Insulin Intensive Therapy: Current Options and Future Opportunities

At an industry-sponsored symposium held in conjunction with the American Association of Clinical Endocrinologists’ 2003 Annual Meeting and Clinical Congress, three leaders in the diagnosis and treatment of diabetes presented the latest information on basal and bolus insulin therapies and their incorporation into clinical practice to achieve enhanced patient adherence, improve therapeutic response and delay or prevent disease complications. Topics covered included current and future therapies using basal insulin, emerging bolus insulin therapies and insulin therapy for patients with unique needs.

This program was supported by an unrestricted educational grant from Novo Nordisk Pharmaceuticals, Inc.

Basal Insulin: Current and Future Therapies?

Insulin therapy has come a long way. Since the first publication announcing its discovery in 1921, many changes have taken place in the formulation and application of the medication.

The American Diabetes Association (ADA) publishes comparisons of onset, peak, and duration of action for insulin formulations. However, there are some questions as to how well they reflect what really happens when patients inject their insulin.

“The profiles of insulin do not differentiate between pharmacokinetic (PK) and pharmacodynamic (PD) properties,” said Irl B. Hirsch, MD, Professor of Medicine at the University of Washington in Seattle. “PK is the rate of appearance in the blood while PD is defined by the time for the biologic effect to occur. The two are not equal.” (Table 1)

As an example, Dr. Hirsch suggested that the rapid-acting insulin analogs might not be all that rapid-acting. Mudaliar and colleagues undertook a PD study using a glucose clamp with regular insulin and insulin aspart administered at .2 units per kilogram. They noted that the appearance of medication in the blood was much sooner that the biological effect which can last much longer. This illustrates the difference between PK and PD (Mudaliar SR, et al. Diabetes Care. 1999;22:1501)

Insulin absorption can vary greatly between individuals and even in the same person from one dose to another. Some of the causes of this variation include the insulin dose, site and depth of injection, exercise, lack of proper mixing of insulins, and poor resuspension.

“We have known for years that the intra-subject coefficient of variation for the same insulin can be 25% from injection to injection in the same person,” said Dr. Hirsch. “But even more amazing is that intersubject variations, between different subjects, can be as much as 50%. This helps to explain why our patients get frustrated when they do everything right yet their day-to-day blood sugars are so different.”

Improper mixing has a major impact on variability of blood sugars. A study by Jehle and others enrolled 109 patients who were given NPH insulin pens and told to resuspend the medication by rolling and tipping the pen 20 times. They found that only 9% of the patients tipped and rolled the pen more than ten times. The insulin content ranged from 5% to 214% and varied by more than 20% in 65% of the cartridges (Jehle PM, et al. Lancet.1999;354:1604).

An early study by Aventis presented to the Food and Drug Administration (FDA) looked at normal volunteers given 0.4 units per kilogram of insulin glargine, NPH or ultralente. They used a euglycemic clamp of 81 mg/dl. The PK data, especially for ultralente insulin, were hard to interpret due to endogenous insulin secretion. They found a tremendous variation across the insulins used but there was less variation with NPH than with insulin glargine (http://www.fda.gov/cder/foi/nda/2000/21081_Lantus_biopharmr.pdf).

“The C-max finding was very interesting,” said Dr. Hirsch. “Not surprisingly, the variation of the peak of the ultralente was the greatest and NPH the least. But there was more variability at peak for insulin glargine which has me scratching my head since I didn’t think it had any peak at all.”

Basal insulin can show major variations in PK profiles both between and within patients. Hypoglycemia often occurs when individual insulin profiles are outside of the mean PK profiles. Thus, variability may be more important than mean PK.

Insulin glargine is the first long-acting insulin analog to be marketed in the United States. Another analog soon to be released is insulin detemir (ID). But is it more predictable than insulin glargine?

When compared with NPH insulin, there was a reduction of 51% in the 24-hour variability and a reduction of 45% in the variability of C-max. Both are statistically significant suggesting lower within-subject variability (Strange P, et al. Diabetes. 1999;48(suppl 1):A103)

“Clinically this means less hypoglycemia with ID when compared to NPH,” said Dr. Hirsch.

Glycemic control is another factor in the decision making process. There was no difference seen in hemoglobin A1C or mean fasting glucose between ID and NPH. This suggests that reduced hypoglycemic risk with equivalent glycemic control with ID (Vague P, et al. Diabetes Care. 2003;26:590).

The pivotal studies for ID show that the standard deviation, a reflection of glycemic variability, is lower with ID than NPH in four of the five studies including both type 1 and type 2 diabetes.

Emerging Bolus Insulin Therapies in the Treatment of Type 1 and Type 2 Diabetes

The goals of intensive insulin therapy in diabetes treatment are clear-cut: maintain near-normal glucose, avoid short-term crises, minimize long-term complications, and improve the patient’s quality of life. Insulin is the most powerful agent available to control glucose.

“The rapid-acting insulin analogs have become important in treatment considerations because they are more convenient,” said Bruce W. Bode, MD, Medical Director of Atlanta Diabetes Associates in Atlanta, GA. “They can be administered at mealtimes, they mimic physiologic insulin, improve postprandial glycemic control, and there is lower risk for late hypoglycemia.”

Studies have shown that analogs such as insulin lispro and insulin aspart are more quickly absorbed than regular insulin. They also dissipate faster, controlling post-meal blood glucose with fewer incidents of hypoglycemia.

The more prolonged action of regular insulin was clearly evident in the pivotal studies for insulin aspart versus regular insulin. There were significantly fewer major hypoglycemic episodes at night among those receiving insulin aspart versus regular insulin. The relative risk of nocturnal hypoglycemia with insulin aspart was 30% and 50% lower compared to regular insulin in these studies (Data on file, Novo Nordisk Pharmaceuticals, Inc., Studies 035/EU and 036/US).

“Most physicians are familiar with treatment-to-target data comparing insulin glargine to NPH,” said Dr. Bode. “What you probably aren’t familiar with is some data that adding a short-acting analog is also very effective.”

Browdos et al, looked at a mealtime bolus of insulin lispro+sulfonylurea (SU) compared to NPH+SU or metformin+ SU. Although all were effective, the insulin lispro combination was more effective than NPH. The combinations reduced hemoglobin A1C levels by 2.3%, 1.9% and 1.9% respectively. However, there was more weight gain seen when the analog was given with meals compared to NPH administration at bedtime (Browdos, et al. Diabetes. 1999;48 (suppl 1): A104).

“Basal insulin use will lower the A1C around 2% by itself,” said Dr. Bode. “Pre-meal bolus will lower A1C by 2.3%.”

The basal/bolus insulin concept uses basal insulin administration to suppress glucose production between meals and overnight. Bolus insulins are given at mealtimes to limit hyperglycemia post-meal.

“I think you start with NPH or insulin glargine for the basal insulin,” said Dr. Bode. “The easiest thing to add is pre-meal insulin aspart or insulin lispro. Your other option is premixed insulin twice a day.”

Concerning oral agents, he suggests that metformin should be continued to lessen weight gain and decrease cardiovascular events. Glitazones can be continued to lessen insulin resistance and improve glucose control, though edema and fluid retention are common side effects.

Dr. Bode also suggested the starting insulin dose for multiple daily injections (MDI) be based on weight. Generally 0.2 times weight in pounds or 0.45 times weight in kilograms are the parameters. The basal dose of insulin glargine or NPH is 40% of the starting dose at bedtime. The bolus of insulin aspart or insulin lispro is 20% of the starting dose at each meal.

In addition to these two calculations, a correction bolus is needed as a supplement. Roughly the equivalent of a sliding scale, the Correction Factor (CF) is the amount of glucose that is lowered by 1 unit of short- or rapid-acting insulin.

CF can be estimated by dividing 1700 by the total daily dose (TDD) of insulin. So if the total dose of insulin is 36U then CF is equal to 1700 divided by 36. One unit of insulin will lower the blood glucose by nearly 50 mg/dl.

“A simpler regimen can be used by employing the continuous subcutaneous insulin infusion (CSII) pump,” said Dr. Bode. “The real advantage of the pump is a variable basal rate.”

Raskin and others looked at a head-to-head comparison of glycemic control in 127 patients using either MDI or CSII. Those on the pump received insulin aspart while those prescribed MDI regimens received insulin aspart and NPH. There were no significant differences seen in glycemic control. They did note significant differences in quality of life issues (Raskin, et al. Diabetes. 2001;50 (suppl 2):A128).

“The advantage of pumps is that you can eat what you want, sleep when you want and work when you want,” said Dr. Bode. “As a result, continuation rates for CSII are >95%.”

He suggested that CSII therapy can be appropriate in every patient with diabetes, including those newly diagnosed with type 1 diabetes. Other indications include A1C >0.5%, those with significant glycemic excursions, and a history of hypoglycemia unawareness or severe hypoglycemia.

To establish a starting dose, reduce the pre-pump total daily dose by 25% to 30%. The basal rate is 45% to 50% of TDD divided 24 for the hourly rate. The bolus dose is based on an individually determined carbohydrate-to-insulin ratio (CIR). CIR is equal to 2.8 times weight in pounds/TDD.

“Insulin remains our most powerful agent in the control of diabetes,” said Dr. Bode. “When used appropriately
in a basal/bolus format, near normal glycemia can be achieved.”

Insulin Therapy for Patients with Unique Needs

The glucose-balancing act in diabetes was originally viewed as a triangle with exercise, diet and insulin making up the corners. Now it is known that stress also impacts on this finely tuned equilibrium. The fourth corner of this square is stress, which has a greater impact on glucose control than the other three corners.

“Stress comes in three forms: physical, psychological and hormonal,” said Lois Jovanovic, MD, Director of Sansum Medical Research Institute in Santa Barbara, CA. “The body responds by secreting stress hormones which are anti-insulin by nature.”

Dr. Jovanovic suggests the following formula for changing insulin dosing due to stress. Weight in kilograms is multiplied by a stress constant based on the amount of stress present. With no stress the constant is 0.6, mild stress increases to 0.7, moderate stress 0.8 and severe stress, including life-threatening stress or ketoacidosis, can be up to 2 units per kilogram per day.

Both the amount and nature of this stress changes over the patient’s lifecycle. A child before puberty requires approximately 0.6 units of insulin per kg per day. When the pituitary begins to secrete the hormones triggering growth and development, the insulin needs rise proportionally.

“There is an old Chinese proverb that says children only grow while sleeping,” said Dr. Jovanovic. “The tropic hormones are secreted in higher amounts while they are asleep and we have to compensate by giving the majority of their insulin overnight.”

“There are also differences related to gender as the hormonal variations of the menstrual cycle change insulin requirements,” said Dr. Jovanovic. “Progesterone is very anti-insulin in nature and insulin requirements go up during the premenstrual or luteal phase of the menstrual cycle. The difference in insulin requirements can be 20% to 30% greater than during the follicular phase.”

As a woman continues through life, there are different sets of stresses and insulin needs. There is a rise in requirements during pregnancy. Career women may need to change their insulin doses based on psychological stress and changes in time cycles and eating patterns associated with work. When reaching menopause, the insulin requirement may actually decrease in those not on hormone replacement therapy.

“All of us should know about the unique needs of diabetes and pregnancy,” said Dr. Jovanovic. “Diabetes is the most common disease associated with pregnancy impacting on more than 85 women per 1,000 pregnancies.”

Normalizing blood sugar levels is the only way to improve diabetes-related outcomes in pregnancy. Glucose freely crosses the placenta, while insulin, unless bound to antibodies, does not. The fetus has its own insulin production capabilities, so too much maternal glucose crossing the placenta causes hyperinsulinemia and insulin resistance syndrome in utero.

It is the peak postprandial response, or highest blood glucose of the day, that predicts pregnancy outcomes. If the peak postprandial response is >120 mg/dl one hour after eating, the risk of macrosomia increases rapidly.

In order to maintain peak postprandial glucose in the target range, it is necessary to match insulin administration to food intake. There is still a need to restrict carbohydrates since it is very hard to keep the postprandial peak glucose in the acceptable range even when using rapid-acting insulins. The glucose goals are the same for gestational diabetes (GDM).

The availability of the fast-acting insulin analogs has been very useful in GDM. Insulin lispro was the first in class to be studied systematically during pregnancy.

To test for safety during pregnancy, Dr. Jovanovic recruited 40 women with GDM who were randomized to receive either human insulin or insulin lispro. The postprandial blood glucose peak with insulin lispro was significantly less than the human insulin group (Jovanovic L, et al. Diabetes Care. 1999;22:1422) (Figure 1).

Insulin aspart was also studied by Jovanovic and colleagues in GDM. Using insulin aspart, the peak postprandial glucose levels were also significantly lower than when GDM used human regular insulin. In this study, the researchers also looked at the second phase of insulin secretion and showed that insulin aspart not only improves the first phase, but also normalizes the second phase of insulin response thus facilitating dietary compliance by eliminating snacking between meals. The insulin aspart group also had a significantly lower IgG anti-insulin antibody response than seen with regular human insulin (Jovanovic L, et al. JAMA. 2001;286:2516).

Dr. Jovanovic noted there have been some case reports raising concerns about using insulin analogs in pregnant patients with type 1 diabetes. One concern is retinopathy associated with insulin lispro.

“It appears that rapid normalization of blood glucose, in those with a predisposing condition of the retina, may result in an increased risk for progression of retinopathy,”said Dr. Jovanovic. “However, it may be rapid normalization and not the insulin that results in retinopathy.”

A second concern with insulin lispro, triggered by a letter to the editor in the New England Journal of Medicine, is fetal malformation. To address these concerns a multicenter trial was undertaken with the results being officially released in the near future. In addition, there is a multinational trial ongoing to test the safety of insulin aspart in type 1 diabetic pregnant women.

There have been no trials in pregnancy for either type 1 women or GDM with the two long-acting insulin analogs, insulin glargine and insulin detemir. There is some in vitro data that might impact on the decision to use in pregnancy prior to appropriate clinical data being released.

Kurtzhals and colleagues showed that an in vitro assay of binding to IGF-1 receptors, insulin aspart had slightly less activity than human insulin, insulin lispro had significantly greater binding, insulin glargine had a six-fold increase in IGF-1 quality and insulin detemir had significantly less (Kurtzhals P, et al. Diabetes. 2000;49:999).

However, another publication by Robert Henry and associates using a different receptor assay found human insulin and insulin glargine have almost no activity up to very large pharmacological doses (Ciaraldi TP, et al. Metabolism. 2002;51:1171).

“The insulin analogs have fulfilled their promise of a better basal, a better bolus and thus a better blood sugar response,” said Dr. Jovanovic. “I am convinced that normal blood sugars before and during all pregnancies complicated by diabetes means a healthy baby.”