Pharmacologic and Surgical Interventions That Prevent or Worsen Type 2 Diabetes

Angiotensin-Converting Enzyme Inhibitors and Angiotensin Receptor Blockers

It is not clear that a reduction in diabetes incidence through the RAAS blockade pathway observed in a number of randomized trials represents a truly preventive effect or simply a delay or a masking effect ( Table 6-1 ). Overall, diabetes incidence was not a prespecified, primary endpoint in any of these prior studies, and there was insufficient evidence to definitively recommend any given angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers (ARBs) in patients at risk of developing type 2 diabetes for the purpose of diabetes prevention. Nevertheless, in the Heart Outcomes Prevention Evaluation (HOPE) trial, ACE inhibitor therapy with ramipril lowered diabetes incidence from 5.4% to 3.6% overall, and in the Studies of Left Ventricular Dysfunction (SOLVD) trial of patients with systolic heart failure, enalapril reduced diabetes incidence from 22% to 6%. ^, The Prevention of Events with Angiotensin Converting Enzyme Inhibition (PEACE) study reported a significant reduction in the incidence of type 2 diabetes from 11.5% to 9.8% in patients with stable CAD treated with trandolapril. ARB therapy significantly decreased diabetes incidence in the Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity (CHARM) study from 7% to 6% and resulted in a nonsignificant decrease in diabetes incidence from 5.3% to 4.3% in the Study on Cognition and Prognosis in the Elderly (SCOPE) trial. ^, In the Valsartan Antihypertensive Long-Term Use Evaluation (VALUE) trial of 10,419 hypertensive patients at high cardiovascular risk, a valsartan-based treatment regimen was associated with a decrease in the incidence of type 2 diabetes from 16% to 13% compared with an amlodipine-based regimen.

There are several potential mechanisms by which ACE inhibitors or other pharmacologic therapies may exert beneficial effects on glycemic control ( Fig. 6-1 ). The activation of the SNS via alpha ₂ -adrenergic receptors impairs insulin secretion and peripheral glucose uptake. Inhibition of the SNS via inhibition of angiotensin II production counteracts this effect. Also, angiotensin II impairs pancreatic blood flow and enhances insulin resistance. However, this pathophysiologic pathway has not been a consistent finding across studies. Angiotensin II also mediates a number of potentially toxic effects within the pancreas, such as islet cell fibrosis and death, oxidative stress, impaired first-phase insulin release, and cytokine release. ACE inhibitors also directly improve insulin sensitivity, primarily in skeletal muscle. This appears partially mediated through the vasodilatory actions of the upregulation of bradykinin and nitric oxide. Inhibition of the RAAS system also may augment the postreceptor activity of insulin, and angiotensin II has also been shown to inhibit adipocyte differentiation in vitro. ^, Through this latter mechanism, ACE inhibition may improve insulin sensitivity through promotion of adipocyte differentiation, yielding increased storage depot for caloric excess. Furthermore, inhibition of the RAAS pathway increases adiponectin and leptin levels, both of which increase insulin sensitivity and promote adipocyte differentiation. ACE inhibitors also raise circulating potassium levels, which counteracts the impairment in insulin secretion seen with hypokalemia. With the exception of effects mediated through the upregulation of bradykinin, all of these potential mechanisms are, in theory, also applicable to ARBs. In addition, telmisartan, a partial agonist of the peroxisome proliferator-activated receptor gamma (PPAR-γ), may lower glucose levels in a similar fashion but with less efficiency, compared with the thiazolidinedione medications.

Beta Blockers

Beta blockers, which decrease SNS activation via beta-adrenergic receptor antagonism, are effective in reducing cardiovascular morbidity and mortality in patients with several conditions, including those who have sustained a myocardial infarction and those with systolic heart failure. Despite these clinical benefits, many physicians are reluctant to prescribe beta blockers because of perceived negative metabolic effects ( Box 6-3 ). However, beta blockers should not be considered a homogenous class of agents. Beta blockers consist of nonvasodilating and vasodilating agents, which differ in terms of their mechanisms of action and effects on glucose and lipid metabolism. Treatment with nonvasodilating beta blockers is associated with an increased propensity of patients with hypertension to develop diabetes. A substudy of the Atherosclerosis Risk in Communities (ARIC) observational cohort study, which included 3804 patients with hypertension, demonstrated that patients treated with nonvasodilating beta blockers had a 28% higher risk of developing diabetes than patients on no pharmacologic treatment for hypertension. Patients receiving thiazide diuretics, ACE inhibitors, or calcium channel antagonists were not at significantly higher or lower risk for subsequent diabetes than untreated patients. Similarly, the Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) randomized trial demonstrated in 9193 patients that the risk of developing diabetes was 25% lower among patients with hypertension and left ventricular hypertrophy who received losartan- based therapy than among patients who received atenolol-based therapy.

Box 6-3

Potential Deleterious Metabolic Effects of Beta Blockers

• Reduced glycemic control

• Masking of hypoglycemia

• Deterioration in insulin resistance

  • Decreased blood flow to muscles, reducing peripheral insulin-stimulated glucose uptake

  • Interference with insulin secretion from pancreatic beta cells

  • Decrement in the first phase of insulin secretion

• Weight gain

• Dyslipidemia

Nonvasodilating beta blockers (atenolol, metoprolol, pindolol, and propranolol) reduce blood pressure in association with a cardiac output reduction and may increase or have no appreciable influence on peripheral vascular resistance. Nonvasodilating beta blockers include first- and second-generation agents. First-generation beta blockers (propranolol) block both beta ₁ - and beta ₂ -adrenergic receptors (nonselective beta blockade), whereas second-generation beta blockers (atenolol and metoprolol) specifically target beta ₁ -adrenergic receptors (cardioselective beta blockade). Nonvasodilating beta blockers significantly decrease insulin sensitivity by approximately 14% to 33% among patients with hypertension. ^, Studies have shown increases in glucose concentrations with the use of atenolol alone, ^, metoprolol, or propranolol, ^, although other studies have shown no changes in glucose levels with atenolol or propranolol. ^, However, glucose levels at a particular timepoint may not reflect long-term changes in glucose metabolism as reflected by hemoglobin A1c (HbA1c). As an example, after 6 months of treatment, once-daily metoprolol did not affect fasting plasma glucose but significantly increased HbA1c levels by a relative increase of 5% from baseline in patients with hypertension.

In contrast, vasodilating beta blockers (carvedilol, labetalol, and nebivolol) reduce peripheral vascular resistance but have little or no effect on cardiac output. Numerous studies have shown that vasodilating beta blockers are associated with more favorable effects on glucose and lipid profiles than nonvasodilating beta blockers. Bisoprolol, a beta ₁ -selective adrenergic blocker, was reported to have a neutral effect on glucose and insulin levels during a glucose tolerance test after 24 weeks of treatment at 5 to 10 mg/day in 13 patients with hypertension.

Although the specific mechanisms have not been identified, several have been postulated to explain the negative effects of nonvasodilating beta blockers on glucose and lipid metabolism, most of which relate to their hemodynamic effects. Treatment with nonvasodilating beta blockers, which block either the beta ₁ -adrenergic receptor or the beta ₁ - and beta ₂ -adrenergic receptors, results in unopposed alpha ₁ -adrenergic receptor activity (which can induce vasoconstriction), decreased blood flow to the muscles, and reduced insulin-stimulated glucose uptake in the periphery. ^, Nonvasodilating beta blockers may also interfere with insulin secretion from pancreatic beta cells. Moreover, beta blockers may decrease the first phase of insulin secretion (potentially an important predictor of diabetes) via impairment of beta ₂ -mediated insulin release. ^, Weight gain also has been noted in patients who received nonvasodilating beta blockers ^, and is closely linked to an increased risk for developing diabetes. Increased peripheral blood flow from the action of vasodilating beta blockers may result in efficient glucose dispersal to the skeletal muscles, thereby facilitating insulin sensitivity. The mechanisms responsible for the beneficial effects of vasodilating beta blockers on glucose and lipid metabolism are not entirely understood but may include alpha ₁ -adrenergic receptor blockade, vasodilation, reduced oxidative stress, anti-inflammatory activity, and lack of weight gain. ^{, ,}

In the Glycemic Effects in Diabetes Mellitus: Carvedilol-Metoprolol Comparison in Hypertensives (GEMINI) trial, the metabolic effects of carvedilol and metoprolol were compared in 1235 patients with type 2 diabetes and hypertension. The use of carvedilol in the presence of RAAS blockade was not deleterious to glycemic control and improved some components of the metabolic syndrome relative to metoprolol. On the other hand, metoprolol significantly increased HbA1c levels from baseline (absolute increase of 0.15%), in contrast to carvedilol (0.02%) with an absolute difference between the groups of 0.13% ( P = 0.004 for carvedilol versus metoprolol). Statistically significant improvement was observed in insulin sensitivity with carvedilol (− 9.1%) but not statistically significant with metoprolol (− 2.0%). In this trial, metoprolol-treated patients experienced a significant weight gain (1.2 ± 0.16 kg) compared with carvedilol-treated patients (0.17 ± 0.19 kg). In another study, a group of patients treated with metoprolol had an increase in body weight of 1.8 kg after 2 months of treatment. In contrast, no significant weight gain was found in the group of patients treated with carvedilol. This is in accordance with the weight gain seen after treatment with beta blockers in large clinical trials. In the presence of heart failure, carvedilol was shown to be associated with improved survival (the Carvedilol or Metoprolol European Trial [COMET]) and with fewer cases of new-onset diabetes compared with metoprolol. ^,

Carvedilol and nebivolol enhance nitric oxide synthesis and thus mediate endothelial-dependent vasodilation. Carvedilol is associated with antioxidant activity, possibly because of stimulation of endothelial nitric oxide production or reduced nitric oxide inactivation. ^, Carvedilol also inhibits low-density lipoprotein cholesterol (LDL-C) oxidation, potentially reducing the accumulation of oxidized LDL-C in vessel walls and subsequent vascular damage. Furthermore, carvedilol has been shown to protect against reactive oxygen species via scavenging of free radicals, suppression of free radical generation, and prevention of ferric ion-induced oxidation. ^, Carvedilol treatment reduced proinflammatory markers, including plasma C-reactive protein and monocyte chemotactic protein 1, in patients with hypertension and diabetes.

Nebivolol is considered to have a neutral effect on metabolic parameters in patients with hypertension, but among 80 patients with hypertension, nebivolol significantly reduced baseline insulin levels and insulin resistance versus metoprolol after 6 months of treatment. In a larger, open-label study involving 328 patients with hypertension, compared with baseline, nebivolol significantly reduced fasting glucose, total cholesterol, and triglyceride levels. Another large, open-label study involving 2838 patients with hypertension and diabetes showed that after 3 months of nebivolol treatment, significant reductions were observed from baseline in fasting glucose, HbA1c, total cholesterol, LDL-C, and triglyceride levels, with an increase in high-density lipoprotein cholesterol (HDL-C). Several small studies have demonstrated that labetalol treatment compared with other antihypertensive therapies is associated with neutral effects on glucose and lipid metabolism in patients with essential hypertension. ^,

Thiazide Diuretics

Thiazide diuretics can result in various undesired biochemical changes, and in general increase glucose and insulin resistance. Thiazides have been shown to increase fasting glucose levels and impair glucose tolerance curves in many long-term studies. The effect on glucose tolerance is usually reversible if the thiazide is stopped, and the effect on blood glucose levels is dose related.

The mechanistic underpinning of these effects is not well understood. Thiazide-induced hypokalemia, as well as effects on other pathophysiologic pathways, may explain these metabolic disturbances, such as increased visceral adiposity, hyperuricemia, decreased glucose metabolism, and pancreatic beta cell hyperpolarization. Whereas many large randomized, prospective clinical trials show an association between thiazide use and increased blood glucose, findings are mixed regarding the association with new-onset diabetes. Many issues must be considered in evaluating these associations in these trials, including the following:

• Most are post hoc findings and were not adequately powered to assess this association.

• New-onset diabetes was defined differently in many studies.

• Many studies had follow-up durations of only a few years, which may not be long enough to fully assess prolonged hyperglycemia.

• Comparing drug classes is difficult because of differing study designs.

A meta-analysis of clinical trials revealed that of all antihypertensives assessed, beta blockers and thiazide diuretics are associated with the highest risk of diabetes. This analysis found that thiazides are associated with higher risk of diabetes than placebo and, along with beta blockers, had the highest risk of all major classes of antihypertensives. The pathophysiologic process accounting for these effects may be mediated through influence on potassium balance and circulating potassium levels. It has been reported that potassium infusions causing more than a 1- to 1.5-mEq/L elevation in plasma potassium enhance insulin release twofold to threefold compared with basal levels. The relationship between potassium and glucose homeostasis is central because many believe that thiazide-induced potassium depletion drives hyperglycemia. Actually, a meta-analysis of 59 clinical studies showed a significant correlation between thiazide-induced potassium depletion and increased blood glucose levels, as well as a correlation between potassium supplementation (or concomitant use of potassium-sparing agents) and attenuation of hyperglycemia. In addition, a secondary analysis of the Systolic Hypertension in the Elderly Program (SHEP) investigated the relationship between serum potassium and thiazide-induced diabetes. The risk for developing diabetes was increased in the first year of thiazide treatment. In addition, independent of drug treatment, each 0.5-mEq/L decrease in serum potassium was associated with a 45% increased risk for development of diabetes throughout the course of the study. A retrospective analysis of an extended follow-up of the SHEP trial was recently reported. After a mean follow-up period of 14.3 years, patients treated with thiazide were more likely to develop new-onset diabetes. However, new-onset diabetes that developed in patients treated with thiazide diuretics was not associated with significantly increased cardiovascular or total mortality.

Alteration in fat distribution is another possible mechanism for thiazide-induced dysglycemia. Patients treated with 25 to 50 mg of thiazide daily had significant reductions in insulin sensitivity, compared with those treated with candesartan or placebo. Serum potassium levels were significantly lower in patients taking thiazide, but levels in all groups remained within normal limits. While taking thiazide, patients also developed a significantly higher hepatic fat content, and a significant correlation was found between hepatic fat content and decreased insulin sensitivity. Whether decreased insulin sensitivity was a result of this visceral fat accumulation or vice versa is not clear. Low-grade inflammation assessed with C-reactive protein was also significantly increased with thiazide treatment, suggesting a possible role for inflammation in the development of insulin resistance. Of clinical interest is that patients with abdominal obesity are more likely to experience new-onset diabetes with thiazide treatment than those without abdominal obesity. ^, In older studies of thiazide diuretics, the dosage (or equivalent in vitro concentration) used was 50 mg of thiazide or more. Today, clinicians do not often prescribe doses greater than 25 mg of thiazide or its equivalent. Of note, findings from the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) diabetes extension study suggest that thiazide-related incident diabetes has less adverse long-term CVD impact than incident diabetes that develops while patients are taking amlopidine or lisinopril.

Calcium Channel Blockers

In 16,176 hypertensive patients with CAD enrolled in the International Verapamil SR Trandolapril Study (INVEST), a verapamil-based treatment regimen was associated with a decrease in the incidence of type 2 diabetes from 8.2% to 7.0% compared with an atenolol-based regimen. Diabetes incidence was not a predefined endpoint in this study, and no adjustment was made for concomitant therapies, which could potentially affect diabetes incidence. In ALLHAT, the risk of incident diabetes at 4 years of follow-up was 9.3% with the calcium channel blocker amlodipine and 7.8% with the ACE inhibitor lisinopril.

High-dose calcium channel blocker therapy can inhibit insulin release, but this effect is generally not seen with usual therapeutic doses. ^, Impaired insulin release appears to be counterbalanced by increased peripheral glucose uptake, such that the predominant effect of these agents is metabolically neutral or favorable. ^, Evidence from animal models suggests that vasodilation and improved peripheral blood flow may explain the potential improvement in insulin sensitivity seen with calcium channel blockade. Thus, calcium channel blockers appear to have a neutral or slightly favorable effect on glucose metabolism.

Niacin

Niacin is known to increase insulin resistance and have adverse effects on blood glucose levels, but to have favorable effects on plasma lipids and lipoproteins. Niacin reduces plasma triglycerides, increases HDL-C, and reduces LDL-C modestly. Concerns have been raised about use of niacin in diabetic patients because of its adverse effects on insulin resistance and blood glucose levels. ^,

Reports from the Assessment of Diabetes Control and Evaluation of the Efficacy Niaspan Trial (ADVENT), the Arterial Disease Multiple Intervention Trial (ADMIT), and the HDL-Atherosclerosis Treatment Study (HATS) have shown that the modest increase in glucose level caused by niacin treatment could be easily counteracted by adjusting the diet, amount of exercise, and dose of glucose-lowering medication. During the follow-up period in the Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglycerides: Impact on Global Health Outcomes (AIM-HIGH) trial, the dose of the study drug was reduced in 6.3% of the patients in the niacin group and 3.4% of the patients in the placebo group ( P < 0.001). Increased glucose level was the primary reason for dose reduction in 5 (0.3%) and 10 (0.6%) patients (placebo versus niacin, respectively). The study drug was discontinued in 25.4% of the patients in the niacin group and in 20.1% of the patients in the placebo group ( P < 0.001). The primary reason for discontinuation because of increased glucose level was reported in 14 (0.8%) and 29 (1.7%) patients (placebo versus niacin, respectively). During HATS, there was a 20% rise in insulin levels in the groups taking niacin. This finding was accompanied by a 2% to 3% increase in fasting glucose levels. In patients with diabetes, glucose levels increased by approximately 15% by 3 months in those receiving niacin, but returned to baseline by 8 months. Changes in glucose-lowering medications were permitted, but no data were provided. Glycemic control among patients with diabetes returned to pretreatment values after 8 months, probably because of better diabetes management. The changes in blood glucose with extended-release (ER) niacin are typically modest and transient and more prevalent in patients with diabetes. ^, On average, the rise in HbA1c levels is small and can be managed by titrating hypoglycemic therapy, but blood glucose levels should be closely monitored in patients with difficult-to-treat diabetes.

The Heart Protection Study 2—Treatment of HDL to Reduce the Incidence of Vascular Events (HPS2-THRIVE) study, a large randomized trial that comprising 25,673 patients tested the use of extended-release niacin and the antiflushing agent laropiprant for the reduction of major vascular events, did not find a significantly reduced risk of major vascular events in patients with well-controlled LDL-C levels. The failure of niacin in the HPS2-THRIVE study was first announced in late December 2012. In light of these findings, the role of extended-release niacin for the prevention of CVD should be reconsidered, given the side effects of niacin including a 25% increased risk of new-onset diabetes and the difficulties in controlling glucose level when patients are taking niacin.

Statins

Statin therapy, particularly high-dose therapy, is associated with increased diabetes risk. ^, These study observations are supported by results from two meta-analyses including over 100,000 participants demonstrating that long-term statin intake was associated with increased risk of new-onset diabetes. ^, Several meta-analyses have been conducted to elucidate the effect of statins on glucose metabolism. ^{, ,} Treatment with statins has been associated with a 9% increase in the risk of developing diabetes without any clear differential effect among individual statins. ^, However, the overall data available strongly suggest that the reduction in CVD events outweighs the minor effect on glucose homeostasis. ^, In contrast, it has been suggested that various statins may affect glucose metabolism differentially. ^,

Pravastatin appeared to improve insulin sensitivity, whereas simvastatin was associated with an adverse effect on glucose metabolism. Also, atorvastatin and rosuvastatin nonsignificantly worsened insulin sensitivity. Rosuvastatin administration in hypercholesterolemic patients with impaired fasting glucose was associated with a dose-dependent increase in insulin resistance. ^,

The mechanisms by which statins may impair glucose metabolism are not fully understood. Several mechanisms may be responsible for these diabetogenic effects ( Box 6-4 ). One possibility is a statin-mediated decrease in various metabolic products of the mevalonate pathway, such as the isoprenoids farnesyl pyrophosphate or geranylgeranyl pyrophosphate. These isoprenoid molecules have been linked with the upregulation of the membrane transport protein glucose transporter 4 (GLUT-4) in 3 T3-L1 adipocytes, thus augmenting glucose uptake. In type 2 diabetic mice and human patients treated for 3 months, atorvastatin impaired glucose tolerance and GLUT-4 expression by inhibiting isoprenoid biosynthesis. In addition, a possible role for the small guanosine triphosphate (GTP) binding proteins as regulators of glucose-mediated insulin secretion by beta cells has been suggested. Statins, by inhibiting 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase, decrease the production of these substances. High doses of lipophilic statins decreased insulin secretion from beta cell lines, mediated either by the inhibition of HMG-CoA reductase or direct cytotoxicity. Therefore the lipophilicity of individual statins may influence their effects on glucose metabolism. Statins, particularly the lipophilic compounds, have been shown to inhibit glucose-induced cytosolic Ca ^{2 +} elevations and insulin secretion as a result of blockade of L-type Ca ^{2 +} channels in rat islet beta cells. Simvastatin attenuates increases in cardiorespiratory fitness and skeletal muscle mitochondrial content when combined with exercise training in overweight or obese patients at risk of the metabolic syndrome. However, this opposes the observed effects of rosuvastatin, which is known to be a hydrophilic molecule. A protective effect of pravastatin, which is a hydrophilic statin, on progression to diabetes has been reported in a post hoc analysis of West of Scotland Coronary Prevention Study (WOSCOPS).

Box 6-4

Potential Mechanisms by Which Statins May Impair Glucose Metabolism

• Decrement in various metabolic products of the mevalonate pathway

  • ↓ Isoprenoid farnesyl pyrophosphate

  • ↓ Geranylgeranyl pyrophosphate

• ↓ Glucose transporter 4 expression

• ↓ Protein isoprenylation and affect on the distribution of several small G proteins

  • ↓ Potentiation of nutrient-induced insulin secretion by bombesin and vasopressin

• ↓ Insulin secretion from beta cell lines

  • Blockade of L-type Ca ^{2 +} channels in animal models