Effect of Antihypertensive Agents on the Development of Type 2 Diabetes Mellitus

Both diuretics and β-blockers are reported to accelerate the appearance of new-onset type 2 diabetes mellitus in patients with hypertension.^{,  ,  ,  ,  ,  ,  ,  ,  ,  ,  ,}  The greater incidence of diabetes in reports comparing diuretics and β-blockers with angiotensin-converting enzyme (ACE) inhibitors may reflect in part the beneficial effects of ACE inhibitors on glucose metabolism.^{,  ,  ,  ,}  Compared with calcium channel blockers (CCBs), which are generally considered metabolically neutral, diuretics and β-blockers are associated with new-onset diabetes mellitus.^{,  ,}  In the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attacks Trial (ALLHAT), the diuretic chlorthalidone was 43% more likely to be associated with diabetes than the ACE inhibitor lisinopril and 18% more likely than the CCB amlodipine. However, these calculations only account for those individuals whose glucose values increased to greater than 126 mg/dL at 4 years not those who started taking the medication for diabetes during follow-up. Nevertheless, diuretics have consistently demonstrated their ability toreduce the cardiovascular mortality in patients with diabetes. ALLHAT concluded that thiazide diuretics comparably reduced all-cause mortality, such as stroke, coronary artery disease, and heart failure. Therefore, diuretics continue to play an important role in the management of hypertension in patients with diabetes, especially as an adjunct to ACE inhibitors and angiotensin receptor blockers (ARBs).

Evidence is accumulating that overcoming insulin resistance with antihypertensive agents that interrupt the reninangiotensin system may prevent or delay the emergence of type 2 diabetes mellitus in patients with essential hypertension^,  (Table 1). The Captopril Prevention Project was the first controlled clinical trial to show that an ACE inhibitor reduces the development of diabetes in patients with hypertension. This trial was designed to compare the effect of ACE inhibition with conventional antihypertensive therapy (β-blocker, diuretic, or both) on cardiovascular morbidity and mortality. The number of patients with newly diagnosed diabetes was 14% lower in the captopril group than in the group receiving conventional therapy. These data were confirmed in the Heart Outcomes Prevention Evaluation trial in which a fixed dose of ramipril or placebo was added to whatever other therapy was prescribed for patients at high risk of cardiovascular events (including β-blockers, CCBs, and diuretics). During the 4.5-year trial, 35% fewer patients in the ramipril group than in the placebo (control) group developed diabetes (3.6% of the 4645 patients in the ramipril group vs 5.4% of the 4652 patients in the placebo group). The same percentage of patients (39%) in the ramipril and placebo groups were receiving a β-blocker at baseline. In addition, a retrospective analysis of data from one site that followed up 292 patients, the Studies of Left Ventricular Dysfunction, showed a relative risk reduction of 74% for developing diabetes during the 2.9-year study period in patients receiving enalapril compared with those receiving placebo (incidence of diabetes, 5.9% with enalapril vs 22.4% with placebo). It has been suggested that, by blocking both kininase II and ACE, ACE inhibitors may increase not only nitric oxide production but also bradykinin, thus improving blood flow to skeletal muscle, properties that should improve insulin-mediated glucose uptake.^, 

Several clinical trials demonstrate that ARBs also have beneficial effects on glucose metabolism that likely are independent of bradykinin-mediated mechanisms.^{,  ,  ,}  The Losartan Intervention for Endpoint Reduction in Hypertension study showed that losartan reduced the relative risk of developing type 2 diabetes mellitus by 25% compared with the β-blocker atenolol. However, since the study did not include a placebo control group, it is likely that the reduction in incident diabetes reflects the net result of both increased insulin sensitivity in the losartan group and increased insulin resistance in the atenolol group. Similar findings relative to placebo were reported in the Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity (CHARM) studies.^,  In CHARM-Overall, a study of 7601 patients followed up for a median of 37.7 months, candesartan (32 mg/d) reduced the relative risk of developing diabetes by 22% compared with placebo. In CHARM-Preserved, which included patients with class II to IV heart failure and an ejection fraction greater than 40% who were followed up for a median of 36.6 months, candesartan reduced the relative risk of developing diabetes by 39% compared with placebo. After 1 year, candesartan had reduced the relative rate of incident diabetes by 88% compared with hydrochlorothiazide in patients with newly diagnosed hypertension in the Antihypertensive Treatment and Lipid Profile in a North of Sweden Efficacy Evaluation study. However, this study followed up relatively few patients (Table 1).

The Valsartan Antihypertensive Long-term Use Evaluation trial demonstrated the advantage of an ARB, valsartan, over a CCB, amlodipine, in reducing the relative risk of new-onset diabetes by 23% in patients with hypertension 50 years or older at high risk of cardiac events who were treated for a mean of 4.2 years. Furthermore, the beneficial effect of valsartan in preventing new-onset diabetes in the Valsartan Antihypertensive Long-term Use Evaluation may have been underestimated, since the valsartan treatment arm was more likely to include patients taking hydrochlorothiazide and other antihypertensive drugs that might negatively affect glucose metabolism than the amlodipine treatment arm. Because amlodipine is considered neutral in its effects on insulin sensitivity and was substantially better than a thiazide diuretic in this regard in the ALLHAT study, drugs in the ARB class may independently improve insulin sensitivity and may have a role in protecting high-risk patients with hypertension from developing diabetes.

The possibility that an ARB can prevent transition from impaired glucose tolerance, which is common in patients with essential hypertension, to type 2 diabetes mellitus is being explored in the Nateglinide and Valsartan on Impaired Glucose Tolerance Outcomes Research study. The objectives of the 2 × 2 factorial design of this large prospective study of more than 9000 patients with impaired glucose tolerance are to determine whether restoration of the early insulin response with nateglinide and/or improvement in insulin sensitivity by blockade of the renin-angiotensin system with valsartan prevents the transition from either impaired glucose tolerance to diabetes or the incidence of cardiovascular disease. In the Diabetes Reduction Approaches With Medication study, 5269 patients with impaired glucose tolerance have been randomized to ramipril or rosiglitazone vs placebo in a 2 × 2 factorial design to determine whether new-onset diabetes can be prevented.

Other ACE inhibitor and ARB studies are assessing the incidence of type 2 diabetes mellitus as a secondary end point. The Ongoing Telmisartan Alone and in Combination With Ramipril Global Endpoint Trial is a double-blind, parallel-group trial with 3 treatment arms—telmisartan, ramipril, and telmisartan plus ramipril—that will include 23,000 persons. It is designed to determine the effect of one or both agents on a composite cardiovascular end point of myocardial infarction, stroke, and hospitalization for heart failure during the 5.5-year follow-up period. Patients unable to tolerate an ACE inhibitor will be entered into a parallel study of telmisartan vs placebo, the Telmisartan Randomized Assessment Study in ACE Intolerant Subjects with Cardiovascular Disease.

Collectively, these ongoing studies are expected to clarify the extent by which inhibition of the renin-angiotensin system can reduce the incidence of new-onset diabetes in patients with impaired glucose tolerance, a group that includes many of the 65 million Americans with essential hypertension.

A number of mechanisms exist by which antihypertensive agents may affect glucose metabolism. Mechanistic studies support and explain the findings from clinical studies that have evaluated the effects of antihypertensive agents on glucose metabolism.

Thiazide Diuretics. Thiazide diuretics appear to worsen glycemic control in a dose-dependent fashion by reducing insulin secretion and peripheral insulin sensitivity.^{,  ,  ,  ,}  Development of hypokalemia (and perhaps hypomagnesemia) appears to be important because the use of potassium supplementation to prevent hypokalemia reduces the occurrence of thiazide-induced glucose intolerance.^{,  ,}  Nevertheless, deterioration in glucose metabolism occurs even with minimal reductions in serum potassium levels.^,  Since ACE inhibitors and ARBs appear to counteract some of the adverse effects associated with the use of thiazide diuretics, including potassium wasting and aldosterone secretion, early combined therapy has been recommended.^{,  ,}  Although combination therapy likely leads to better blood pressure control, head-to-head randomized controlled trials are lacking in determining the metabolic outcomes of such combinations.

β-Blockers. Therapy with β-blockers has been shown to inhibit pancreatic insulin secretion and peripheral glucose utilization.^{,  ,  ,  ,  ,  ,}  Weight gain, decreased skeletal muscle blood flow, and unopposed stimulation of β₂-receptor-mediated glycogenolysis are also potential mechanisms by which β-blockers might exert adverse effects.^{,  ,  ,  ,  ,  ,}  Interestingly, much attention is being paid to third-generation β-blockers, such as carvedilol and nebivolol, that possess vasodilator actions through such proposed mechanisms as nitric oxide release, antioxidant effects, and Ca²⁺ blockade. Basic and clinical studies suggest that these drugs may improve insulin sensitivity^{,  ,  ,  ,}  or are at least metabolically neutral in patients with impaired glucose tolerance.^,  In the Glycemic Effects in Diabetes Mellitus: Carvedilol-Metoprolol Comparison in Hypertensives trial, the comparison of carvedilol to metoprolol in the presence of reninangiotensin system blockade on glycemic control showed that carvedilol was superior in improving insulin sensitivity, preserving hemoglobin A_1c levels, and preventing the progression to microalbuminuria, despite similar blood pressure responses in patients with diabetes mellitus and hypertension. Therefore, third-generation β-blockers may become the preferred antiadrenergic blood pressure drugs in patients with insulin resistance.

Calcium Channel Blockers. Calcium channel blockers have been proven effective as adjunct therapy in diabetic patients with hypertension. Furthermore, CCBs have been shown to reduce insulin resistance or new-onset diabetes among patients with metabolic syndrome, seeming to be intermediate in effectiveness between ACE inhibitors and ARBs (which reduce the incidence of diabetes mellitus) and thiazide diuretics and β-blockers (which increase it).^{,  ,  ,  ,}  Both dihydropyridine and long-acting nondihydropyridine^,  calcium antagonists have shown metabolic benefits, with effects on insulin sensitivity and insulin secretion. These agents may improve insulin sensitivity by exerting vasodilatory action in insulin-sensitive tissues without stimulating sympathetic nervous activity, by preventing the inhibition of glucose transporters and glycogen synthase by calcium, or through antioxidant effects.

Centrally Acting Agents. Centrally acting antihypertensive agents are not widely used because of a relatively high incidence of adverse effects. Most central adverse effects appear to be through the α₂-receptor. The use of an α₂-agonist such as clonidine has been limited by adverse effects, including central nervous system effects, sexual dysfunction, and dry mouth. The α₂-agonists appear to have minimal effects on lipid profiles but may inhibit pancreatic β-cell insulin secretion, thereby impairing glucose metabolism.

Imidazoline receptor agonists such as moxonidine are selective for the I1-imidazoline receptor in the sympathetic vasomotor centers with little effect at the central α₂-receptor. Therefore, the adverse effect profile is favorable compared with other centrally acting agents. Moxonidine has been shown to diminish sympathetic activity, and it reduces arterial pressure by lowering systemic vascular resistance without affecting heart rate and cardiac output. Furthermore, drugs such as moxonidine may exert favorable metabolic effects, including improved insulin sensitivity.^,  Animal studies have suggested that this may be due to decreased vasoconstriction in insulin-sensitive tissues such as skeletal muscle and improved insulin signaling, which leads to increased glucose uptake. The Moxonidine and Ramipril Regarding Insulin and Glucose Evaluation study will compare the effects of moxonidine and the ACE inhibitor ramipril and the combination of both drugs on metabolic and hemodynamic variables in patients with hypertension and impaired fasting glycemia.

α₁-Blockers. α-Adrenergic blockers have also been shown to have beneficial metabolic effects. Indeed, the selective α₁-blockers, such as prazosin, terazosin, and doxazosin, are the only class of antihypertensive agents that appear to have the combined effect of improving insulin sensitivity, raising high-density lipoprotein cholesterol levels, and lowering low-density lipoprotein cholesterol levels.^{,  ,  ,}  However, in 2000, use of the α-blocker doxazosin arm of the ALLHAT study was discontinued early (median follow-up, 3.3 years) based on interim comparisons with the diuretic chlorthalidone. Particularly worrisome was a doubling of risk of congestive heart failure. Furthermore, α-blockers are associated with troublesome adverse effects, including dizziness, headache, and weakness. In a prospective trial in which 6 different antihypertensive drugs were compared, the α-blocker prazosin was found to have the highest incidence of adverse effects. Therefore, α-blockers are generally not considered first-line therapy in essential hypertension, even in patients with diabetes or metabolic syndrome.

Patients with essential hypertension frequently have impaired glucose tolerance due to increased insulin resistance. This is related in part to alterations in insulin action at the level of skeletal muscle tissue, which comprises 40% of body mass and is the predominant site of insulin-stimulated glucose utilization (Table 2). With aging and sedentarylifestyles, which are associated with an increased propensity to hypertension, skeletal muscle mass and insulin sensitivity are reduced. Furthermore, an association exists between blood pressure and the proportion of type II fibers in skeletal muscle, which are less sensitive to insulin than type I fibers. Interestingly, ACE inhibitors^,  and ARBs may influence skeletal muscle toward a greater percentage of type I fibers.

Insulin resistance in skeletal muscle is associated with increased intracellular lipid and fat interspersed between muscle fibers,^,  which is more prevalent with aging and obesity.^,  It has been proposed that angiotensin II has differential effects on the lipid storage capacity of adipose and skeletal muscle tissues such that triglyceride accumulation occurs in muscle. Interestingly, renin-angiotensin system activity also appears to affect recruitment and differentiation of adipocytes,^{,  ,}  and recent studies suggest that certain ARBs may interact with the nuclear hormone receptor peroxisome proliferator-activated receptor γ independent of angiotensin II receptors.^{,  ,}  This interaction may represent a potential mechanism by which ARBs improve insulin sensitivity, since peroxisome proliferator-activated receptor γ is the cellular target for the insulin-sensitizing thiazolidinedione drugs.

Increased skeletal muscle blood flow and resulting improvements in insulin delivery may be important mechanisms by which attenuation of the renin-angiotensin system improves glucose uptake and metabolism in insulin-sensitive tissues.^,  One potential mechanism would be via the inhibiting effects of ACE inhibitors on kininase II, thereby increasing bradykinin and the subsequent enhancement of nitric oxide production. Other potential mechanisms of renin-angiotensin system blockade are improved vascular sensitivity to insulin and improved endothelial function.^{,  ,}  Moreover, both ACE inhibitors and ARBs likely increase skeletal muscle blood flow by diminishing the vasoconstricting effects of angiotensin II, particularly at the level of the small arteriole.

Although ACE inhibition and receptor blockade un-doubtedly diminish the effect of angiotensin II on vascular tissues, angiotensin II appears to have direct effects on skeletal muscle tissue. For example, angiotensin II directly infused into the interstitial space of canine muscle causes insulin resistance for glucose uptake that cannot be attributed to hemodynamic changes. Thus, ACE inhibitors and ARBs likely improve insulin-mediated glucose disposal in hypertensive and insulin-resistant conditions by augmenting blood flow to insulin-sensitive tissues and by having direct effects on skeletal muscle glucose uptake capacity (Table 2).

Insulin-dependent skeletal muscle glucose transport^,  involves the binding of circulating insulin to the plasma membrane (sarcolemma) receptor, which results in the activation of a cascade of phosphorylation steps. Ultimately, the pathway leads to the translocation of the insulin-sensitive glucose transporter (GLUT4) to the sarcolemma, facilitating glucose uptake into the cell. Evidence is accumulating that angiotensin II interferes with insulin signalingin skeletal muscle. First, studies support the existence of a physiologically functional paracrine skeletal muscle renin-angiotensin system,^,  including angiotensin I, ACE,^{,  ,  ,}  and angiotensin receptors.^,  Second, animal^{,  ,}  and human studies have shown that long-term ACE inhibition or angiotensin receptor blockade increases whole-body insulin sensitivity under experimental conditions. Moreover, our laboratory and others have shown that skeletal muscle glucose uptake is diminished in transgenic hypertensive rats (Ren-2) that overproduce tissue angiotensin II and that this is prevented with ARB treatment.

Among the mechanisms by which angiotensin II might inhibit insulin signaling pathways are increases in the generation of reactive oxygen species or alterations in production of nitric oxide. Studies indicate that angiotensin II increases superoxide production in skeletal muscle, which has also been observed in muscle from diabetic mice. In contrast, blockade of the angiotensin II type 1 receptor by ARBs is reported to greatly reduce muscle superoxide production in rodents.^,  It is well documented that angiotensin II increases generation of reactive oxygen species in vascular cells primarily through the activation of membrane-bound NADPH oxidase.^{,  ,  ,  ,}  However, it remains uncertain whether the angiotensin II-associated increase in skeletal muscle oxidative stress is mediated by NADPH oxidase as depicted in Figure 1 or by other pathways that produce reactive oxygen species.

The effects of oxidative stress on insulin signaling have been demonstrated in insulin-resistant animal models in which increased protein carbonyl levels in skeletal muscle and liver and increased superoxide content in skeletal muscle are evident. Indeed, treatment of these animals with antioxidants such as α-lipoic acid or the superoxide dismutase-catalase mimetic tempol improved whole-body glucose tolerance and insulin-stimulated skeletal muscle glucose uptake. Interestingly, angiotensin II may act as a negative regulator of GLUT4 gene expression in skeletal muscle, whereas long-term angiotensin II receptor antagonism or ACE inhibition increases GLUT4 protein content in skeletal muscle.

Alternatively, recent evidence suggests a role for nitric oxide in skeletal muscle glucose metabolism. For example, nitric oxide synthase inhibitors induce insulin resistance, whereas nitric oxide donors stimulate glucose uptake.^{,  ,  ,}  Furthermore, endothelial nitric oxide synthase knockout mice exhibit insulin resistance. The nitric oxide synthase isoforms neuronal and endothelial are expressed in skeletal muscle. Moreover, the production of nitric oxide in skeletal muscle can modulate muscle superoxide production. The potential role of angiotensin II in the generation of reactive oxygen species, impaired insulin signaling, nitric oxide-related mechanisms, and GLUT4 expression and translocation are presented in Figure 1.