Insulin is the cornerstone of pharmacotherapy for the estimated 0.73 to 1.46 million persons with type 1 diabetes mellitus in the United States. While the peak incidence of Type 1 diabetes mellitus is around the time of puberty, about 25% of cases will present after 35 years of age. As progressively more aggressive targeted glycemic, blood pressure, and LDL cholesterol treatment strategies impact both microvascular and macrovascular diabetes complications and comorbidities, there is an ever increasing need for effective insulin therapy strategies for treatment of the individual with Type 1 diabetes mellitus. In conjunction with lifestyle and self-care management, meticulous attention, by both the patient and the health-care provider, to the insulin regimen prescribed through the use of either multiple daily insulin dosing or a continuous subcutaneous insulin infusion (CSII) pump will enable attainment of HbAlc and blood glucose targets.

That an HbAlc of < 7% can be achieved in the person with Type 1 diabetes mellitus has been clearly demonstrated in the Diabetes Control and Complications Trial (Diabetes Control and Complications Trial). Furthermore, this study demonstrated that improved glycemic control with intensive insulin therapy in patients with Type 1 diabetes mellitus leads to graded reduction in retinopathy, nephropathy, and neuropathy on complications. The Epidemiology of Diabetes Interventions and Complications (EDIC) follow-up study of Diabetes Control and Complications Trial subjects has more recently demonstrated that intensive insulin therapy also reduces cardiovascular morbidity and mortality.

As early as 1993, the Diabetes Control and Complications Trial research group recommended that intensive diabetes treatment be instituted in most individuals with Type 1 diabetes mellitus, unless contraindications to doing so existed. Furthermore, the Diabetes Control and Complications Trial demonstrated that intensive therapy is most effective in preventing complications when introduced during the first 5 years of diabetes. In 303 subjects with early Type 1 diabetes mellitus and residual beta-cell function who were randomly assigned to intensive or conventional therapy, those receiving intensive therapy were slower to lose residual beta-cell function than the conventional therapy group (risk reduction 57%) (4). In addition, intensive therapy in those with residual beta-cell function resulted in a lower HbAlc, a 50% reduction in risk for retinopathy progression, and a lower risk for severe hypoglycemia compared to those who received intensive therapy but did not have residual beta-cell function. It seems abundantly clear that intensive therapy should be implemented as early as possible, and be maintained for as long as possible in Type 1 diabetes mellitus.

The health-care provider must tailor an individualized insulin regimen for each person with Type 1 diabetes mellitus to enable targeted blood glucose control. In order to be successful in this regard, it is essential to have an understanding of the normal physiologic pattern of insulin secretion, the currently available insulin preparations and safe and effective methods for both initiating and adjusting the insulin therapy regimen. This chapter will provide a discussion of each of these key elements of insulin therapy for adults with Type 1 diabetes mellitus.

PHYSIOLOGIC INSULIN SECRETION

While the person with Type 1 diabetes mellitus has an absolute deficiency in ability to secrete insulin from the pancreatic beta cell, it is helpful to understand physiologic insulin secretion as we prescribe an insulin regimen to optimally meet the daily needs of the Type 1 diabetes mellitus patient.

Basal Insulin

The normal concentration of insulin measured by radioimmunoassay in the peripheral venous plasma of fasting humans who do not have diabetes is 0 to 70 (iIJ/mL (0-502 pmol/L). Basal insulin secretory profiles reveal a pulsatile pattern of hormone release, with small secretory bursts occurring about every 9 to 14 minutes, superimposed upon greater amplitude oscillations of about 80 to 150 minutes. The amount of insulin secreted in the basal state averages 1 U/hr.

The Prandial Insulin Response

Meals, particularly those incorporating carbohydrates and/or other nutritional stimuli of insulin secretion may induce up to a 4- to 10-fold increase in insulin secretion when compared to the basal state, which usually lasts for 2 to 3 hours before returning to the baseline. Rise in blood glucose concentration following intravenous administration of glucose cause a burst in secretion that peaks within 3 to 5 minutes and subsides within 10 minutes and is known as “first phase” insulin release (FPIR). If the blood glucose concentration remains high, then the rise in insulin secretion is sustained in a second-phase of insulin release. The average amount of insulin secreted per day in a normal human is about 40 U (287 nmol).

Loss of pulsatile insulin secretion is one of the earliest signs of beta-cell dysfunction in patients destined to have Type 1 diabetes mellitus. By the time of diagnosis, beta-cell insulin secretion is negligible to absent. Therefore, one should assume that the individual with Type 1 diabetes mellitus has absolute insulin deficiency and will always require exogenous insulin therapy to prevent ketogenesis and uncontrolled hyperglycemia due to gluconeogenesis.

When prescribing insulin for the person with Type 1 diabetes mellitus, one will attempt to mimic these physiologic basal-bolus patterns of insulin secretion. In order to safely and effectively do so, the prescriber must have a sound knowledge of currently available insulins.

TYPES OF INSULINS

With advances in recombinant DNA technology, it is now possible to produce large quantities of insulin with an amino acid structure identical to that of human insulin using strains of genetically altered Escherichia coli bacteria or yeast. All forms of insulin have identical physiologic effects. Insulins differ in their rapidity of time to onset of action, the time from subcutaneous injection to peak of action, and their duration of action.

When insulin is injected, six monomers are associated in a hexameric form. The time it takes for the hexamer to dissociate into monomers, which can be absorbed across the capillary basement membrane is a strong determinant of the time of onset of action, peak levels in the circulation, and duration of action. For example, regular insulin must first dissociate into dimers, then into monomers, a process that takes 30 to 60 minutes following administration of a subcutaneous shot. This phenomenon accounts for the need to dose regular insulin 30 to 45 minutes prior to a meal if it is to attenuate the postprandial glycemic excursion. On the other hand, rapid-acting insulin analogs dissociate more quickly into monomeric form following injection. This results in their shorter time to onset of action and ability to dose with the meal, or even at the end of a meal.

COMPONENTS OF THE PHYSIOLOGIC INSULIN REGIMEN

Insulins are divided for practical purposes into two broad categories, basal and bolus, based on their pharmacokinetics. Physiologic insulin replacement attempts to mimic normal insulin secretion patterns, and is used to meet an individual’s total daily insulin requirement that consists of the sum of basal, prandial, and correction dose insulin requirements.

Basal insulin refers to exogenous insulin per unit of time necessary to prevent unchecked gluconeogenesis and ketogenesis. It provides a constant background level of insulin that controls blood glucose overnight while the patient sleeps and between meals when they are not eating and the meal bolus insulin action has waned. When dosed appropriately, basal insulin should not cause hypoglycemia if/when the patient does not eat or ingests less food than was anticipated during a meal. In treating Type 1 diabetes mellitus, basal insulin needs will most commonly be met by: injection of once daily insulin glargine; once or twice daily insulin detemir; or by rapid-acting or regular insulin delivered subcutaneously via an insulin pump.

The term bolus insulin incorporates both prandial and correction doses of insulin. Bolus insulin is preferentially provided as one of the rapid-acting insulin analogs, e.g., aspart, glulisine and lispro, or may be provided as short-acting regular insulin. Prandial or meal insulin refers to insulin which covers the postmeal glycemic excursion. Efforts are made to match meal insulin doses to anticipated carbohydrate intake, which will be achieved either by a consistent carbohydrate meal plan or by “carbohydrate counting.” The latter refers to counting the number of grams of carbohydrate to be taken in a meal and calculating an appropriate dose of insulin to take with the food. An individualized carbohydrate to insulin ratio is based upon an estimate of known insulin sensitivity. (Further details are discussed below in the section on pattern management.)

Correction- or supplemental-dose insulin is used to treat hyperglycemia that occurs before or between meals despite administration of routine daily doses of basal and prandial insulin, and is taken in addition to these standing doses. When the patient with diabetes is ill or stressed, total daily insulin requirements commonly increase. This increase in insulin requirement is a result of release of insulin counter-regulatory hormones, predominantly cortisol and catecholamines, and to a lesser extent glucagon and growth hormone, which are released in the physiologic endogenous stress response. If correction-dose insulin is needed at bedtime, it should be administered at a reduced dose compared to other times of day to reduce risk of nocturnal hypoglycemia.

Table Salient Features of Insulin Preparations

Patients with diabetes mellitus 1 have absolute insulin deficiency and therefore require basal insulin replacement at all times to prevent diabetic ketoacidosis, even when they are unable to eat. Withholding basal insulin from the patient with Type 1 diabetes mellitus results in a rapid rise in blood glucose, by as much as 29 to 60 mg/dL/hr, with accompanying onset of ketonemia in approximately 2 to 3 hours, leading inevitably to diabetic ketoacidosis.

Basal Insulins: Long-Acting and Intermediate-Acting Insulins

Even during an overnight fast, the normal pancreas continues to secrete insulin. Basal insulin suppresses hepatic glucose production and ketogenesis and maintains near normoglycemia in the fasting state. When administered subcutaneously, basal insulins have a delay of 2 to 6 hours from time of injection into the subcutaneous depot that determines their individual time to onset and duration of action. In the setting of Type 1 diabetes mellitus, subcutaneous basal insulin is most commonly used in combination with bolus prandial insulin doses administered prior to each meal in the multiple daily insulin regimen. Basal insulins for subcutaneous injection may be broadly categorized into long-acting and intermediate-acting insulins. The salient features of each of the currently available basal insulins will now be overviewed.

Long-Acting Insulins

Insulin Glargine (Lantus). Glargine is a recombinant human insulin analog. It differs from human insulin in that asparagine at position A21 is replaced by glycine, and two arginines are added to the C-terminus of the beta chain. Because of these changes, insulin glargine is soluble in an acidic environment and forms a stable hexamer precipitate in the neutral pH environment upon injection into subcutaneous tissue. The hexamer precipitate allows for a delay in the onset of action and a constant release of insulin over a 24-hour period with no pronounced peak. It thus serves to provide basal insulin action over the course of a day. The mechanism of action of glargine is similar to that of human insulin, and on a molar basis its glucose-lowering effects are similar to those of human insulin. Because glargine is provided in an acid solution, it cannot be mixed with other forms of insulin as it would alter their absorption profiles. Its acidity also accounts for discomfort with injection in a small proportion (2.7%) of users. In clinical trials in patients with Type 1 diabetes mellitus, glargine when compared to twice daily NPH insulin has been associated with a reduced risk of hypoglycemia (particularly nocturnal hypoglycemia). Hypoglycemia is less likely to occur with once daily glargine dosing when it is taken in the morning. In about 10% to 20% of patients with diabetes mellitus 1, glargine must be taken twice daily to provide 24-hour coverage of basal insulin needs. In a smaller proportion of patients, there may be a modest peak approximately 2 hours after injection. In a comparison study between insulin glargine and detemir in adults with type 2 diabetes mellitus, insulin glargine showed a better adjusted HbAlc than that seen with insulin detemir (6.92% vs. 7.13%, respectively, p = 0.035).

Insulin Detemir (Levemir)

Also, a recombinant human insulin analog, detemir, which has a 14-carbon fatty acid (myristic acid) covalently bound to lysine at position B29 and threonine at position B30, is omitted. Fatty acid acylation enhances detemir’s affinity to albumin. Albumin binding allows for a protracted duration of effect predominantly via delayed absorption from the subcutaneous adipose tissue depot at the injection site. Detemir’s duration of action is longer than that of neutral protamine Hagedorn (NPH) insulin (Table 1). In one study, a detemir dose of 0.29 U/kg provided the same effect as 0.3 U/kg NPH, but with a longer duration of action (16.9 hours vs. 12.7 hours, respectively). The duration of action for insulin detemir increases dose dependently from 5.7 hours at a low dose (0.1 U/kg) to 23.2 hours at a high dose (1.6 U/kg). Detemir’s duration of action in some cases is less than 24 hours. This is particularly so when the total daily insulin requirement is low (< 0.1 U/kg/day) as may be the case in Type 1 diabetes mellitus. Therefore in persons with Type 1 diabetes mellitus, particularly those who are lean, detemir may need to be dosed twice daily to effectively meet basal insulin requirements.

Detemir has been shown to effect glycemic control in several controlled noninferiority clinical trials in Type 1 diabetes mellitus patients on a basal-bolus regimen when used either twice or once daily. Detemir is associated with less risk of hypoglycemia, particularly nocturnal hypoglycemia when compared to NPH insulin.

Data have demonstrated consistently across clinical trials to date in diabetes mellitus 1 that patients treated with detemir have less weight gain than those using NPH. Data regarding weight comparison for detemir versus glargine have not yet been published for Type 1 diabetes mellitus. For type 2 diabetes mellitus, it has recently been reported in a large observational (PREDICTIVE) study’s German subgroup analysis that modest clinical weight reduction (0.8 ± 0.2 kg, p < 0.0001) was observed when patients were transitioned from a glargine ± oral agent regimen to detemir. In addition, less weight gain was observed when detemir was compared to glargine for 26 weeks in a recent report from the 2006 International Diabetes Foundation meeting in adults with type 2 diabetes mellitus (+1.3 kg vs. 2.6 kg, respectively). The mechanism underlying detemir’s modestly favorable weight effects have not been elucidated to date. It has been suggested that enhanced activity in the brain may suppress appetite and/or that it may exhibit greater effects in the liver than in the periphery, thus restoring a more physiological mode of insulin action.

Intermediate-Acting Insulin

NPH insulin (Humulin N; Novolin N). NPH or isophane insulin is a crystalline suspension of insulin with protamine and zinc. Combination with protamine and low concentrations of zinc enhance the aggregation of insulin into dimers and hexamers after subcutaneous injection. A depot is formed after injection and the insulin is released slowly, providing an intermediate-acting insulin with a slower onset of action and a longer duration of activity (12-16 hours) than that of regular insulin. The duration of action of NPH insulin is variable; rarely some patients may require only one NPH injection daily; while others require three or more injections daily. NPH insulin is equipotent to the other basal insulins. NPH has variable absorption and peaks both of which can predispose to hypoglycemia, particularly when a meal is delayed or food intake is curtailed. For these reasons, NPH insulin is not commonly used in an multiple daily insulin regimen for Type 1 diabetes mellitus.

Bolus Insulins: Rapid-Acting Insulin Analogs and Regular and Inhaled Insulins

Rapid-Acting Insulin Analogs

Rapid-acting insulins are generally preferred as the bolus insulin of choice in intensive glycemic control regimens. Their rapid time to onset of action allows injection immediately before meals, whereas regular insulin must be given 30-45 minutes before meals to optimally match the glycemic excursions after a meal. The rapid-acting analogs glulisine and lispro are also indicated for injection at the end of or up to 15 minutes following a meal, which confers the potential for increased flexibility in meal scheduling and allows the person with Type 1 diabetes mellitus to take the meal-time insulin following eating. This latter feature is particularly useful when the caloric intake for a given meal is not certain, e.g., when eating out or when ill, as the analog may be dosed after the meal to match actual carbohydrate intake.

Conversely, the one clinical setting in which rapid-acting analogs may not be preferred is when they are to be given to meet all of the patient’s insulin requirements for a period of time, e.g., during acute illness, or when nothing is to be taken by mouth after midnight for a procedure the next day or when there will be a prolonged NPO period following surgery after an insulin drip is discontinued. The rapid-acting analogs have a relatively shorter duration of action (up to 4 hours) when compared to regular insulin. It is therefore preferable to continue basal insulin or use an insulin drip under such circumstances when at all possible. If this cannot be done, then rapid-acting insulin analogs must be dosed every 4 hours to prevent diabetic ketoacidosis, or the patient can take regular insulin every 6 hours until the usual basal insulin regimen can be resumed. This consideration is more likely to be an issue of concern in the hospital rather than in the outpatient setting. (Further discussed under sick day adjustments below.)

A meta-analysis of 42 randomized controlled trials (involving 5925 patients with Type 1 diabetes mellitus) that compared rapid-acting insulin analogs to regular insulin showed only a minor benefit of the rapid-acting insulin analogs in terms of HbAlc reduction. A moderate increase in the dose of basal insulin may be required when a patient is switched from regular insulin to a rapid-acting insulin for premealtime dosing, in order to meet insulin requirements between meals when the action of the rapid-acting analog has waned that were previously being met by the tail of action of regular insulin.

Regular insulin and the rapid-acting analogs are equipotent. In clinical trials comparing regular insulin to the rapid-acting insulin analogs, improvements in overall glycemic control have been similar; however, the rapid-acting insulin analogs may be superior to regular insulin in improving overall glycemic control when they are used via CSII.

The rapid-acting insulins are also particularly useful in addressing unexpectedly high blood glucose levels (e.g., between meals or in the setting of stress) because they will lower glucose levels more rapidly and without the prolonged effect of regular.

The teratogenicity and long-term safety profile of rapid-acting insulins in pregnancy are unknown, except for insulin aspart, which has been recently granted a category B pregnancy rating for Type 1 diabetes mellitus by the Food and Drug Administration (FDA).

Insulin Lispro (Humalog). It is also of recombinant DNA origin. It is Lys (B28), Pro (B29) insulin. The effect of this amino acid rearrangement is to reduce the capacity of the insulin to self-aggregate in subcutaneous tissues, resulting in behavior similar to that of monomeric insulin. This allows more rapid absorption from the subcutaneous depot following injection. Given intravenously, the pharmacokinetic profiles of lispro and human regular insulin are similar. Lispro was the first available rapid-acting insulin analog that closely matches circulating insulin levels to the time course of the increase in plasma glucose seen after ingestion of a carbohydrate-rich meal. Frequency of hypoglycemia is lower with premeal lispro than with regular insulin (6.4 episodes/30 days vs. 7.2 episodes/30 days, respectively). The rapid onset of action of insulin lispro is not blunted by mixing with NPH insulin just before injection. A meta-analysis of patients with Type 1 diabetes mellitus found that the incidence of severe hypoglycemia was 30% lower in patients treated with insulin lispro (« = 2327) when compared to regular insulin (« = 2339). Insulin lispro can also be administered via external CSII pumps. Pharmacodynamically, insulin lispro has an onset of glucose-lowering activity in 5 to 15 minutes and reaches mean peak plasma concentrations at 60 minutes when given subcutaneous. It has a duration of action of about 2 to 4 hours. After subcutaneous administration, the half-life of insulin lispro is about 1 hour. Intermittent subcutaneous injections of insulin lispro may be given within 15 minutes prior to or immediately after a meal because of its fast onset of action.

Insulin Aspart (Novolog). It differs from human insulin by substitution of aspartic acid for proline at position B28. This substitution also leads to a more rapid onset and duration of action analogous to those seen with insulin lispro when compared to regular insulin. Insulin aspart is administered by subcutaneous injection and is also approved for delivery via external CSII pump. Insulin aspart has an onset of action of about 15 minutes. It is therefore given immediately before meals (start meal within 5-10 minutes after injection). Insulin aspart has a peak glucose lowering effect at 60 minutes and exhibits a duration of action of roughly 2 to 4 hours. The half-life of insulin aspart following subcutaneous injection is about 80 minutes (34). On January 29, 2007 the FDA approved a pregnancy category B rating for insulin aspart [rDNA origin] injection, indicating that adequate clinical studies of its use in pregnant women have not revealed increased risks to the fetus. The approval was based on data from a study conducted at 63 sites in 18 countries (« = 322), showing that changes in glycated hemoglobin and rates of maternal hypoglycemia were comparable with insulin aspart and human regular insulin. Although the study was not large enough to evaluate the risk for congenital malformations, the use of insulin aspart compared with human regular insulin yielded fewer preterm deliveries (P < .053), consistently low rates of major hypoglycemia, a decreased risk for neonatal hypoglycemia (glucose < 2.6 mmol/L) requiring treatment, and reduced risks to the fetus. Outcomes with insulin aspart are comparable to those of human regular insulin.

Insulin Glulisine (Apidra™). It is produced by recombinant DNA technology utilizing a nonpathogenic laboratory strain of E. coli (K12). Insulin glulisine differs from human insulin in that asparagine at position B3 is replaced by lysine and the lysine at position B29 is replaced by glutamic acid. Insulin glulisine may be mixed with NPH insulin (Apidra should be drawn into the syringe first). Insulin glulisine is administered by subcutaneous injection and can be used for administration via external CSII pumps. Insulin glulisine has an onset of action of approximately 5 to 15 minutes, and also has a peak glucose lowering effect at 1 hour. The apparent half-life of insulin glulisine after subcutaneous administration is 42 minutes, compared to 86 minutes for regular human insulin. Intermittent subcutaneous injections of insulin glulisine may be given within 15 minutes before to 20 minutes after starting a meal.

Short-Acting Insulin [Regular Insulin (Humulin R; Novolin R)]

Regular Insulin. It consists of zinc insulin crystals in monomeric form in a clear solution. After subcutaneous injection it tends to self-associate, first into dimers and then into hexamers that must then dissociate prior to absorption as only the monomers and dimers can be absorbed to any appreciable degree (42). This results in a 30- to 60-minute delay in the time to its onset following subcutaneous injection, which practically speaking limits its flexibility in terms of convenience of time of administration relative to meals for the patient. Furthermore, since the peak glycemic response to a mixed meal is between 2 to 4 hours after ingestion, regular insulin may peak too late to allow targeted control of postprandial hyperglycemia. Finally, there is also a potential for hypoglycemia to develop as a late sequelae some hours after a meal, due to regular’s longer duration of action, which

often further limits ability to titrate it to tight postmeal blood glucose goals. Regular insulin can be administered via the intravenous, intramuscular, or subcutaneous routes, and it is used in CSII pumps.

The onset of action of regular insulin (100 U/mL) after subcutaneous administration begins approximately 30 minutes after injection with maximal effects occurring 2 to 4 hours later. The apparent plasma half-life following subcutaneous administration is approximately 1.5 hours with a duration of action of 6 to 8 hours. subcutaneous regular insulin must be given 30 to 45 minutes before a meal to allow matching of the insulin action to the postprandial blood glucose rise.

Regular insulin is approved for intravenous administration. When given intravenous, its onset of action is within 15 minutes with maximal effects occurring 15 to 30 minutes after injection. The plasma half-life of intravenous regular insulin is approximately 5 to 6 minutes and its duration of action is 30 to 60 minutes. This short half-life and duration of action has important practical implications. When a patient with Type 1 diabetes mellitus is being treated with an intravenous insulin infusion, e.g., for diabetic ketoacidosis, it is imperative to give the first shot of subcutaneous insulin at the time of drip discontinuation with sufficient lag time prior to stopping the intravenous to allow time to onset of the insulin that was given subcutaneous. This step must be taken, as there is rapid dissipation of the intravenous insulin’s action when the drip is stopped, in order to prevent gluconeogenesis and ketogenesis.

Finally, regular insulin has been developed for administration via inhalation (see below) and is in product development for possible delivery as an oral spray for absorption via the buccal mucosa.

Inhaled Insulin (Exubera). Inhaled insulin has been approved for use in the management of Type 1 diabetes mellitus in adults. It causes a rapid rise in serum insulin concentration (similar to that which occurs after subcutaneous aspart or glulisine are injected, and faster than that seen with subcutaneous regular insulin). It has a slightly longer duration of action than the rapid analogs.

The FDA has to date approved one inhaled insulin product and delivery device, Exubera. In this system, insulin powder is packaged in a foil blister, which is inserted into the device. A 1-mg capsule of Exubera provides the equivalent of about 2.7-3 U of insulin; the 3-mg capsule provides about 8 U. When the device is activated, the blister is pierced and the insulin powder is dispersed into a cloud in a chamber, which the patient then inhales through a mouthpiece. The bioavailability of this inhaled insulin preparation is approximately 10% to 20% that of a subcutaneously injected insulin dose. Decreased bioavailability is due to a combination of factors: loss of the insulin powder (~30%) through retention in the blister and the inhalation device, deposition in the oropharynx (~20%), and the tracheobronchial tree (~ 10%). Forty percent of an inhalation is delivered to the alveolar spaces from where it passes across the alveolar capillary membrane to the circulation. Exhalation of particles, breakdown by enzymes, and elimination by macrophages also have some impact on bioavailability. Bioavailability may vary among insulin delivery systems, amongst patients, and even within the same patient.

In a 6-month randomized trial of Type 1 diabetes mellitus patients (mean age 29 ± 14) in which premeal inhaled (« = 163) was compared with subcutaneous regular insulin (« = 165), mean glycosylated hemoglobin was reduced to a similar degree in the inhaled and subcutaneous insulin groups [-0.3% and -0.1%, respectively; adjusted difference -0.16% (CI -0.34 to 0.01)], with a similar percentage (23.3% in the inhaled insulin group vs. 22% in the subcutaneous group) of subjects achieving Ale < 7. Although 2-hour postprandial glucose reductions were comparable between the groups, fasting plasma glucose levels declined more in the inhaled than in the subcutaneous insulin group [the mean adjusted change in FPG was -35 mg/dL in the inhaled group, whereas in the subcutaneous group, there was a slight increase in FPG (4 mg/dL); adjusted treatment group difference -39.53 mg/dL (CI -57.50 to -21.56)]. Inhaled insulin was associated with a lower overall hypoglycemia rate [9.3% (inhaled) vs. 9.9% (subcutaneous) (risk ratio [RR] 0.94 [CI 0.91-0.97])] but higher severe hypoglycemia rate [6.5% vs. 3.3% (RR 2.00 [CI 1.28-3.12])] when compared to regular insulin before meals.

A systematic review of six randomly controlled trials comparing inhaled insulin with rapidly acting injections (three in Type 1 diabetes mellitus and three in type 2 diabetes mellitus) concluded that glycemic control was equivalent, but that patient satisfaction and quality of life was greater with inhaled insulin. These studies were limited in length (12 and 24 weeks), therefore long-term safety and/or pulmonary effects could not be established.

Clinical trials of to date, have been designed as noninferiority studies, therefore ability of inhaled insulin to enable attainment of glycemic goals (A1C < 6.5-7.0%) known to be effective in preventing long-term complications has not yet been fully assessed.

From the nonpulmonary perspective, both intradose variability in insulin absorption and the difficulty in making precise dose adjustments [Exubera allows variation by 1 mg (three regular insulin equivalent units) at a time] may preclude the use of inhaled insulin for Type 1 diabetes mellitus patients managed with intensive insulin regimens. Two other inhaled insulin delivery devices that are in development and clinical trials testing at present will each use a novel delivery system that may allow increased flexibility in dosing moving forward. The Type 1 diabetes mellitus patient who uses inhaled insulin also needs to take subcutaneous basal insulin.

Safety issues related to Exubera may be classified as nonpulmonary and pulmonary. Nonpulmonary risks include hypoglycemia, which is similar to that seen with use of other insulins. Several studies have found an increase in insulin antibodies with inhaled insulin, compared with those receiving subcutaneous insulin. Patients with Type 1 diabetes mellitus had higher levels of antibodies than those with type 2 diabetes mellitus. The presence of insulin antibodies had no correlation with HbAlc level, change in insulin dose, or incidence of hypoglycemia. The clinical significance of the presence of these insulin-binding antibodies is not yet established.

With regards to pulmonary effects, inhaled insulin has been under intense and ongoing scrutiny in terms of potential impact on pulmonary function as it has moved through the development, clinical trials, and approval processes. The most common pulmonary symptom associated with inhaled insulin is a nonproductive cough, that is reported more frequently in patients taking inhaled insulin than in those in the comparison group receiving subcutaneous insulin or oral agents [risk ratio, 3.52 (CI, 2.23-5.56); 16.9% vs. 5.0%, respectively]. There were no differences between patients with Type 1 diabetes mellitus or type 2 diabetes mellitus. Cough occurred within seconds to minutes after administration of inhaled insulin; it was mild and was not associated with changes in pulmonary function. Cough was noted early in the treatment course (within the first month) and diminished in frequency and severity over time. Cough may be seen in up to 21% of persons using Exubera inhaled insulin, compared to 4% to 8% for patients not using it.

Diabetes mellitus is known to affect the lung. The underlying mechanism(s) that cause change are unclear. Mediators of inflammation, such as IL-1, IL-6, and TNF are associated with insulin resistance. Reduction in inflammatory markers with tight glucose control has been reported implicating diabetes itself as a cause of systemic inflammation. It is possible that this inflammatory process is involved in the pathophysiology of diabetes related lung disease. It has also been postulated that lack of insulin in lung tissue causes increased oxidative stress and the production of free radicals. Pulmonary function test changes are seen in diabetes patients. When compared with nondiabetic adults, individuals with diabetes mellitus have some reduction in pulmonary function, namely lower average values for FVC and FEVi. Glycemic control is not felt to be as important to this deterioration as is the duration of the diabetes. Sandier and associates demonstrated that the lower mean DLCO/alveolar ventilation in diabetes patients was associated with a lower pulmonary capillary blood volume, hypothesized to be due to pulmonary microangiopathy, premature lung aging, or glycosylation-induced alterations in hemoglobin-carbon monoxide reaction rates (74).

In clinical trials, Type 1 diabetes mellitus patients receiving inhaled insulin had a decline in FEVi from baseline when compared to those in a comparison group treated with subcutaneous regular insulin [weighted mean difference, -0.031 L (CI -0.043 L to -0.020 L)]. The modest decline in FEVi seen with inhaled insulin was statistically significant. The decrease in FEVi was slowly progressive over the first 6 months but stabilized in studies of up to 2 years’ duration. Among patients with Type 1 diabetes mellitus, inhaled insulin was associated with a greater decrease in DLCO from baseline than was subcutaneous insulin [weighted mean difference, — 0.902 mL/min/mmHg (CI, -1.546 to -0.258 mL/min/mmHg)]). The decline in diffusing capacity of the lung for carbon monoxide (DLCO) was evident in studies of 24 weeks duration or less, although there was no difference in the 2-year study. In a 12-week crossover trial, DLCO returned to baseline after patients were switched back to subcutaneous insulin. Among patients with type 2 diabetes mellitus, there was no difference in DLCO from baseline between the inhaled insulin group and the comparison group in studies up to 2 years in duration. The modest decrease in DLCO does not have any recognized clinical correlates.

Before starting inhaled insulin, a baseline spirometry should be obtained. Following initiation of inhaled insulin therapy, repeat spirometry is recommended at 6 months and then yearly as long as there is no deterioration in pulmonary function.

Inhaled insulin is contraindicated in patients with any degree of pulmonary compromise. Pathology of the lung as well as other exogenous factors play a crucial role in the absorption, delivery, and systemic exposure of inhaled insulin. Several pulmonary conditions impact systemic exposure to inhaled insulin. In chronic smokers, the alveolar-capillary membrane is more permeable, increasing absorption of insulin by two- to fivefold; chronic obstructive pulmonary disease increases exposure by about 50%. Asthma decreases it by 20% to 30%. Acute smoking attenuates absorption, perhaps due to reversible constriction of the airways. Active smokers should not be started on inhaled insulin, and previous smokers must demonstrate at least 6 months of abstinence. It is also known that passive smoking decreases exposure to inhaled insulin by 20% to 30%.

DETERMINANTS OF INSULIN EFFICACY

Factors Determining Absorption of Subcutaneously Administered Insulin in the Ambulatory Patient

Understanding variables that influence rates of absorption of insulin from the subcutaneous injection depot enables one to develop a clear understanding of how they will act to impact blood glucose and of the importance of consistent timing of doses with regards to time of day and to meals. The degree of absorption of any insulin dose, both among patients and in the same patient, can vary from day to day by as much as 25% to 50%, leading to unexplained fluctuations in glycemic control. This effect is greatest with long-acting insulins and least with regular, lispro, aspart, and glulisine insulin. There is some suggestion that day-to-day variability of absorption is less with insulin determir compared to glargine, but the clinical significance of this observation has not been established.

Factors that influence insulin absorption include: the time course of dissociation of injected insulin in the subcutaneous depot, size of the subcutaneous depot, site of the injection, subcutaneous blood flow, impact of exercise on uptake from the injection site, presence of hypertrophy or atrophy in the subcutaneous injection site, and the presence of anti-insulin antibodies.

Time Course of Dissociation

As has already been mentioned, the time it takes for the insulin to dissociate into monomers, which can be absorbed directly across the capillary basement membrane is a strong determinant of its time to onset of action, to peak levels in the circulation, and its duration of action. For example, regular insulin must first dissociate into dimers, then into monomers, a process that takes 30 to 60 minutes following administration of a subcutaneous shot and is maximal 2 to 4 hours following delivery of the dose. On the other hand, rapid-acting insulin analogs dissociate more quickly into monomeric form following injection, which results in their shorter time to onset of action and an ability to dose these analogs with a meal.

Insulin Type

The type of insulin administered determines the time of onset, peak activity, and duration of action of subcutaneous administration.

Size of the Subcutaneous Insulin Depot

Variability in absorption is increased and net absorption is reduced with increasing size of the subcutaneous depot. While it is not common for the adult with Type 1 diabetes mellitus to be on high doses of insulin, in the patient who is taking large number of units of insulin in a given dose, e.g., over 50 to 100 U in a single injection, it is preferred to split the shot into two equally divided doses to decrease the size of the depot, thereby promoting efficacy and reducing absorption variability.

Injection Technique

Both the angle of needle entry and the depth of penetration affect the rate of insulin absorption. Very shallow insertion can cause a painful intradermal injection that will not be well absorbed. In comparison, a perpendicular injection in a lean area may result in an intramuscular injection, from which absorption is more rapid.

The recommended insulin injection technique is to use an area of the body in which about 2.5 cm (1 in.) of subcutaneous fat can be pinched between two fingers. The syringe, with a 0.5-inch microflne (27 G) or ultraflne (29 G or 31 G) needle, is inserted perpendicular to the pinched-skin up to the hilt and the insulin is then injected. The needle should be held in place for several seconds before being withdrawn to avoid insulin leakage after withdrawal of the needle.

Site of Injection

Potential sites for injection are the upper arms, abdominal wall, thighs, and buttocks. Insulin is absorbed most rapidly from the abdominal wall, slowest from the leg and buttock, and at an intermediate rate from the arm. At any of these sites, the rapidity of insulin absorption varies inversely with subcutaneous fat thickness.

Rate of Subcutaneous Blood Flow

The degree of absorption is also impacted by the rate of subcutaneous blood flow. Insulin absorption is reduced by smoking and increased by any increase in skin temperature induced by such things as exercise, saunas or hot baths, and local massage. These variations are more marked with regular and rapid-acting insulins than with long-acting insulins.

POTENTIAL COMPLICATIONS OF INSULIN THERAPY

PHYSIOLOGIC REPLACEMENT THERAPY INSULIN REGIMENS

INSULIN DOSING ADJUSTMENTS AND PATTERN MANAGEMENT

GUIDELINES FOR DOSING CORRECTION/SUPPLEMENTAL INSULIN

SPECIFIC PRACTICAL GUIDELINES FOR PATTERN MANAGEMENT

Core Insulin Adjustment Guidelines

Core insulin adjustment guidelines will address basic recommendations for correction of hypoglycemia, hyperglycemia, variations in food intake or level of physical activity, and days when the patient is sick or stressed. When glycemic control is suboptimal and both hyperglycemia and hypoglycemia are present, one should first address hypoglycemia and correct it. This approach is recommended for several reasons. First and foremost, the short-term hypoglycemia is a safety issue. In addition, if hyperglycemia is due to rebound from hypoglycemic episodes, e.g., nocturnal hypoglycemia leading to high fasting blood glucose or to sequential extra correction doses of insulin, then increasing the insulin dose to treat highs will only exacerbate the tendency for hypoglycemia, perpetuating a vicious cycle of lows and highs.

Adjusting for Hypoglycemia

In evaluating episodes of hypoglycemia, one must first establish whether the lows are explained or unexplained as this will impact whether or not insulin doses need to be adjusted as the corrective action of choice. An exploration of variables that may be causing the hypoglycemia should be undertaken. Is the hypoglycemia explained by a decreased food intake, e.g., skipped meal or bedtime snack; an increase in the number of insulin doses taken, e.g., serial correction doses to treat a high; an increase in the number of units of insulin taken in a dose, e.g., a large correction dose; or by an increase in physical activity? If the explanation was an isolated occurrence, then the corrective action is to try and avoid the circumstances that caused it, e.g., to carry a snack when it is likely a meal will be skipped. If it is known that the explanation is going to be an ongoing phenomenon, e.g., beginning of an effort to lose weight through a cut in caloric intake or initiation of a regular exercise program, then the responsible insulin is adjusted downward to avoid further recurrences, per the general guidelines for basal and bolus insulin adjustment discussed earlier in this section.

Guidelines for Treating Hypoglycemia

Simple carbohydrate is taken to treat hypoglycemia. The patient should be advised to avoid indiscriminately ingesting large quantities of food or calories-containing beverages (such as regular soda or juice) in response to symptoms of hypoglycemia, as this will contribute to subsequent hyperglycemia. In general for blood glucose of 51 to 70 mg/dL, treatment with 10 to 15 g of fast-acting carbohydrate is recommended; blood glucose less than or equal to 50 mg/dL is treated with 20 to 30 g. blood glucose should be retested 15 minutes after carbohydrate ingestion and repeat treatment taken as needed, based upon the blood glucose result. The patient should also be advised to eat a more substantial snack or a meal that was missed or is late following initial treatment of a hypoglycemic reaction. This will prevent a recurrence. Once blood glucose is more than 70 to 80 mg/dL, the patient can generally safely take an appropriate prandial dose to cover carbohydrate intake with the next scheduled meal to be eaten.

Adjusting for Hyperglycemia

Insulin doses will be adjusted upward when a pattern demonstrating hyperglycemia at a given time of day is present for 2 to 3 days in a row at the same time of day and the hyperglycemia is unexplained by increased food intake, inactivity, or the somogyi phenomenon (rebound hyperglycemia).

If the hyperglycemia is explained by an increased food intake or a decline in physical activity, it is preferable to correct the underlying lifestyle indiscretion rather than to raise the insulin dose(s). If fasting hyperglycemia is present, and particularly if fasting hyperglycemia is seen in association with wide variation in blood glucose values, including the presence of normal and/or lower values, one must exclude the possibility that the highs represent rebound in response to nocturnal hypoglycemia. This distinction is accomplished by asking the patient to check a blood glucose reading between 2 and 3 AM to see if it is normal. If the overnight blood glucose is high, then it is appropriate to adjust the basal insulin dose upward to move blood glucose levels toward the desired target range. If this value is low, it demonstrates that nocturnal hypoglycemia with subsequent rebound is the likely cause of the fasting highs, and the appropriate insulin adjustment is a reduction by 10% to 20% in the basal insulin that is acting at this time of night, e.g., once daily glargine or detemir dose or the evening dose of twice daily dosed insulin NPH or detemir.

Adjusting Insulin for Variations in Food Intake

In order that postprandial blood glucose levels will be optimized, it is necessary for the provider and the patient to have an understanding of the relationship between the caloric content, and in particular the carbohydrate content of the meal, as the latter is the major contributor to the postprandial glycemic excursion. Prandial insulin dose will be matched with the anticipated or actual carbohydrate content of the meal. In the consistent carbohydrate meal plan, the number of grams of carbohydrate included in a given meal from day to day will be kept constant, thus allowing a prespecifled insulin dose prescribed to be taken with the meal to control the postmeal glucose excursion. The nutritionist diabetes educator will typically provide a meal plan that incorporates in the range of 30 to 45 g of carbohydrate with each of breakfast and lunch and 45 to 60g of carbohydrate daily with dinner, depending on the patient’s total daily caloric intake. The bedtime snack will contain 15 or more grams of carbohydrate if it is to be taken with a dose of insulin.

The second method whereby insulin doses are matched to carbohydrate content of the meals is carbohydrate counting in which a predetermined insulin-to-carbohydrate ratio is matched to the premeal anticipated or postmeal known carbohydrate content. Carbohydrate counting requires the patient to count grams of carbohydrate and estimate insulin doses based on carbohydrate intake. An average of 1 U of short- or rapid-acting insulin will dispose off 10 to 15 g (one starch equivalent) of carbohydrate, with a range of 0.5 to 2.0 U. In the adult with Type 1 diabetes mellitus, where the total daily insulin requirement is typically not very high and insulin resistance is not present, it is generally safe to start with an insulin-to-carbohydrate ratio of 1 U of insulin for every 15 g of carbohydrate. This method allows flexibility in the content of each meal. The patient can increase the amount of insulin taken with the meal, e.g., if eating out, or to reduce it in the event that a meal will be small or one does not feel like eating.

The insulin-to-carbohydrate ratio will account for insulin sensitivity relative to the postprandial glycemic excursion. It is important to note that insulin-to-carbohydrate ratios can vary with time of day, and that they are affected by stress, illness, and variations in physical activity. One should also note that the dawn phenomenon often induces a state of relative insulin insensitivity in the early morning, in which case it may be necessary to provide one insulin-to-carbohydrate ratio for the patient to take with breakfast, e.g., 1/10 and another ratio for the other meals of the day, e.g., 1/12 or 1/15 to appropriately match each to individual requirements.

Several formulae may also be applied to calculate an individual insulin-to-carbohydrate ratio: the 450 or 500 rule and the weight method. The 500 or the 450 rule may be used when a dose of insulin given before a meal results in postprandial blood glucose levels in the target range. The insulin-to-carbohydrate ratio by the 450 or 500 rule is calculated as follows:

•   Rapid-acting insulin (aspart, glulisine, or lispro)-to-carbohydrate ratio = 500 divided by TDDI.

•   Regular insulin-to-carbohydrate ratio = 450 divided by TDDI.

As an example, if the individual TDD is 50 U and the patient uses a regimen with prandial rapid-acting then the insulin-to-carbohydrate ratio would be 500 divided by 50 or 1 U of analog for 10 g of carbohydrate.

Table Weight-Based Insulin to Carbohydrate Ratios

The weight method for determining the insulin-to-carbohydrate ratio uses assignment of a ratio from a table based upon the patient’s weight to provide the insulin-to-carbohydrate ratio.

Whichever method of matching insulin-to-carbohydrate content of the meal is used, it is important to assess the impact of the dose on postmeal blood glucose levels by checking a finger-stick value 60 to 90 minutes after a dose of rapid-acting insulin analog has been given with a meal or 2 hours after the meal if regular insulin is provided as the prandial insulin. If this value is high, and a consistent carbohydrate diet is being used, the premeal insulin bolus dose will be raised by 10% to 20%. If the postmeal blood glucose is above target and if an insulin-to-carbohydrate ratio is used to determine the premeal dose, the ratio will be increased, typically in 2 to 5 g increments, e.g., from 1/15 to 1/12 or 1/10, depending upon the rise in magnitude of the blood glucose after the meal. Conversely, when blood glucose following a meal is lower than desired, the number of units of insulin given per gram of carbohydrate prior to the meal will be reduced.

Adjusting Insulin for Changes in Activity/Exercise

Increased levels of physical activity, including formal exercise, impact blood glucose control by promoting movement of glucose into glycogen stores in the peripheral tissues. The entry of glucose into skeletal muscle is increased during exercise via an insulin-independent increase in the number of GLUT 4 transporters in muscle cell membranes. This increase in glucose entry persists for several hours after exercise and regular exercise training can produce prolonged periods of time where insulin sensitivity is increased. Exercise can precipitate hypoglycemia in diabetes not only because of the increase in muscle uptake of glucose but also because absorption of injected insulin is more rapid during exercise. Patients with diabetes will often need to either take in extra calories or reduce their insulin dosage when they exercise. If body weight is a concern, it is preferable to lower insulin doses in anticipation of exercise rather than to ingest extra calories to prevent hypoglycemia. Determination of the optimum insulin regimen for the patient with diabetes mellitus 1 for exercise will be facilitated by careful blood glucose monitoring in the periexercise period until insulin requirements are determined and appropriate insulin doses for these times have been determined. blood glucose testing is recommended before, during, and after the activity to monitor the patient’s response to exercise. Exercise-induced hypoglycemia may occur many hours after the activity as glycogen stores are repleted. This is particularly true following intense exercise, such as weight lifting and/or prolonged periods of aerobic exercise such as long-distance running or biking.

If exercise is planned, insulin dosages will be adjusted in anticipatory fashion in order to decrease risk for hypoglycemia either during or following the period of increased physical activity. It is not uncommon for the patient with Type 1 diabetes mellitus who exercises regularly to have one insulin regimen for exercise days and another for days on which exercise is not undertaken. Premeal rapid-acting or regular insulin can be reduced 25% to 50% for moderate levels of planned postprandial activity. If the activity is strenuous, the patient may need additional carbohydrate along with the reduction in premeal insulin. Patients using insulin pumps can temporarily lower the basal rate by 20% to 40% for sustained periods of exercise, particularly those lasting over 60 minutes. A reduction in the basal rate by 25% during postexercise hours may also be necessary to avoid postexercise hypoglycemia. Suspending the basal rate for more than one hour is not recommended in the insulin deficient patient who has Type 1 diabetes mellitus as ketogenesis may develop.

If exercise is unplanned, ingestion of additional carbohydrate will be necessary (15-30 g of carbohydrate for every 30 to 45 minutes of moderate exercise). It is also important to note that it is always necessary for the patient with diabetes to maintain adequate hydration during exercise as dehydration has a negative impact on insulin sensitivity.

Adjusting Insulin for Illness or During Periods of Stress: Sick Day Rules

Stress and illness clearly impact glycemic control. In the patient with Type 1 diabetes mellitus, release of insulin counterregulatory hormones under such circumstances will typically lead to hyperglycemia. Indeed progressive development of hyperglycemia without other aggravating factors may indicate that an illness, e.g., urinary tract infection or viral syndrome is in its prodromal stages. Careful questioning of the patient about symptoms that suggest underlying illness is part of a thorough assessment under these circumstances. It is necessary for the patient to have a plan of action to enable glycemic control on sick days.

Diabetes education for the adult patient with Type 1 diabetes mellitus should include a thorough grounding in the general principles of sick day management, which includes the following:

•   Checking finger-stick BGs every 4 hours if not eating, or before meals and bedtime if eating discrete meals.

•   Check 2 to 3 AM blood glucose if running high at bedtime or nocturnal hypoglycemia is suspected.

•   Take all usual prescribed doses of basal insulin glargine or detemir, with adjustment in basal dose(s) as described below.

•   Reduce NPH insulin doses by 1/3 to 1/2 if food intake curtailed (to prevent hypoglycemia when the NPH peaks).

•   Take usual meal insulin doses if eating well with a correction insulin dose if premeal blood glucose is high; if not eating well, decrease meal insulin dose by 50% and consider taking at the end of the meal after assuring that the food is eaten.

•   Maintain hydration.

•   Check urine for ketones.

•   If vomiting and cannot keep food or liquids down, go to the emergency room.

Table Practical Guidelines for Insulin Adjustment in Adults with Type 1 Diabetes Mellitus^a

Multiple daily insulin insulin dose adjustments are made to minimize hyperglycemia on sick days or during periods of stress.

•   If all BGs are running high, basal insulin dose(s) will be adjusted upward by 20% increments.

•   Correction or supplemental doses of insulin will be given in addition to basal and prandial insulin doses when hyperglycemia is present. Correction or supplemental doses of insulin will be determined by calculation of a correction dose based on the total daily insulin requirement, e.g., by the rule of 1800 or as 10% of the TDDI if urine is dip negative for ketones and as 20% of the total daily insulin requirement if there are ketones in the urine (as discussed earlier), or by use of a correction dose scale.

•   As with all insulin doses prescribed, it is essential to review the impact of sick day doses given to an individual patient and to revise these doses per BGs obtained to assure that they lower blood glucose appropriately while minimizing risk of hypoglycemia.

CSII dose adjustments

•   When the patient is being treated with an insulin pump, the sick day insulin dose adjustments are to increase the basal insulin rate by about 50% and the bolus insulin doses by 20%.

HOSPITAL MANAGEMENT

The principles of using physiologic insulin replacement to mimic normal insulin secretion patterns that are used in the outpatient with Type 1 diabetes mellitus generally are applied in hospital management of adults with Type 1 diabetes mellitus as well.