These pages are best viewed with Netscape version 3.0 or higher or Internet Explorer version 3.0 or higher. When viewed with other browsers, some characters or attributes may not be rendered correctly.

ORIGINAL ARTICLE

Overview of Current Therapeutic Options in Type 2 Diabetes

Rationale for combining oral agents with insulin therapy

John B. Buse, MD, PHD

The overriding goals of the treatment of type 2 diabetes are first to relieve symptoms and then to prevent acute and chronic complications of the disease. The former goal can be largely achieved by maintaining the fasting and premeal glucose levels substantially below the renal threshold for glucose, ~200 mg/dl. The latter goal is more difficult to achieve because it requires long-term patient and provider attention to the principles of self-care, a healthful lifestyle, cardiovascular risk factor management, and the more intensive glycemic goal of achieving near-normal glucose control. In most patients with diabetes, near-normal glycemia can be achieved with single oral agents, combinations of oral agents, or insulin. However, a significant proportion of patients, particularly those with more longstanding disease, do not achieve near-normal glucose levels despite large doses of insulin or combinations of oral agents. This is documented in a recent study in which maximum doses of troglitazone and micronized glyburide achieved a HbA_1c level of 7% in 41% of subjects and <8% in only 60% of patients (1) and in the Veterans Affairs Cooperative Study in Type II Diabetes (VACSDM), in which some patients did not achieve a HbA_1c<7% despite maximal educational intervention and large doses of insulin delivered as multiple injections (2).

This inability to control diabetes in some cases despite intensive oral or injected therapy is likely due to the dual pathophysiology of type 2 diabetes—a combination of insulin resistance and a progressive insulin secretory deficit. Clinical experience and clinical trial evidence suggest that combinations of insulin and oral agents provide for improvements in glycemic control and/or other endpoints thought to be important in diabetes management. After a brief discussion of the rationale for insulin–oral combination therapy, clinical trial data regarding these combinations are reviewed and then a framework for considering potential strategies that could drive the use of insulin and oral antidiabetic agents is presented.

RATIONALE— Type 2 diabetes has been classically thought of as a condition that can be managed initially with diet and exercise. Later, with the progressive failure of insulin secretory capacity, oral agents are generally used to promote insulin secretion (sulfonylureas and repaglinide), to improve insulin action in the liver (metformin), to improve insulin action in muscle and fat (troglitazone), or to delay the absorption of carbohydrates from the meal, allowing the delayed secretion of insulin to catch up with rapid carbohydrate absorption (acarbose or miglitol). In recent years, combinations of oral agents have been used to attack the pathophysiology of diabetes at multiple points in cases where insulin secretion is still moderate. Nevertheless, most patients with diabetes, given enough time and in the context of generally inadequate attention to lifestyle issues, will become truly insulin deficient and require insulin therapy for management, as their insulin secretory reserve/capacity is insufficient to drive insulin action in target tissues despite therapy with insulin-sensitizing agents. In the U.K. Prospective Diabetes Study (UKPDS), fully 22% of patients randomized to metformin and 30% of patients randomized to sulfonylurea were switched to insulin therapy by 6 years of study (3). Until recently, most patients were merely switched to insulin therapy and took one, then two or more injections a day. Purely from a theoretical basis, insulin supplementation in the form of injected insulin should still synergize with agents that provide for an improvement in insulin action in the setting of insulin resistance. The difficult issue from a clinical perspective is that insufficient clinical trial data are available to prove one way or another whether combinations of drugs provide a greater long-term benefit than single agents given more or less similar glycemic control. Thus, we are left to rationalize various patterns of practice in the absence of the ideal dataset to drive clinical decision-making.

First, we should agree on reasonable goals for glycemic control in patients with type 2 diabetes. As discussed elsewhere in the issue, the lessons of the Diabetes Control and Complications Trial among patients with type 1 diabetes (4), the Kumamoto study of nonobese patients with type 2 diabetes (5), and the UKPDS of more classic type 2 diabetes (6,7) suggest that improvements in glycemic control will provide for a ~20% reduction in the rate of development and progression of microvascular and macrovascular complications for each 1% decline in HbA_1c.

In any case, there is reasonable agreement that in patients with significant life expectancy (>5–10 years), treatment with a goal of achieving near-normal glucose control would provide for a significant reduction in microvascular complications. Roughly speaking, this near-normal control would correspond to HbA_1c levels of <7%, premeal capillary plasma glucose levels between 80 and 140 mg/dl, and 1- to 2-h postprandial capillary plasma glucose levels <180 mg/dl (8).

Review of reported monotherapy studies of oral agents document that few patients randomized to a study generally achieve such goals, regardless of the agents selected. Similarly, intensive insulin treatment in patients who have failed sulfonylureas has shown that truly massive doses of insulin are generally required. The VACSDM is very informative in that regard. In that study, male subjects between the ages of 40 and 69 years with type 2 diabetes treated with insulin or who had failed sulfonylurea were exposed to an intensive systematic educational program and phased treatment by algorithm with a goal of achieving a near-normal HbA_1c. At entry, subjects were treated with an average daily insulin dose of 23 U and had a HbA_1c of 9.3%. A total of 75 patients were randomized to the intensive treatment algorithm and were started on a single dose of bedtime intermediate insulin. With titration over an average of 9 months to an average dose of 61 U at bedtime (1 SD = 38 U), 12% of the patients achieved near-normal glycemia, though the entire group had a reduction of hemoglobin A_1c of 1.4% to a mean of 7.9%. The remaining 66 subjects went on to receive bedtime insulin and daytime sulfonylurea (BIDS therapy) with an improvement in HbA_1c from 7.8 to 7.4% over an average of 11 months despite no significant change in insulin dose. BIDS combination therapy provided for an additional 24% of the cohort achieving near-normal glycemic control. The remaining 48 patients who were not yet optimally controlled went on to receive two injections per day of insulin with an increase in the total daily insulin dose from 82 to 116 U per day over 7 months. HbA_1c was reduced in that group on average from 7.7 to 7.4% (NS), and an additional 31% of the patients achieved near-normal HbA_1c. Fully one-third of the initial cohort did not achieve the therapeutic goal of near-normal glycemia despite intensive education, lifestyle intervention, and two large doses of insulin. In the final phase of the protocol, the average patient had his or her total daily insulin dose increased from 113 to 133 U of insulin (1 SD = 80 U) divided in three or more injections, with an improvement of the average HbA_1c from 7.2 to 6.9% (NS). As a testament to the excellent lifestyle interventions, there was no excess of severe hypoglycemia or weight gain in the intensively treated group as compared with the group randomized to standard care. The conclusion of the study was that "intensive stepped insulin therapy...is effective in maintaining near-normal glycemic control for >2 years." However, the major benefit in regard to glycemic control was seen in the initial phase of therapy with a bedtime dose of insulin, with some modest benefit to BIDS therapy. Although 64% of patients went on to two or more injections a day of insulin and most achieved near-normal glycemic control, the additional benefit in reduction of HbA_1c was modest and not statistically significant. In a way, this result could be expected. Patients who fail to achieve reasonable glycemic control despite a large dose of bedtime insulin are by definition quite insulin resistant. To overwhelm such insulin resistance, truly massive doses of insulin are required, and the mechanism of action of insulin at these doses may be distinct from the usual insulin-receptor mediated mechanisms. The VACSDM suggests a benefit to the combination of insulin and sulfonylurea, which will be expanded upon later, and a rationale for using insulin-sensitizing agents in patients who do not respond to moderate doses of insulin.

INSULIN-SULFONYLUREA COMBINATION THERAPY— The theoretical benefits of combinations of insulin and sulfonylurea are not obvious. The major action of sulfonylureas is to promote insulin secretion in response to glucose. Insulin therapy supplements insulin production. Thus, the combination of the two merely provides a dual mechanism for increasing insulin levels. Nevertheless, substantial clinical trial data suggest that insulin in combination with sulfonylurea at least provides an insulin-sparing effect, reducing the total daily dose of insulin required to achieve a given level of glycemic control, and generally provides for a modest improvement in glycemic control (9,10). Some have criticized these observations as being inadequate to compensate for the additional cost of sulfonylurea therapy (11). It is incontrovertible that substantial numbers of patients achieve adequate control on the combination of insulin and sulfonylurea but do not or cannot with a single injection of insulin, as demonstrated in the VACSDM discussed above. Thus, BIDS therapy can provide greater convenience for patients by avoiding a second, generally morning, injection. There is a clear rationale to using the combination as a bridge in therapy in moving a patient from sulfonylurea therapy to insulin therapy. Maintaining the sulfonylurea creates the opportunity for a nonthreatening discussion with patients that a bedtime insulin dose will supplement their prior therapy, often making insulin therapy seem like less of a revolutionary change in their management. If it later becomes clear that two shots a day are necessary and that the sulfonylurea is not providing substantial benefit, it can be stopped then. Combination therapy has been associated with less weight gain in some studies and with a lower risk of daytime hypoglycemia in others. These benefits are thought to result from the more physiological release of insulin during the day in response to sulfonylurea, driven by increases in glucose with meals as opposed to the continuous absorption of insulin from subcutaneous depots in patients treated with daytime injections of insulin. Furthermore, the use of sulfonylurea in combination with insulin does allow for a ~30% reduction in the total daily insulin dose required to achieve a given glycemic target.

BIDS therapy, classically morning sulfonylurea combined with bedtime intermediate-acting insulin, is the best studied form of insulin-sulfonylurea combination therapy. Many other combinations have been examined and are arguably effective. There are limited clinical trial data to suggest that presupper mixed insulin (either premixed or patient-mixed) may work as well or better, particularly in obese patients who consume large evening meals and snack before bed. In the more obese patient, the larger insulin dose that can be delivered allows for substantial insulin action through the night, even when the dose is given at supper, because of the longer duration of action of insulin associated with larger doses of insulin.

A reasonable approach in patients in whom sulfonylurea alone fails is to maintain them on either long-acting or intermediate-acting sulfonylurea at a quarter-maximum to half-maximum dose and add insulin at bedtime. We ask patients to start with 10 U at bedtime and increase the dose by 5–10 U every few days while monitoring glucose in the morning on a daily basis. When the blood glucose reaches 120–150 mg/dl in the morning, we ask them to slow the pace of upward titration of the insulin dose until they achieve morning premeal capillary blood glucose levels that are generally <100 and >70 mg/dl. In patients with BMIs >30 and/or those who are unwilling to give up big evening meals, snacking, and sedentary evenings, we often start with a higher dose of 70/30 insulin before supper with a similar patient-driven algorithm of insulin dose titration. Particularly in patients treated with presupper 70/30, but also with patients treated with bedtime NPH, intermittent monitoring of blood glucose at mid-sleep helps confirm that asymptomatic hypoglycemia is not a significant problem. In patients with type 2 diabetes in general and in the setting of BIDS therapy in particular, severe and asymptomatic hypoglycemia is not as common or severe as in type 1 diabetes. Furthermore, checking mid-sleep blood glucose is not a major inconvenience for most older patients, as many are getting up to void anyway.

Glimepiride is the only sulfonylurea with a Food and Drug Administration (FDA) approved indication for combination therapy with insulin (12,13). There is no reason to assume that there are lesser benefits to combination therapy with other sulfonylurea agents, most of which have been studied in combination with insulin. Repaglinide is not addressed in this manuscript because no data on its use in combination with insulin are available, though there is no reason to believe that it may not provide benefits similar to those of sulfonylureas in this regard.

INSULIN-METFORMIN COMBINATION THERAPY— The combination of metformin and insulin is in fairly widespread use in the U.S. and has recently been approved by the FDA. The rationale for using insulin in combination with metformin is that combining insulin with an agent that is known to sensitize the liver to the action of insulin should be of additive if not synergistic benefit in reducing blood glucose, particularly in the fasting state. Furthermore, it is attractive to think that the documented effect of metformin to minimize weight gain in the setting of sulfonylurea therapy may likewise provide benefit in minimizing weight gain associated with insulin therapy. In fact, these effects of metformin in combination with insulin have been demonstrated in several randomized clinical trials and anecdotal experience, though published studies to date are in general either very small, very short, uncontrolled, or in abstract form (14–18). In general, studies in which metformin is added to insulin in poorly controlled patients have demonstrated an insulin-sparing effect of metformin with a reduction of total daily insulin dose and circulating insulin levels on the order of 25%, a reduction in HbA_1c on the order of 1–2%, and in some studies, a reduction in weight gain. In one of the larger and longer of these studies, which is fully published, Giugliano et al. (15) randomized 50 obese subjects with type 2 diabetes who were poorly controlled (HbA_1c 12%) despite large doses of insulin (mean 90 U per day) to metformin or placebo and followed them for 6 months. The metformin-treated patients had a 1.8% reduction in HbA_1c with a 25% reduction in average total daily insulin dose, an 18% reduction in LDL, a 13% increase in HDL, and a significant reduction in blood pressure. Uncontrolled studies have documented an ability to withdraw substantial numbers of patients from insulin to treatment with metformin alone or with a combination of metformin and sulfonylurea, with a simultaneous improvement in glycemic control and weight loss (19).

Thus, two basic approaches are commonly used and supported by limited clinical trial data. The first is the addition of insulin to patients treated with metformin or with metformin plus sulfonylurea to improve glycemic control. The second is the addition of metformin, with or without sulfonylurea, to insulin treatment with an aim to improve glycemic control, reduce the insulin dose, and/or minimize weight gain. Either approach should be undertaken with gradual titration over a period of several weeks to first normalize fasting glucose and then premeal and postprandial glucose. Care to avoid use of metformin in patients with relative or absolute contraindications (renal, cardiac, hepatic, or pulmonary insufficiency and/or advanced age or alcoholism) is always a primary concern to avoid the rare but potentially fatal complication of lactic acidosis.

ACARBOSE-INSULIN COMBINATION THERAPY— Combination therapy with insulin and acarbose has been recently approved by the FDA in the U.S. The rationale for use of this combination is that acarbose therapy can reduce the postprandial rise in glucose by delaying absorption of dietary carbohydrate and perhaps reduce weight gain like its effect in minimizing weight gain in sulfonylurea-treated patients (20). In three U.S. and Canadian studies, a moderate reduction in HbA_1c (0.4–0.7%) and variably significant reductions in weight, triglycerides, and insulin dose have been demonstrated in insulin-treated patients with type 2 diabetes randomized to acarbose versus placebo (21–23). In a smaller study conducted in Japan with either acarbose or another -glucosidase inhibitor (miglitol), a reduction of HbA_1c at 6 months was significant in patients who consumed a diet rich in carbohydrates (>50%) but not in subjects consuming diets with <50% of calories from carbohydrates (24).

Acarbose can be added to insulin therapy to improve postprandial glycemic excursions in the patient who consumes a diet rich in carbohydrates. The use of the monomeric insulin analog lispro may be an acceptable alternative (25). Likewise, insulin, particularly bedtime insulin, could be added to the regimen of an acarbose-treated patient with persistent fasting hyperglycemia. Patients treated with acarbose should be counseled to start with a low dose and increase the dose slowly to minimize flatulence as the major adverse event.

TROGLITAZONE-INSULIN COMBINATION THERAPY— The addition of troglitazone to the treatment regimen of insulin-treated patients with type 2 diabetes has been well studied and is effective in reducing insulin dose and improving glycemic control. This combination is approved by the FDA and was the initial indication for which troglitazone was approved in 1997. The rationale is to combine an insulin-sensitizing agent whose predominant effect is an improvement in the action of insulin in peripheral tissues, namely muscle and fat, with exogenous (e.g., peripheral) insulin administration. The benefit of this combination has been well established through twin studies that attempted to isolate the insulin-sparing effect of troglitazone from the effect of troglitazone to improve glycemic control in subjects with type 2 diabetes treated with insulin.

In the first study, patients with typical type 2 diabetes who had previously failed sulfonylurea therapy and were treated with 30–150 U per day of insulin with fasting C-peptide levels >0.5 nmol/l were randomized to placebo or 200 or 400 mg of troglitazone for 26 weeks in a double-blind fashion (26). They were subjected to a forced insulin dose reduction algorithm where the insulin dose was reduced by a quantity equal to 25% of the baseline dose at each study visit when the average fasting capillary blood glucose was <95% of the baseline value. Through this procedure, the average insulin dose in patients treated with 400 mg of troglitazone was reduced by 57%, along with a 0.4% reduction in HbA_1c, compared with an 18% reduction in total daily insulin dose in the placebo group without significant reduction in HbA_1c. This documents that troglitazone has a significant insulin-sparing effect.

Of greater clinical relevance, the second study randomized a similar group of patients to placebo or 200 or 600 mg of troglitazone and employed a treatment algorithm through which insulin dose was reduced only for a fasting serum glucose <90 mg/dl, two consecutive fasting serum glucose measurements between 90 and 109 mg/dl at study visits, or two fasting blood glucose measurements <100 mg/dl on two consecutive days at home (27). In the 600 mg troglitazone group, there was a 1.4% reduction in HbA_1c versus placebo despite a need to reduce the insulin dose by 28%. Thus, troglitazone is effective in improving glycemic control in insulin-treated patients. The major areas of concern raised in these reports are the 3.6 kg weight gain that occurred in the subjects treated with 600 mg of troglitazone during the first 16 weeks of the study and the substantial weight gain in an extension phase of the first study. Minimal increases in fasting lipid values were documented, as well.

Thus, troglitazone can reasonably be used in attempting to reduce the insulin dose in well-controlled patients treated with insulin or to improve glycemic control in poorly controlled patients treated with insulin. Further studies to examine the ability of a combination of troglitazone and sulfonylurea to improve glycemic control and reduce insulin dose in insulin-treated patients are underway. It is recommended that troglitazone be initiated at 200 mg once daily with food in insulin-treated patients and that the dose be increased at 2- to 4-week intervals depending on response, with the realization that the maximum effect of troglitazone may take 12 weeks to be expressed. Insulin dose reduction of 10–25% is recommended when the fasting plasma glucose is reduced <100–140 mg/dl. Patient counseling regarding potential rare severe hepatic toxicity is necessary and helps with patient compliance regarding the required monitoring of alanine aminotransferase levels for the first 8 months, every other month for the balance of the 1st year, and intermittently thereafter.

SUMMARY AND CONCLUSIONS— Adequate randomized prospective clinical trial data exist to support insulin combination therapy aimed at improving glycemic control for troglitazone and metformin, with more modest improvements documented for acarbose. Adequate randomized prospective clinical trial data also exist to support combination therapy aimed at reducing insulin dose for sulfonylurea, metformin, and troglitazone, again with more modest benefit seen with acarbose. Conflicting data suggest that metformin in particular, but also sulfonylurea and acarbose, may have modest effects in minimizing weight gain in patients treated with insulin.

Thus, as would be expected with five classes of oral antidiabetic agents with completely different mechanisms of action, each seems to provide potential benefit in insulin-treated patients with type 2 diabetes. By the end of the decade, essentially all the potential three-drug combinations involving insulin, as well as many higher-order regimens, should be reported. Very limited studies and anecdotal reports suggest that they all will provide more or less additive benefit. There are new short-acting monomeric insulin analogs and long-acting insulin analogs available and in clinical development. There are essentially no data on their use in combination with oral agents, but there is substantial rationale to suggest that they may have special benefits in certain populations.

For the purist, there is certainly a rationale for continued use of insulin as the sole antidiabetic drug in the management of diabetes. Insulin is without a doubt the best studied, most clearly safe drug for the therapy of diabetes and the only agent that has been documented to reduce complications. Concerns that insulin therapy in the management of type 2 diabetes may predispose patients to cardiovascular disease clearly exist based on epidemiological data in nondiabetic populations (28). However, there was no evidence of increased atherogenesis associated with insulin therapy in the UKPDS, and in fact, the Diabetes Mellitus Insulin-Glucose Infusion in Acute Myocardial Infarction (DIGAMI) Study (29) suggests that more aggressive insulin therapy in the peri–myocardial infarction period is associated with a substantial reduction in mortality, though the study raises as many questions as it answers.

Nevertheless, for the realists among practicing providers, there is substantial patient preference not to use insulin as a first-line therapy, driven to a large extent by a fear of needles that is ingrained from childhood. This concern seems to be waning with improved insulin delivery systems and may disappear entirely if oral or inhaled insulin become clinical realities. Thus, the question becomes: What strategies can be employed in clinical practice to guide decision-making regarding insulin combination therapy? I believe there are several rational approaches to help in clinical decision-making, based purely on theoretical and practical issues, as outcomes-based research is not available to guide us at this time.

Minimum cost strategy
Perhaps the most compelling strategy for using insulin in the treatment of diabetes is to minimize drug costs. The majority of patients with diabetes are over age 65, and most have Medicare as exclusive coverage for health insurance. As a result, many patients are not reimbursed for prescription drugs. The least expensive drug for the treatment of diabetes are the sulfonylureas. Thus, a minimum-cost strategy would almost certainly start with sulfonylurea and progress to the addition of bedtime or presupper insulin. As demonstrated in the VACSDM, if moderate doses of bedtime insulin do not allow one to reach glycemic targets, massive insulin doses and multiple injections will likely be required. Though insulin is relatively inexpensive, it is not free; there certainly would begin to exist a rationale for adding metformin, acarbose, or troglitazone once an insulin dose of ~1 unit per kilogram is reached. At that level, the cost to double the dose would exceed $30 dollars per month.

Minimum weight gain strategy
Weight gain associated with improved glycemic control is classically believed to be due to improved hydration, increased fat mass as a result of the loss of glycosuria, increased caloric intake in the face of symptomatic or subconscious hypoglycemia, and in the case of troglitazone, perhaps as a change in the biology of the fat cell. In any case, weight gain is an issue of considerable concern to most clinicians and of overwhelming concern to many patients. A strategy to minimize weight gain would almost certainly begin with metformin or acarbose in a patient consuming a high-carbohydrate diet, with the addition of the other agent if one was inadequate. It is as yet unclear what the relative effects of troglitazone, sulfonylurea, and repaglinide are with respect to weight and thus which would be preferable to add next in such a strategy. As sulfonylureas seem to have a modest weight gain–sparing effect in combination therapy with insulin, sulfonylurea would be added either simultaneously or before insulin administration in such a strategy.

Minimum injection strategy
The perceived unremitting desire of many patients to avoid insulin injections at any cost is a common and disturbing challenge to clinicians and our national effort to contain health care costs nationally. Nevertheless, it is an extremely strong motivating force for many patients. In this strategy, sulfonylureas, metformin, acarbose, and troglitazone could be added in any order. Insulin, probably as a bedtime dose to minimize the inconvenience, would only be added as salvage therapy. Though this is an important issue in day-to-day clinical practice, efforts to defuse it are critical to achieve good population-wide diabetes control, as it is a major impediment to achieving near-normal glycemia. Insulin should be listed among therapeutic choices in all discussion of antidiabetic agents, and providers should include an objective assessment of its documented efficacy. When approached in this manner, it is shocking how many patients will actually pick insulin as their first choice for therapy.

Minimum peripheral insulin strategy
The concern over the potential atherosclerotic effects of exogenous insulin has been widely touted in the lay press and by the pharmaceutical industry. The relationship between circulating insulin levels and cardiovascular risk in nondiabetic populations is incontrovertible. However, at least two models could explain the observation. In one, an unrecognized factor drives insulin resistance, with its concomitant elevation in circulating insulin, as well as accelerated cardiovascular disease. In this model, reducing insulin resistance and thus insulin levels would have no effect on vascular disease. Thus, insulin would be epidemiologically associated with atherosclerosis but not causative. In the alternative model, insulin resistance drives hyperinsulinemia, which drives atherosclerosis. In such a condition, insulin-sensitizing agents might reduce vascular disease if they did not simultaneously sensitize the atherosclerotic process with the target tissue of interest (muscle, fat, and/or liver).

In any case, this clinical strategy is increasingly driving clinical decision-making in the U.S. Practitioners need to critically analyze this issue while also recognizing that there are many benefits and barriers to using insulin. The barriers include 1) perceived patient preference, which, as discussed, is generally relatively plastic and amenable to discussion; 2) greater physician/staff effort to educate patients on insulin administration; and 3) concern regarding weight gain and hypoglycemia.

In any case, this strategy is analogous to the minimum injection strategy except that the order is perhaps critical, with troglitazone having the greatest efficacy in reducing insulin levels, metformin second, and acarbose third, with sulfonylurea increasing peripheral insulin levels less than injected insulin does. Arguably, long-acting potent sulfonylureas, such as glipizide-GITS (gastrointestinal therapy system) and glimepiride, may be associated with slightly lower insulin levels than other sulfonylureas. Similarly, repaglinide should be associated with lower fasting and premeal insulin levels than other sulfonylureas from a theoretical perspective, but published comparative data are not yet available. Again, in this strategy, injected insulin would be used only as salvage therapy.

Minimum patient effort strategy
Unfortunately, there are still many patients who for a variety of reasons are interested in or only capable of making a minimal effort with regards to their diabetes. For them, a major issue is the frequency of taking pills. This is arguably a bigger clinical problem than unwillingness to take insulin because it is more difficult to recognize and probably a bigger contributor to nonadherence to prescribed therapy. Questioning patients about their pill-taking history and their realistic ability to comply with prescribed frequency is important. In general, taking a sulfonylurea and/or troglitazone once a day (in the morning or evening) requires the least effort. The effort in taking troglitazone is increased by the need to monitor liver function frequently and its improved bioavailability when taken with food. On the other hand, this potential drawback to troglitazone therapy is mitigated in patients on monotherapy by the lack of a need for blood glucose self-monitoring to help recognize hypoglycemia or to determine the cause of minor and common symptoms like nervousness, fatigue, and sweating. Taking bedtime insulin is actually relatively well accepted by patients driven by this consideration, as they can do it at their leisure while getting ready for bed and they can be reminded to take their medicine by a spouse or significant other, who is more likely to be present at bedtime than in the morning. It is also worth remembering that a single dose of metformin at bedtime can be highly effective, as can a single dose of acarbose with the main meal in patients who eat sparingly except at one meal during the day.

In conclusion, there is overwhelming agreement in the medical community that diabetes therapy aimed at maintaining near-normal levels of glycemic control is of benefit in patients with significant life expectancies. There is substantial agreement that insulin is the best studied, most effective drug in the treatment of diabetes. There is a consensus that patients generally are moderately adverse to insulin treatment, at least as first-line therapy of diabetes. There is concern, though essentially no clinical trial data, that suggests that exogenous insulin may have some pro-atherogenic effects, at least in comparison with insulin-sensitizing agents, in the management of diabetes. Unfortunately, that leaves us with a clinical reality that a substantial number of patients and providers will be left with a great deal of uncertainty regarding the most appropriate course of action in the management of diabetes. Combinations of insulin and oral antidiabetic agents can provide for excellent glycemic control in the vast majority of patients, and yet there are huge gaps in even the short-term clinical trials necessary to support their use and no large-scale prospective randomized outcomes studies to suggest their benefit. For now, we are left with a variety of strategies that constitute guesses as to optimal care. Greater understanding is necessary to drive informed clinical decision-making in the future.

Acknowledgments— The excellent secretarial and administrative support of Laura Wagner, Katie Thompson, Dawn DaQuano, and Jennifer O'Lear is greatly appreciated. The counsel and critical thinking driven by Joe Largay, Elizabeth Bruntlett, and Camille Izlar was essential in the development of these concepts.

References
1. Rezulin [package insert]. Morris Plains, NJ, Parke-Davis, 1998

2. Abraira C, Colwell JA, Nuttall FQ, Sawin CT, Nagel NJ, Comstock JP, Emanuele NV, Levin SR, Henderson W, Lee HS: Veterans Affairs Cooperative Study on glycemic control and complications in type II diabetes (VA CSDM): results of the feasibility trial. Veterans Affairs Cooperative Study in Type II Diabetes. Diabetes Care 18:1113–1123, 1995

3. United Kingdom Prospective Diabetes Study Group: United Kingdom Prospective Diabetes Study 24: A 6-year, randomized, prospective, controlled trial comparing sulfonylurea, insulin and metformin therapy in patients with newly diagnosed type 2 diabetes that could not be controlled with diet therapy. Ann Intern Med 128:165–175, 1998

4. The Diabetes Control and Complications Trial Research Group: The effect of intensive treatment of diabetes on the development and progression of the long-term complications of insulin-dependent diabetes mellitus. N Engl J Med 329:977–986, 1993

5. Ohkubo Y, Kishikawa H, Araki E, Isami S, Motoyoshi S, Kojima Y, Furuyoshi N, Shichiri M: Intensive insulin therapy prevents the progression of diabetic microvascular complications in Japanese patients with non-insulin-dependent diabetes mellitus: a randomized prospective 6-year study. Diabetes Res Clin Pract 28:103–117, 1995

6. UK Prospective Diabetes Study Group: Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 352:837–853, 1998

7. UK Prospective Diabetes Study Group: Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet 352:854–865, 1998

8. American Diabetes Association: Clinical Practice Recommendations. Diabetes Care 20 (Suppl. 1):S23–31, 1997

9. Riddle MC: Combined therapy with a sulfonylurea plus evening insulin: safe, reliable and becoming routine. Horm Metab Res 28:430–433, 1996

10. Johnson JL, Wolf SL, Kabadi UM: Efficacy of insulin and sulfonylurea combination therapy in type 2 diabetes. Arch Intern Med 156:259–264, 1996

11. Peters AL, Davidson MD: Insulin plus sulfonylurea agent for treating type 2 diabetes. Ann Intern Med 11:45–53, 1991

12. Riddle MC, Schneider J: Beginning insulin treatment of obese patients with evening 70/30 insulin plus glimepiride versus insulin alone: Glimepiride Combination Group. Diabetes Care 21:1052–1057, 1998

13. Amaryl [package insert]. Kansas City, MO, Hoechst Marion Roussel, 1997

14. Golay A, Guillet-Dauphine N, Fendel A, Juge C, Assal JP: The insulin-sparing effect of metformin in insulin-treated diabetic patients. Diabetes Metab Rev 11:s63–s67, 1995

15. Giugliano D, Quatraro A, Consoli G, Minei A, Ceriello A, De Rosa N, D'Onofrio F: Metformin for obese, insulin-treated diabetic patients: improvement in glycaemic control and reduction of metabolic risk factors. Eur J Clin Pharmacol 44:107–112, 1993

16. Aviles-Santa L, Sinding J, Raskin P: The effects of metformin in poorly controlled insulin-treated type 2 diabetes mellitus (Abstract). Diabetes 47 (Suppl. 1):A89, 1998

17. Jaber LA, Slaughter RL: Efficacy of insulin-metformin combination therapy in patients with type 2 diabetes mellitus (Abstract). Diabetes 47 (Suppl. 1):A88, 1998

18. Bergenstal R, Johnson M, Whipple D, Noller D, Boyce K, Roth L, Upham P, Fish L, Debold R: Advantages of adding metformin to multiple dose insulin therapy in type 2 diabetes (Abstract). Diabetes 47 (Suppl. 1): A89, 1998

19. Bell DSF, Mayo MS: Outcome of metformin-facilitated reinitiation of oral diabetic therapy in insulin-treated patients with non-insulin-dependent diabetes mellitus. Endocr Pract 3:73–77, 1997

20. Wolever TM, Chiasson JL, Josse RG, Hunt JA, Palmason C, Rodger NW, Ross SA, Ryan EA, Tan MH: Small weight loss on long-term acarbose therapy with no change in dietary pattern or nutrient intake of individuals with non-insulin-dependent diabetes. Int J Obes Relat Metab Disord 21:756–763, 1997

21. Chiasson JL, Josse RG, Hunt JA, Palmason C, Rodger NW, Ross SA, Ryan EA, Tan MH, Wolever TM: The efficacy of acarbose in the treatment of patients with non-insulin-dependent diabetes mellitus: a multicenter controlled clinical trial. Ann Intern Med 121:928–935, 1994

22. Coniff RF, Shapiro JA, Seaton TB, Hoogwerf BJ, Hunt JA: A double-blind placebo controlled trial evaluating the safety and efficacy of acarbose for the treatment of patients with insulin-requiring type 2 diabetes. Diabetes Care 18:928–932, 1995

23. Kelly DE, Bidot P, Freedman Z, Haag B, Podlecki D, Rendell M, Schimel D, Weiss S, Taylor T, Krol A, Magner J: Efficacy and safety of acarbose in insulin-treated patients with type 2 diabetes. Diabetes Care 21:2056–2061, 1998

24. Hara T, Nakamura J, Koh N, Sakakibara F, Takeuchi N, Hotta N: An importance of carbohydrate ingestion for the expression of the effect of -glucosidase inhibitor in NIDDM. Diabetes Care 19:642–647, 1996

25. Feinglos MN, Thacker CH, English J, Bethel MA, Lane JD: Modification of postprandial hyperglycemia with insulin lispro improves glucose control in patients with type 2 diabetes. Diabetes Care 20:1539–1542, 1997

26. Buse J, Gumbiner B, Mathias NP, Nelson DM, Whitcomb RW, for the Troglitazone Study Group: Troglitazone use in insulin-treated type 2 diabetic patients. Diabetes Care 21:1455–1461, 1998

27. Schwartz S, Raskin P, Fonseca V, Graveline JF for the Troglitazone and Exogenous Insulin Study Group: Effect of troglitazone in insulin-treated patients with type II diabetes mellitus. N Engl J Med 338:861–866, 1998

28. Despres JP, Lamarche B, Mauriege P, Cantin B, Dagenais GR, Moorjani S, Lupien PJ: Hyperinsulinemia as an independent risk factor for ischemic heart disease. N Engl J Med 334:952–957, 1996

29. Malmberg K, Ryden L, Efendic S, Herlitz J, Nicol P, Waldenstrom A, Wedel H, Welin L, on behalf of the DIGAMI Study Group: Randomized trial of insulin-glucose infusion followed by subcutaneous insulin treatment in diabetic patients with acute myocardial infarction (DIGAMI study): effects on mortality at 1 year. J Am Coll Cardiol 26:57–65, 1995

From the Diabetes Care Center, Division of Endocrinology, Department of Medicine, University of North Carolina School of Medicine, Chapel Hill, North Carolina.

Address correspondence and reprint requests to John B. Buse, MD, Diabetes Care Center, Division of Endocrinology, Department of Medicine, University of North Carolina School of Medicine, CB 7172, Burnett-Womack 647, Chapel Hill, NC 27599. E-mail: jbuse@med.unc.edu.

Received for publication 6 July 1998 and accepted in revised form 21 December 1998.

J.B.B. has served on an advisory panel for Eli Lilly; holds stock in Warner Lambert; has received grant support from Eli Lilly, Novo Nordisk, Pfizer, Hoechst-Marion-Roussel, Bristol-Myers Squibb, Bayer, Glaxo-Wellcome, and Parke-Davis; and has received honoraria or consulting fees from Eli Lilly, Novo Nordisk, Pfizer, Hoechst-Marion-Roussel, Bristol-Myers Squibb, Bayer, Parke-Davis, and Glaxo-Wellcome.

Abbreviations: BIDS, bedtime insulin and daytime sulfonylurea; FDA, Food and Drug Administration; UKPDS, U.K. Prospective Diabetes Study; VACSDM, Veterans Affairs Cooperative Study in Type II Diabetes.

A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

This article is based on a presentation at a conference organized by the Indiana University Diabetes Research and Training Center. The conference and the publication of this article were made possible by an unrestricted educational grant from Eli Lilly and Company.

Copyright © 1999 American Diabetes Association
Last updated: 3/99
For ADA Related Issues contact CustomerService@diabetes.org

For Technical Issues contact webmaster@diabetes.org