Advances in Diabetes for the Millennium: Drug Therapy of Type 2 Diabetes

Abstract

There are many new orally administered agents to treat type 2 diabetes. Sulfonylureas and meglitinides stimulate insulin secretion. Metformin has been joined by thiazolidinediones to reduce insulin resistance. Disaccharidase inhibitors slow glucose uptake after a meal. Beta-3 agonists and agents that augment glucagon-like peptide activity are promising new agents in the effort to not only control glucose levels but also restrain weight gain. The future treatment of diabetes will require multiple drugs working in concert to normalize blood glucose.

Introduction

In the 1920s, insulin injections joined dietary treatment and exercise as the cornerstones of therapy of diabetes. In the early 1950s, the advent of sulfonylureas and phenformin added an oral route to a reduction of blood sugar levels. The University Group Diabetes Program (UGDP) was a study originally designed to assess the effect of these new agents on the vascular complications of diabetes. Non-insulin-dependent diabetic patients were randomized into groups treated with a placebo plus diet, a fixed dose of tolbutamide, a fixed dose of insulin, a fixed dose of phenformin, or a sliding scale of insulin doses on the basis of fasting glucose levels. At the end of 7 years, the study was stopped when excess cardiovascular mortality was discovered in the tolbutamide group with excess overall mortality as well in the phenformin group.[1]

The publication of the UGDP results led to a ban on the use of phenformin. Sulfonylurea use was not officially banned, but its use was strongly discouraged in favor of diet and insulin treatment of diabetes. The furor surrounding the UGDP results dealt a severe blow to research on antidiabetic pharmaceuticals in the United States. However, research continued outside the United States, and in the past 10 years a host of new oral hypoglycemic agents have become available to treat type 2 diabetes. With so many new choices, there is often considerable confusion about which agent or combination of agents is optimal for a given patient.

Agents Which Stimulate Insulin Secretion

Sulfonylureas stimulate the production and release of insulin by binding to a receptor site on the membrane of the pancreatic beta cell. Binding blocks the opening of ATP-dependent potassium channels, which leads to a depolarization of the membrane, leading to an influx of calcium. These events result in an increased production of insulin by the beta cell.

The evolution of the third-generation agents glipizide and glyburide was a major advance over the older sulfonylureas.[2] They are 20-50 times more potent than previous sulfonylureas on a milligram basis. They have a longer biological action than all preceding agents except for chlorpropamide, with a much lower incidence of adverse reactions, such as hyponatremia and reactions to alcoholic beverages. They have low protein binding, so that they have fewer drug interactions. Glimepiride (Amaryl) was developed more recently and differs from glyburide in several ways.[3] It is more potent, but behaves more like glipizide than glyburide with a good postprandial insulin response and a lower incidence of hypoglycemia than glyburide. A single daily dose of 8 mg is maximal, with very little added benefit from twice-daily administration of this dose level.

The major side effect of the sulfonylureas is hypoglycemia. Hypoglycemia is usually associated with reduced oral intake or prolonged exercise, and is more common with longer-acting sulfonylureas than with short-acting agents, such as tolbutamide.

The newer meglitinides, although not chemically sulfonylureas, increase insulin production by a similar mechanism, at the ATP-dependent potassium channels. They are much shorter-acting. Typically taken at the beginning of a meal, they induce an insulin surge, which fades rapidly, thus reducing the risk of later hypoglycemia. Repaglinide was the first such agent introduced.[4] Recently, nateglinide, a D-phenylalanine derivative that appears to be even shorter-acting, has been introduced. There is no added insulin release with these agents over a maximal dose of sulfonylurea. There is a potential advantage in using these agents in situations in which hypoglycemia may have significant risk, such as the elderly and renal and coronary disease patients. The short action of these agents reduces the risk of hypoglycemia, although not entirely eliminating it. The disadvantage of use of these agents is the need for multiple daily doses.

Metformin

Metformin is a biguanide that has been marketed in Europe for 30 years. It reduces hepatic glucose production and increases peripheral glucose utilization. The mechanism of action is still poorly understood.[5] The degree of glucose lowering induced by metformin in non-insulin-dependent patients is similar to that of glyburide.[6] Furthermore, when added to glyburide treatment, metformin produced a further substantial reduction in glucose levels.[7] Additionally, it decreases the release of free fatty acids from adipose tissue and lowers the cholesterol and triglyceride levels.

The most serious complication of biguanide use is lactic acidosis, which can be fatal. Fortunately, the incidence of lactic acidosis with metformin use is low (1 case per 33,000 patient-years).[8] The risk of lactic acidosis is increased in patients with renal disease. A serum creatinine of 1.5 mg/dL is the suggested upper limit on use of this agent. The risk of lactic acidosis is also increased with dehydration and with the use of radiologic contrast dye. Metformin should be stopped at the time of the radiographic contrast procedure and not restarted for 48 hours. Although lactic acidosis is very rare, a much more common problem with metformin is a high incidence of gastrointestinal complaints. One out of 3 patients will experience problems ranging from mild heartburn to significant diarrhea. Patients do tend to become more tolerant of metformin with time, so that, in some cases, one can reduce the dose and achieve a lower level of gastrointestinal distress.

Metformin is contraindicated in congestive heart failure and is relatively contraindicated in the elderly.

Unlike insulin and sulfonylurea treatment, metformin does not encourage weight gain. In fact, some patients lose weight on metformin therapy. Metformin is effective when given twice daily. An extended-release, once-daily preparation has recently been introduced.

Disaccharidase Inhibitors

Type 2 diabetes results from resistance to insulin effects coupled with a relative deficiency of insulin secretion. The most characteristic abnormality of insulin production is a reduction in the early-phase release of insulin from the pancreas.

Absorption of carbohydrates requires the eventual breakdown of disaccharides to form single sugars by the enzymes in the brush border of the small intestine. Disaccharidase inhibitors, such as acarbose and miglitol, effectively compensate for defective early-phase insulin release by inhibiting the breakdown of disaccharides to monosaccharides in the intestinal epithelium. Consequently, there is delayed and decreased absorption of these sugars.[9,10] Thus, there is a lower glycemic peak, permitting the diminished early-phase insulin secretion to cope more effectively with glucose disposal. The result is a decrease in postmeal glucose peaks in diabetic patients.

The efficacy of acarbose and other disaccharidase inhibitors is limited by the adverse reactions caused by a large amount of nonabsorbed disaccharides in the intestinal tract. This situation is one of effective malabsorption with the attendant symptoms of flatulence, abdominal discomfort, and diarrhea. As the dose of the disaccharidase inhibitor is increased, the level of nonabsorbed disaccharides rises, leading to worsening malabsorption symptoms. However, increased disaccharide concentration leads to the induction of disaccharidases in the jejunum and ileum. Eventually, this induction of new enzymes results in a slower, smoother absorption of disaccharides. The slower absorption is still effective in reducing postprandial glucose levels, but with fewer malabsorptive symptoms. Therefore, disaccharidase inhibitors must be started at a very low dose, with small increments over time. When started at a low dose with slow increases, the adverse reactions are minimized. Even so, the gastrointestinal adverse reactions of acarbose or miglitol occur in up to 40% of patients. Despite these limiting adverse reactions, the disaccharidase inhibitors have an advantage in terms of safety. They do not cause hypoglycemia. They do not undergo renal excretion, so that they are safe in patients with a modest elevation of serum creatinine.

The disaccharidase inhibitors are effective as single agents for the treatment of diabetes and are effective in combination with sulfonylureas or insulin.

Thiazolidinediones

This class of agents, like metformin, works not by increasing insulin secretion but, rather, by increasing insulin sensitivity. However, metformin and thiazolidinedones have different mechanisms of action because they synergistically improve glycemic control when given together.[11] Thiazolidinediones appear to activate peroxisome proliferator-activated receptor (PPAR)-gamma, which is involved in the metabolism of lipids and the differentiation of adipocytes. There is an interaction with the retinoid X receptor (RXR) to produce an activated heterodimer.[12]

Unlike other antidiabetic agents, the thiazolidinediones have a very slow onset of action. Although effects begin within 2 weeks, the maximal benefit of treatment is not seen for about 3 months.[13] When combined with insulin or with sulfonylureas, the onset and peak effect occur more rapidly, perhaps within 4 weeks.[14,15]

Troglitazone was the first thiazolidinedione to reach the market. Unfortunately, troglitazone showed hepatic toxicity.[16] In a small number of patients, severe liver damage occurred. As a result of this liver toxicity, troglitazone was withdrawn from the market. Rosiglitazone and pioglitazone followed troglitazone. Neither agent is toxic to the liver.[17] Rosiglitazone is the most potent thiazolidinedione with a maximal effective dose of 8 mg daily[18] as compared with 600 mg of troglitazone and 45 mg of pioglitazone. Carcinogenesis has been a concern with these agents in animal studies. Troglitazone produced lipoangiosarcomas in mice. Pioglitazone was associated with bladder cancer in rats. Rosiglitazone has shown no animal carcinogenesis in preclinical studies.

Thiazolidinediones are effectively used as single agents, but their relatively slow onset of action means that other agents are generally preferred as the first treatment of poorly controlled diabetes. Thiazolidinediones are very effective in combination use with other agents. Rosiglitazone or pioglitazone reduce HbA1c by about 1% in patients treated with either a maximal dose sulfonylurea or a maximal dose metformin, or with insulin treatment.[19-23] This same reduction of HbA1c appears to hold when a thiazolidinedione is added to an existing combination of metformin and glyburide.[24]

All thiazolidinediones cause weight gain. This weight gain is partially due to fluid retention. An increase in plasma volume results in a small drop of about 1% in hematocrit. In some susceptible patients, fluid retention may trigger congestive heart failure. This phenomenon occurs far more frequently in insulin-treated patients receiving a thiazolidinedione. There is also increased adiposity, although some studies suggest relative sparing of visceral fat. All thiazolidinediones cause a slight increase in low-density lipoprotein (LDL) levels and a substantial increase in high-density lipoprotein (HDL) levels. Thus, the LDL-to-HDL ratio actually decreases. There is also a slight lowering of blood pressure. As single-use agents, the thiazolidinediones do not cause hypoglycemia. They are entirely safe in patients with renal impairment. Animal studies suggest an increase in heart size in thiazolidinedione-treated animals. Careful, long-term echocardiographic studies in troglitazone-, rosiglitazone-, and pioglitazone-treated patients have shown no adverse cardiac effects.[22]

The most exciting aspect of thiazolidinedione therapy appears to be the suggestion of a long-acting effect of these agents. Both sulfonylurea- and metformin-treated patients experience a gradual loss of efficacy over time, the phenomenon of so-called "secondary failure." In contrast, thiazolidinedione effects appear to be maintained over longer periods. The most long-lasting studies have been carried out with rosiglitazone. Rosiglitazone-treated patients maintain stable HbA1c levels for over 2 years (Figure), whereas glyburide-treated patients, after first showing rapid improvement, then experience a progressive rise in HbA1c from initial improved values.

Figure 1.

Maintenance of HbA1c with thiazolidinedione treatment.

Future Agents

There has been an influx of new agents in the past several years. Many more are likely to follow in the near future. The evolving understanding of the PPAR system is leading to increased drug discovery in this area. The isoxazolidinediones lack the thiol group but affect the same system. The challenge is clearly to achieve reductions in insulin resistance without triggering fluid retention and weight gain.

The goal of achieving glucose reduction without weight gain is clearly desirable. The disaccharidase inhibitors and metformin are the principal agents to accomplish this goal. Efforts are under way to attempt to selectively activate those portions of the complex of the PPAR system and the retinoic acid dimer that lead to increased insulin sensitivity and improved lipid metabolism without triggering the differentiation of adipocytes, which occurs with current-day thiazolidinediones. PPAR-gamma ligands, which also behave as partial PPAR-alpha or PPAR-delta agonists, may also be found later to be useful in regulating dyslipemia as well as glucose levels. Ligand agents known as rexinoids are also being tested to determine whether they interact with the RXR component of the PPAR-gamma-RXR heterodimer to modify insulin sensitivity and dyslipemia.

Amylin is a 37 amino acid polypeptide secreted by the beta cell concurrently with insulin. Amylin slows gastric emptying and inhibits postprandial glucagon secretion. Amylin concentrations, like those of insulin, are reduced in type 1 diabetic patients. The amylin analogue pramlintide has been developed as an injectable agent for treatment. Although the reduction of HgbA1c by pramlintide is modest, there have been reductions in weight in both type 1 and type 2 diabetic patients taking this agent.[25]

Beta-3 agonists are under study as antidiabetic agents. Their effect appears to be mediated by enhanced thermogenesis and increased glucose uptake.[26] It is hoped that these agents may reduce glucose levels while decreasing adiposity.[27]

Glucagon-like peptide 1 (GLP-1) is a hormone secreted by the intestines during food absorption. It has multiple effects, including the stimulation of insulin release, the suppression of glucagon release, and a reduction in appetite.[28] GLP-1 is rapidly degraded in the circulation by dipeptidyl peptidase IV. Various analogues of GLP-1 have been developed that are relatively resistant to degradation. One agent, exenatide, has shown promise in reducing glucose levels and weight in type 2 diabetic patients and is in active development.[29] GLP-1 and its analogues must be administered by injection. The alternative approach is to inhibit dipeptidyl peptidase IV to conserve endogenous GLP-1. Several agents are under study and have the benefit of oral administration rather than injection.[30]

Changes in Drug Management of Type 2 Diabetes

For many years, there were few pharmaceutical options for the treatment of type 2 diabetes. Now, new sulfonylureas, metformin, the disaccharidase inhibitors, the thiazolidinediones, and meglitinides have rapidly become available. It is hoped that the new agents will lead to improved diabetic control and a lower incidence of diabetic complications, and, ultimately, to lower mortality. The Diabetes Control and Complications Trial in the United States and the Swedish Diabetes Intervention Study have pointed to a reduction of microvascular disease in type 1 diabetes with an improvement in glycemic control.[31,32] However, the majority of diabetic patients have type 2 diabetes. Large-vessel disease affecting the coronary, cerebral, and peripheral arteries is a much more significant source of morbidity and mortality in this older population than microvascular disease.[33] Although improved glycemic control is associated with a reduced risk of coronary artery disease, concurrent improvement also occurs in the associated factors of obesity, hypertension, and dyslipidemia. It is not entirely clear how much of the reduction in cardiovascular risk is due to better glucose levels as opposed to the concurrent improvement in the associated risk factors.[34,35]

The results of the United Kingdom Prospective Diabetes Study (UKPDS) have shed some light on these issues.[36,37] This study, which began in 1977, was much larger than the old UGDP study with over 4000 patients at 23 centers randomized to either diet therapy alone or pharmacologic therapy with either sulfonylurea, insulin, or metformin. Only obese patients were treated with metformin. There was an effort to isolate the additional effect of hypertension in their population by the addition of either captopril or atenolol in a substudy. The study concluded in 1997, after a median duration of randomized treatment of 11 years (6-20 years). After 9 years, the difference in HbA1c between the conventional and the intensively treated groups was 0.9%, with no differences between the insulin, sulfonylurea, and metformin groups. In the pharmacologically treated groups, there was a 25% risk reduction for microvascular end points, which included retinal disease and renal disease. The benefit of pharmacologic therapy on the macrovascular end points was much smaller. There were no significant differences between the groups in the risk of fatal myocardial infarction, heart failure, angina, stroke, amputation, or death from peripheral vascular disease. There was a significant risk reduction in episodes of sudden death and a borderline difference in the incidence of nonfatal myocardial infarction favoring pharmacologic therapy. In contrast, the results of the substudy of antihypertensive treatment were far more impressive. There was a 32% reduction in diabetes-related mortality, a 44% reduction in stroke, and a 37% reduction in microvascular disease.

The most important result of the UKPDS was to show the equivalence of sulfonylureas, metformin, and insulin therapy in the end points of the study, including diabetic complications and overall mortality. Thus, the findings of the UGDP study were not duplicated. However, the complexity of the study design has led to much controversy over the interpretation of the results. In particular, metformin therapy showed a clear superiority in outcomes over sulfonylureas and insulin in obese patients. However, patients in whom metformin was added to sulfonylurea treatment in an attempt to improve glycemic control actually showed higher mortality than if sulfonylureas were simply maintained. These contradictory findings leave some uncertainty in the interpretation of the results. One of the major lessons of the UKPDS was to demonstrate that treatment of non-insulin-dependent diabetes with a single agent is not sufficient to attain the target goal of normalization of HbA1c. Patients in the UKPDS started at an HbA1c level of about 7%. Although the level of attained benefit in the pharmacologically treated group as compared with the diet-treated group was 0.9%, there was still a deterioration over 10 years to 7.9%, with no advantage for any one pharmacologic group.[38] Clearly, these levels are far from the target of normal HbA1c.

No one drug is capable of normalizing HbA1c in the vast majority of patients. This is particularly true in view of the progressive deterioration in control demonstrated in monotherapy in the UKPDS. However, each class of drugs shows additive benefits when added to other classes. Fortunately, metformin and the thiazolidinediones each reduce insulin resistance by different synergistic mechanisms. The combination of metformin and rosiglitazone has shown particular strength in combined treatment. The addition of sulfonylurea adds increased insulin secretion to the benefits of decreased insulin resistance. There is now increasing usage of multiple drugs in the treatment of type 2 diabetes. This change in physician perspective is due to acceptance in the medical community of the belief that even limited abnormalities of serum glucose are harmful. The UKPDS demonstrated a definite advantage of pharmaceutical management. In addition, many of the new drugs offer less risk of hypoglycemia than the sulfonylureas. Metformin, disaccharidase inhibitors, and thiazolidinediones do not provoke hypoglycemia unless coupled with the use of sulfonylureas or insulin. The new insulin secretagogues have a lower incidence of hypoglycemia because their action is mainly during and shortly after a meal.

The use of 3 or more drugs in combination is even now becoming commonplace. Oral hypoglycemics have always offered a needleless alternative to insulin therapy. Thus, patients have accepted regimens with multiple oral agents as a refuge from insulin treatment. However, it is very useful to support the effects of oral insulin sensitizers, such as metformin and the thiazolidinediones, with an augmentation of serum insulin levels. In this respect, sulfonylureas can only achieve limited improvement in endogenous insulin secretion.

There are 2 developments that may help increase the acceptability of insulin treatment. One is the development of glargine insulin. This synthetic insulin achieves very stable baseline insulin levels over a 24-hour period. Thus, a once-daily injection may suffice to augment insulin levels to a range in which insulin sensitizers can then achieve normoglycemia.[39] The second promised change in insulin treatment is inhaled insulin. The extensive lung surface permits a significant absorption of insulin. Several studies have now shown that inhalation of insulin before a meal can reduce glucose excursions as much as does injected insulin. Type 1 diabetic patients still require injections to provide a basal insulin level, which is then augmented by inhaled insulin. However, type 2 diabetic patients typically have significant basal insulin secretion. The use of inhaled insulin at meals could provide the necessary boost to allow insulin sensitizers, such as metformin and the thiazolidinediones, to maintain glucose control. If inhaled insulin does not show long-term deterioration of lung function, then the combination of inhaled insulin with insulin sensitizers will have a significant impact on glucose control.

Diabetes is one of the most common health problems, particularly of the aging population. Although its eradication is not likely with the use of drugs, we anticipate that multiple drug regimens will lead to improved glucose control in the diabetic population. It is expected that the new regimens will permit normal HbA1c levels to be achieved in most if not all patients. It is hoped that normalization of glycemic levels will then lead to a marked reduction in the diabetic complications that have afflicted so many people.