Diabetes mellitus represents a collection of genetically determined disorders characterized by hyperglycemia resulting from defects in insulin secretion, insulin action, or both which affect the metabolism of carbohydrates, proteins, and fats, associated with a relative or absolute insufficiency of insulin secretion (Fig.1) that is defined by hyperglycemic levels. Currently, nearly 17 million Americans are affected and the incidence is on the rise as it closely follows the nation's obesity crisis.

Figure 1. Diabetes is a heterogeneous group of diseases. IMD; Immune Mediated Diabetes (Type-1A).

Type-1 diabetes comprises those forms of diabetes that are primarily due to insulin deficiency. Type-2 diabetes comprises those forms that result from a primary defect in insulin resistance (often associated with obesity), coupled with a relative insulin deficiency. Eventually, progressive loss of ß-cell function in type-2 diabetes creates an absolute insulin deficiency, since long standing type-2 diabetes becomes complicated by glucose (glucosamine) and lipid mediated toxicities. These complications decrease the patients' abilities to secrete insulin to the point where it becomes a diagnostic confusion with type-1 diabetes.

Population studies have defined cut-off levels of glycemia that are eventually associated with increased micro-vascular disease, such as retinopathy. Two replicate fasting levels that exceed 126 mg/dl (>7 mmol/L) are diagnostic in the absence of symptoms. The new 2003 ADA’s definition of the cutpoint for normal fasting blood glucose levels was dropped from 110 mg/dl to 100 mg/dl, meaning that a value of 100 mg/dl or above would lead to a diagnosis of impaired fasting glucose (IFG), which is included in the term pre-diabetes. Persons with impaired fasting plasma glucose (FPG) levels (FPG= 100-125 mg/dl (5.66.9 mmol/l) and/or with impaired glucose tolerance test (IGT) (2 hour post-load glucose 140-199 mg/dl (78.8 mmol/L-11.1 mmol/L) are at risk of diabetes and should be observed periodically to detect progression to overt diabetes (1a). Replicate, two-hour glycemic responses >200 mg/dl (>11.1 mmol/L) after a standard oral glucose tolerance test also indicates diabetes. However, this stage is often reached before the fasting glucose levels rise. Indeed post-prandial hyperglycemia may precede fasting hyperglycemia by months to years. Thus, the reliance on only fasting glucose levels as recommended by the ADA expert committee is flawed (Table 1).

In this paper, we present our views on diabetes treatment and support our novel approach using early aggressive insulin replacement therapy.

TYPE-I DIABETES

Insulin treatment

The immune mediated form of type-1 diabetes (Type-1A or T1ADM) accounts about 10-15% of patients with diabetes. It is most common among Caucasian races and is rare in pure blood African-Americans, albeit the latter frequently develop an atypical form of maturity onset diabetes of youth or MODY. T1ADM results from the autoimmune destruction of the b-cells of the pancreas which leads to absolute insulin deficiency. It can be diagnosed in patients with diabetes in the presence of autoantibodies to islet cells and to islet protein enzymes and/or to insulin (Fig.2).

Figure 2. Different phases of natural history of immune mediated diabetes (Type-1A) and possible treatments.

However, an insulin deficient type of diabetes in the presence of the high-risk HLA phenotypes may be taken as presumptive, but not absolute, evidence. Adult onset IMD is characterized by a more gradual decline in insulin secretion compared to children, a situation readily confused with type-2 diabetes unless autoantibodies to islet cell cytoplasm (ICA) and/or to glutamic acid decarboxylase (GAD₆₅) and/or to the tyrosine phosphatases named insulinoma associated antigens (IA-2 and IA-2ß) can be detected. The term latent autoimmune diabetes of adults or LADA is used by some to describe this entity.

The other forms of T1DM have no known etiologies (Type-1B). There is no evidence of autoimmunity and it is not HLA associated. It accounts only a small percentage of those with T1DM.

The effect of treatment on the honeymoon period in Type-1A Diabetes

Various studies have demonstrated the beneficial effect of intensive insulin therapy to protect pancreatic b-cell function through the induction of b-cell "rest" in newly diagnosed T1ADM patients (2, 3). Numerous studies including the Diabetes Control and Complications Trial (DCCT) have shown that preservation of b-cell function in patients with T1DM diabetes results in better glycemic control and fewer end-organ complications (158-161). Use of exogenous insulin in Type-1A diabetes may provide "rest" for the b cells and preserve endogenous insulin secretion. Two studies by Kobayashi et al. (2, 4) of adult patients with IMD have shown that CSII (continuous subcutaneous insulin infusion) therapy preserves cell function over time compared with sulfonylurea therapy in ICA positive patients. It has also been reported that diazoxide, a K+-channel opener which inhibits the release of insulin, can preserve endogenous insulin in diabetic animal models and in insulin treated ICA positive patients. This supports the ß-cell "rest" theory (5-11); however, the recently concluded DPT-1 trial in which insulin was given to autoantibody positive relatives of patients with type-1 diabetes showed no protection against progression to diabetes. Arguably, this trial may not have provided sufficient insulin to induce effective ß-cell rest; however, the results were not encouraging that higher doses would be effective.

Pancreatic ß-cell "rest" or suppression of b cell function at the time of diagnosis of IMD may render insulin-producing cells less susceptible to immunological destruction because of their lowered expressions of type-1 diabetes autoantigens, such as insulin, GAD₆₅ and IA-2, and IA-2ß on resting ß cells. This would have the effect of making the b cells immunologically "invisible" to the immune system, which had been reacting against them (12). Recently, Schnell et al. (13) demonstrated that high dose IV insulin infusions and intensive insulin therapy, as initial treatment for newly diagnosed patients with type-1 diabetes, were equally effective in preserving insulin secretary capacity after a one year follow up period. Protection of the pancreatic ß-cells against complete destruction allows for some endogenous insulin secretion (appropriate to physiological signals), which is important to the reduction of acute metabolic disturbances and the maintenance of metabolic control.

Preventing complications: intensive treatment

The DCCT trial (14) and a prospective Swedish study (15) and a meta-analysis of 16 other randomized trials of intensified therapy in type-1 diabetes (16) documented that tight diabetes control can prevent complications (Fig. 3).

Figure 3. Good glycemic control reduces incidence of complications.

Different regimens are not equal in their abilities to preserve ß-cell function. Diabetes control is important to achieve from the time of diagnosis, even in children. Whereas a balanced lifestyle is important to childhood development and self esteem, we argue that this is best achieved by normalizing glycemia as much as possible without undue hypoglycemia or an unreasonable number of insulin injections. In our experience, this is best accomplished by continuous subcutaneous insulin infusion (CSII) (17). The same opinion was supported by several other studies (18, 19).

Though there are several ways to achieve tight glycemic control, our experience differs from that of Tsui et al. (20) and Reeves et al. (21), who concluded that different regimens of insulin treatment are similar in their improvement of overall blood glucose control, reduction in hemoglobin A1c (HbA1c), frequency of hypoglycemic events, and impact on the quality of life. However, our experience with newly diagnosed patients with IMD treated by CSII versus multiple daily injections (MDI) indicates better diabetes control with CSII at lower doses of insulin to keep the same level of HbA1c with less hypoglycemia, and for more convenient management (17).

Principles of insulin therapy: the different types of insulin and treatment schedules

Patients with IMD lack sufficient pancreatic insulin to maintain normoglycemia. Thus, management of IMD primarily focuses on adequate insulin replacement matched with food intake, as modified by exercise. Insulin need only be matched in units per gram of carbohydrate consumed, albeit some protein sources may have a significant glycemic index. Basal insulin, even in the short-term absence of food, is required throughout the whole day to prevent the development of a starvation state. Both short- and rapid-acting insulins as well as long-acting insulin preparations are needed to mimic the pattern of insulin delivery that normally controls blood glucose levels in non-diabetic individuals (22, 23). Throughout the day, short acting insulin is given to normalize blood glucose levels and to cover carbohydrates consumed during meals. Currently, we use CSII with lispro (Humalog) or aspartate (Novolog), or once or twice daily glargine (Lantus) insulin to mimic basal insulin secretion, in addition to intermittent Humalog injections, based upon glycemic corrections and carbohydrate food boluses throughout the day. In the stimulated phase after meals, insulin levels increase within minutes and peak at 15-30 minutes. Levels fall of to basal values within 2 hours (Figs.4 and 5).

Figure 4. Physiologic Insulin Secretion: 24-Hr Profile.

Figure 5. Plasma glucose is normally maintained in a narrow range.

Since the common daily diet includes three meals per day, short acting insulins should be given at least three times daily to prevent excessive hyperglycemic excursions. The dose depends upon the level of glycemia before the meal (SEE EXAMPLE BELOW). The difference between the measured blood glucose and the target of 120 mg/dl is used to calculate the correction bolus dose. This may range from 1-10 units for 100 mg/dl blood glucose, depending upon the age and body size of the patient. The starting correction dose begins with an estimation of the total average daily dose as divided into 1,500. The accuracy of this calculation is then modified by serial blood glucose level experiences. Meal boluses are calculated from an estimation of the carbohydrate content in grams and an individual factor relating insulin dosage to the amount of carbohydrate to be consumed. The range is from 1-5 units to cover 50 grams of carbohydrate. The starting carbohydrate bolus can be estimated by 500 divided by the total daily dose of insulin, but will need to be modified based on individual post-prandial glycemic responses. For infants, the intermittent, short acting insulin may be given after the meal when food consumption is unpredictable. In addition to the three main meals, additional amounts of short acting insulin may be taken at any and all times to cover snacks, and to reduce blood glucoses as necessary at bedtime. Correction boluses should not be given more often than at 3- hour intervals.

Most newly diagnosed patients with type-1 diabetes can be started on 0.25 to 0.5 units of insulin per kilogram. Adolescents often need more, but the dose can be adjusted every few days based on symptoms and blood glucose measurements. There are several ways to calculate basal and bolus doses (24, 25). The total daily dose of insulin can be reduced to 80% if the patient on MDI is in good control with a HbA1c < 7.5% and for HbA1c > 7.5%, start with 100% of TDD.

Example: How to Estimate an Insulin-to-Carbohydrate Ratio for 50 kg patient with a total daily dose 1 unit/kg/day.

1. Total daily dose of insulin for such patient will be 50 unit/day (1 unit/kg/day for 50 kg patient)
2. Insulin-to-Carbohydrate Ratio

500/50 =10 (therefore give 1 unit per every 10 grams carbohydrate eaten)

How to estimate correction factor for 50 kg patient with a total daily dose 1 unit/kg/day.

1500 =30 (therefore 1 unit of short acting insulin given will decrease blood
50 glucose by 30 mg/dl )

3. Target blood glucose-- we use 150 mg/dl in the beginning while training and later 120 mg/dl after patient is comfortable with the insulin schedule.
Example: In case of blood glucose 300 mg/dl and target blood glucose 150 mg/dl:
Correction factor = 300 -150/30 = 5 unit of short acting insulin ( therefore 5 unit of short acting insulin will decrease blood glucose to 150 mg/dl).

We confirm the accuracy of these ratios by frequent blood glucose testing.
Insulin injections are usually given subcutaneously because injections into the peri-umbilical area have the most rapid and consistent absorption kinetics (26-28).

The goal of treatment is to achieve a fasting blood glucose (FPG) concentration between 80 and 110 mg/dL, postprandial glucose between 100 and 140 mg/dL, and glycosylated hemoglobin (HbA1C) below 6.5 % (29-31).

Several insulin preparations are available: rapid-acting insulin (lispro or Humalog), (aspart or Novolog), short-acting preparations (regular insulin), long-acting insulins (neutral protamine Hagedorn [NPH], Lente insulins) and ultra-long-acting insulins (Ultralente, Glargine [Aventis Pharmaceuticals, Parsipanny, NJ]) insulins (Table 2).

In addition to insulin preparations presented in Table 2, there are pre-mixed (short and long acting) insulin preparations e.g. 70% NPH/30% regular (Humulin 70/30; Novolin 70/30), 75% lispro-protamine (NPL)/25% lispro (Humalog Mix 75/25) and 70% aspart-protamine/30% aspart (Novolog 70/30). Premixed insulins are convenient for some patients, especially the elderly. However, the ratio of insulins cannot be changed and this may lead to inadequate glycemic control or unexpected hypoglycemia episodes. Thus in our practice, we seldom use pre-mixed insulins except in some elderly patients.

Continuous subcutaneous insulin infusion (SCII)

Since its’ introduction in the late 1970s, CSII (insulin pump) has become an increasingly popular option for diabetes management. Because CSII is the most appropriate physiological regimen of insulin replacement, we prefer to use CSII in our practice. Today, the yearly increase rate of patients-treated with CSII is about 40% in USA. There are multiple models available and their operations continue to change with advancing technology. They contain multiple programs including basal and temporary basal rates, correction and carb boluses. They can be programmed according an individual’s life style. Pre-programmed CSII automatically gives basal insulin (unit/hour) according to that individual’s requirements and in addition temporary basal rates can be programmed for exercise or for inter-current infections. Again bolus doses can be automatically calculated by modern insulin pumps to help reduce calculation errors. Insulin pumps are as small as a pager (Figure 6) and can hold 2-3 day supply of rapid acting insulin (Humalog or Novolog) and deliver insulin via a catheter to the sub-cutaneous tissues. In CSII, the infusion site is best changed every 2-3 days to avoid skin infections and clogging up of the catheters.

Figure 6. Insulin pump and infusion set. (With permission of Animas Corp.)

The advantages of CSII are that insulin is taken only when needed, and not in an anticipatory fashion as with long acting insulins, insulin boluses are only taken as needed, no special diets as required, hypoglycemic episodes are minimal and the system is convenient and portable (38, 39). This increased flexibility has the greatest impact on quality of life (170).

Whereas CSII has traditionally been reserved for adolescents and adults, it is gaining more widespread acceptance in children especially in infants and toddlers (162-164). Although parental supervision is maximum in this age group to monitor BG levels, and to give multiple insulin injections, it is usually difficult to achieve near-normal metabolic control in this population by MDI because of the extremely small insulin doses required due to marked insulin sensitivity, erratic dietary habits with unpredictable food intake, and the varied activity level during the day and day to day, and frequent infections in this age group. As a result, infants face hypoglycemia and hyperglycemia episodes more often than older children. Therefore, CSII appears to be the ideal treatment option for this age group. In the past, there were concerns about safety and suitability of CSII in these young patients. To date, our experience and the multiple studies demonstrated that CSII is a safe and effective treatment for optimizing glycemic control which minimizes hypoglycemia episodes (165-169). Most modern pumps have a child lock feature.

We have found that CSII may even help with compliance where this has been a management problem and not the converse.

The one down side of CSII is that since only short acting insulin (Humalog effects are gone within 3 hours) are taken, any blockage (a kinked or damaged cathether) or pump failure or forgetting to put the pump back on after showers etc can lead to rapid onset of hyperglycemia and a rapid loss of control of diabetes. Adolescents, especially girls who use pumps, complain that the treatment is uncomfortable, embarrassing, or unpleasant, particularly when bathing or having sexual intercourse (40). The disadvantages of CSII although few, need to be explained before this form of therapy is begun. If the patient is not a candidate for aggressive insulin management (and a few are not), we modify their MDI regimen accordingly to the life style of the patient.

Continuous Glucose Monitoring System (CGMS)

Current approach to monitor blood glucose is limited with premeal and bedtime glucose measurements achieved four to five times daily. This approach, however, will not assess postprandial state and overnight blood glucose levels particularly nocturnal hypoglycemia. CGMS gives more detailed information on glycemic control with respect to the time of meals, impact of insulin dosages, exercise and overnight glucose profile. It measures subcutaneous interstitial glucose every 5 minutes over 48-72 hour period and the information can be downloaded for analysis. Today, CGMS has been used in patients with high variability of glucose values, or for assessment of glycemic control at night to optimize insulin therapy and metabolic control in patients with CSII (171, 172).

Multiple Daily Injections (MDI)

The next best regimen is intensive insulin treatment with long acting insulin substituting for basal insulin by CSII and short acting MDI before meals and as necessary with additional short-acting insulins at bedtime. While the choice of regimen is an individual matter of patient and physician preferences, we exclusively use glargine with short acting lispro or aspart in our practice. We prefer glargine (Lantus) insulin to all other available long-acting insulins to provide our basal insulin requirements because it provides day-long basal insulin without significant peaks of action (32, 33) (Fig.7), such as confound NPH.

Figure 7. Glargine vs. NPH Insulin in Type 1 DM.

However glargine does have a shallow peak of action, while in about a third of our patients, the single dose does not cover a 24-hour period (34), (35).

Glargine cannot be mixed in the same syringe with another form of insulin in a single syringe due to pH incompatibilities, and must be given in a syringe by itself. The burden of many injections of short acting insulin each day can be reduced by use of an insulin "pen", which is a convenient way of carrying multiple doses in a single dispenser. Both Humalog and Novolog pens are in wide use while a Glargine pen has only recently been introduced in the US.

Monomeric insulins, such as lispro (Humalog) and aspart (Novolog), have fewer episodes of hypoglycemia as compared with regular insulin. They are effective in normalizing post-prandial blood glucose levels (36). A meta-analysis of 8 large multi-center trials representing over 1,400 patient-years of insulin treatment revealed that severe hypoglycemia occurred in 3.1 percent of patients during lispro treatment compared with 4.4 percent while taking regular insulin (37).

The twice-daily injection regimen, consisting of regular or Humalog/Novolog insulin and intermediate-acting insulin (NPH or Lente) as basal insulin, used to be the standard of care. The downside of this is that the morning dose of intermediate-acting insulin usually is often not sufficient to prevent a post-lunch rise in blood glucose. Moreover, the intermediate-acting insulin administered before the evening meal may not be sufficient to induce normoglycemia the next morning unless a larger dose is given, which increases the risk of hypoglycemia during the night (Fig.8). Unfortunately, this happened more often than we and our patients were aware of in the past. We do not use premixed insulin preparations because of variability in their actions (41). Ultra-lente has irregular absorption properties and so we seldom use this insulin either in our practices.

Figure 8. The effect of basal insulin given as NPH, lente insulin, or insulin glargine, and with monomeric insulin before each meal.

Treatment of associated autoimmunities

The clinical associations between IMD itself and other autoimmune diseases are well established. The presence of other organ-specific autoantibodies suggests that patients with IMD have a generalized tendency toward autoimmunity involving multiple endocrine glands and specific organs. All patients with IMD should be optimally be screened for the presence of adrenal, celiac disease related, steroidal cell, gastric parietal, and thyroid relevant autoantibodies at the time of their diagnoses, which itself should be confirmed as the IMD form of type-1 diabetes by ICA, GAD₆₅A, IA-2A, and IAA testing when clinical presentations are not classical for type-1 diabetes (Table 3).

IMD-immune mediated diabetes, GAD, glutamic acid decarboxylase; H+, K+-ATPase, the parietal cell protein pump; IAA, insulin autoantibody; ICA, islet cell antigen; TSH, thyroid-stimulating hormone; 21-OH, P450 steroidogenic enzyme 21-hydroxylase; 17-OH, 17a-hydroxylase; P450scc, P450 side-chain cleavage enzyme; IA-2, members of protein tyrosine phosphatase; AADC, aromatic L-amino acid decarboxylase, ANA-antinuclear antibodies, SMA-smooth muscle antibodies, LKM-antibodies against liver/kidney microsomes, SLA - anti-soluble liver protein antibodies.

Family members of a proband with IMD should also undergo these autoantibody studies, especially when the proband is found positive. Positive autantibodies should be followed for the relevant hormonal evaluations and treatments. Positive thyroid autoantibodies should be followed by thyroid function tests and coexisting hyper-or hypothyroidism treated. Similarly, the finding of positive 21-hydroxylase autoantibodies should be followed by the screening of serum electrolytes, recumbent renin and PM ACTH levels (Fig. 8). However it is outside of the scope of this article to cover all of these associated disorders.

Figure 9. Diagnostic work-up for association between immune-mediated (type 1) diabetes mellitus (IMD) and other autoimmune diseases.

TYPE-2 DIABETES

Treatment of type-2 diabetes

T2DM is a metabolic disease that characterized by both impaired insulin secretion and insulin resistance. Today, better understanding the pathophysiology of the disease leads to develop more effective treatment options. At present, at least six different major forms of therapy are available for T2DM: lifestyle modifications of diet and exercise, insulin replacement therapy, insulin secretagogues, biguanides, meglinitides, a-glucosidase inhibitors, and PPAR-g agonists. These options provide a huge number of possible combinations. To achieve good glucose control and prevent complications of diabetes, patients are usually on multiple drug combinations. These combinations, however, may cause side-effects which can limit the efficacy of treatment. However, it is impossible to know which combinations produce the best long-term outcomes for individual patients, and it more clinical trial data is needed to address this issue.

The goals of treatment are to achieve physiologic control of blood glucose (BG) (HbA1c 6.5%, pre-prandial blood glucose 110 mg/dl and post-prandial BG 140 mg/dl) and to prevent/reduce complications and mortality (140). We believe that insulin sensitizers are the most optimal initial pharmacological agents with respect to safety and efficacy. We use metformin, but up to, 40% of patients may have transient G-I upsets. While PPAR-?g agonists can lead to fluid retention and weight gain. We often add a small dose of a PPAR-?g such as Avandia 4mgs to our patients who are unable to take full doses of meformin or are unable to be controlled by metformin alone. Exactly which additional agents to add once a patient fails to be controlled by this regimen remains unclear.

Conventional therapy for type-2 diabetes has taken a stepwise approach. First we prescribe lifestyle modification, then oral agents as monotherapy, and then combinations of oral agents. Sulphonylureas add to glycemic control, although at a risk of provoking hypoglycemia, since insulin secretion is promoted whether food is taken or not. This is an increasing problem in the aged. Another possible problem arises from the concern that such agents may shorten the period whereby appreciable amounts of insulin are capable of being secreted, as was the case in the UPD trial. Another approach is to add an a-glucosidase inhibitor such as acabose. These agents are poorly tolerated and lead to flatulence. Thus the dosage needs to be carefully built up to improve compliance. When oral therapy fails, exogenous insulin is often prescribed as the last resort, an approach that we believe is flawed since at the time of initial presentation of type-2 diabetes, 50% of patients already have significant macrovascular complications. We therefore advocate early consideration of insulin therapy to better control hyperglycemia and retain insulin secretory capacity.

Such an approach is supported by recent advances in understanding of the progressive nature of type-2 diabetes, gained from UKPDS, Kumomota and other studies. These studies proved the possibility of prevention of complications when the HbA1c is less than 6.5%.

Insulin therapy

It remains unclear when to initiate insulin replacement in T2DM. In the face of progressive impairment in insulin secretion with ongoing ß-cell dysfunction leading to insulinopenic phase, a large percentage of patients with T2DM will eventually fail on oral therapy for intensive glycemic control and will require insulin therapy. Insulin therapy can be started as an initial therapy when diet/exercise alone fails. It was stated by a recent American Diabetes Association (ADA) consensus that "if glycemia goals are not achieved with combination therapy, then treatment with insulin is indicated". We hold that the benefits of good glycemic control which ultimately can only be achieved with insulin therapy in T2DM patients, albeit there is reluctance from both patients and physicians to initiate insulin therapy (78, 152). Previous studies showed better glycemic control when insulin is added to the oral anti-diabetic regimen and there are multiple studies that have shown better glycemic control can be achieved with insulin combination therapy than with insulin alone (153). Combination therapies allow use of reduced daily insulin replacements.

Several studies, have documented that intensive insulin therapy for up to 4 weeks improves insulin sensitivity in T2DM and in insulin resistant subjects, as measured by the glucose-insulin clamp method (86). In obese, non-insulin dependent diabetics, control of hyperglycemia for 1 month, led to improvements in both insulin secretion and action that persisted for at least 2 weeks after cessation of therapy (6-11, 81, 87, 88). Insulin therapy decreased hepatic glucose production and improved endogenous insulin secretion. The mechanism for this improvement in insulin sensitivity is presumably reduced glucose/glucosamine or lipid mediated pancreatic cell toxicities from improved glucose control.

In T2DM glargine insulin has been shown to produce less hypoglycemia than NPH insulin with less weight gains (79, 80). However such patients have a primary problem with excessive post-meal glycemia that long acting insulins do not often solve. Most require MDI form the outset of insulin replacement therapy. The studies of CSII in adults with T2DM revealed significant improvement in endogenous insulin and C-peptide secretion, reduction in hepatic glucose output, improved insulin sensitivity, and significant improvement in HbA1c comparing to twice-daily injections of regular and NPH (81, 82). Ryan et al. demonstrated that in newly diagnosed T2DM with elevated fasting glucose levels, a 2- to 3-week course of intensive insulin therapy by MDI can successfully lay a foundation for prolonged good glycemic control (78a). Studies with CSII treatment have shown that transient CSII can also induce long term glycemic control in newly diagnosed T2DM patients. These results could be due to improvement of b cell function, especially the restoration of first phase insulin secretion, could be the responsible for the remissions seen (155, 156). Li et al. demonstrated that short-term (2 weeks) CSII treatment in newly diagnosed T2DM patients with severe hyperglycemia induced adequate glycemic control and the patients stayed euglycemic without requiring anti-diabetic agents which is similar to the honeymoon period in T1DM (157).

Our own preferred approach is to add insulin therapy after diet, exercise and combination of insulin sensitizers have failed to keep normoglycemia and Hb1Ac levels to a normal 6.5%. We use the same principles of insulin therapy as discussed above in T1DM, giving preferences to intensive multiple short acting insulin injections (Humalog or Novolog) plus glargine (Lantus) insulin plus glargine or by CSII. Many patients find that CSII is more convenient than MDI when it has been established that they need multiple doses of insulin each day to control their blood glucose levels.

As mentioned earlier, the issue of when to initiate insulin treatment in T2DM is controversial. However, above observations clearly indicate that early insulin treatment in T2DM patients may preserve endogenous insulin secretion by improving b cell function. In some, the improved endogenous insulin secretion coupled to lowered insulin requirements may permit them to come off insulin therapy after some time, albeit relapses are eventually common with time.

Preventing complications: intensive insulin treatment

The principal aim of treatment of T2DM is to prevent complications. The large study United Kingdom Prospective Diabetes Study (UKPDS) (83) revealed that diabetes is a progressive disease with complications that are directly proportional to the level of glycemic control. Concomitant with the inexorable decline in endogenous insulin secretion in the UKPDS was a progressive increase in hyperglycemia, and HbA_1C levels regardless of the mode of treatment given. Thus, over the course of 15 years of T2DM, the proportion of patients able to use oral agents alone significantly declines, with most (some 90%) requiring exogenous insulin treatment (84, 85). With over 10 years of follow-up, intensive therapy resulted in an absolute 1% reduction in HbA_1C value over conventional therapy (Fig. 10). The 11% difference in HbA_1C was associated with a 12% lower risk in aggregate diabetes outcomes, with most of the reduction based on a 25% reduction in micro-vascular disease such as in retinopathy and nephropathy (Fig. 11). Increased atherosclerotic disease in T2DM may depend more on concomitant dyslipidemia and hypertension than on glycemic control.

Figure 10. The effect of intensive treatment results in reduction of HbA1C compared to conventional treatment.

Figure 11. UKPDS: Benefits of glycemic control in type 2 diabetes.

To delay or prevent T2DM by improving insulin sensitivity

The relationship between insulin sensitivity and insulin secretion predicts that the disposition index (DI), a measure of b cell compensation for insulin resistance, is the best measure of effective b cell function. As persons develop worsening insulin sensitivity, those who develop diabetes fail to show sufficient compensatory increase in insulin secretion to overcome heir insulin resistance. Both short-term and long-term changes in insulin sensitivity and in glycemia may affect the risk of developing diabetes. Therefore, it is possible to produce b cell rest by improving insulin sensitivity; the b cell "rest" hypothesis. The Troglitazone in Prevention of Diabetes (TRIPOD) study, which delayed or prevented onset of T2DM in high-risk Hispanic women (89), appeared to confirm the "b cell rest hypothesis." The diabetes rate decreased in the troglitazone-treated group to 5.5%, as compared to 12.5% in the placebo group (Fig.12).

Figure 12. Cumulative incidence rates of type-2 diabetes in women who returned for at least one follow-up visit after randomization.

The theory of b cell rest was invoked from studies on children and adolescents. Several studies revealed that metformin and diet may act synergistically to limit weight gain and improve glucose tolerance (90-92). The recently completed Diabetes Prevention Program (93) showed that metformin could delay or prevent the onset of T2DM. The 3-yr cumulative incidence of diabetes in the group overall was 28.9% in the placebo group, 21.7% in the metformin-treated group, and 14.4% in the intensive lifestyle group. However the intensive lifestyle intervention was more effective than metformin. We argue that as a result of these studies, and given the natural history of progression of IGT to diabetes, that the earlier use of insulin sensitizers and insulin replacement therapy will likely preserve endogenous insulin reserves (Fig.13). Ultimately, it should reduce long-term diabetes-associated complications.

Figure 12. Our approach to the treatment of evolving diabetes.

Medications

INSULINSENSITIZERS

Metformin: Metformin is approved for the treatment of T2DM in children as in adults, but is the drug of choice for the treatment of insulin resistance in the absence of diabetes too. Some have suggested that it is the gastrointestinal side effects of the drug that accounts for much of its action. However the drug is effective in T2DM without weight loss, being found to reduce hepatic glucose output and increase insulin sensitivity in muscles amongst other actions. Metformin has various mechanisms of action in insulin resistance. For detailed explanations, the reader should refer to the following references: (82-99). Metformin is safe in cases of treatment of insulin resistance in pediatric patients (90-92, 94, 95) and in pregnant woman to decrease extreme hyper-androgenemia and improve pregnancy outcomes (96, 97). Furthermore, obese patients in the UKPDS who were assigned initially to receive metformin rather than sulfonylurea or insulin therapy had a decreased risk of any diabetes-related endpoint and mortalities from all causes (83).

Our experience in treating obese children and adolescents with metformin is likewise very positive. We begin with 500mg or 850mg once daily with the evening meal and as tolerated, add a second dose with breakfast and than a third with lunch to an eventual dose of 850 mg TID in adolescents or adults. Extended release metformin is available making it unnecessary to take a midday dose. In children, the size of the metformin pill causes compliance problems and the use of a metformin suspension as in Riomet can be very useful. The much touted lactic acidosis from metformin use appears to have largely evaporated in recent years. We still do not give the agent to patients with chronic renal or congestive heart failure because of this possible side effect. We also do not routinely continue metformin throughout pregnancy out of prudence, and give MDI or CSII as needed until delivery.

Thiazolidinediones (TZDs) [The peroxisome proliferator activated receptor g-agonists (PPAR-gs): They are a group of ligand-activated transcription factors that govern numerous biological processes, including energy metabolism, cellular proliferation, and inflammation (98). PPAR-g agonists are effective as insulin sensitization but are less useful in patients who are trying to lose weight. The PPAR-g isotype is mainly expressed in adipose tissue where it stimulates adipogenesis and lipogenesis. It is the target of a group of anti-diabetic drugs called thiazolidinediones. These PPAR-g agonists have been shown to inhibit FFA release from adipocytes, increase FFA uptake and storage in adipocytes, increase adipocyte triglyceride synthesis and storage by induction of adipocyte glycerol kinase, decrease inflammatory proteins and adhesion molecules, decrease cytokine production, improve lipid oxidation, decrease 11b HSD type-1, reduce intra-myocellular lipids, reduce muscle insulin resistance (99-101), decrease PAI-1 expression in endothelial cells (102) and decrease testosterone levels in insulin resistance females (103).

There is no clinical experience with the use of thiazolidinediones (TZDs) in the prevention or treatment of T2DM in obese children. One reservation with their use has been with the instances of fatal hepato-toxicity seen with the prototype agent troglitazone, resulting in the FDA withdrawing the agent. This side effect has not been implicated with later agents of this class such as pioglitazone or rosiglitazone, albeit it should be monitored for. Thiazolidinediones as mentioned above commonly cause weight gains, although this appears to result from the accumulation of subcutaneous, rather than visceral fat (104). Other medications indicated for treatment of T2DM, such as a-glucosidase inhibitor acarbose (105), lipase inhibitors (106), sulfonylureas, meglitinides are not used widely in our pediatric practice as they are relatively ineffective and have significant side effects.

INSULIN SECRETAGOGUES

Sulfonylureas: We believe that the use of sulphonylureas in children with T2DM should be minimal. A typical initial sulfonylurea regimen consists of 2.5-5 mg of glipizide or glyburide taken 30 minutes before breakfast with another before dinner. Amaryl is a 24-hour sulphonylurea that can be given once daily at 2-8 mgs dosing and thus is the one that is used preferentially in our clinic. Sulfonylureas directly stimulate the KATP channel subunit containing the cytoplasmic binding sites for both sulfonylureas and ATP and result in the closure of the KATP channel and insulin secretion. However, as mentioned, hypoglycemia induced by a long-acting sulfonylurea may be severe and is a frequent problem (107) especially in the elderly. Our additional concern is that, such agents might enhance progression to b cell failure.

Glucagon-like peptide-1 (GLP-1): GLP-1 is one of the intestinal pro-glucagon-derived peptides synthesized from pro-glucagon in the lower gut, mainly distal ileum and colon (141). It is an incretin hormone secreted in response to food intake rich in fat and carbohydrate (142). GLP-1 stimulates insulin secretion, inhibits glucagon secretion, regulates food intake by delaying gastric emptying and induces satiety. In addition, it stimulates islet cell proliferation and differentiation while inhibiting apoptosis (142, 143). It is a promising drug for the future in the management of T2DM. However, GLP-1 is rapidly metabolized by the dipeptidyl peptidase-4 (DPP-4) enzyme. To overcome this, gGLP-1 receptor agonists and DPP-4 inhibitors have been developed and clinical trials are undergoing. Studies with GLP-1 receptor agonist Exenatide which has recently become available has shown significant decrease in BG levels resulting in a reduction of HbA1c with minimum side effects (nausea, mild hypoglycemia) (144, 145). In addition there was a marked weight loss in subjects treated with GLP-1. One relative problem with the agent is that it must be taken by injection before meals.

MEGLITINIDES

Meglitinides are structurally different than sulfonylureas, but act similarly by regulating ATP-dependent potassium channels in pancreatic b cells, thereby increasing insulin secretion. Repaglinide (PRANDIN) is a short-acting glucose-lowering drug with similar efficacy to the sulphonylureas. Repaglinide appears to act via different receptors than sulfonylureas. It is less effective than glyburide at higher blood glucose concentrations (108). Hypoglycemia is the most common adverse effect. However it can be given pre-prandially to reduce the hyperglycemic excursions following food ingestion.

Nateglinide (Starlix) is a new meglitinide analogue and a derivative of d-phenylalanine (173). Nateglinide mimics physiologic insulin secretion dynamics seen in healthy individuals by increasing early phase insulin secretion into the portal vein and in that way increases hepatic glucose uptake as well as hepatic glucose suppression (174). In contrast to sulfonylureas and repaglinide, nateglinide is a more potent agent to restore early phase insulin release with less hypoglycemia episodes (175). When administered before meals, nateglinide rapidly acts at the same pancreatic ß- cell K+ATPase channel as sulfonylureas and repaglinide but dissociates from the receptor within seconds (176). Therefore, delayed hyperinsulinemia and an increased risk of hypoglycemia are unlikely with nateglinide. Clinical trials have demonstrated that nateglinide can reduce postprandial hyperglycemia and thereby improve glycemic control (176-178).

a-GLUCOSIDASE INHIBITORS

Acarbose and miglitol are members of the a-glucosidase inhibitors. They inhibit the upper gastrointestinal enzymes (alpha-glucosidases) that convert carbohydrates into monosaccharides leading to slow absorption of glucose. The slower rise in post-prandial blood glucose concentrations improve glycemic control without increasing the risk for weight gain or hypoglycemia. Acarbose as well decrease LDL cholesterol and increase HDL cholesterol (109). Miglitol has similar efficacy (110). The main side effects of these drugs are flatulence and diarrhea. Slow increases in dosage minimize these adverse effects (111). Hepatic injuries have been reported with acarbose as well (112).

LIPASE INHIBITORS (ORLISTAT)

Orlistat (Xenical) inhibits pancreatic and gastric lipases, blocking absorption of approximately 30 percent of ingested fat (113). The agent is sometimes given as part of a regimen to induce weight loss and will contribute to correcting triglyceride elevations from the dyslipidemia associated with insulin resistance and T2DM.

Currently available oral antidiabetic agents are summarized in Table 4.

Non-invasive insulin treatment in the future

Insulin therapy is central to the treatment of people with T1DM. Intensive insulin therapy, in particular, is associated with better long-term clinical outcomes. Originally, insulin was administered intramuscularly. It soon became clear that subcutaneous injections were just as effective but considerably less traumatic. Over the years, researchers have suggested transdermal, oral, buccal, nasal, and pulmonary routes of administration as alternatives (146, 147, 148, 149). High permeability and large surface area of the lungs makes pulmonary insulin a viable alternative to injections. Very rapid absorption of insulin after inhalation mimics the time activity profile of fast acting insulin, and thus it is appropriate for pre-meal administration. It appears comparable to subcutaneous insulin on glycemic parameters for both T1DM and T2DM patients (179-181). Several pulmonary insulin delivery systems are in various stages of development. So far the results are promising but there are still long term concerns about safety and tolerability (150, 151). The possible risk of lung damage in particular is of concern, however to date there has emerged little evidence that this concern is valid. In children, the subcutaneous route is the only one recommended at present. There is a long history of attempts to develop novel routes of insulin delivery that are both clinically effective and tolerable. Nasal and oral insulins have been used in trials attempting to prevent T1DM in high risk persons.

Anti-obesity drugs

Obesity not only in adults but also in children has become an epidemic disease and a major public health problem worldwide. Obesity is associated with serious medical complications including T2DM, dyslipidemia, and hypertension. Especially, there is a strong correlation between obesity and the early onset of T2DM in the children. T2DM was known as an adult type of diabetes. However, over the past decade, there is an alarming increase in T2DM as a new diagnosis of diabetes in particularly adolescents. Prevention of obesity is a logical first step for the prevention of developing diabetes. The usual approach is to start with lifestyle modification (diet and age appropriate exercise programs). When lifestyle intervention fails, the last resource is pharmacotherapeutic agents for prevention of weight gain. However, because of their serious side effects many anti-obesity and appetite suppressant drugs have been abandoned. Today, sibutramine and orlistat are the only antiobesity drugs approved for children ³ 16 year and ³ 12 year respectively. Both sibutramine and orlistat found to be effective in decreasing BMI, HbA1C, improving dyslipidemia, fasting and postprandial BG levels (182-184). Sibutramine inhibits serotonin re-uptake and induces premeal hypophagia. The common side effects are mild elevation of blood pressure (182, 183). It may induce depression, anxiety and insomnia. Orlistat inhibits gastric and pancreatic lipase and thereby reduces triglyceride and cholesterol absorption. Orlistat side effects are mainly gastrointestinal problems (184). Fat soluble vitamin deficiency (A, D, and E) may be seen with orlistat.

Metformin is also effective in significant weight loss in obese adolescents (185).

Prevention and treatment of diabetic complications

Besides insulin replacement therapy for IMD and type-2 diabetes, co-existing hypertension and dyslipidemia should be aggressively treated.

ANGIOTENSIN CONVERTING ENZYME (ACE) INHIBITORS

Angiotensin converting enzyme (ACE) inhibitors are widely used for treatment of hypertension and microalbuminuria, and protection of the kidney against diabetes provoked damage (114-116). In animal studies, aldosterone blockade prevented myocardial remodeling and reduced myocardial and renal damage (117-119). Diabetes and hypertension promote development of atherosclerosis and renal impairment (120, 121). This understanding was reflected by recent guidelines published by the American Diabetes Association and National Kidney Foundation specifying that in patients with diabetes, BP should be lowered to <130/80 mm Hg in an attempt to prevent cardio-vascular events and preserve renal function (122). Two recent studies - Irbesartan Diabetes Nephropathy Trial (IDNT) (123) and Reduction of Endpoints in Non-insulin diabetes mellitus with the Angiotensin II Antagonist Losartan (RENAAL) (124) - demonstrated that blockade of the renin-aldosterone system protects the kidney from damage. The combination of ACE inhibitors and angiotensin II type-1 receptor blocker (ARB) has been proven beneficial for urinary albumin excretion (125-128). We use enalopril (Vasotec) at 5-20 mgs daily depending upon the age of the patient.

Poorly controlled diabetes induces rise in hepatic VLDL output and triglyceride levels. There is also a rise in total cholesterol, since 20% of VLDL is cholesterol. Whereas there may be a modest rise in LDL-cholesterol, a low level of the protective HDL-cholesterol is fairly constant with this atherogenic lipid profile. With severe (>500mgs/dl) and/or chronic hypertriglyceridemia, pancreatitis may result. This is a serious problem with a mortality of some 20% with an acute attack. Diet reduced in animal fat and administration of fibrates (eg gemfibrozil) should be given to combat established hypertriglyceridemia (129).

Fibrates lower triglycerides as mediated through the PPAR-a transcription factor, mainly in liver where it has an important role in FA oxidation, gluconeogenesis, and amino acid metabolism. Pretreatment of endothleial cells with a PPAR-a agonist (fenofibrate) reduced markers of inflammation such as vascular cell adhesion molecule-1 (VCAM-1) expression, CRP, fibrinogen, PAI-1 and IL-6 (130-132). The American Diabetes Association recommends use of the agents in children for elevated triglyceride level =150 mg/dl, to enhance efforts to maximize blood glucose control and achieve desirable weight. If triglycerides are =500 mg/dl, a significantly increased risk of pancreatitis is present, and treatment with a fibric acid medication should be given (129).

Statins inhibit 3-hydroxy-3-methylglutaryl CoA (HMG-CoA) reductase, the rate-limiting enzyme in the mevalonate pathway through which cells synthesize cholesterol. To compensate for decreased synthesis and to maintain cholesterol homeostasis, cells, particularly hepatocytes, increase the expression of LDL receptors, which increases the uptake of plasma LDL, the main carrier of extra-cellular cholesterol, resulting in lower plasma LDL concentrations. Decreased plasma LDL levels reduce the progression of atherosclerosis and may even lead to the regression of preexisting atherosclerotic lesions. Statins have important immunomodulatory effects as well, and are able to decrease the recruitment of monocytes and T cells into the arterial wall and inhibit T cell activation and proliferation in vitro (133, 134). If after 6 months of optimized blood glucose control and dietary intervention there is no significant improvement in lipid parameters, intervention based on LDL is proposed by American Diabetes Association in children (129):

• LDL 100-129 mg/dl: maximize non-pharmacologic treatment.
• LDL 130-159 mg/dl: "consider" medication, basing the treatment decision on the child's complete CVD risk profile, including assessment of blood pressure, family history, and smoking status.
• LDL =160 mg/dl: begin medication.

Low doses of aspirin inactivate the enzyme cyclooxygenase, which catalyzes the conversion of arachidonic acid to prostaglandins G₂ and H₂. These prostaglandins are precursors of thromboxane, a potent platelet pro-aggregant and vasoconstrictor. Low doses of aspirin (81 mg/day) are preferred. Aspirin should be used in diabetic individuals over the age of 30 years who are at high risk for cardiovascular events (78).

We will not discuss in this review benefits of various combinations of oral agents and oral agents and insulin, albeit most of them have additive beneficial effect. Our own approach is to start with low carbohydrate low animal fat diet, daily anerobic exercise plus metformin to tolerance. If normoglycemia is not achieved, we often add a small dose of Avandia (4mg) to our patients who are unable to take full doses of metformin or are unable to be controlled by metformin alone. In the cases where combination of insulin sensitizes fail to keep normoglycemia and a Hb1Ac less than 6.5%, we add insulin (glargine plus intermittent Humalog or CSII) early to provide intensive insulin management.

T1DM WITH UNDERLYING INSULIN RESISTANCE

IMD with underlying insulin resistance and obesity is a clinically confusing entity that is often misunderstood. Too often, a child or adolescent presenting with symptoms of insulinopenia in the presence of obesity and acanthosis nigricans will be given the diagnosis of type-2 diabetes, even when they develop DKA. However, positive islet cell autoantibodies will delineate a distinct group of IMD developing in obese children who happen to have coincidental insulin resistance syndrome. It may be that the latter condition dictates an earlier onset of IMD than would otherwise be the case. Currently four classic autoantibodies are available for confirmation of immune nature of the disease: islet cell antibodies (ICA), glutamic acid decarboxylase antibodies (GAD₆₅A), insulin antibodies (IAA), and tyrosine phosphatase antibodies (IA-2) (135, 136). Hathout et al (137) measured islet cell autoantibodies in phenotypic type-2 diabetic patients and found 8.1 % to be ICA positive, of which 30.3 % were GAD₆₅A positive and 34.8 % were IAA positive. The epidemic of obesity has become as pronounced in childhood as in adulthood where it leads to the development of insulin resistance much earlier. The combination phenotype should not mislead physician. We advocate that assay of islet cell autoantibodies in any patient who presents with insulinopenic symptoms of diabetes since it creates a management problem in the face of insulin resistance, since the diabetes is more severe with rapid deterioration to absolute insulinopenia and larger than normal requirements for daily insulin replacement (138, 139).

A possible explanation may be that increased insulin secretion as a compensation for the degree of insulin resistance may induce an increase in the quantitative expression of certain antigenic determinants in ß-cells and them more susceptible to immune mediated destruction.

Many physicians select a therapy according to the clinical phenotype of the patient. However, it is necessary to determine islet cell auto-antibodies in every patient with diabetes, and when found to prompt the initiation of intensive insulin therapy much earlier than would otherwise be the case. The routine recommendation of diet and exercise and early use of insulin sensitizers are clearly indicated as well to such subgroup of patients, according the same principles we have already discussed (Fig. 13).

Figure 14. Algorithm for the treatment of different forms of diabetes. IMD; immune mediated diabetes (Type-1A).

Future approaches: There are many attempts being made to induce immunological tolerance to islet cell antigens to treat newly diagnosed type-1 diabetes. Since none are yet viable, they will not be covered herein. However, several companies are investing in methods to measure blood glucose levels through implantable glucose sensors or by devices that sense blood glucose changes through the skin. The hope is that development of such sensors which function in the long term would create one half of a closed loop system, feeding glucose information to insulin pumps that are already widely used. Similarly, there are many attempts being made through industry to create safe appetite suppressing agents which would impact of the problem of insulin resistance syndrome and type-2 diabetes, since 90% of the latter patients are both obese and insulin resistant. One of the more interesting areas lies in hormonal perturbations, affecting GLIP-1, gastric Ghrelin or intestinal peptide YY (NYY).

CONCLUSION

The prognosis of type-1 diabetes continues to improve with advances in home blood monitoring, basal insulins with modest peaks of action and insulin delivery systems exemplified by insulin pumps. Further progress is awaited from implantable blood glucose sensing devices. To our minds, type-2 diabetes has emerged as the more serious form of diabetes, while prevention of it through attention to the predisposing factors are looming increasingly important to our public health. The US obesity epidemic continues unabated, with ever increasing numbers of the nation's obese children becoming irreversibly obese adults, replete with the insulin resistance in all of its' burgeoning complications, notably of progressive atherosclerotic disease, hypertension and type-2 diabetes. The only rational long term solution must lie in the realization that the epidemic has its' genesis in childhood and thus it must be that the interventional focus should be placed in early life. Long term therapeutic trials that can show the long-term benefits of aggressive prevention and intervention, initially targeting highly prone ethnicities, are urgently needed.

ACKNOWLEDGMENT

We remain grateful to our patients who ultimately teach us most of what we may know.