Presumed loss of clinical equipoise between the agents being investigated

To determine whether using a long-acting insulin analog at the time of diagnosis, instead of intermediate-acting insulin, affects the rate of loss of the body's ability to make insulin in children with newly diagnosed type 1 diabetes.

BACKGROUND: Type 1 diabetes (or insulin dependent diabetes mellitus (IDDM)) is a very common disorder affecting 1/400 persons by the age of 18 years and over 1 million people in the United States alone. An autoimmune disorder affecting the endocrine pancreas, it causes, in the untreated state, high blood glucose levels and ketosis. Over longer periods of time it can contribute to growth failure, malnutrition, and even death. It is responsible for a disproportionate percentage of the diabetes associated retinopathy and nephropathy seen in adults with diabetes (although type 1 diabetes only accounts for a small percentage of the total populace of patients with diabetes). Although insulin allows most children with this disease to grow and develop normally, it will not provide a permanent cure.

Many techniques have been studied to attempt to prevent the onset or progression of the autoimmune destruction of islets associated with type 1 diabetes, including a variety of immunomodulatory drugs, nicotinamide and small doses of insulin itself. Thus far nothing has proven completely successful. Some investigators have suggested that intensive control early in the disease process will slow the progression of the disease, but these results remain controversial. Ultimately, any treatment aimed at the prevention or cure of diabetes must have as its goal the preservation of the insulin secretory capacity of the pancreas.

In recent years, a number of insulin analogs have been developed which have different time-action profiles in humans. Currently, all children diagnosed with type 1 diabetes are placed on a combination of a long-acting insulin (such as insulins detemir or glargine) or a moderate-acting insulin (such as NPH), and a short acting insulin (such as lispro, aspart or regular insulin) as soon as they demonstrate they are able to resume a regular diet after their initial presentation. Each of the long- and moderate-acting insulins are given once or twice daily and the short-acting insulins at least twice and up to 4-5 times daily.

Normally, the pancreas secretes a basal level of insulin at rest, and then when challenged with a carbohydrate load, responds with an acute surge of insulin release. Intuitively, one might surmise that a pharmaceutical preparation that more closely mimics the normal physiological profile of the pancreas might be beneficial over the long term to pancreatic beta cell function. Levemir (insulin detemir) and Lantus (insulin glargine) are two relatively new insulin analogs known for their consistent and reproducible absorption profiles and steady time-action profiles. In comparison, insulin NPH is a product that had been the standard of care for children with diabetes for many years until the availability of the newer analogs, but is characterized by a peaking profile (with onset of action 2-4 hours after injection and peak effect 8-10 hours after injection). The combination of an insulin which mimics the basal insulin production of a pancreas at rest (i.e., glargine or detemir), together with a short-acting insulin with meals which mimics the bolus production of insulin generated in response to a carbohydrate load (regular, lispro or aspart), might allow for smoother blood glucose trends when compared to a combination of two "peaking" insulins (such as NPH and aspart).

In addition to the therapeutic benefit of better blood glucose control, we hypothesize that the maintenance of a steady baseline level of insulin with use of the newer insulin analogs may contribute to a longer period of pancreatic rest in the child newly diagnosed with diabetes. Many children with diabetes enter what is known as a "honeymoon phase" shortly after their diagnosis, which is characterized by relative ease of blood glucose control and relatively lower insulin requirements. This period represents a time during which the body is still able to make some insulin on its own.

We have retrospective data to suggest that children started on insulin glargine at diagnosis do, in fact, achieve significantly better glycemic control than age-matched children started on insulin NPH. This was assessed by HgbA1c measurements, which averaged a full percentage point lower at their 9 month visit for the glargine treated patients (Figure 1). A failure to see a significant improvement in patients switched from NPH to glargine well after diagnosis suggests that there is something to the initiation of treatment with glargine: perhaps the reason for improved control is a prolonged honeymoon period, in which the preservation of a small amount of innate insulin production allows for easier disease management.

What effect, if any, our current treatment modalities - specifically the choice of the longer acting insulin - have on the preservation of innate insulin secretory capacity remains unknown. If treatments aimed at the prevention or cure of diabetes are to maximize this secretory capacity, optimizing the insulin regimen may be imperative, and lack of attention to this parameter may confound trials of other interventions.

CONCISE SUMMARY OF PROJECT: Children aged 6-18 years who have been diagnosed with type 1 diabetes within the past 2 weeks will be randomized and placed in one of three treatment groups. 24 children will be randomized to each of three treatment arms differentiated by the choice of the longer-acting insulin (i.e. either detemir, glargine or NPH). All children will be treated with insulin aspart as the short-acting insulin to be used in combination with their longer-acting insulin. The insulin secretory capacity of the pancreas will be measured and compared between the groups by measuring C-peptide levels following a mixed meal tolerance test (using Boost) at 1, 6 and 12 months after diagnosis. Current standard of care as practiced at our institution is for these children to be seen every 3 months by a physician or advanced practice nurse at Children's Medical Center, Dallas. At each of these visits the children will have HgbA1c values measured and total daily insulin doses recorded. These will be secondary outcome measures for the purpose of this study.

Inclusion Criteria:

• Newly diagnosed type 1 diabetes within 1 week of diagnosis
• Age 6 - 18 years
• Care provided at Children's Medical Center, Dallas

Exclusion Criteria:

• Actual treatment with oral drugs influencing beta cell function or blood glucose levels (e.g. oral hypoglycemic agents)
• Actual treatment with drugs influencing insulin sensitivity (e.g. Metformin, or systemic steroids)
• Significant concomitant disease likely to interfere with glucose metabolism (children with active bacterial infections at the time of diagnosis must be cured prior to entry)
• Expected poor compliance
• Pregnancy
• Any other condition that by the judgement of the investigator may be potentially harmful to the patients

Subjects who completed randomization into one of the three treatment arms

1 year

Interpretation of the data is limited by the small sample sizes due to early study termination