Optimizing Insulin Therapy: New Approaches to MDI Therapy

The Need to Optimize Insulin Therapy in Patients with Type 2 Diabetes

The AACE considers type 2 diabetes to be an under-recognized but very serious disease that must be treated as aggressively as type 1diabetes. Moreover, as reported by Jaime A. Davidson, MD (University of Texas Southwestern Medical Center, Dallas), the current explosion in the prevalence of type 2 diabetes has reached the point at which it exceeds 10% in ten states, and is diagnosed with increasing frequency in children and adolescents. Cardiovascular death due to type 2 diabetes continues to increase, indicating that glycemic control in this disease remains inadequate.

Of the approximately $132 billion annual cost of diabetes care in the United States, 80% is expended on the treatment of end-stage complications such as stroke, myocardial infarction, amputation, and renal failure. It is estimated that whereas the average annual health cost of an individual without diabetes is approximately $2,500, the mean cost to the health system among diabetic patients in $13,000. For individual patients, the cost of care correlates with the concentration of glycosylated hemoglobin (HgbA1c) when it remains above the target of 6.0% established by the AACE. The respective glycemic control targets of the American Diabetes Association (ADA) and the AACE appear in Table 1. The overall treatment goal is to achieve and maintain the lowest A_1c possible without unacceptable hypoglycemia.

Although it was previously thought that A_1c was primarily dependent on the fasting plasma glucose, both
fasting and postprandial excursions of hyperglycemia are important components of generating A1c.

One of the important findings of the United Kingdom Prospective Diabetes Study (UKPDS) was that type 2 diabetes is a progressive disease involving genetic predisposition to insulin resistance and/or beta cell dysfunction and ultimate
failure, leading to impaired glucose tolerance and diabetes. Unfortunately, by the time the disease is diagnosed, it has typically been progressing for 10 to 12 years, resulting in as much as a 50% loss of beta cell function at diagnosis. Although animal studies suggest that some oral antidiabetic agents may prevent additional beta cell dysfunction, evidence from the UKPDS indicates that with oral therapies, progressive loss continues over the 10 years following diagnosis, and that the progression correlates with the steady loss of response to oral therapy (UKPDS 16. Diabetes. 1995;44: 1249). In addition, if the A1c is very high at presentation, oral therapy will be unable to control it to target. It is apparent, therefore, that many patients will require exogenous insulin for preventing dramatic glucose excursions earlier than has previously been understood. In support of this conclusion, it has been demonstrated that compared with normal subjects, individuals with type 2 diabetes experience delayed and blunted first-phase insulin responses to all meals (Polonsky KS et al. N Engl J Med. 1996; 334:777). They also experience prolonged hyperglycemia (Coates PA et al. Diabetes Res Clin Pract. 1994;26: 177).

In light of the foregoing findings, is the care of Americans with type 2 diabetes improving? One analysis of National Health and Nutrition Examination Survey (NHANES) data for the years 1997 through 2000 indicates that the number of diagnosed patients who remain untreated has fallen. Among those who are under treatment, there has been a decline in those taking insulin monotherapy and an increase in the number taking oral medication alone. However, there has also been a three-fold increase in patients taking insulin plus oral antidiabetic agents, though the absolute number remains small (Sayadah S et al. JAMA. 2004; 291:335).

But treatment does not necessarily translate into adequate control. In a study of 1997 NHANES data, for example, in which A_1c of 7.0% was the therapeutic target, 38% of patients taking oral monotherapy were adequately controlled compared with 27% taking insulin alone (Harris MI et al. Diabetes Care. 1999;22:403). Dr. Davidson suggested that this disparity results from insufficiently aggressive insulin treatment consisting of 10 units of NPH insulin or glargine at bedtime with no further adjustment. Although all patients with type 2 diabetes will eventually require insulin to achieve treatment goals, UKPDS data make it clear that almost half of patients require it early and that early initiation induces the most effective achievement of goal (UKPDS 33. Lancet. 1998;352:837). In the Treat-to-Target study, patients with poorly controlled type 2 diabetes were randomized to glargine or NPH insulin in addition to their oral medications. Over a period of 18 weeks, the percentage of patients not achieving the treatment goal (A_1c 7.0%) decreased from 97.5% to 33.8%, with no significant difference between the two types of insulin (Riddle M et al. Diabetes Care. 2003;26:3080).

For some patients, aggressive insulin therapy may not achieve the A_1c target when treated for elevated plasma glucose alone. These patients require additional short-acting insulin to treat the postprandial glucose excursions that contribute to high A_1c concentrations. A recent French study utilizing continuous plasma glucose demonstrated a significant correlation between fasting glucose and A_1c when the latter was greater than 8.5%. As the level falls below 8.5%, however, the majority of the A_1c is attributable to postprandial hyperglycemia (Monnier L et al. Diabetes Care. 2003; 26:881). Thus to achieve the therapeutic goal for A_1c in a patient with a concentration in the range of 7% to 8.5%, appropriate treatment consists of basal insulin plus a short-acting insulin analog for postprandial control. It is currently estimated, however, that 80% of all insulin-treated patients with type 2 diabetes are not receiving intensified therapy and, therefore, are not achieving and maintaining the 6.5% A_1c concentration recommended by the AACE. Consequently, despite their treatment with insulin, those patients remain exposed to the long-term risk of potentially life-threatening complications.

Intensive Diabetes Self-Management:
Implementing Insulin Therapy with Your Patients

As his segue from Dr. Davidson’s presentation, Richard M. Bergenstal, MD (International Diabetes Center, Minneapolis) noted that “just moving to insulin is not enough. It is how we use the insulin and how effectively we use it that determines how we control type 2 diabetes.” In addition to appropriate pharmacologic intervention, however, Dr. Bergenstal emphasized the importance of education for patient self-management of diabetes. He also stressed the role of dietitians for nutritional counseling and for preparing patients to match carbohydrate intake to insulin therapy. An important principle of self-management is that the patient and clinician form a contract that commits the patient to comply with all aspects of a plan based on an understanding of the contribution of each of its components to overall glucose control.

Oral antidiabetic agents alone or in combination will decrease A_1c concentrations in most patients, but will infrequently bring them to the target level of 6.5% recommended by the AACE. A critical decision in the management of diabetes, therefore, is the appropriate point for initiating insulin in combination with oral agents. A study for which patients are currently being recruited will attempt to evaluate the efficacy of insulin when introduced before the onset of type 2 diabetes and in its early stages.

Following failure of oral therapy, the addition of insulin typically improves A_1c performance whether the oral agent be metformin, a sulfonylurea, or a thiazolidinedione (TZD) (Riddle MC et al. Diabetes Care. 1998;21:1052; Yki-Jarvinen et al. Ann Intern Med. 1999; 130:360; Rubin et al. Diabetes. 48[suppl 1]:A110). Similarly, the addition of an oral agent to the treatment of a patient who is not achieving goal by insulin monotherapy may also improve A_1c levels (Bergenstal RM et al. Diabetes. 1998; 47[suppl 1]:A89). These combinations also have the benefit of less weight gain than insulin alone.
In the Treat-to-Target study, which compared NPH insulin with glargine, both in combination with oral agents, 57% of patients reached the therapeutic goal of 7.0% A_1c targeting fasting plasma glucose only (Riddle MA. Diabetes Care. 2003;26:3080). “This is an excellent improvement,” Dr. Bergenstal said, “but to address the remaining 43% of patients, we need to solve the contribution of postprandial hyperglycemia to A_1c elevation.” People with diabetes need both basal and mealtime insulin for optimum control of blood glucose. In some individuals, basal insulin combined with oral agents for mealtime coverage will achieve glycemic target levels. In other cases, however, where there is less insulin reserve, one must supply both exogenous basal and bolus insulin therapy, mimicking normal physiology as closely as possible.

These therapeutic insulin options were compared in the Initiation of Insulin to Reach A_1c Target (INITIATE) study in which 210 patients with type 2 diabetes were randomized to receive either glargine every 12 hours plus metformin with or without a TZD or the analog insulin aspart 70/30 before breakfast and dinner plus metformin with or without a TZD. All subjects had A_1c levels of 8.0% or greater. Significantly more patients reached A_1c targets of either <6.5% or <7.0% using the mix of 70/30 analog plus oral agents compared with basal insulin plus oral agents.

In another study involving 169 patients with either type 1 or type 2 diabetes comparing regular human insulin with insulin aspart, patients randomized to receive insulin aspart experienced 9.1% and 12.2% lower plasma glucose concentrations 1.5 hours after breakfast and dinner, respectively. A_1c levels were not significantly different (Boehm BO et al. Diabetic Med. 2002; 19:393). The significance of these findings for potential prevention of complications during long-term treatment has not been established.

The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial, currently in progress under National Heart, Lung and Blood Institute auspices with a goal of 10,000 enrolled patients with type 2 diabetes, will address whether optimal glycemic control (A_1c<6.0%) can reduce diabetes-related cardiovascular disease. The ACCORD trial will also evaluate the importance of obtaining a systolic blood pressure of less than 120 mmHg and of treating the dyslipidemia profile of elevated triglycerides and low HDL-cholesterol often seen in type 2 diabetes.

Based on current clinical evidence, as the ability of the pancreas to provide both basal insulin and bolus secretion at mealtimes decreases, aggressive basal/ bolus intervention is appropriate. Typically this consists of combination oral medication insulin (glargine or NPH insulin) therapy for addressing basal glycemic control with the addition of bolus insulin analog therapy such as aspart or lispro for stabilizing postprandial hyperglycemia. Among the oral agents, metformin is often preferred for weight control and TZDs for improved stability of glycemic control and for treating one of the primary defects of type 2 diabetes: insulin resistance. Both agents appear to reduce cardiovascular risk factors as measured by PAI-1s and C-reactive protein, two probable surrogate indicators of cardiovascular morbidity.

For intensified insulin therapy, CSII may be the best model of the basal/bolus concept even when oral agents are continued. Insulin analogs are both safe and efficacious when delivered in this manner rather than by MDI. When administered by pump, insulin aspart has been shown in one study to be associated with less crystallization and, therefore, with less clogging than with buffered regular insulin (p<0.05), as well as with a 30% relative reduction in hypoglycemic incidents compared with both buffered regular insulin and insulin lispro (p<0.05) (Bode B and Strange P. Diabetes Care. 2001;24:69). In an open-label, parallel-group trial of 127 pump-naïve adult patients, intensified insulin therapy was observed to be equally effective when insulin aspart was administered by CSII or by MDI (Raskin P et al. Diabetes Care. 2003;26:2598). In an eight-point comparison of the two methods self-monitored by patients, however, MDI administration was associated with statistically superior glycemic control at one time point: 90 minutes following breakfast (Raskin P. Diabetes. 2001;50[suppl 2]:A128). In general, patients with type 2 diabetes prefer CSII to MDI. The broader role of CSII in type 2 diabetes management merits further study.

New Basal Approaches to MDI Therapy

The development of human insulin analogs has provided clinicians with insulins that have more physiologic time-action profiles that more closely mirror physiologic insulin secretion by beta cells. Although basal insulin analogs have taken longer to develop than the rapid-acting analogs used for postprandial glycemic control, they are currently evolving as new options for treatment. Martin J. Abrahamson, MD (Harvard Medical School and Joslin Clinic) outlined the ideal pharmacodynamic properties of a basal insulin as they appear in Table 1. The ideal basal insulin should provide a continuous low-level and peakless concentration of insulin over a 24-hour period that decreases the risk of hypoglycemia. In a small study evaluating pharmacologic profiles, glargine was shown to be a peakless insulin while NPH insulin and ultralente had definite peaks after approximately 4 to 8 and 6 to 12 hours, respectively, following administration (Lapore M et al. Diabetes. 2000;49:2142).

There are two design approaches to the development of soluble basal insulin analogs. The first, of which glargine is an example, involves solubility at a slightly acidic pH in order to modify the isoelectric points of the insulin peptide. When this insulin is injected subcutaneously, it forms microprecipitants that are absorbed slowly. The alternative method is acylation of the insulin peptide using hydrophobic residues. This approach, used in the design of detemir insulin, involves solubility at a neutral pH and no microprecipitants. The lengthy action duration of insulin detemir results from slow absorption due to its binding with plasma albumin in subcutaneous tissue as well as in the intravascular space. Ninety-eight percent of the insulin is bound in this manner, leaving 2% as monomeric insulin to bind to the insulin receptor and exert its effect.

Clinical trials confirm that detemir insulin is absorbed significantly more predictably than NPH insulin (Strange P et al. Diabetes. 1999[suppl 1]:A103) and lowers fasting plasma glucose more consistently than NPH insulin in patients with type 1 diabetes (Vague P et al. Diabetes Care. 2003;26:590; Russell-Jones D et al. Diabetologia. 2000[suppl 2]:A51) and type 2 diabetes (Haak T Diabetes. 2003; 52[suppl1]: A120). Use of insulin detemir is also associated with less absorption variability than glargine (Heise TC et al. Diabetes. 2003;52[suppl 1]:A121).

When compared with NPH and regular insulin regimens, detemir insulin plus aspart use is associated with significantly less variation in mean plasma glucose levels (p<0.001) over a period of 20 hours, superior reduction in mean A_1c concentrations (p<0,05) after 18 weeks of treatment, and a significant reduction in the relative risk for hypoglycemia (p=0.014) (Hermansen K et al. Diabetologia. March 12, 2004 e-pub). Similar results have been reported from a study comparing insulin detemir plus aspart with NPH insulin plus aspart (Howe P et al. Diabetes. 2003;52[suppl 1]:A122).

In a long-term study comparing detemir insulin with NPH as basal insulin in subjects with type 2 diabetes suboptimally controlled by oral agents, detemir was associated with a significantly lower risk (p=0.016) of nocturnal hypoglycemia and weight gain (p<0.001) (De Leeuw I et al. Diabetologia. 2002;45 [suppl 2]:A257). Table 2 summarizes the features of insulin detemir compared with NPH insulin.