Diabetes and obesity: therapeutic targeting and risk reduction – a complex interplay


  • Kevin Niswender

    Corresponding author
    1. Department of Medicine, Division of Diabetes, Endocrinology and Metabolism, Tennessee Valley Healthcare System and Vanderbilt University School of Medicine, Nashville, TN, USA
      Dr Kevin Niswender, MD, PhD, Department of Medicine, Division of Diabetes, Endocrinology and Metabolism, Tennessee Valley Healthcare System and Vanderbilt University School of Medicine, Nashville, TN, USA.
      E-mail: kevin.niswender@vanderbilt.edu
    Search for more papers by this author

Dr Kevin Niswender, MD, PhD, Department of Medicine, Division of Diabetes, Endocrinology and Metabolism, Tennessee Valley Healthcare System and Vanderbilt University School of Medicine, Nashville, TN, USA.
E-mail: kevin.niswender@vanderbilt.edu


Obesity is a major risk factor for the development of diabetes and predisposes individuals to hypertension and dyslipidaemia. Together these pathologies increase the risk for cardiovascular disease (CVD), the major cause of morbidity and mortality in type 2 diabetes mellitus (T2DM). Worsening trends in obesity and T2DM raise a serious conundrum, namely, how to control blood glucose, blood pressure, and lipids when many antidiabetic agents cause weight gain and thereby exacerbate other cardiovascular risk factors associated with T2DM. Further, evidence suggests that some established antihypertensive agents may worsen glucose intolerance. Many patients who are obese, hypertensive, and/or hyperlipidaemic fail to achieve blood pressure, lipid and glycaemic goals, and this failure may in part be explained by physician reluctance to utilize complex combination regimens for fear of off-target effects. Thus, a clear need exists for clinicians to understand the risks and benefits of different pharmacologic, and indeed non-pharmacologic, options in order to maximize treatment outcomes. While intensive lifestyle modification remains an elusive gold standard, newer diabetes targets, including the incretin axis, may offer greater cardiovascular risk reduction than other antidiabetes therapies, although definitive clinical trial data are needed. The glucagon-like peptide-1 (GLP-1) receptor agonists exenatide and liraglutide and the dipeptidyl peptidase-4 (DPP-4) inhibitors sitagliptin and vildagliptin effectively lower HbA1c; exenatide and liraglutide reduce weight and blood pressure and improve lipid profiles. Sitagliptin and vildagliptin are weight neutral but also appear to improve lipid profiles. Integration of incretin therapies into the therapeutic armamentarium is a promising approach to improving outcomes in T2DM, and perhaps even in reducing complications of T2DM, such as co-morbid hypertension and dyslipidaemia. Additional long-term studies, including CVD end-point studies, will be necessary to determine the appropriate places for incretin-based therapies in treatment algorithms.


The cardiometabolic risk-factor cluster, including insulin resistance, hyperglycaemia, hypertension, low high-density lipoprotein cholesterol (HDL-C), and increased levels of very low-density lipoprotein triglycerides (VLDL-TG), are highly prevalent in type 2 diabetes mellitus (T2DM) and tremendously increase risk for cardiovascular disease (CVD) in diabetes. While overall cardiovascular death rates are declining, death rates because of CVD have not shown similar improvements in the obese and diabetic populations [1–3]. A recent prospective cohort study conducted among Canadian men found that the risk of cardiovascular death was almost four times greater among men with T2DM than among men with neither diabetes nor CVD [4]. Results from a population-based retrospective cohort study (n = 9.4 million ) reported that age-adjusted rates for acute myocardial infarction (MI) and all-cause mortality were approximately 2–4 times higher (depending upon age) in men and women with diabetes than in those without diabetes or recent acute MI [5]. With the dramatic worldwide rise in obesity prevalence, this constellation of cardiometabolic risk factors, variously termed metabolic syndrome, syndrome X, or ‘cardiometabolic risk’ , is increasingly recognized as a therapeutic target per se. Major clinical studies have convincingly shown the pivotal role of obesity, particularly abdominal obesity, in the development of prediabetes and T2DM, hypertension, dyslipidaemia and CVD [6–8].

Because the link between T2DM and obesity is so profound, the term ‘diabesity’ has additionally been applied to this high-risk cohort [9]. In the United States, 54.8% of persons with diabetes are obese, and 85.2% are overweight or obese [10]. For each unit of increase in the body mass index (BMI; reflecting ∼2.7 to 3.6 kg body weight), the risk for developing diabetes increases by ∼12% [11]. In particular, central adiposity is strongly correlated with insulin resistance and dyslipidaemia, even in persons whose BMI is not markedly elevated, as described early on in thin-appearing Japanese Americans [12–14]. A waist circumference of ≥ 94 cm in men and ≥ 80 cm in women of European descent (cutoffs vary somewhat for persons of Asian ethnicity) is considered indicative of visceral adiposity, and therefore of a heightened risk of metabolic syndrome and CVD [15].

Hypertension, Dyslipidaemia and Cardiovascular Disease

Hypertension is highly correlated with both obesity and diabetes. In a random population sample study, Staessen et al. found that each 10-kg increase in body weight was associated with a 2-mm Hg increase in systolic and diastolic blood pressure (SBP and DBP) in men and a 3-mm Hg increase in women [16–18]. The incidence of hypertension is five times higher among people who are obese than among individuals of normal weight [19]. An estimated 75% of adults with diabetes have blood pressure levels ≥ 130/80 mm Hg or use antihypertensive medication [20]. Persons with hypertension are 2.5 times more likely to develop T2DM, implying that obesity is a shared risk factor for both conditions [21]. Clinical data support the existence of an inverse relationship between insulin sensitivity and blood pressure, whereas a direct correlation has been posited between insulin resistance, and resultant hyperinsulinaemia, and elevated arterial pressure [22–26].

Abdominal obesity and T2DM are often accompanied by a characteristic pattern of plasma lipid abnormalities: low levels of HDL-C; numerically normal but structurally abnormal low-density lipoprotein cholesterol (LDL-C), favouring a predominance of small, dense, highly proatherogenic particles; and elevated levels of triglyceride-rich VLDL particles [27]. This so-called obesity dyslipidaemia profile, highly correlated with T2DM, increases the risk of atherosclerosis and subsequent CVD [27].

Hypertension, dyslipidaemia and atherosclerosis, among other pathophysiologic mechanisms, amplify the association between T2DM and elevated cardiovascular risk. The cardiovascular mortality rate for patients with T2DM is much higher than in those without diabetes (75 vs. 36%) [9,28]. National Health and Nutrition Examination Survey (NHANES) III data show that persons ≥ 50 years of age who have T2DM but no additional components of metabolic syndrome have a prevalence of coronary heart disease of 7.5%. This risk is similar to non-diabetic individuals who have had a prior cardiovascular event; Haffner et al. found the incidence of first MI in persons with diabetes to be identical to the incidence of recurrent MI in non-diabetic subjects who have a history of prior MI [29]. However, other cross-sectional and cohort studies have concluded that patients with established coronary heart disease remain at greater risk for death from all causes, cardiovascular death, and hospitalization for MI than do patients with T2DM but without coronary heart disease, and this is particularly true for newly diagnosed T2DM patients [30]. Among persons in the same age group who have metabolic syndrome in addition to diabetes, the prevalence of coronary heart disease is 19.2% [8]. Co-morbid hypertension and diabetes are associated with a 2.5- to 7.2-fold increase in mortality compared with healthy individuals, and diabetic nephropathy increases mortality 37-fold [31]. Finally, an intriguing recent prospective study suggests that MI (or the metabolic stress thereof) may be the presenting feature of diabetes; indeed diabetes, or impaired glucose tolerance, diagnosed at the time of MI is common, again illustrating this relationship [32].

T2DM, Obesity, Hypertension and Dyslipidaemia: A Complex Management Paradigm

Clearly, obesity, diabetes and additional risk factors for CVD, including hypertension and dyslipidaemia, are intricately and pathophysiologically interlinked, which has significant implications for clinical care. As Steno-2 has shown, treating the whole patient – all co-morbid pathophysiologies with combinations of behaviour modification and pharmacotherapy – can significantly reduce risk of cardiovascular and microvascular events compared with standard treatment protocols. Multimodal therapy should be considered the gold standard for risk reduction, but effective management of multiple pathophysiologies is complex [33]. Although processes of care and intermediate outcomes in T2DM have improved nationally in the past decade, data from NHANES suggest that 43% of individuals with T2DM still have poor glycaemic control [34]. Even in patients whose glycaemia is currently controlled, the progressive worsening of diabetes characterized by continual loss of β-cell function and, often, ongoing weight gain necessitates therapeutic dose escalation or outright regimen change [35]. The American Diabetes Association (ADA) recommends that patients with T2DM maintain a BMI ≤ 25 kg/m2, primarily through diet and exercise [36]. The benefits of intentional weight loss, especially loss of intra-abdominal fat, are well established. Weight reduction improves insulin sensitivity and helps restore β-cell function [37,38]. One of our best studies to date in diabetes risk, the Diabetes Prevention Program (DPP), clearly indicates that an intensive lifestyle intervention, aimed not just at controlling weight, but improving dietary intake and encouraging physical activity, significantly reduces the risk to progression to diabetes in those already at exceptionally high risk [39]. More recently, the Look AHEAD (Action for Health in Diabetes) study reported an initial weight loss of 8.6% of total body weight at 1 year in patients receiving intensive lifestyle intervention (i.e. comprehensive diet and physical activity counselling in 42 group and individual sessions) compared with a 0.7% initial weight loss at 1 year in the control intervention (three educational sessions) group. Greater weight loss was correlated with significantly greater improvement in HDL-C and reductions in glycaemia, blood pressure (SBP −6.8 mm Hg/DBP −3.0 mm Hg, from baseline), and triglycerides (p < 0.001) [40] and, perhaps, represents a gold standard for care of persons with diabetes and co-morbid metabolic syndrome. This level of intervention, however, is not feasible in many practice environments. It will be critical to identify determinants of success and potentially distil such an intervention to a more practical approach, or, alternatively, reform health systems such that this level of intervention becomes feasible.

Nonetheless, weight loss may reduce SBP by 5–20 mm Hg [17,18] and is associated with a less atherogenic lipid profile, including decreased triglyceride levels, increased HDL-C levels, and larger, less dense LDL-C particles [41]. Improving dietary intake, particularly reductions in dietary sodium, also shows beneficial blood pressure effects. Patients on the Dietary Approaches to Stop Hypertension (DASH) diet showed significant (p < 0.05) reductions in SBP for each level of sodium intake reduction (i.e. high, intermediate or low) compared with a control diet typical of caloric and food choice intake in the United States. There was also a significant difference in SBP between the high-sodium and low-sodium phases of the control diet, and between the high-sodium phase of the control diet and the low-sodium phase of the DASH diet (p < 0.001 for both comparisons) [42]. Finally, a 12-year observational study in patients with diabetes found that intentional weight loss was associated with a 28% reduction in CVD and diabetes-related mortality. The greatest reductions in mortality (approximately 33%) occurred in persons who lost between 20 and 29 pounds [43]. A definitive link between weight loss and reduced mortality in T2DM is, however, difficult to prove as studies are limited by the involuntary weight loss resulting from disease progression and complications of the heart and kidneys [41].

Benefits such as these, however, are difficult to achieve and even more difficult to sustain in a clinical trial setting, to say nothing of real-world clinical practice [44]. Meta-analyses of clinical trials implementing non-pharmacologic strategies for weight reduction, including dietary counselling, physical activity and reduced-calorie diets report modest weight reductions (1–6 kg), which tend to diminish over time as lost weight is regained [45–47]. A meta-analysis of the effects of long-term non-pharmacologic weight-loss interventions in patients with T2DM noted some benefit with very-low-calorie diets (3-kg weight loss) [45]. Pharmacologic interventions with weight-loss agents fare little better. Meta-analyses of clinical trials featuring pharmacologic interventions (sibutramine and orlistat) for weight loss report average reductions of 2.7–4.7 kg, but often the average weight reduction does not reach the 5–10% of total body weight-loss threshold. The meta-analysis conducted by Padwal et al. found that sibutramine was associated with small but significant elevations in both SBP and DBP; a summary of blood pressure data in orlistat trials (Rucker et al.) found a small but significant reduction in both SBP and DBP. Attrition rates tend to be very high in these trials, ∼30 to 45% [48,49].

Bariatric surgical procedures, including reversible laparoscopic gastric banding and non-reversible procedures such as jejunal and gastric bypass (e.g. Roux-en-Y gastric bypass) have resulted in substantial weight loss, metabolic improvement, and, often, blood pressure improvement (although the latter parameter is not always assessed in clinical trials) [50–53]. A systematic review and meta-analysis of 621 studies in which 135 200 obese patients with T2DM were treated with bariatric surgery, reported a mean weight loss of 38.5 kg (56% of excess body weight), and 78.1% of patients experienced complete resolution of diabetes. These improvements were maintained for at least 2 years [54]. Bariatric surgery is often indicated for the morbidly obese with multiple metabolic and cardiovascular risk factors. Complications, both peri- and postoperative, include esophagitis, esophageal dilation, pouch dilation and band leakage, which are associated with gastric banding, and anaemia (resulting from malabsorption of calories and various nutrients), nephrolithiasis, and gastric perforation, which are associated with gastric bypass surgery [55–58]. The need for a second procedure is common in gastric banding. In a study of Swedish adjustable gastric banding in patients ≥ 50 years of age, 34% of patients required re-operation [55].

Bariatric surgery aside, for many patients, facilitating permanent weight loss remains a major challenge in the clinical management of T2DM, as adherence to dietary recommendations and weight-loss regimens is limited [59,60]. Further, in the setting of the need to improve glycaemic control, therapies such as sulfonylureas (SUs), thiazolidinediones (TZDs), and some formulations of insulin, are associated with significant weight gain [61]– a powerful negative re-inforcer for those recently diagnosed. Newer insulin analogue options may attenuate this effect: use of the basal insulin analogue detemir results in significantly less weight gain and in some weight loss in the heaviest individuals [62]. Exercise and physical activity, which may help to maintain weight in healthy persons, are often ineffective in patients with T2DM, who are more likely to be obese and to have multiple chronic conditions that curtail adherence to such a regimen [63,64]. The effects on weight, lipids, and blood pressure associated with many commonly prescribed antidiabetic agents clearly complicate treatment by reinforcing clinical inertia, which has been recognized as a primary cause of poor glycaemic control [65,66]. Quite simply, physicians and patients resist intensifying therapy as needed [67,68], often because of fear of weight gain and its potential impact on the disease process [69].

Impact of Antidiabetic Agents on Weight and Cardiovascular Parameters

The current standard of care for T2DM mandates control of elevated glucose levels, as well as targeted treatment of hypertension and hyperlipidaemia, to reduce morbidity and mortality; weight reduction and weight control are also strongly recommended [36]. Unfortunately, some diabetes therapies result in significant weight gain, confounding glycaemic control as well as blood pressure and lipid control. Thus, devising optimal treatment regimens for patients with T2DM and co-morbidities requires an understanding of the risks and benefits of individual diabetes and cardiovascular drugs.

The ADA recommends a target HbA1c <7% [36], which is in accordance with the United Kingdom Prospective Diabetes Study (UKPDS) Group's findings that treatment to HbA1c 7.0 vs. 7.9% was associated with a 12% lower risk of any diabetes-related endpoint (e.g. sudden death, death from hyper- or hypoglycaemia, fatal or non-fatal MI, heart failure, stroke, etc), a 25% decrease in microvascular disease, and a 10% lower risk of diabetes-related mortality [70]. The ADVANCE (Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled Evaluation) trial showed significant (p = 0.01) reductions in the incidence of combined major macrovascular and microvascular events with intensive HbA1c control vs. standard control (18.1 vs. 20.0%, respectively) [71]. Because the ACCORD (Action to Control Cardiovascular Risk in Diabetes) trial recently showed an increased risk of cardiovascular mortality in high-risk patients treated to an HbA1c target of 6.4% (vs. 7.5% in the less intensively treated cohort) [72], as well as a higher rate of death from any cause in the intensively treated cohort, there has been some confusion about the appropriate glycaemic goal. Preliminary recommendations made in consideration of the ACCORD results continue to designate an HbA1c target of 7.0% as appropriate [73]. It has been speculated that perhaps more important than the numerical HbA1c target is the therapeutic approach employed to achieve such control [73,74]. Indeed, a recent meta-analysis of large clinical trials (UKPDS, PROactive, ADVANCE, VADT and ACCORD) confirms the supposition that intensive glycaemic control will reduce cardiovascular events, and that therapeutic choices may be relevant to outcomes [75].

Most oral diabetes medications produce absolute reductions in HbA1c levels of between 1 and 2% when given as monotherapy; the magnitude of effect is highly dependent upon HbA1c at initiation of therapy [65,76]. Further, most oral diabetes medications have minimal effects on blood pressure (<5-mm Hg reduction) and lipids, while effects on weight may vary considerably across oral classes (table 1) [61,65,70,76–125]. These medications will be individually reviewed next.

Table 1.  Cardiovascular, metabolic and glycaemic effects of antihyperglycaemic therapies in type 2 diabetes.
AgentBlood pressureFasting lipidsNon-traditional CV markersEffect on β-cell functionΔ Body weight (kg)Expected decrease in HbA1c(%)
SBP (mm Hg)DBP (mm Hg)Triglycerides (mg/dl)HDL-C (mg/dl)LDL-C (mg/dl)
  1. BNP, B-type natriuretic protein; CV, cardiovascular; DBP, diastolic blood pressure; HDL-C, high-density lipoprotein cholesterol; HOMA-B, homeostasis modal assessment-B; LDL-C, low-density lipoprotein cholesterol; PAI-1, plasminogen activator inhibitor-1; SBP, systolic blood pressure; TNF-α, tumour necrosis factor-α; TZDs, thiazolidinediones.

Metformin+7 to −5 [65,89]0 to −5 [65,89]−8% [89]+2.6% [89]−8% [89]Decreases TNF-α and PAI-1; improves endothelial function [96,99]No clear effect [113,120]Weight neutral or −1 to −2 [76,94]∼1.5 [61,89]
Sulfonylureas−5 to +7 [65,70]0 to −5 [65,70]Effects not knownEffects not knownEffects not knownEffects not knownNo positive effects [121]; may increase β-cell apoptosis [114]+1 to +5 [65,76]∼1.5 [61,78]
Non-sulfonylurea secretagogues (meglitinides)−4 [91]Effects not knownEffects not knownEffects not knownEffects not knownImproves endothelial function [97]Improves HOMA-B [122]+0.7 to +2.4 [81,98]0.5−1.5 [61]
Alpha-glucosidase inhibitors−1 mm Hg [87]−3 mm Hg [87]−10% [95]Effects not knownEffects not knownEffects not knownNo positive effects [123,124]Weight neutral [61]0.5−0.8 [61,79]
TZDs−5 [65,103]−4 [103]−14% [82]+11% [82]+1.2 [82]Improves endothelial dysfunction; decreases PAI-1 [77,90]Improves HOMA-B and PR:I levels [115]+1 to +5 [65]−0.5 to −1.4 [61]
InsulinVaries by formulation; no effect to +2 [70]No clinically relevant effects [70]No clinically relevant effects [93]No clinically relevant effects [93]No clinically relevant effects [93]Effects not knownImproves HOMA-B [125]+1 to +4 [70,92,103,104]1.5−3.5 [61]
Exenatide−1 [84]−3 [84]−16% [84]+12% [84]−1.4% [84]Effects not knownImproves HOMA-B and PR:I levels [101,111,112,126]−4.4 [84]−0.8 to −1.1 [84,101]
Liraglutide−8 [108]−3 [108]−22% [108]No consistent changes among Rx groups [108]No consistent changes among Rx groups [108]Decreases PAI-1 and BNP levels [88]; inhibits TNF-αin vitro[80]Improves HOMA-B and PR:I levels [110,117,118]−4.4 [107]−0.8 to −1.4 [108,109]
Sitagliptin−2 to −3 (in non-diabetic hypertensive patients) [100]−2 to −3 (in non-diabetic hypertensive patients) [100]−16.9% [86]+2.0% [86]−0.8% [86]Effects not knownImproves HOMA-B and PRI:I 83]Weight neutral [83]−0.6 to −0.8 [83]
VildagliptinNot significant [85]−2.0 [85]−5.8% [105]+7.7% [105]−0.4 [105]Effects not knownImproves HOMA-B and PRI:I [116]Weight neutral [83,105]−0.15 to −1.1 [102,105,106]


Antihyperglycaemic therapy is often initiated with metformin [65,76], in part because this agent is associated with modest weight loss or no weight gain in addition to reliably reducing HbA1c (table 1) [127–129]. Because metformin acts on hepatic glucose production rather than by stimulating β-cells [130,131], hyperinsulinaemia tends to decline, which, theoretically, may confer some cardiovascular advantage [76]. Blood pressure reductions associated with metformin, however, have been minimal (no effect to <5-mm Hg, systolic/diastolic) [65,89]. Other non-glycaemic benefits ascribed to metformin include improvement in lipids (perhaps as a result of liver insulin sensitization), reduced plasma levels of the procoagulant factor plasminogen activator inhibitor-1 (PAI-1) [76,89,132], as well as improvement in vascular reactivity, endothelial function and microvascular function [76,99,133]. Choy et al., analysing data from a number of large clinical trials, found that metformin moderately reduced the risk of new-onset heart failure compared with diet modification or rosiglitazone [134]. Cross-sectional analysis of data on Military Health System beneficiaries with T2DM enrolled in a health maintenance organization, found that average annual incidence of congestive heart failure and acute MI was lowest in patients taking metformin and highest in those on insulin [135]. Traditionally, however, metformin has not been recommended for patients with established heart failure or recent acute MI because of the heightened risk of lactic acidosis [129,136]. Yet a meta-analysis of eight controlled studies evaluating various antidiabetic treatments found that metformin was not associated with harm in patients with heart failure and diabetes, presumably as long as renal function was carefully considered [137]. The major adverse effects associated with metformin are gastrointestinal, which may limit compliance (although adherence rates to metformin tend to be good; a Scottish study of 1500 patients with diabetes on metformin monotherapy reported an adherence rate of 80% after 6 months of follow-up) [65,138]. With careful dose titration, most patients can tolerate this drug without problem. An additional potential concern is the impaired absorption of group B vitamins and folate, which may lead to increased plasma homocysteine levels and the progression of vascular disease [129,139,140].


SUs are a common alternative or add-on to metformin and act by stimulating pancreatic β-cell insulin secretion in a glucose-independent manner. SUs are associated with a weight gain of 1–5 kg [65,76], but do not appear to have a clear marked impact on SBP and DBP (changes ranging from <5-mm Hg decrease systolic/diastolic to a 7-mm Hg increase systolic) (table 1) [65,70]. Despite a clear association with hypoglycaemia, SUs have been a mainstay of antidiabetic therapy; ADA guidelines commonly recommend SUs as second-line therapy in patients failing on metformin monotherapy [36]. However, a meta-analysis of nine observational studies (numbers of participants ranging from 910 to 39,721; mean follow-up times ranging from 2.1 to 7.7 years) found that combination therapy with metformin and an SU significantly increased the relative risk of cardiovascular hospitalization or mortality (fatal and non-fatal events) compared with diet therapy, metformin monotherapy or SU monotherapy [141,142]. Thus, the weight loss (or weight neutrality) and improvement in some cardiovascular risk factors associated with metformin may be negated when combined with SU, which is associated with weight gain [142].Another possible explanation for this finding is the propensity of SU to cause hypoglycaemia and the possibility that metformin may impede recovery from hypoglycaemia as a result of its downregulation of hepatic glucose production. Hypoglycaemia may increase the risk of cardiovascular abnormalities, including ischaemia and cardiac arrhythmia [142]. The majority of these data described, however, are derived from observational studies making establishment of a causal linkage less definitive.

Non-sulfonylurea Secretagogues (Meglitinides)

Non-sulfonylurea secretagogues are distinguished from SUs by their short metabolic half-lives and consequent brief episodic stimulation of insulin secretion. Non-sulfonylurea secretagogues have not been as extensively studied as SUs; consequently, less is known about their beneficial or adverse cardiovascular impact [76]. Because they target postprandial glycaemia, non-sulfonylureas may produce less hypoglycaemia and weight gain than the SUs [76,98]. In a randomized trial of 109 patients with T2DM, a slight decrease in SBP(–4 mm Hg) was reported after 12 weeks with nateglinide in comparison to placebo [91].

Alpha-glucosidase Inhibitors

The alpha-glucosidase inhibitors, which mitigate postprandial glucose levels by partially blocking intestinal carbohydrate absorption, are not associated with weight gain or hypoglycaemia; gastrointestinal, side effects are common with this class, however [76]. Small reductions in triglycerides and postprandial insulin levels have been reported [95]. In the STOP-NIDDM (STOP Noninsulin Dependent Diabetes Mellitus) trial, the alpha-glucosidase inhibitor acarbose was associated with significantly reduced systolic (–0.97 mm Hg) and diastolic (–2.88 mm Hg) blood pressure in patients with impaired glucose tolerance. Additionally, acarbose therapy was associated with a 49% relative risk reduction in cardiovascular events and a 34% relative risk reduction in new cases of hypertension over a mean of 3.3 years [87]. These data may support recent interest in minimizing postprandial metabolic excursions as a mechanism to attenuate CVD risk. Recent evidence suggests that endothelial dysfunction may be a link between postprandial glucose excursions and cardiovascular events. However, a recent trial found that 20 weeks of acarbose treatment had no effect on fasting or postprandial endothelial function [143]. In addition, a meta-analysis of 41 trials (n = 8130) of alpha-glucosidase inhibitor therapy in T2DM reported no clinically relevant effects on lipids or body weight [79]. Long-term effects of alpha-glucosidase inhibitors on microvascular and macrovascular complications associated with diabetes have not been studied [129], however, the recently initiated Acarbose Cardiovascular Evaluation (ACE) trial is designed to determine whether acarbose therapy can reduce CV-related morbidity and mortality in patients with impaired glucose tolerance and co-morbid CVD or acute coronary syndrome and should, thus, shed some light upon this issue [144].


TZDs provide benefit in patients with glycaemic disorders by decreasing insulin resistance in peripheral tissues [130]. TZDs may also protect or restore β-cell function and thus, like the incretin therapies discussed later, may alter the natural history of the disease process (table 1) [145]. Common adverse effects associated with TZD therapy are weight gain, and increased risk of oedema and heart failure, particularly when TZDs are used in combination with insulin [76,134]. Pre-existent cardiac dysfunction appears to play an important role in subsequent heart failure episode on TZDs. Both pioglitazone and rosiglitazone have a black box warning regarding incidence and exacerbation of congestive heart failure in some patient groups [146], and ADA treatment guidelines state that rosiglitazone is not a recommended option when use of TZDs is considered [61]. Weight gain is also potentiated by concomitant insulin therapy and may also increase with TZD dosage [147]. In a population-based study of patients with T2DM enrolled in a health plan, weight gain in the TZD arm averaged 5 kg after 1 year of therapy (with a 0.9% reduction in HbA1c) compared with an average weight gain of 1.8 kg in the SU arm (with a 1.6% reduction in HbA1c) and an average weight loss of 2.4 kg in the metformin arm (with a 1.2% reduction in HbA1c) [148]. TZDs have also been associated with bone loss and increased susceptibility to fracture, particularly in women; reduced bone formation as a result of TZD-activation of peroxisome proliferator-activated receptor-gamma (PPAR-γ) has been proposed as the central mechanism of action [149,150]. A study of TZD use in older patients with T2DM reported progressively greater bone loss for each year of TZD use in women but not in men [151]. Despite their association with weight gain, TZDs (particularly pioglitazone and troglitazone, the latter since recalled because of liver toxicity) appear to mediate lipid improvements, particularly increased HDL-C levels and decreased triglycerides [82,152], although rosiglitazone is known to increase apolipoprotein B concentrations [153–156].

A prospective, randomized comparison of the metabolic effects of pioglitazone or rosiglitazone in patients with T2DM (n = 127) previously treated with troglitazone reported significant improvement in lipids in the pioglitazone group (p < 0.01), but no improvement with conversion to rosiglitazone [157]. A retrospective review across 605 primary care practices of the medical records of patients with diabetes treated for at least 1 year with either pioglitazone or rosiglitazone reported significantly (p < 0.001) greater improvement across a range of lipid parameters (triglycerides, total cholesterol, LDL-C, but not HDL-C) with pioglitazone vs. rosiglitazone [158].

A beneficial effect of TZD therapy on blood pressure in patients with T2DM has been reported. A meta-analysis of 37 clinical trials showed a mean reduction of 4.70 mm Hg in SBP and 3.79 mm Hg in DBP in subjects taking TZDs (table 1) [103]. This finding was confirmed by a more recent meta-analysis of comparative efficacy and safety of oral antidiabetics, which found a TZD-related blood pressure reduction benefit of ∼5 mm Hg [65]. TZD therapy has been associated with improvements in a number of vascular parameters implicated in the development of atherosclerosis [159], such as enhancement of fibrinolysis [160]. Blaschke et al. suggest a role for TZDs as vasoprotective agents in diabetes and in vitro data indicate these agents may play a moderate role in the inhibition of vascular inflammatory mediators [161–163] Improvements in endothelial function [164] and reductions in vascular inflammation and vascular smooth muscle cell proliferation have also been reported in animal models of TZD therapy [76,165]. Claims have been made for a pioglitazone-associated benefit on incidence of ischaemic cardiovascular events. In the PROactive trial, which was designed to determine whether pioglitazone had an effect on macrovascular morbidity and mortality, the composite of all-cause mortality, non-fatal MI (including silent MI), stroke, acute coronary syndrome, endovascular or surgical intervention in the coronary or leg arteries, and amputation above the ankle (the primary endpoint) was not significantly different in pioglitazone-treated patients vs. the placebo arm. The main secondary endpoint (composite of all-cause mortality, non-fatal MI, and stroke), however, showed a significant difference (p = 0.027) in favour of pioglitazone [146,166]. Two recent meta-analyses are also of relevance here: in the first, by Lincoff et al., which included the PROactive trial, pioglitazone was associated with a significantly reduced risk of death, MI, and stroke, although risk of serious heart failure but not mortality associated with serious heart failure was increased by pioglitazone; in the second, by Mannucci et al., which did not include PROactive data, pioglitazone was associated with reduced all-cause mortality, although the effects on risk of non-fatal coronary events were not significant [167,168].

Rosiglitazone is currently under some degree of clinical suspicion as a result of findings from a meta-analysis from 42 trials; rosiglitazone was associated with a significantly increased odds ratio for MI (1.43;p = 0.03) and for death from cardiovascular causes (1.64;p = 0.06) [169]. A similar meta-analysis found no such risk associated with pioglitazone therapy [167]. Results from the RECORD (Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes) trial found that rosiglitazone was similar to its comparators (metformin or SU) with respect to cardiovascular hospitalization rates and cardiovascular mortality, a finding also confirmed in a recent meta-analysis [170,171]. However, rosiglitazone was associated with a significantly greater risk for heart failure.


A gradual decline in β-cell function will eventually cause most patients with T2DM to initiate insulin therapy to achieve recommended HbA1c targets [172]. Administration of insulin to patients with T2DM is one of the most effective strategies to lower HbA1c and to reduce risk of microvascular and macrovascular complications [173]. In fact, early insulinization and then discontinuation after achieving glycaemic control, as reported in a recent study, appears to have significantly influenced insulin requirements at 12 months; ∼50% of those who received insulin vs. oral therapy remained off insulin at 12 months [125]. Nonetheless, weight gain remains a major concern with insulin therapy [70,174], and the mechanisms involved are complex and incompletely understood. Some weight gain is likely to be a by-product of reversal of the glucosuria and state of negative energy balance that is characteristic of uncontrolled diabetes; or it could be the result of metabolic rate reduction following reduction in hepatic glucose output [175]. More comprehensible to patient and physician alike is the profound hyperphagic response to hypoglycaemia and even ‘defensive snacking’, a behaviour presumably learned as a response to the imperfect pharmacodynamics of even the best insulin analogues [175,176]. Apprehension of possible weight gain may lead to ‘psychological insulin resistance’[176–178] clearly affecting both patients and healthcare professionals [178]. Further, the weight gain associated with improved glycaemic control upon intensification of therapy is a potent form of negative ‘biofeedback’. Newer, more predictable and tolerable insulin analogue formulations may reduce the incidence of defensive snacking and of psychological insulin resistance and are certainly associated with less weight gain than older formulations [173]. Clinical trial data and meta-analyses suggest that weight gain and hypoglycaemia associated with insulin may vary according to formulation, although the findings remain somewhat equivocal. Singh et al. found little additional benefit associated with long-acting insulin analogues vs. neutral protamine Hagedorn (NPH) insulin [179]. Other meta-analyses have reported a reduced risk of hypoglycaemia with long-acting insulin analogues [180,181]. Monami et al. observed that the newer basal insulin analogue, insulin detemir, although not insulin glargine, has a weight-sparing effect compared with NPH insulin; a finding also noted in the PREDICTIVE (Predictable Results and Experience in Diabetes through Intensification and Control to Target: An International Variability Evaluation) study and other trials [62,180,182].

Insulin and Blood Pressure.

In individuals with T2DM and in those without, insulin resistance and hyperinsulinaemia are associated with hypertension and increased cardiovascular risk [183–187]. A recent large meta-analysis of studies of circulating levels of three insulin markers – fasting insulin, non-fasting insulin, and pro-insulin – and risk of coronary heart disease found that the strongest coronary heart disease, risk correlation was with pro-insulin levels [188]. The relationship of insulin therapy and hypertension in patients with T2DM has not been intensively investigated; consequently, not much is known either about the putative impact of any insulin-associated weight gain on blood pressure elevations [189]. Randeree et al. retrospectively assessed 80 subjects with T2DM with secondary failure to diet and maximum oral therapy who were initiated on insulin therapy. Significant increases in systolic (+17 mm Hg) and diastolic (+8 mm Hg) blood pressure were observed, and weight gain was also significant (a mean of ∼6 kg). However, patients in this trial were also treated with maximum doses of oral antidiabetics, which may have contributed to blood pressure increases [182]. Conversely, the UKPDS did not observe a correlation between blood pressure elevation and insulin therapy [70]. Where weight gain is moderate, insulin therapy appears to have little impact on blood pressure. A study wherein 476 subjects with T2DM were titrated to twice-daily insulin detemir or NPH insulin for 24 weeks found that while weight increased slightly (a mean of 1.2 kg for insulin detemir and of 2.8 kg for NPH), no trend for blood pressure change was noted for either agent [190]. A comparison of insulin glargine vs. rosiglitazone in T2DM patients poorly controlled on metformin plus sulfonylurea reported no significant changes in SBP with either agent; a slight but significant (p < 0.01) decrease in diastolic pressure was noted with rosiglitazone [191]. Finally, a trial of short duration (2 weeks) to investigate the effect of NPH insulin on blood pressure and other parameters in obese, but non-diabetic hypertensive patients found that insulin infusion actually had a small blood-pressure-lowering effect [192].

Combination Antihyperglycaemic Therapy

Because T2DM is characterized by progressive loss of β-cell function [113], monotherapy with conventional oral antidiabetic agents often fails to maintain glycaemic control. In the UKPDS study, after 3 years, only 50% of patients had adequate glycaemic control with a single drug, and after 9 years, monotherapy provided adequate glycaemic control for only 25% of patients [193]. Stepwise protocols suggest beginning with a single agent (metformin); rapidly (within 1, 2 months) titrating to maximum tolerated dose if the patient is not controlled; then instituting combination therapy if titration to maximum dose of the single agent remains ineffective [61,69]. The hope is that combination therapy involving two or three different classes of drugs with distinct mechanisms of action will not only improve glycaemic control but will also result in lower overall drug dosing, thereby minimizing adverse events [76,194]. Treatment selection should be guided by the goal of lowering HbA1c to target levels and individualized for each patient according to treatment response and other clinical factors, including presence of co-morbidities [76]. Achieving these therapeutic goals may become daunting when agents are combined that, individually, are associated with potential adverse effects, such as weight gain or cardiovascular risk. The challenge of optimizing combination diabetes therapy is intensified in patients whose co-morbid conditions could be treated with agents unfavourably affecting glycaemic parameters.

Impact of Antihypertensive Agents on Glycaemic Control

According to the ADA, the target blood pressure for subjects with T2DM is <130/80 mm Hg; for those subjects with T2DM plus nephropathy it is 125/75 mm Hg [36,195]. Every 10-mm Hg reduction in SBP reduces the risk for any complication related to diabetes by 12% [196]. A meta-analysis that evaluated the efficacy of major antihypertensive drug classes [thiazide diuretics, beta-blockers, angiotensin-converting enzyme (ACE) inhibitors, angiotensin receptor blockers (ARBs), and calcium channel blockers (CCB)] found that standard doses of those agents applied as monotherapies reduced SBP and DBP by ∼8 to 10 mm Hg and ∼4 to 7 mm Hg, respectively (table 2) [197–203]. Most antihypertensive therapies appear to be weight neutral; some classes have been associated with greater benefit, or fewer off-target effects, in patients with co-morbid T2DM or at risk for T2DM.

Table 2.  Physiologic targets, metabolic and glycaemic effects, and blood pressure-lowering efficacy of major antihypertensive classes for patients with T2DM.
ClassPhysiologic targets [198]Metabolic effects [198]Glycaemic effectsAverage blood pressure reduction [197]
  1. LDL, low-density lipoprotein.

  2. *Average reduction obtained with standard doses over 24 h, as calculated in a meta-analysis of randomized, controlled studies [197].

  3. Includes patients treated with diuretics as monotherapy and in combination with other hypertensive agents.

  4. Adapted with permission from Wofford MR, Smith G, Minor DS. The treatment of hypertension in obese patients. Curr Hypertens Rep 2008; 10: 143–150.

Angiotensin-converting enzyme inhibitorsRenal-angiotensin-aldosterone systemDecreased insulin resistanceHbA1c–0.15% [200]–8.5/–4.7 mm Hg*
 Decreased aldosteroneDecreased plasma leptin  
 NatriuresisDecreased plasma insulin  
Angiotensin receptor blockersDecreased total peripheral resistanceDecreased insulin resistanceHbA1c–0.4% [203]–10.3/–5.7 mm Hg*
  Decreased plasma leptin  
  Decreased body mass index  
Beta-blockersSympathetic nervous systemDecreased insulin sensitivityHbA1c–0.11% [202]–9.2/–6.7 mm Hg*
 Decreased heart rate, cardiac outputIncreased triglycerides  
 Decreased renin release   
Calcium channel blockersDecreased total peripheral resistanceDecreased plasma leptinHbA1c–0.5% [200]–8.8/–5.9 mm Hg*
 NatriuresisDecreased plasma insulin  
  Increased sympathetic activity  
DiureticsDiuresisDecreased insulin sensitivityHbA1c +1.3% [201]–8.8/–4.4 mm Hg*
 NatriuresisIncreased LDL cholesterolFBG +0.22 mmol/l to +0.33 mmol/l[199] 
 Decreased cardiac outputTriglycerides  
 Decreased total peripheral resistance   

Renin-angiotensin-aldosterone System Blockade: Angiotensin-converting Enzyme Inhibitors and Angiotensin II Receptor Blockers

Antihypertensive agents that target the renin-angiotensin- aldosterone system (RAAS) are particularly relevant for this discussion and may be the preferred therapy for patients with co-morbid T2DM [204]. ACE inhibitors lower blood pressure by inhibiting the formation of angiotensin II (Ang II), the bioactive product of the RAAS and a powerful vasoconstrictor. ARBs competitively bind to the angiotensin type 1 receptor, preventing its activation by Ang II [205,206]. A number of large-scale clinical trials – Heart Outcomes Prevention Evaluation (HOPE), the Captopril Prevention Project (CAPPP), Candesartan in Patients with Heart Failure Overall Programme (CHARM), and Valsartan Antihypertensive Long-Term Use Evaluation (VALUE) – evaluating ACE inhibitors and ARBs in patients with hypertension and/or CVD have reported significant reductions in new-onset diabetes vs. both placebo and comparator therapies (e.g. beta-blockers or thiazide diuretics) [207,208]. More recently, the Diabetes Reduction Assessment with Ramipril and Rosiglitazone Medication (DREAM) study, conducted in patients with CVD or hypertension, found that treatment with the ACE inhibitor ramipril did not significantly reduce the rate of new-onset diabetes vs. placebo. However, ACE-inhibitor therapy significantly improved the rate of regression to normoglycaemia [209]. A number of possible mechanisms have been proposed to explain the beneficial effect of RAAS blockade on glycaemic parameters: for example, enhancement of insulin sensitivity in skeletal muscle and adipocytes, enhancement of insulin signalling, and reduction of hepatic glucose production [208,210–212]. Accordingly, some promise exists that adoption of RAAS blockade in high-risk patients may abrogate the transition to type 2 diabetes, ultimately influencing the choice of diabetes control agent that may, in turn, aggravate or further alleviate CVD risk.


The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood Pressure (JNC 7) has recommended thiazide diuretics as a preferred initial agent for the treatment of hypertension [213], but in patients at risk for diabetes or with co-morbid T2DM, the utility of thiazide diuretics is less clear-cut. Citing the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) data, JNC 7 noted that therapy with a diuretic in patients with concomitant diabetes appeared to be more cardioprotective than ACE-inhibitor therapy. Development of glucose intolerance and/or diabetes, however, was higher in patients treated with chlorthalidione therapy than was observed with CCB or ACE-inhibitor therapy. The observation seemed to be that thiazide diuretics may worsen metabolic disorders but may have other, compensatory benefits [213,214]. Several large clinical trials evaluating thiazide diuretic therapy in hypertension have reported an association between those agents and the development of glucose intolerance and diabetes [214–218]. Finally, it is worthy of mention that thiazide use in the obese and diabetic patient population has the potential to worsen and/or cause hypertriglyceridaemia in a subset of sensitive patients, thereby worsening their dyslipidaemia [219,220].

Calcium Channel Blockers

CCBs have shown beneficial effects on high blood pressure as well as reductions in the complications of hypertension [213]. CCBs have also been associated with impaired insulin release, but this effect appears to be balanced by an increase in peripheral glucose uptake [210,221,222]. In the large, long-term Nurses' Health Study (NHS) I and II and the Health Professionals Follow-up Study (HPFS), CCBs were not associated with an increased risk of diabetes [217]. A review of 22 clinical trials (∼143 000 patients without diagnosed diabetes at randomization), 17 of which enrolled hypertensive patients, 3 high-risk patients, and 1 patient with heart failure found that the association of new-onset diabetes and an antihypertensive agent was lowest for ARBs and ACE inhibitors, followed by CCBs, beta-blockers, and diuretics in rank order [223]. A review of six trials (n = 99 000 patients ) in which CCBs were compared with beta-blockers or diuretics reported a similar finding [224].


Use of beta-blocker therapy has been problematic for treatment of hypertension in T2DM, because some of the older non-selective or β-1 selective agents may impair glucose metabolism, leading to hyperglycaemia [225–227]. Beta-blockers have been associated with weight gain [228,229] and insulin resistance, which may have a deleterious effect on glycaemic control in patients with T2DM [210]. Several large clinical studies have reported an increased risk of developing diabetes in patients receiving a beta-blocker compared with an ACE inhibitor, a CCB, or an ARB [207,217,230]. In contrast, newer beta-blockers that cause vasodilatation, such as carvedilol and pindolol, may not have harmful effects on glucose metabolism and, in fact, may improve insulin sensitivity [210]. Despite concerns regarding the potential risks of beta-blockers in patients with diabetes, these drugs have a clear mortality benefit in patients with stages B and C heart failure [134]. However, the recently concluded Anglo-Scandinavian Cardiac Outcomes Trial, which compared an amlodipine-based antihypertensive regimen vs. an atenolol-based antihypertensive regimen in patients with co-morbid hypertension and T2DM, was terminated early as a result of significant mortality- and stroke-associated benefits in the calcium channel blocker cohort [231].

Pharmacologic Regimens: Progressive Complexity, Safety and Compliance

Despite an increasing appreciation for the cardiovascular risks associated with hypertension co-morbid with T2DM, results of the NHANES and Behavioral Risk Factors Surveillance System surveys show that one in three persons with diabetes has poor blood pressure control [232]. Blood pressure may be particularly difficult to control in this population: the target is lower because the risk for cardiovascular complications is higher; in addition, a number of the pathophysiologic mechanisms implicated in the etiology of hypertension are potentiated in patients with co-morbid diabetes (e.g. upregulation of the RAAS) [233].

Treatment of co-morbid diabetes and hypertension is more difficult than treating either condition in isolation; however, the absolute risk reduction resulting from blood pressure lowering is consistently greater in hypertensive patients who have T2DM vs. those who do not [204]. Blood pressure reduction by any means is therefore crucial, although treatment selection should be guided by a recognition of the potential advantages and risks associated with available antihypertensive classes [204].

Treatment guidelines for T2DM continue to emphasize the benefits of lipid-lowering therapies in reducing cardiovascular risk. A meta-analysis has shown that the use of cholesterol-lowering drugs for either primary or secondary cardiovascular prevention in T2DM is associated with a 21% reduction in major coronary events (coronary heart disease death, non-fatal MI, or myocardial revascularization procedures) [234]. The American Association of Clinical Endocrinologists (AACE) has recommended aggressive management of dyslipidaemia in patients with diabetes, setting a goal of <100 mg/dl for LDL-C (<70 mg/dl for patients with co-morbid coronary artery disease), >40 mg/dl for HDL-C in men, >50 mg/dl for HDL-C in women, and <150 mg/dl for triglycerides. The guidelines emphasize the importance of lifestyle modification, use of statins as initial pharmacologic therapy, and the introduction of additional agents (e.g. niacin, fibrates, ezetimibe) in patients who fail statin monotherapy [235]. Adherence to statin therapy is nonetheless poor in patients with diabetes; a recent study indicated that only 44% of patients being treated for dyslipidaemia in a diabetes management programme achieved LDL-C <100 mg/dl [236]. Treatment burden may be a component of poor patient adherence to lipid-lowering regimens, particularly if patients have multiple co-morbidities and are on multiple-medication regimens for each co-morbid condition. Therapies that combine benefits – LDL-C reduction and improved glycaemic control, such as is claimed for colesevelam – would have utility in those patients [237].

Implementation and intensification of antihypertensive and dyslipidaemia therapy in patients with T2DM are similar to protocols governing antihyperglycaemic therapy. Each follows a treat-to-target paradigm in which additional drugs are introduced, in the case of treatment failure, to achieve prespecified goals. The need to manage multiple metabolic and cardiovascular risk factors consequently presents a situation of formidable complexity for the physician and patient. Strategies to simplify overall pharmacologic treatment in T2DM and co-morbid hypertension (and co-morbid dyslipidaemia) are desirable, such as the use of single-pill formulations containing multiple drugs or of single agents with multiple therapeutic effects.

Incretin-based Therapies: Effects on HbA1c, Weight and BP

A recently developed therapeutic approach to T2DM has followed from an improved understanding of the essential role played by the incretin hormones in glucoregulation [238]. Glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP), secreted in the gut in response to a meal, play pivotal roles in the stimulation of glucose-dependent insulin secretion. The incretin hormones also promote β-cell proliferation and inhibit apoptosis in animal studies, potentially facilitating augmentation of β-cell mass. Both peptides have very short half-lives, as a result of rapid degradation by the serine protease dipeptidyl peptidase-4 (DPP-4) [238]. Nonetheless, the incretin effect (i.e. augmentation of insulin secretion in the presence of elevated glucose levels) may be responsible for as much as two-thirds of the insulin released following a meal [239]. In T2DM, the incretin effect is significantly blunted [240,241] suggesting that deficient incretin physiology may play a contributory role in the pathogenesis and progression of T2DM.

In addition to its potent effects on insulin secretion, GLP-1 mediates a number of other physiologic actions that may offer therapeutic advantages in T2DM, including reduction of endogenous glucose production [242] via inhibition of glucagon secretion in a glucose-dependent manner [243,244]; improvement in β-cell function [243,245,246]; and enhanced insulin-independent glucose disposal in the peripheral tissues [247,248]. GLP-1 has been shown to promote satiety, delay gastric emptying, and significantly reduce caloric intake, implicating a role for this hormonal system, at least in therapeutic doses, in weight loss [249]. The effects of GLP-1 on appetite appear to be complex, potentially involving a central GLP-1 pathway in hypothalamic sites involved in the regulation of energy homeostasis [250], but certainly peripheral mechanisms, including inhibition of gastric emptying, also are involved [251–253]. Beneficial effects on the vascular system have also been posited as preliminary data suggest GLP-1 may improve endothelial function in patients with stable coronary artery disease and T2DM [254].

Incretin Therapies

Two types of incretin-based therapy, GLP-1 receptor agonists and DPP-4 inhibitors, have been developed for the treatment of T2DM. DPP-4 is a widely expressed cell surface enzyme that rapidly degrades endogenous GLP-1, rendering it inactive. Sustained inhibition of DPP-4 elevates levels of intact, bioactive GLP-1, resulting in insulin secretion (in a glucose-dependent manner) and plasma glucose reduction [255]. Data suggest an equal or potentially greater efficacy of incretin-based therapies, particularly GLP-1 receptor agonists, to lower HbA1c compared with other antidiabetes therapies – importantly, with a low risk of hypoglycaemia. DPP-4 inhibitors have been associated with a less robust glycaemic effect. (table 1). Other benefits include weight reduction (GLP-1 receptor agonists) or weight neutrality (DPP-4 inhibitors) (table 1). Preliminary data suggest that incretin agents may decrease blood pressure and improve triglyceride levels [108,256].

GLP-1 Receptor Agonists


Exenatide, an exendin-based GLP-1 receptor agonist currently available in the United States, is a synthetic 39-amino acid peptide [257] with 53% homology to human GLP-1 [258]. It is derived from the protein exendin-4, which is found in the saliva of the Gila monster lizard. Exenatide is administered subcutaneously twice daily within 60 min of breakfast and dinner [257]. In a 30-week trial, patients receiving exenatide 5 or 10 µg twice a day while on background therapy of SU, metformin, or SU + metformin experienced a mean decrease of HbA1c from baseline of 0.9%; a 52-week open-label extension of this trial reported a mean HbA1c reduction of 1.1% [84] A 30-week, four-arm trial included an acclimation phase of 4 weeks with patients on 5 µg of exenatide twice a day before the dose was increased to 10 µg of exenatide twice a day in one arm, or maintained at 5 µg. Two placebo arms received dummy injections. All patients were maintained on background metformin. At 30 weeks, HbA1c changes from baseline were −0.78% (10 µg), −0.40% (5 µg) vs. +0.08% for placebo [112]. A similarly designed trial with a background therapy of SU, metformin, or the combination reported HbA1c changes from baseline of −0.8 and −0.6% for 10 and 5 µg, respectively, vs. +0.2% for placebo [259].

Exenatide 5 or 10 µg twice daily was associated with a dose-dependent mean weight loss of up to 2.8 kg at 30 weeks [112], which increased to 4.4 kg at week 82 in an open-label trial extension [84]. Weight reductions were greatest in persons with the highest baseline BMI and in those taking metformin, with lesser reductions occurring in those taking a SU or combination metformin-SU [84].

Exenatide therapy appears to induce beneficial effects on blood pressure, although the clinical trial data remain somewhat ambiguous. In the 30-week trials, changes in blood pressure were not reported [112]. In the 82-week uncontrolled extension trial, however, SBP and DBP decreased by a mean of 1.3 mm Hg and 2.7 mm Hg, respectively [84]. In patients with the greatest weight reductions, SBP decreased by 3.9 mm Hg and diastolic blood pressure by 4.4 mm Hg. A pooled analysis of three trials of adjunctive treatment with exenatide 5 or 10 µg bid showed a mean decrease in SBP and DBP of 2.6 and 1.9 mm Hg, respectively, at week 104, suggesting sustained improvement in blood pressure [260]. Changes in lipid parameters at 82 weeks included decreased triglyceride (–38.6 mg/dl), LDL-C (-1.6 mg/dl), and apolipoprotein B (–1.1 mg/dl) levels and an increase in HDL-C (+4.6 mg/dl) [84]. The greatest improvement in cardiovascular risk factors was observed in patients who had the greatest weight loss, suggesting that the weight effects mediate much of the improvements [84].


Liraglutide is a once-daily, subcutaneously administered human GLP-1 analogue with 97% primary amino acid homology to native GLP-1 [261], recently approved in Europe and under evaluation by the FDA in the United States. In a 14-week study, liraglutide monotherapy in doses of either 1.25 or 1.9 mg/day decreased HbA1c by 1.40 to 1.45% from baseline [108]. A phase 3 monotherapy study of liraglutide 1.2 or 1.8 mg/day decreased HbA1c by 0.84 to 1.14%, respectively, at week 52. In a subgroup of drug-naïve patients with T2DM, HbA1c reductions of up to 1.6% were achieved with liraglutide monotherapy [109]. In 26-week, phase 3, combination therapy trial, HbA1c decreased by 1.00 to 1.33% with liraglutide 1.8 mg/day plus glimepiride and/or metformin [110,117,118,262]. Changes in HbA1c for comparator regimens were +0.09% for metformin alone, +0.23% for glimepiride alone, −0.24 to −0.98% for metformin plus glimepiride, −0.79% for exenatide in combination with metformin and/or glimepiride, and −1.09% for insulin glargine in combination with metformin plus glimepiride [110,117,118,262]. In combination with metformin and rosiglitazone, liraglutide 1.8 mg/day reduced HbA1c by 1.5%, compared with a reduction of 0.5% with metformin and rosiglitazone [119]. In a 26-week trial vs. exenatide (10 µg bid) in patients uncontrolled on metformin and/or SU, liraglutide (1.8 mg qd) produced a mean reduction from baseline HbA1c of 1.12 vs. 0.79% for exenatide [262].

Like exenatide, liraglutide has been associated with dose-dependent weight loss (–1.0 to −3.2 kg) [109,117, 119,262], including when used in combination with SU, which itself is associated with weight gain [110,118]. However, one combination study with SU found that patients receiving low doses (0.6 and 1.2 mg) of liraglutide experienced slight weight gain (<1 kg) [118]. Patients whose BMI exceeded 35 kg/m2 derived the greatest absolute benefit (weight loss up to 4.4 kg) [107]. When added to metformin therapy, liraglutide significantly reduced weight by up to 2.8 kg (p < 0.0001) and subcutaneous fat by up to 8.5% (p < 0.05) compared with a combination of glimepiride and metformin; in addition, a trend towards decreased visceral fat was noted in patients receiving liraglutide plus metformin compared with those receiving glimepiride and metformin [117,263]. In a 52-week monotherapy trial, liraglutide was associated with a 2.05-kg mean weight reduction from baseline (1.2 mg once-daily) and a 2.45-kg mean reduction from baseline (1.8 mg once daily) vs. a 1.12-kg mean gain from baseline with glimepiride [109].

Liraglutide was compared with exenatide in a 26-week trial in which both agents were added to metformin and/or an SU. The liraglutide group experienced greater reductions in HbA1c (p < 0.0001); weight-loss differences did not achieve statistical significance but were numerically greater for liraglutide vs. exenatide (−3.2 kg, liraglutide; −2.87 kg, exenatide; p = ns ). Improvements in triglyceride levels were significantly greater in liraglutide-treated patients (p = 0.0485) and minor hypoglycaemia was significantly more common in the exenatide group (p = 0.0131). The incidence of nausea was similar, but less persistent in the liraglutide group [262].

In the 14-week study with liraglutide monotherapy, dose-dependent cardiovascular benefits included a mean reduction in SBP of 7.9 mm Hg and in diastolic blood pressure of 3 mm Hg [108]. In 26-week studies of liraglutide combined with metformin plus an SU or metformin plus a TZD, liraglutide therapy was associated with SBP reductions from 4.5 (vs. insulin with metformin plus SU) to 6.7 mm Hg (vs. placebo with metformin plus rosiglitazone) [110,119,256]. SBP reductions occurred within 2 weeks, antedating changes in body weight and suggestive of a blood pressure reduction independent of weight loss. The mechanism underlying the apparent larger blood pressure effects of liraglutide vs. exenatide is not yet understood.

Although published data to date are scarce on the effects of liraglutide on lipid parameters, Vilsboll et al., in the 14-week trial noted above, observed a significant reduction in triglycerides vs. placebo (−22%;p = 0.0110) when liraglutide was administered as monotherapy to patients with T2DM [108].

DPP-4 Inhibitors

Sitagliptin is the only DPP-4 inhibitor currently available in the United States. It is indicated for the treatment of T2DM together with a regimen of diet and exercise [264]. In a placebo-controlled study, sitagliptin monotherapy, in doses of 100 and 200 mg/day (100 mg is the recommended dose [264]), showed placebo-subtracted reductions in HbA1c of 0.79 and 0.94% (−0.61 to −0.76 least square mean change from baseline) [83]. The largest decrease occurred in patients whose baseline HbA1c exceeded 9% [83]. When added to metformin (≥ 1500 mg/day), sitagliptin (100 mg/day) provided significant reductions in HbA1c (mean −0.65%, range −0.77 to −0.57) vs. placebo (mean −0.02, range −0.15 to +0.10) in a 24-week trial [86]. Generally small but significant decreases in total cholesterol (–2.8%), triglycerides (–16.9%), non-HDL-C (–4.8%), and the triglyceride-to-HDL ratio (–19.4%) were associated with sitagliptin therapy; in addition, a significant increase in HDL-C (2.0%) was reported. Initial combination therapy of sitagliptin (100 mg/day) and metformin (1000 to 2000 mg/day) provided placebo-subtracted changes in HbA1c from −1.57 to −2.07% at 24 weeks, with no increased risk of hypoglycaemia; it is important to note that in this trial sitagliptin was not added on to metformin failures [86,265]. Significant reductions in body weight (–0.6 kg to −1.3 kg from baseline) were noted with sitagliptin–metformin combination therapy but not with sitagliptin monotherapy [265]. Sitagliptin appears to produce small but statistically significant reductions (−2 to −3 mm Hg) in SBP in hypertensive patients [100].

Vildagliptin is a DPP-4 inhibitor currently available in Europe but not in the United States. Vildagliptin monotherapy, in doses of 50 mg qd, 50 mg bid, or 100 mg qd (50 mg bid, once in morning and once in evening is the recommended dose), reduced HbA1c by 0.5 to 1.1% during 24-week trials in drug-naïve patients with T2DM [102,105,106]. A trial of vildagliptin in combination with pioglitazone produced dose-dependent decreases in fasting plasma glucose (FPG) of −2.4 to −2.8 mmol/l [105]. In patients inadequately controlled with metformin alone, the addition of vildagliptin 50 mg twice daily resulted in a significant adjusted mean reduction (–1.1%) in HbA1c vs. placebo over 24 weeks [85].

Vildagliptin carries little or no risk of hypoglycaemia and appears to be weight neutral [102]. Effects on lipid parameters include decreased triglycerides (–5.8%) and increased HDL-C (+7.7%) [105]. In a study that reported blood pressure outcomes, vildagliptin 50 mg twice daily was associated with reductions in SBP and DBP; the reduction in diastolic pressure from baseline (adjusted mean change: −2.0 mm Hg) achieved statistical significance compared with placebo (p = 0.034) [85]. Peripheral oedema has been reported in approximately 5% of patients [105,266].

Incretin Therapies: Safety Considerations

Although DPP-4 inhibitors have been reported in clinical trials to be generally safe and well tolerated, increased risk of upper respiratory and urinary tract infections have been observed [83,267]. Results of a meta-analysis suggest an increased risk of urinary tract infection with DPP-4 inhibitors (risk ratio of 1.5 vs. comparators) [268]. As DPP-4 is a membrane-spanning protein expressed on many cell types, including lymphocytes, long-term safety data on the potential effects of DPP-4 inhibitors on immune function are desirable [269].

Adverse events data associated with GLP-1 receptor agonist therapy indicate an excellent safety profile with a low risk of hypoglycaemia or oedema. Gastrointestinal symptoms, including nausea, vomiting and diarrhoea, are common, transient and dose dependent with initiation of GLP-1 receptor agonist therapy. In phase 3 studies of exenatide 5–10 µg twice daily added to metformin and/or a sulfonylurea, incidence rates were in the range of 36–51% for nausea, 10–15% for vomiting, and 9–17% for diarrhoea. Nausea occurred most often during the first 8 weeks of treatment with exenatide and decreased in frequency thereafter; in each study, incidence was around 30% during weeks 4–8 of treatment but had declined by approximately half in weeks 28–30. The higher dose (10 µg) was associated with more frequent nausea [111,112,259]. Phase 3 studies of liraglutide 0.6–1.8 mg once daily, used alone or in combination with metformin and/or sulfonylurea therapy, reported a nausea incidence of 10–29%. Nausea was most common in the early treatment phase, diminishing to <10% by week 4. Liraglutide was associated with vomiting in 4–12% of patients and diarrhoea in 8–19% [109,110,117–119,262].

Cases of acute pancreatitis have been reported in patients with T2DM who were treated with exenatide, and the product label cautions prescribers to be vigilant for signs and symptoms of acute pancreatitis [257]. A recently published report of a claims-based safety surveillance system designed to assess risk of acute pancreatitis with either exenatide or sitagliptin found no difference in risk for acute pancreatitis between initiators of exentatide or sitagliptin compared with initiators of metformin or glyburide [270]. Patients with T2DM have a three times higher risk of developing pancreatitis than the general population [271]. Currently available clinical trial data indicate that the incidence rate among subjects using liraglutide or a comparator is in line with what one would expect in a type 2 diabetes population [109,110,117–119]. The present number of cases remains very small; to date there have not been sufficient numbers of cases to be able to determine if there is any association between the development of acute pancreatitis and treatment with liraglutide [272,273].

In preclinical rodent studies, liraglutide induced c-cell hyperplasia, c-cell adenoma, and, at the highest doses, c-cell carcinoma. These lesions were preceded by both acute and chronic increases in calcitonin, a biomarker of c-cell activation and c-cell mass. Similar findings did not occur in non-human primates at an exposure of 64-fold that of the human dose of 1.8 mg [272,273].

In vitro experiments support a mechanism of action for c-cell neoplasia through the GLP-1 receptor; and in rodents this receptor is functional. However, in humans and non-human primates, the c-cell GLP-1 receptor does not appear to be functional. Results of an extensive calcitonin screening programme, carried out in over 5000 individuals included in the liraglutide development programme, found no data to suggest that liraglutide activates the human c-cell. The cumulative data thus strongly suggest that while rodent c-cells are sensitive to activation by GLP-1 agonists, human and non-human primate c-cells are not (Data on file).

Use of Incretin Therapies in Patients with T2DM

Incretin therapies have the potential to address the need for effective antihyperglycaemic therapies with neutral or possibly even beneficial impact on cardiovascular risk parameters associated with T2DM [274]. Initial clinical trial data indicate that use of incretin agents in T2DM improves glycaemic control and may result in either weight neutrality or weight loss, a decrease in blood pressure, and improved lipid parameters [84–86,102,105,108,110,119,256,260,263]. In the absence of head-to-head clinical trial data, individualized patient goals, together with the degree of cardiovascular risk co-morbidity and tolerance of potential side effects, will likely determine appropriate selection of GLP-1 receptor agonists or DPP-4 inhibitors.

Incretin therapies may help overcome barriers impeding treatment to target [274] and assist providers and patients in achieving goals [69]. Pleiotropic effects of these agents could even assist patients with co-morbid hypertension and dyslipidaemia attain treatment goals. The effects on β-cell survival observed in animal models suggest that incretin therapy may be disease-altering, in terms of reversing the decline in β-cell function characteristic of T2DM. The use of incretin-based therapy has yet to be fully integrated into clinical guidelines; however, AACE guidelines suggest sitagliptin may be one therapeutic option for initiating antihyperglycaemic monotherapy [274] and may be given in combination with metformin or a TZD [235]. The most recent ADA consensus algorithm for the treatment of T2DM includes GLP-1-agonist based treatments in Tier 2, noting that these agents may have particular utility in situations where hypoglycaemia may be of great concern [61]. As more clinical experience is gained with GLP-1 receptor agonists and DPP-4 inhibitors, appropriate use of these agents as therapeutic options will become more fully defined.


The close association between overweight and obesity, T2DM, hypertension, and an atherogenic lipid profile lead to a significantly increased risk of cardiovascular morbidity and mortality. Despite the evidence that weight loss provides significant benefits, persons with T2DM are often unable to reach or maintain a target BMI of <25 kg/m2, an HbA1c of <7%, and blood pressure and lipid targets through lifestyle changes alone. In fact, the majority of patients gain rather than lose weight, especially during intensification of glycaemic control. Selection of appropriate agents to treat co-morbid T2DM, hypertension and dyslipidaemia therefore becomes challenging. Care must be taken to consider potential adverse impacts upon weight and other cardiovascular risk parameters. Similar concerns should govern selection of antihypertensives, some of which have been recognized to effect weight and glucose regulation adversely. Metformin is a common first-line antidiabetic therapy because it is weight neutral or may promote modest weight loss. As β-cell function declines, additional therapy is usually required. SUs, TZDs and insulin are effective choices for improving glucose control, but clinical trial data and meta-analyses suggest that these agents may be associated with weight gain. While cardiovascular end-point trial data are not available for many agents, weight gain is potentially associated with increased cardiovascular risk, and for some patients and clinicians may be an additional consideration in therapeutic decision-making. Newer antihyperglycaemic options include incretin-based therapies. Two agents, exenatide and sitagliptin, are available in the United States; liraglutide is available in Europe, and it and other long-acting GLP-1 receptor agonists as well as newer DPP-4 inhibitors are under regulatory review in the United States or are in development. The GLP-1 receptor agonists exenatide and liraglutide have shown clear improvements in glycaemic control and weight loss in T2DM, with added benefits on blood pressure and lipids, although the ultimate impact of these benefits on cardiovascular endpoints requires confirmation in clinical trials. Integration of incretin therapy into a therapeutic regimen designed to control glycaemia with beneficial effects on blood pressure, dyslipidaemia and obesity in patients with T2DM appears to be a promising new approach to improving patient outcomes.


I would like to thank Robert McCarthy, PhD, of AdelphiEden Health Communications for providing medical writing and editorial services supported by Novo Nordisk. I would also like to acknowledge the significant contributions made by the peer reviewers of this manuscript.

Conflict of Interest

Dr Niswender has received a consulting fee, investigator-initiated research funding, and is a site Principal Investigator for a Novo Nordisk weight-loss trial.