to have diabetes in 2050.11 As prevalence increases, the cost burden will increase; therefore, diabetes-related CER will become more important. Over years of disease progression, a patient with T2DM may receive numerous agents to achieve and maintain glycemic control, and may experience complications.

CER may be the best way to consolidate and interpret data on the many agents involved, and thereby guide rational treatment decisions. Examples can be observed in other countries since CER is still in its infancy in the United States. However, it must be noted that while the underlying goals are the same, there are differences, notably that CER discourages the use of cost-utility analyses (ie, outcomes measured in “dollars per quality-adjusted life-year [QALY]”); this differs from practices in other countries. For example, in the United Kingdom and Australia, cost-utility findings have influenced diabetes drug approvals.5 After cost-utility studies found that insulin glargine had only slight advantages over neutral protamine Hagedorn (NPH) insulin, glargine was approved by the Australian government only if at a greatly lower price than initially proposed. In the United Kingdom, glargine was approved only for type 1 diabetes and a small subset of patients with T2DM, a tailoring of application likely to contain costs.

In 2007, AHRQ released a Comparative Effectiveness Review of oral antidiabetic drugs (OADs) for adults with T2DM.9 The review was designed to answer several key

questions about differences in OAD efficacy and safety.9 Besides such diabetes-specific metrics as glycosylated hemoglobin (A1C) levels, key questions addressed blood pressure, serum lipids, microvascular and macrovascular complications, quality-of-life (QOL) measures, and life-threatening and minor adverse events (AEs).9 To enable better treatment individualization, this review investigated safety differences across populations defined by race/ethnicity, age, sex, and comorbidities.9

The AHRQ review found that most OADs had similar efficacy in reducing A1C and all combination therapies increased efficacy compared with monotherapy—but only by 1%.9 Several other beneficial effects were found. Metformin consistently promoted weight loss; depending on the study, acarbose promoted weight loss or was weight neutral.9 Lowdensity lipoprotein (LDL) cholesterol levels decreased with metformin and second generation sulfonylureas, high-density lipoprotein (HDL) cholesterol increased with thiazolidinediones (TZDs), and triglycerides decreased with pioglitazone but increased with rosiglitazone.9 TZDs consistently increased LDL; rosiglitazone increased LDL more than pioglitazone. 9 The available OADs had minimal to no effect on blood pressure.9

Differences in AEs also became clear. TZDs were associated with LDL increases, risk of congestive heart failure, weight gain, edema, and anemia.9 Second generation sulfonylureas were associated with hypoglycemia and weight gain, and metformin with gastrointestinal (GI) AEs.9 Despite widespread concerns, the review did not find significant differences in the incidence of lactic acidosis between metformin and other OADs,9 an example of a CER finding that may not emerge with more traditional research.

The AHRQ review found insufficient evidence to make conclusions about OAD outcomes in cardiovascular and peripheral vascular disease, retinopathy, nephropathy, neuropathy, and health-related QOL.9 Furthermore, insufficient data existed to draw conclusions about differences in OAD safety and efficacy across demographic groups and co-morbid conditions,9 making questions about long-term outcomes and treatment individualization priorities for future CER. Nevertheless, the differences found between and within classes may help guide rational treatment decisions based on individual patient risk factors, preferences, and other characteristics.

CER in Diabetes: Incretin-Based Agents

The most recent innovations in marketed T2DM therapies include 2 classes of drugs focused on the incretin system: glucagon-like peptide-1 (GLP-1) receptor agonists (eg,

exenatide, liraglutide), which stimulate insulin secretion, decrease glucagon secretion, and slow gastric motility and emptying12,13; and dipeptidyl peptidase-4 (DPP-4) inhibitors (eg, sitagliptin, saxagliptin), which work by blocking the enzymatic degradation and inactivation of GLP-1 and glucose-dependent insulinotropic polypeptide (GIP), both of which are crucial in glucose homeostasis.14,15 Additional information about incretin-based therapies appears in the article by Calabrese16 in this supplement. Head-to-head studies have compared the incretin-based agents with older therapies and sometimes with each other. Thus, CER can speed evaluation of the new medications and help establish appropriate use in specific patient populations.

DPP-4 Inhibitors

A 24-week randomized, double-blind trial of 1091 patients with type 2 diabetes compared the DPP-4 inhibitor sitagliptin 50 mg twice daily metformin 500 mg twice daily, sitagliptin 50 mg twice daily metformin 1000 mg twice daily, metformin 500 mg twice daily, metformin 1000 mg twice daily, sitagliptin 100 mg once daily, and placebo (Table 1).17 Although all active treatments produced significant reductions in A1C and fasting plasma glucose (FPG) levels (P <.001), sitagliptin alone was not more effective than either dose of metformin alone.17 However, the combinations of sitagliptin and metformin had greater efficacy in reducing A1C (P <.01) and FPG (P <.01) levels than the corresponding metformin monotherapy dose or sitagliptin monotherapy.17 Homeostasis model assessment of beta cell function (HOMA-B) showed greater improvements with the combinations (P <.05) and little difference between sitagliptin and lower-dose metformin monotherapies.17

A 52-week, randomized, double-blind noninferiority study compared sitagliptin 100 mg/day with the sulfonylurea glipizide (5-20 mg/day; mean dose, 10.3 mg/day) in 1172 patients already receiving metformin (at least 1500 mg/day) (Table 1).18 Mean reduction in A1C level from baseline was -0.67% in both sitagliptin and glipizide groups.18 The proportion of patients who reached an A1C level less than 7% and the changes in FPG levels were also similar in the 2 groups (63% sitagliptin, 59% glipizide), but glipizide was associated with greater improvement in HOMA-B.18 Body weight decreased significantly with sitagliptin (least squares mean change from baseline, -1.5 kg) and increased significantly with glipizide ( 1.1 kg relative to baseline).18 Hypoglycemia occurred in 4.9% of patients given sitagliptin and 32.0% of patients given glipizide.18 It is important to note that the hypoglycemia limited the ability to titrate glipizide and limited the total dose of glipizide to approximately 10 mg/day (the intent was to titrate it up to 20 mg daily).

Table 1

Another trial compared sitagliptin 100 mg/day with the TZD rosiglitazone 8 mg/day or placebo when added to metformin (at least 1500 mg/day) ().19 A total of 273 patients with type 2 diabetes were randomized to treatment. Sitagliptin and rosiglitazone were associated with similar decreases in A1C and increases in HOMA-B.19 Greater reductions in FPG levels and insulin resistance (by homeostasis model assessment of insulin resistance [HOMA-IR]) were achieved with rosiglitazone than sitagliptin (FPG, -24.5 for rosiglitazone vs -11.7 for sitaglipin; HOMA-IR, -2.1 at week 18 for rosiglitazone compared with 0.3 for placebo and -0.5 for sitagliptin).19 Body weight decreased slightly with sitagliptin (-0.4 kg) and placebo (-0.8 kg) but increased with rosiglitazone (1.5 kg).19

A randomized, double-blind, double-dummy crossover trial compared sitagliptin 100 mg/day with the injected GLP-1 receptor agonist exenatide (5 μg twice daily for 1 week, then 10 μg twice daily for 1 week) in 94 patients on a stable metformin regimen (Table 1).20 Changes in FPG levels were similar with the 2 agents, but exenatide produced significantly greater reductions (P <.0001) in postprandial plasma glucose (PPG) levels, which were confirmed after crossover by decreases observed in patients switched to exenatide and increases in patients switched to sitagliptin.20 Beta cell function assessed by insulinogenic index and insulin secretion rate improved significantly with exenatide versus sitagliptin.20 Compared with sitagliptin, exenatide significantly slowed gastric emptying and reduced caloric intake and body weight (-0.8 ± 0.2 kg vs -0.3 ± 0.2 kg),20,21 but was associated with more nausea and vomiting (34% vs 12% and 24% vs 3%, respectively).20 Among the safety analysis population (n = 95), 4 patients withdrew during exenatide treatment (2 due to AEs) and 4 during sitagliptin treatment (1 from AEs).20

Within the past 2 years, head-to-head studies have shown that combination treatment with saxagliptin metformin is comparable to sitagliptin metformin in safety and efficacy, 22 promotes weight loss versus gain with less hypoglycemia when compared with glipizide metformin,23 and increases antihyperglycemic efficacy over monotherapy metformin24 or uptitrated glyburide25 (Table 1).

The preceding studies added valuable evidence that sitagliptin and saxagliptin are effective antihyperglycemic agents with a relatively low risk of AEs. More studies are needed for a comprehensive CER evaluation of the DPP-4 inhibitors.

GLP-1 Receptor Agonists: Liraglutide

The large, multicenter, 26-week and 52-week Liraglutide Effect and Action in Diabetes (LEAD) trials compared liraglutide, a human GLP-1 receptor analog, injected subcutaneously once daily, with other antidiabetic medications as either monotherapy or added on to 1 or 2 OADs; some of these trials had extensions for an additional 1 to 3 years (Table 2). Please refer to the article by Bode26 in this supplement for further information.


A 52-week, double-blind, double-dummy monotherapy trial showed significantly greater A1C level reductions with liraglutide 1.2 or 1.8 g/day than glimepiride 8 mg/day.27 (A1C decreased by 0.51%, 0.84%, and 1.14% from baseline for glimepiride, 1.2-mg, and 1.8-mg liraglutide groups, respectively; P <.0001 for 1.8-mg liraglutide). Significantly more patients given liraglutide reached an A1C level less than 7% compared with those given glimepiride (43% and 51% of 1.2-and 1.8-mg groups, respectively, compared with 28% of the glimepiride group).27 Body weight decreased with liraglutide and increased with glimepiride (P = .0001).27


In a double-blind, double-dummy, oral and injected placebo-controlled trial, patients receiving metformin 1g twice daily were randomized to liraglutide 0.6, 1.2, or 1.8 mg/day, glimepiride 4 mg/day, or placebo.28 Liraglutide 1.2- and 1.8-mg doses were noninferior to glimepiride in A1C reduction, but all doses of liraglutide were superior to metformin alone. FPG reductions were similar for liraglutide and glimepiride groups; PPG reductions were similar for glimepiride and 1.2- and 1.8-mg liraglutide. Patients in the glimepiride group gained weight, whereas patients in all liraglutide groups lost weight.28 Approximately 3% of patients in the liraglutide and placebo groups had minor hypoglycemia versus 17% of those in the glimepiride group.28 GI AEs were more common with liraglutide (35%-44%) than with glimepiride (17%).28


Patients were randomized to receive glimepiride 2 to 4 mg/day combined with liraglutide 0.6, 1.2, or 1.8 mg/day or rosiglitazone 4 mg/day.29 Glimepiride doses were determined by the highest doses approved in participating countries and could be adjusted between 2 to 4 mg/day because of hypoglycemia or other AEs.29 All arms were placebo-controlled.29 Rapid reductions in A1C were seen in the liraglutide groups (particularly 1.2- and 1.8-mg groups; -1.1% from baseline) compared with the rosiglitazone (-0.4%) or placebo ( 0.2%) groups. All active groups were superior to placebo (P <.0001); the 2 higher doses of liraglutide were superior to rosiglitazone (P <.0001; 0.6-mg dose noninferior). Greater decreases were seen in FPG with 1.2- and 1.8-mg liraglutide versus rosiglitazone (P <.006). Body weight decreased with liraglutide 1.8 mg and increased with rosiglitazone.29


The LEAD-4 study was not an example of CER, because it compared the addition of liraglutide with placebo, not another active agent.30


Patients were randomized to metformin 1 g twice daily and glimepiride 4 mg/day combined with liraglutide 1.8 mg/day, placebo, or open-label insulin glargine.31 Compared with glargine or placebo, liraglutide reduced A1C levels significantly (A1C reduction from baseline was 1.33% for liraglutide, 0.24% for placebo, and 1.09% for insulin glargine) and was associated with greater weight loss (-1.8 kg vs 1.6 kg for insulin glargine).31 Systolic blood pressure decreased with liraglutide and placebo and increased with glargine.31


An open-label trial enrolled patients receiving maximally tolerated doses of metformin, sulfonylurea, or both.32 Patients were stratified by previous OADs and randomized to liraglutide 1.8 mg/day or exenatide 10 μg twice daily.32 Compared with exenatide, liraglutide was associated with significantly greater decreases in A1C (-1.12% vs -0.79% for exenatide) and FPG levels, and more patients achieved an A1C level less than 7% (54% vs 43% in the exenatide group).32 Weight loss was similar in both groups.32 Compared with liraglutide, more patients given exenatide had minor hypoglycemia and nausea that was more persistent.32

Comparison With Sitagliptin

In an open-label trial in 665 patients with T2DM, sitagliptin (100 mg/day) or the GLP-1 receptor agonist liraglutide (1.2 or 1.8 mg/day) was added to metformin (at least 1500 mg/day) (Table 2).33 Reductions in A1C from baseline were found to be superior among patients given liraglutide (P <.0001) vs sitagliptin (-1.5% for 1.8-mg dose; -1.24% for 1.2-mg dose; -0.90% for sitagliptin). Significantly more patients achieved target A1C levels (<7%, P <.0001) with liraglutide versus sitagliptin and mean decreases in FPG were greater for patients taking liraglutide (P <.0001). Weight loss was greater in the liraglutide treatment group: -3.38 kg for 1.8-mg dose, -2.86 kg for 1.2-mg dose, and -0.96 for sitagliptin. Liraglutide (at both doses) was associated with significant improvements in HOMA-B cell function (P <.0001), C-peptide concentration, and proinsulin-to-insulin ratio compared with sitagliptin. Transient nausea occurred more often with liraglutide than sitagliptin; the incidence of minor hypoglycemia was similar among all 3 groups.33

GLP-1 Receptor Agonists: Exenatide

In addition to LEAD-6 and the sitagliptin trial, other exenatide studies (Table 2) included an open-label trial comparing exenatide 10 μg twice daily with titrated insulin glargine in patients receiving metformin and a sulfonylurea (stable doses).34 Both agents reduced A1C levels similarly (1.11%) and were associated with comparable rates of hypoglycemia. 34 Exenatide was associated with greater reductions in PPG excursions and body weight (-2.3 kg), but also more GI AEs.34 Glargine was associated with greater reductions in FPG levels (P <.001), but also weight gain (1.8 kg).34

Another open-label trial (Table 2) compared exenatide 10 μg twice daily and insulin glargine in patients on stable metformin or sulfonylurea monotherapy.35 Both added agents achieved similar glycemic control (P <.001), but greater FPG level reductions occurred with glargine (fasting serum glucose, P <.001) and lower PPG excursions occurred with exenatide.35 Exenatide was associated with weight loss, but more GI AEs; glargine was associated with weight gain and hypoglycemia.35

A third open-label trial (Table 2) compared exenatide (5-10 μg twice daily) with biphasic insulin aspart in patients who continued on metformin/sulfonylurea treatment.36 Glycemic control and AE results were similar to those in the glargine trials.36 Exenatide showed noninferiority to insulin aspart with respect to glycemic control (A1C difference from baseline, -1.04 for exenatide and -0.89 for insulin aspart). The rates of hypoglycemia were similar across treatment groups.36

A new, longer-acting formulation of exenatide is designed to be injected once weekly (vs twice daily in the currently marketed version).37 The new formulation has yet to receive FDA approval, and the FDA asked the manufacturer to complete a study on cardiovascular effects.37 Nevertheless, some efficacy studies have been completed, including CER studies.

One trial compared long-acting exenatide (2 mg injected once weekly) with a DPP-4 inhibitor (sitagliptin, 100 mg/day) and a TZD (pioglitazone, 45 mg/day) in patients treated with metformin (Table 2).38 The trial included both oral and injected placebo controls.38 A1C levels and body weight decreased significantly more with exenatide 2 mg once weekly than sitagliptin 100 mg/day or pioglitazone 45 mg/day (-1.5% vs -0.9% for sitaglipin and -1.2% for pioglitazone). Weight loss in the exenatide group was -2.3 kg versus -0.8 kg for sitagliptin and 2.8 kg for pioglitazone.38 The most frequent AEs were GI symptoms (exenatide and sitagliptin) and upper respiratory tract infection and peripheral edema (pioglitazone).38

An open-label trial compared long-acting exenatide (2 mg once weekly) with titrated insulin glargine in adults receiving OADs (metformin or metformin/sulfonylurea) (Table 2).39 A1C levels decreased more with exenatide (-1.5%) than insulin glargine (-1.3%), and more patients given exenatide achieved an A1C level less than 7% (60% vs 48% of insulin glargine users).39 Body weight changes were -2.6 kg for exenatide and 1.4 kg for insulin glargine. Significantly more patients given exenatide than those given glargine discontinued the study because of AEs, including injection-site reactions and nausea.39

The preceding studies contribute to the body of CER evidence about the effects of GLP-1 receptor agonists and other injected and oral therapies, which may help guide treatment choice for individual patients with specific characteristics (eg, obesity, inadequate glycemic control with oral medications, intolerance to GI symptoms). Five of the 7 CER trials that included exenatide were open-label, thereby decreasing the weight of evidence. Additional fully blinded and controlled studies with different agents are needed. The National Institutes of Health lists many other trials comparing GLP-1 receptor agonists with other agents that are in the recruiting or active stages, including at least 1 that compares liraglutide with long-acting exenatide and 1 that compares liraglutide with albiglutide, a new once-weekly GLP-1 receptor agonist.40,41

Opportunities and Challenges for CER Specific to Diabetes Treatment

With the epidemic of type 2 diabetes continuing and the cost of diabetes medications rising (the cost doubled between 2001 and 2007 in the United States), the need for CER in comparing treatments and strategies is great.42 As such, study designs and hypotheses should mirror the types of real-life decisions faced by managed care practitioners and those who determine medical policies. Retrospective analyses can generate results faster than prospective clinical trials, speeding important evidence to decision makers; however, it should be noted that in systematic reviews, these types of studies are often discounted in favor of clinical trial data. Moving forward, the role of this type of data and information needs to be better understood. In addition, industry- or privatelyfunded research can limit studies to 1 or a few agents, whereas government funding opens CER to every diabetes drug available.

CER can help guide individualized, patient-centered T2DM treatment and potentially reduce costly trial-anderror therapy. CER can determine which agents have complementary effects in combination, and which may be redundant for a given patient. Further, well-designed CER can document relationships between specific treatments and long-term outcomes and determine for patients with comorbid diseases and risk factors the appropriateness of agents with effects beyond glucose control (eg, cardiovascular effects, weight loss). Finally, CER can determine which agents may be most appropriate for patients with specific safety concerns.

An urgent need exists to establish and use uniform metrics in all diabetes research and to identify subpopulations that will benefit from specific therapies. Standard variables may include A1C, FPG, and PPG levels, hypoglycemia incidence, weight gain/loss, blood pressure, lipid levels, microvascular and macrovascular outcomes, patient satisfaction, QOL, and beta cell function. The way results are defined and measured must be consistent for accurate comparisons across studies. For example, definitions of severe and minor hypoglycemia and GI AEs should be uniform.

Head-to-head studies of monotherapy and combination therapy with similar and different mechanisms of action are needed to generate the best evidence. Placebo control is desirable, but complicates head-to-head studies including injectable and oral agents because both methods of administration must be controlled. CER can accommodate results of open-label studies by giving them less weight in systematic analyses, but questions of relative efficacy could remain unanswered. Studies with insulin as a comparator also complicate CER because of the many formulations and dosing regimens and the need for individualized dosing adjustments throughout the study.

More systematic reviews are needed that consolidate and analyze data from studies of different agents and drug classes. More research is needed to evaluate treatment effects in different demographic groups. More long-term outcome studies are needed. Although clinical trials typically last weeks, important diabetes outcomes relate to efficacy over years. Immediate CER needs in T2DM may be served by claims data analyses, retrospective chart reviews, and systematic reviews of existing evidence.


The recent embrace of CER opens new avenues for research that can clarify best practices in T2DM treatment. By creating an evidence base for best medications for specific patient populations, CER can enlighten decision makers in government, insurance, managed care, and clinical practice, and build a foundation of agreement for current practice and set the stage for future advances in care. Research and development of “me-too” drugs and devices, clinical choices of ineffective drugs, and assumptions that the most (or least) expensive therapy is best may all decrease. The application of CER to clinical practice can potentially reduce costs for diabetes treatment and improve public health.

Author Affiliation: Department of Medicine: Endocrinology, Diabetes & Clinical Nutrition and Harold Schnitzer Diabetes Health Center, Oregon Health & Science University, Portland, OR.

Funding Source: Financial support for this supplement was provided by Novo Nordisk.

Author Disclosure: Dr Ahmann reports consultancy/advisory board membership at Novo Nordisk. Dr Ahmann has received grants from Amylin, GlaxoSmithKline, Lilly, Mankind, and Medtronic as well as honoraria and lectureship fees from the American Diabetes Association, Lilly, Medical Education Resources, and Merck.

Authorship Information: Concept and design; acquisition of data; analysis and interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; administrative, technical, or logistic support; and supervision.

Address correspondence to: Andrew Ahmann, MD, Oregon Health & Science University, 3181 SW Sam Jackson Prk Rd, OP05-DC, Portland, OR 97239. E-mail:

1. IOM (Institute of Medicine). Initial National Priorities for Comparative Effectiveness Research. Washington, DC: The National Academies Press; 2009.

2. Kaiser Family Foundation. US health care costs: background brief. Available at: Accessed February 23, 2011.

3. Kaiser Family Foundation. Prescription drug costs: background brief. Available at: Prescription-Drug-Costs/Background-Brief.aspx#footnote5. Accessed February 23, 2011.

4. Sox HC, Helfand M, Grimshaw J, et al. Comparative effectiveness research: challenges for medical journals. Am J Managed Care. 2010;16(5):e131-e133.

5. Mushlin AI, Ghomrawi HMK. Comparative effectiveness research: a cornerstone of healthcare reform? Trans of the American Clinical and Climatological Association. 2010;121:141-155.

6. Civan A, Köksal B. The effect of newer drugs on health spending: do they really increase the costs? Health Econ. 2010;19(5):581-595.

7. United States Department of Health and Human Services. Comparative Effectiveness Research Funding. 2009. Available at: Accessed October 21, 2010.

8. Sox HC. Comparative effectiveness research: a progress report. Ann Intern Med. 2010; 153(7):469-472.

9. Bolen S, Wilson L, Vassy J, et al. Comparative effectiveness and safety of oral diabetes medications for adults with type 2 diabetes. Comparative effectiveness review number 8. Agency for Healthcare Research and Quality, US Department of Health and Human Services, AHRQ Publication No. 07-EHC010-EF. July 2007. Available at: Accessed October 21, 2010.

10. National Diabetes Fact Sheet, 2011. Available at: Accessed February 23, 2011.

11. Centers for Disease Control and Prevention. Number of Americans with diabetes projected to double or triple by 2050. United States Department of Health and Human Services. October 22, 2010. Available at: Accessed October 25, 2010.

12. Victoza [prescribing information]. Princeton, NJ: Novo Nordisk, Inc; 2010.

13. Byetta [prescribing information]. San Diego, CA: Amylin Pharmaceuticals, Inc; 2010.

14. Onglyza [prescribing information]. Princeton, NJ: Bristol-Myers Squibb; 2011.

15. Januvia [prescribing information]. Whitehouse Station, NJ: Merck & Co, Inc; 2011.

16. Calabrese D. Differentiating incretin-based therapies for population-based health care. Am J Manag Care. 2011;17:S52-S58.

17. Goldstein BJ, Feinglos MN, Lunceford JK, Johnson J, Williams-Herman DE, for the Sitagliptin 036 Study Group. Effect of initial combination therapy with sitagliptin, a dipeptidyl peptidase-4 inhibitor, and metformin on glycemic control in patients with type 2 diabetes [published correction appears in Diabetes Care. 2008;31(8):1713-1714]. Diabetes Care. 2007;30(8):1979-1987.

18. Nauck MA, Meininger G, Sheng D, Terranella L, Stein PP: Sitagliptin Study 024 Group. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor, sitagliptin, compared with the sulfonylurea, glipizide, in patients with type 2 diabetes inadequately controlled on metformin alone: a randomized, double-blind, non-inferiority trial. Diabetes Obes Metab. 2007;9(2):194-205.

19. Scott R, Loeys T, Davies MJ, Engel SS; Sitagliptin Study 801 Group. Efficacy and safety of sitagliptin when added to ongoing metformin therapy in patients with type 2 diabetes. Diabetes Obes Metab. 2008;10(10):959-969.

20. DeFronzo RA, Okerson T, Viswanathan P, Guan X, Holcombe JH, MacConell L. Effects of exenatide versus sitagliptin on postprandial glucose, insulin and glucagon secretion, gastric emptying, and caloric intake: a randomized, cross-over study. Curr Med Res Opin. 2008:24(10):2943-2952.

21. Campbell RK, Cobble ME, Reid TS, Shomali ME. Glucoselowering effects of incretin-based therapies. J Fam Pract. 2010;59(9 suppl 1):S5-S9. Available at: Accessed September 29, 2010.

22. Scheen AJ, Charpentier G, Ostgren CJ, Hellqvist A, Gause-Nilsson I. Efficacy and safety of saxagliptin in combination with metformin compared with sitagliptin in combination with metformin in adult patients with type 2 diabetes mellitus. Diabetes Metab Res Rev. 2010;26(7):540-549.

23. Goke B, Gallwitz B, Eriksson J, Hellqvist A, Gause-Nilsson I; D1680C00001 Investigators. Saxagliptin is non-inferior to glipizide in patients with type 2 diabetes mellitus inadequately controlled on metformin alone: a 52-week randomized controlled trial. Int J Clin Pract. 2010;64(12):1619-1631.

24. Jadzinsky M, Pfutzner A, Paz-Pacheco E, Xu Z, Allen E, Chen R; CV181-39 Investigators. Saxagliptin given in combination with metformin as initial therapy improves glycaemic control in patients with type 2 diabetes compared with either monotherapy: a randomized controlled trial. Diabetes Obes Metab. 2009;11(6):611-622.

25. Chacra AR, Tan GH, Apanovitch A, Ravichandran S, List J, Chen R; CV181-040 Investigators. Saxagliptin added to a submaximal dose of sulphonylurea improves glycaemic control compared with uptitration of sulphonylurea in patients with type diabetes: a randomized controlled trial. Int J Clin Pract. 2009;63(9):1395-1406.

26. Bode B. Liraglutide: A review of the first once-daily GLP-1 receptor agonist. Am J Manag Care. 2011;17:S59-S70.

27. Garber A, Henry R, Ratner R, et al; LEAD-3 (Mono) Study Group. Liraglutide versus glimepiride monotherapy for type 2 diabetes (LEAD-3 Mono): a randomized, 52-week, phase III, double-blind, parallel-treatment trial. Lancet. 2009;373(9662):473-481.

28. Nauck M, Frid A, Hermansen K, et al, for the LEAD-2 Study Group. Efficacy and safety comparison of liraglutide, glimepiride, and placebo, all in combination with metformin, in type 2 diabetes. Diabetes Care. 2009;32(1):84-90.

29. Marre M, Shaw J, Brandle M, et al, on behalf of the LEAD-1 SU study group. Liraglutide, a once-daily human GLP-1 analogue, added to a sulphonylurea over 26 weeks produces greater improvements in glycaemic and weight control compared with adding rosiglitazone or placebo in subjects with Type 2 diabetes (LEAD-1 SU). Diabetic Med. 2009;26:268-278.

30. Zinman B, Gerich J, Buse JB, et al; the LEAD-4 Study Investigators. Efficacy and safety of the human glucagon-like peptide-1 analog liraglutide in combination with metformin and thiazolidinedione in patients with type 2 diabetes (LEAD-4 Met TZD). Diabetes Care. 2009;32(7):1224-1230.

31. Russell-Jones D, Vaag A, Schmitz O, et al; Liraglutide Effect and Action in Diabetes 5 (LEAD-5) met SU Study Group. Liraglutide vs insulin glargine and placebo in combination with metformin and sulfonylurea therapy in type 2 diabetes mellitus (LEAD-5 met SU): a randomised controlled trial. Diabetologia. 2009;52(1):2046-2055.

32. Buse JB, Rosenstock J, Sesti G, et al; LEAD-6 Study Group. Liraglutide once a day versus exenatide twice a day for type 2 diabetes: a 26-week randomized, parallel-group, multinational, open-label trial (LEAD-6). Lancet. 2009;374(9683):39-47.

33. Pratley RE, Nauck M, Bailey T, et al; 1860-LIRA-DPP-4 Study Group. Liraglutide versus sitagliptin for patients with type 2 diabetes who did not have adequate glycaemic control with metformin: a 26-week, randomized, parallel-group, open-label trial. Lancet. 2010;375(9724):1447-1456.

34. Heine RJ, Van Gaal LF, Johns D, Mihm MJ, Widel MH, Brodows RG; GWAA Study Group. Exenatide versus insulin glargine in patients with suboptimally controlled type 2 diabetes: a randomized trial. Ann Intern Med. 2005;143(8):559-569.

35. Barnett AH, Burger J, Johns D, et al. Tolerability and efficacy of exenatide and titrated insulin glargine in adult patients with type 2 diabetes previously uncontrolled with metformin or a sulfonylurea: a multinational, randomized, open-label, two-period, crossover noninferiority trial. Clin Ther. 2007;29(11):2333-2348.

36. Nauck MA, Duran S, Kim D, et al. A comparison of twice-daily exenatide and biphasic insulin aspart in patients with type 2 diabetes who were suboptimally controlled with sulfonylurea and metformin: a non-inferiority study. Diabetologia. 2007;50(2):259-267.

37. Amylin, Lilly and Alkermes announce receipt of complete response letter from FDA for Bydureon [press release]. San Diego, CA: Amylin Pharmaceuticals, Inc, Indianapolis, IN: Eli Lilly and Co, and Waltham, MA: Alkermes, Inc; October 19, 2010. Available at: Accessed March 8, 2011.

38. Bergenstal RM, Wysham C, Macconell L, et al; DURATION-2 Study Group. Efficacy and safety of exenatide once weekly versus sitagliptin or pioglitazone as an adjunct to metformin for treatment of type 2 diabetes (DURATION-2): a randomized trial. Lancet. 2010;376(9739):431-439.

39. Diamant M, Van Gaal L, Stranks S, et al. Once weekly exenatide compared with insulin glargine titrated to target in patients with type 2 diabetes (DURATION-3): an open-label randomized trial. Lancet. 2010;375(9733):2234-2243.

40. National Institutes of Health. List results: clinical trials with liraglutide. Available at: Accessed October 25, 2010.

41. National Institutes of Health. List results: clinical trials with exenatide. Available at: Accessed October 25, 2010.

42. Nathan DM. Time for clinically relevant comparative effectiveness studies in type 2 diabetes. Ann Intern Med. 2011;154(2):131-132.