Insulin as a strategy to optimize glycemic control in patients with type 2 diabetes

November 14, 2008
Stephen A. Brunton, MD, Bobby Levy, MD

From Cabarrus Family Medicine, Concord, NC.

Cardiology Review® Online, November 2008, Volume 25, Issue 11

Strategic Alliance Partnership | <b>FAST: Foundation for Angelman Syndrome Therapeutics</b>

Treatment of type 2 diabetes should achieve and maintain euglycemia, thereby preventing complications from this progressive disease. Current antidiabetic therapies should be a part of a multimodal management program that includes diet, exercise, and blood pressure and lipid control. Oral antidiabetic drugs are still first-line therapy for type 2 diabetes, but intensification of therapy, including starting insulin, should occur every 2 to 3 months as needed to achieve euglycemia. The first insulin added is typically a basal insulin, which is effective in lowering fasting plasma glucose (FPG). A persistently elevated glycated hemoglobin (HgbA1C) level despite near or complete normalization of FPG, however, indicates postprandial hyperglycemia. In these cases, the addition of bolus insulin is required to reduce postprandial glucose (PPG). Several approaches to initiate and titrate insulin can be used based on FPG, PPG, HgbA1C, and patient factors.

Numerous clinical studies have clearly shown that a glycated hemoglobin (HgbA1C) level of less than 7%, as recommended by the American Diabetes Association (ADA),1 significantly reduces the incidence of microvascular complications such as retinopathy, nephropathy, and neuropathy, in patients with type 2 diabetes mellitus.2-4 The benefit of tight glycemic control on macrovascular complications such as myocardial infarction (MI) and stroke is more questionable based on the results of 3 clinical trials designed to address this question.5-7 Nevertheless, achieving and maintaining long-term glycemic control in patients with type 2 diabetes is an important yet challenging endeavor.

Based on data from the UKPDS (United Kingdom Prospective Diabetes Study), at the time of diagnosis of type 2 diabetes, approximately 50% of pancreatic beta-cell function typically has been lost.8 More recent evidence suggests that the loss may be as high as 80%.9 The progressive decline in insulin secretion and the worsening of peripheral insulin resistance make it increasingly difficult to maintain long-term glycemic control with oral antidiabetic drugs (OADs), even when using combination therapy. Consequently, individualized treatment using the full range of treatment options, including insulin, is essential.

Adding insulin to OAD treatment has been shown to further reduce hyperglycemia. For example, when once-daily glargine in combination with various OAD therapies was compared with increasing OAD doses in adults age 65 years or older, HgbA1C decreased 1.5% and 0.6%, respectively, at the end of 24 weeks, while fasting plasma glucose (FPG) decreased 29% and 15%, respectively.10 Hypoglycemia was observed less frequently in patients treated with glargine. Reductions in HgbA1C have also been observed with the addition of twice-daily biphasic 30/70 human insulin mixture or the administration of neutral protamine Hagedorn (NPH) at bedtime in patients who were suboptimally controlled on a sulfonylurea alone or in combination with metformin.11 The addition of insulin to a glucagon-like peptide 1 receptor agonist or dipeptidyl peptidase-4 inhibitor has not been investigated, but may be beneficial because both cause reductions in FPG and postprandial glucose (PPG).

Although the value of insulin therapy in reducing cardiovascular risk has been established, conflicting results have been observed. For example, once-daily or basal-bolus insulin therapy decreased the risk of microvascular complications by 12% in the UKPDS.12 After acute MI, in-hospital insulin infusion followed by administration of outpatient subcutaneous insulin improved long-term survival, with a reduction in mortality of 11% in the DIGAMI (Diabetes Mellitus, Insulin Glucose Infusion in Acute Myocardial Infarction) study.13 These results were not confirmed by DIGAMI 2, which compared acute insulin-glucose infusion with and without insulin-based long-term glucose control with routine metabolic management according to local practice.14 The incidence of all-cause mortality and glycemic control was comparable among treatment groups, which may have resulted from lower than expected long-term use of multiple insulin doses per day in the intensive group and low patient recruitment in this trial.

Although the benefits of insulin therapy have been demonstrated in the literature, there is considerable resistance by patients and physicians to using insulin. Helping patients understand that a rise in HgbA1C is a part of the progression of type 2 diabetes and not a personal shortcoming is critical to achieving effective patient-centered disease management. It is equally essential to educate patients on the benefits and role of insulin in managing their diabetes.

Given the impact of type 2 diabetes on cardiovascular health, this article will examine insulin’s role in achieving glycemic control in patients with type 2 diabetes and review approaches to initiating and titrating insulin doses. The importance of PPG control and the rationale and various approaches for intensifying insulin therapy are also discussed.

Goals of therapy

The primary goal of therapy for type 2 diabetes is to reduce the risk of complications related not only to glucose, but also to lipids, blood pressure, weight, and smoking (Table 1).1,15 Keeping low-density lipoprotein cholesterol below 100 mg/dL (and <70 mg/dL for very high-risk patients) and blood pressure less than 130/80 mm Hg are critical components of complete diabetic care. Routine discussions of the importance of exercise and smoking cessation are also essential in preventing cardiovascular disease and the other diabetic complications. For patients who have several comorbidities, treating those that are easiest to control first may be prudent to avoid overwhelming the patient.

After initiating therapy to lower blood glucose levels, therapy should be titrated based on blood glucose monitoring over the next 2 to 3 months until target glycemic goals have been achieved.15 If glycemic goals have not been reached during this treatment period, a more intensive therapy regimen should be initiated and again titrated over the next 2 to 3 months until glycemic goals are achieved.15

Another important goal is to avoid treatment-associated hypoglycemia, which is more common as HgbA1C normalizes. Severe hypoglycemia can be prevented in many patients through patient education about symptom recognition and initiation of acute treatment. Despite such measures, some patients, especially those with longstanding diabetes, will develop hypoglycemic unawareness; thus, regular reminders of hypoglycemia symptoms and what to do when they occur are encouraged for all patients with diabetes.

Evaluating and establishing patient adherence to the medical regimen is yet another important goal. Patient-centered office visits that are culturally competent and allow the patient to discuss the impact of their disease on their life often leads to major treatment breakthroughs with patients who have had a difficult time controlling their diabetes. Supporting patients and trying to tailor the treatments that are cost-effective and with decreased frequency of dosing are also simple measures to improve patient adherence and outcomes.

Hyperglycemia treatment approaches

Traditionally, management of type 2 diabetes has involved the use of at least 1 OAD. This approach may no longer be appropriate. Evidence-based guidelines developed by the American College of Endocrinology (ACE)/American Association of Clinical Endocrinologists (AACE) recommend an HgbA1C level of 6.5% or lower.15,16 In treatment-naïve patients, monotherapy with an OAD is recommended if the HgbA1C level is elevated but less than 7%. If the HgbA1C level is 7% or higher, a combination of OADs or an OAD with basal and/or bolus insulin is recommended. Insulin should also be considered in patients with elevated fasting plasma glucose levels or exaggerated PPG excursions regardless of the HgbA1C level.

Noninsulin antidiabetic drugs

Oral and injectable noninsulin drugs, including alpha-glucosidase inhibitors, exenatide, glinides, metformin, pramlintide, sitagliptin, sulfonylureas, and thiazolidinediones (TZDs), are available to treat type 2 diabetes. Among the many factors to be considered, the glucose-lowering capability of specific agents is most important.17 When used as monotherapy, there is some variability among the OADs in their ability to lower HgbA1C, with alpha-glucosidase inhibitors lowering levels by 0.5% to 0.8%, glinides by 1% to 1.5%, metformin by 1.5%, sulfonylureas by 1.5%, and TZDs by 0.5% to 1.4%.8

Other factors and characteristics of the available agents must also be considered.1,17 For example, careful attention to renal and hepatic function is essential when choosing an OAD and deciding its dose. In addition, treatment-related hypoglycemia is an important consideration and is more common with the older sulfonylurea agents. The glinides, sulfonylureas, and TZDs promote weight gain; alpha-glucosidase inhibitors, metformin, and sitagliptin have neutral effects on weight; and exenatide and pramlintide promote weight loss. Gastrointestinal side effects are common with exenatide, metformin, and alpha-glucosidase inhibitors. The TZDs have a neutral or beneficial effect on atherogenic lipids, with pioglitazone having a more beneficial effect than rosiglitazone. TZDs can cause peripheral edema and can exacerbate congestive heart failure. OADs vary widely in cost. Metformin and sulfonylureas are less expensive than other OADs.

When the use of 1 or more OADs fails to control hyperglycemia, is initiation of insulin preferred to adding another oral agent?

One of the many findings of the UKPDS was that 53% of patients treated with sulfonylurea monotherapy required the addition of insulin within 6 years to achieve glycemic control.18 Based on this finding, treatment guidelines issued by the ACE/AACE, as well as by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes, include the addition of insulin to single or combination OAD therapy for patients who do not achieve the targeted HgbA1C goal.16,17 In fact, had these treatment guidelines been in place at the time of UKPDS, the percentage of sulfonylurea-monotherapy—treated patients who required the addition of insulin would have been higher than the 53% observed in UKPDS. The importance of achieving rapid glycemic control cannot be overemphasized. The AACE, in fact, recommends intensifying treatment every 2 to 3 months until glycemic control is achieved.15

The risks and benefits of glucose-lowering treatment to HgbA1C levels below current guidelines have recently been reported. The ACCORD (Action to Control Cardiovascular Risk in Diabetes) trial investigated whether intensive therapy aimed at lowering the HgbA1C level to less than 6%, compared with standard therapy to achieve an HgbA1C of 7.0% to 7.9%, would reduce cardiovascular events in high-risk patients with type 2 diabetes who had established cardiovascular disease or additional cardiovascular risk factors.5 An interim analysis at 3.5 years found that the mortality rate in the intensive-therapy group was slightly higher than that of the standard-therapy group, with no significant difference in the incidence of major cardiovascular events. Within weeks, the results of the ADVANCE (Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled Evaluation) yielded similar results. ADVANCE showed that intensive therapy to achieve an HgbA1C level less than 6.5% reduced major microvascular events. In fact, the all-cause mortality rate in the intensive group in ACCORD was lower than in ADVANCE. The results of a third study, the VADT (Veterans Affairs Diabetes Trial), were released at the 2008 ADA meeting and showed no significant impact of intensive therapy on cardiovascular mortality.7 Although the mortality rate in ACCORD was higher in the intensive arm than in the standard arm, it was lower than that observed in the ADVANCE, VADT, and UKPDS trials. Until the results of these trials are analyzed further, both the AACE and ADA have separately recommended not modifying the glycemic goals outlined in their respective guidelines.19-21

What are the implications of these trials until further analyses are complete? First, it appears that targeting an HgbA1C lower than 7% in high-risk patients with type 2 diabetes does not reduce cardiovascular risk. Second, HgbA1C goals and other glycemic targets in high-risk patients with type 2 diabetes should be individualized. Third, they reinforce the importance of early identification of patients with type 2 diabetes and the initiation of aggressive management of hyperglycemia at the outset as the best approach to minimize microvascular and macrovascular complications.

Insulin preparations

Human and analog are the 2 types of insulin currently available. Although the recombinant human insulins have virtually eliminated insulin allergy and immunemediated lipoatrophy, they do not mimic the physiologic pattern of endogenous insulin. The insulin analogs, however, more closely mimic the basal (baseline) and prandial (mealtime) phases of physiologic insulin (Figure 1).22 Consequently, the prandial or rapid-acting insulin analogues (aspart, glulisine, and lispro) can be administered immediately before meals, whereas regular human insulin must be administered 30 to 60 minutes before meals. Because of their short duration of action, rapid-acting insulin analogs reduce the risk of hypoglycemia compared with regular human insulin.23 The basal insulin analogues (detemir and glargine) have an onset similar to NPH, but they exhibit a long, relatively flat time-action profile that often effectively controls FPG for 24 hours.22 The incidence of hypoglycemia with the basal insulin analogues is reduced compared with NPH. Basal insulin analogues also cause significantly less weight gain than NPH. Head-to-head comparisons show weight gains of 1.0 kg versus 1.8 kg and 1.2 kg versus 2.8 kg for detemir and NPH, respectively, and 0.4 kg versus 1.4 kg for glargine versus NPH, respectively.24-26

Initiation of insulin therapy

Insulin is the most effective agent available for lowering glucose, typically reducing HgbA1C by 1.5% to 2.5%. For patients with very elevated HgbA1C levels, the reductions can be much higher if titrated properly. Nonetheless, there are significant physician and patient issues associated with initiating insulin therapy, many of which are related to anticipated complications such as hypoglycemia and weight gain, the complexity of insulin dosing, concerns with the injection, the implied severity of the disease, and the perceived lack of insulin efficacy.27,28 Physician concerns can be addressed by appropriate education and experience with insulin therapy.

To mitigate patients’ concerns, which may stem from misperceptions, patient education beginning at the time of diagnosis is crucial. The progressive decline of pancreatic beta-cell function, the advantages and disadvantages of the available treatments, and the importance of self-monitoring are examples of key issues to address with patients. Helping patients to become familiar with the various insulin devices by having them self-administer saline can be especially helpful to lessen their concerns about insulin therapy. Familiarizing patients with the range of medications and devices available and assuring them that their treatment plan will be based on their needs and preferences are crucial. These techniques address patients’ concerns about insulin therapy and strengthen the collaborative relationship between the physician and patient, enabling the patient to effectively self-manage their disease.

When initiating treatment with insulin, which, if any, OADs should be continued?

When insulin is initiated late in the course of type 2 diabetes, a patient has little pancreatic betacell function remaining; thus, the use of drugs that act principally by stimulating the release of insulin (eg, sulfonylureas and glinides) are of little or no value. Drugs that act via other mechanisms, particularly by reducing insulin resistance, should be continued to enhance the action of insulin and to lower the total daily insulin dose. Metformin should be continued, as it has the added advantage of limiting the potential weight gain generally caused by insulin.

Approaches to insulin dosing

Treatment with insulin is usually initiated with an intermediate-acting basal insulin (eg, NPH) or a long-acting basal analogue (eg, detemir or glargine) to target fasting hyperglycemia. There are several approaches to selecting the initial dosage regimen. One approach is to begin treatment with a single daily dose (10 to 12 units or 0.1 to 0.2 unit/kg) of NPH insulin or basal insulin analog at bedtime.29 The dose can be increased every 3 to 7 days based on FPG as follows:

FPG >120 mg/dL

= Increase by 2 units

FPG >140 mg/dL

= Increase by 4 units

FPG >160 mg/dL

= Increase by 6 units

This approach is similar to that used in the treat-to-target trials, with slightly different FPG values used based on a patient’s response to the previous adjustment.25,30

In the glargine versus NPH treat-to-target trial,30 significant reductions in HgbA1C were observed in both the glargine and NPH groups, with approximately 60% of patients in each group achieving an HgbA1C level of 7% or lower at the end of 24 weeks. Significantly more patients in the glargine group attained this goal without developing nocturnal hypoglycemia (33% vs 27%; P <.05). Similar results were observed in the detemir versus NPH treat-to-target trial.25 Among the approximately 70% of patients in both groups who achieved an HgbA1C level of 7% or lower, significantly more patients in the detemir group achieved this goal without hypoglycemia (26% vs 16%; P = .008).

Another approach to determining the initial dose of basal insulin is to calculate a patient’s basal requirement, which is approximately 50% of the total daily insulin requirement.29 Given that the total daily insulin requirement averages 0.6 to 0.7 unit/kg, the initial basal insulin requirement for a 70-kg person would be as follows:

Initial basal insulin requirement

(0.5)(0.7 unit/kg per day)(70 kg) ≈25 units/day

Physician-directed versus patient-adjusted basal insulin

Traditionally, the standard of care has been physician-directed changes to the patient’s insulin dose. The PREDICTIVE™ 303 (Predictable Results and Experiences in Diabetes through Intensification and Control to Target: An International Variability Evaluation 303) study compared this approach with patient self-adjustment in a primary care setting.31 Patients self-adjusted their dose of basal insulin (detemir) every 3 days based on the mean of 3 adjusted FPG values using the following algorithm:

Mean FPG <80 mg/dL

Decrease basal insulin dose by 3 units

Mean FPG 80 to 110 mg/dL

No change in basal insulin dose

Mean FPG >110 mg/dL

Increase basal insulin dose by 3 units

At 26 weeks, the decrease in HgbA1C and FPG from baseline was significantly greater in the group that self-adjusted (HgbA1C, 8.5% to 7.9% vs 8.5% to 8.0%, P = .0106; FPG, 175 to 141 mg/dL vs 174 to 152 mg/dL, P <.001). The incidence of overall hypoglycemic events was significantly lower in the standard-of-care group (2.61 vs 4.58 events/patient per year; P <.001), although the rate of major hypoglycemia was similarly low in both groups (0.26 vs 0.20 event/patient per year).

Blood glucose monitoring while on insulin

Titration of insulin is typically based on self-monitoring of blood glucose. How often should self-monitoring be performed?

According to the ADA,1 daily self-monitoring of blood glucose is important for patients with type 1 diabetes to prevent hypoglycemia. The ADA does not provide a standard recommendation for self-monitoring of blood glucose except to say that the frequency and timing of the monitoring should be determined by the needs and goals of the individual patient. More specific recommendations have been made by an international panel of diabetes experts convened by the World Health Organization (Table 2).32 Physicians should advocate for the importance of checking both fasting and 2-hour postprandial levels and review diabetic goals during each patient visit.

Importance of postprandial glucose

The HgbA1C level reflects glycemia over the preceding 2 to 3 months as well as fasting and postprandial glucose. When the HgbA1C level is high (>8.4%), FPG is the major determinant of HgbA1C, whereas PPG is the major determinant when HgbA1C levels decrease to less than 7.3%.33 Postprandial hyperglycemia should be suspected when HgbA1C levels remain significantly elevated despite normalization or near-normalization of FPG.

Above what plasma glucose level does endothelial cell dysfunction begin to occur?

Plasma hyperglycemia higher than 180 mg/dL for 4 hours has been found to begin a cascade of events that comprise the oxidative stress pathways. This stress promotes endothelial cell dysfunction, leading to microvascular and macrovascular complications and atherosclerosis, with its risks for stroke and mortality. Oxidative stress appears to be related to acute changes rather than chronic changes in glucose levels; it is not related to HgbA1C or FPG.34,35 As shown by the DECODE (Diabetes Epidemiology: Collaborative Analysis of Diagnostic Criteria in Europe) study,36 the overall risk of death is higher with increased PPG levels than with increased FPG levels (Figure 2). In those with impaired fasting glucose, the increase in PPG results in a linear increase in all-cause mortality. These data seem to conflict with the ACCORD trial findings, but are consistent with early analyses of the ADVANCE trial. These limited and conflicting data suggest that there may be a mortality benefit by reducing daily glycemic variability and PPG excursions, perhaps through the oxidative stress pathway. To reduce PPG excursions, therapy must be directed at processes that regulate PPG.33,37,38 Because insulin is the most effective agent currently available to lower PPG, the remainder of this article will focus on this agent.

Intensifying insulin therapy

What change should be made to the insulin regimen of basal-insulin—treated patients who demonstrate a near-normal FPG and elevated HgbA1C and PPG levels?

PPG takes on greater importance and needs to be addressed more aggressively when the FPG is normal but the HgbA1C is elevated. The addition of a bolus or prandial insulin is the appropriate next step, as these formulations are more effective in lowering postprandial hyperglycemia. A further increase in the basal insulin dose is inappropriate because basal insulin is more effective in lowering FPG than PPG. In addition, as FPG is normalized, the risk for hypoglycemia increases. Table 3 provides examples of 3 different insulin dosing algorithms, which will be discussed in greater detail.


Using a treat-to-target design, the INITIATE (Initiation of Insulin to Reach HgbA1C target) study compared the efficacy and safety of glargine once daily at bedtime with premixed biphasic aspart 70/30 (BIAsp) twice daily before breakfast and supper, with titration guided by self-monitoring of blood glucose (Table 3).39 Patients were insulin-naïve and had an HgbA1C of 8% or higher on metformin with or without other OADs. The overall reduction in HgbA1C was 2.8% in the BIAsp group and 2.4% in the glargine group (P <.01) at 28 weeks. The change in FPG values from baseline to 28 weeks was the same for each group (125 mg/dL). For breakfast and supper, the mean change in prandial plasma glucose (postprandial minus preprandial) was less for BIAsp than for glargine (breakfast, 34 vs 55 mg/dL, P <.01; supper, 19 vs 42 mg/dL, P <.05). For lunch, the change was less for glargine than for BIAsp (33 vs 45 mg/dL, P >.05). Overall postprandial hyperglycemia was approximately 25% lower for the BIAsp group than for the glargine group (Figure 3). The overall rate of minor hypoglycemia (documented plasma glucose <56 mg/dL with or without symptoms) was 3.4 events/patient-year in the BIAsp group and 0.7 events/patients- year in the glargine group (P <.05). Only 1 episode of major hypoglycemia was reported, which occurred in the glargine group. Mean body weight increased 5.4 kg in the BIAsp group and 3.5 kg in the glargine group.

PREFER study

Another treat-to-target study, the PREFER study, compared premixed BIAsp30 twice daily to basal-bolus therapy with detemir once or twice daily and insulin aspart 3 times daily before meals.40 Patients with type 2 diabetes had an HgbA1C lower than 7% despite treatment with OADs alone or in combination with glargine or NPH insulin. BIAsp and detemir doses were titrated to a plasma glucose level of 126 mg/dL or lower before breakfast and dinner, while insulin aspart was adjusted to a PPG level of 180 mg/dL or lower. All OADs were discontinued. Overall, about 50% of BIAsp30-treated patients and 60% of basal-bolus—treated patients achieved an HgbA1C of 7% or lower after 26 weeks. Patients previously treated with insulin had the greatest response, especially those subsequently treated with basal-bolus insulin; this was likely because of a greater reduction in PPG.

Similar results were observed in a comparison of BIAsp30 and NPH insulin administered twice daily before breakfast and dinner in patients with type 2 diabetes whose blood glucose levels had not been optimally controlled with OADs, NPH insulin, or the combination of OADs and NPH.41 After 16 weeks of treatment, reduction in HgbA1C was 0.6% in both groups. Mean daily postprandial glycemic exposure was less in the BIAsp group than in the NPH group (mean difference, 12 mg/dL; P <.001). The incidences of minor and major hypoglycemia were similar in both groups. Another trial found that basal-bolus therapy provided better glycemic control than premixed insulin in patients previously treated with glargine plus OADs.42 Basal-bolus—treated subjects received glargine at bedtime plus mealtime lispro, while premixed-insulin–treated patients received lispro (50/50) for 24 weeks. The HgbA1C reduction was significant in both groups, but was greater in the basal-bolus group (2.1% versus 1.8%, respectively). This difference resulted from lower fasting and morning 2-hour PPG levels in the basal-bolus group.

These studies demonstrate the importance of controlling postprandial hyperglycemia. Reduction in PPG results in further lowering HgbA1C beyond that achieved with therapy aimed at FPG alone.

Basal-bolus versus premixed insulin

The INITIATE39 and PREFER40 studies also show the 2 general approaches used to intensify insulin therapy. Basal-bolus therapy uses 2 different insulin preparations (a basal and a bolus insulin), thereby allowing more precise titration targeted specifically at FPG or PPG. The principal disadvantages of this approach are that the 2 insulin preparations and doses may be confused by the patient and that more injections are needed daily. Combining the basal and bolus insulins into 1 premixed insulin preparation eliminates the potential for this confusion and generally reduces the number of daily injections required. Premixed insulins are available in a variety of combinations, based on insulin type (human or analogue) and concentration (eg, 75/25, 70/30, 50/50), to meet the needs of most patients. Premixed insulin also allows intensification of treatment from 1 daily dose to multiple daily doses using the same insulin preparation; however, the use of premixed insulin makes it difficult to target titration to FPG and PPG concurrently. In addition, patients need to maintain a regular meal and exercise pattern with this approach to avoid hypoglycemia.

Dosing of bolus insulin

In most cases, a daily dose of basal insulin of approximately 0.6 to 0.7 unit/kg is needed (~40-50 units/day) to achieve the target FPG of 90 to 130 mg/dL. When adding bolus insulin, several approaches can be used to calculate the initial dose.29 The choice of which approach to use depends on the PPG and HgbA1C levels, the risk of hypoglycemia, and the patient’s ability to reliably self-monitor their blood glucose levels and titrate the insulin dose. The easiest approach should be followed first, especially if the patient is expected to have difficulty with adherence.

The easiest approach may be to administer 10% of the patient’s total daily dose of basal insulin as bolus insulin before the largest meal of the day, while reducing the basal dose by 10%. Bolus doses can be added before other meals as well, generally starting with breakfast. Each bolus dose can be calculated as 10% of the total daily insulin dose; however, the basal dose and other bolus doses need to be adjusted accordingly. For example, if the patient is taking a total of 43 units of basal insulin per day, the initial total daily dose of bolus insulin would be approximately 4 units, and the basal dose would be reduced by 4 units. Generally, the bolus insulin dose is increased by 1 to 3 units/day every 2 to 3 days based on the PPG level. The FPG should also be monitored to ensure no asymptomatic hypoglycemia.

The second approach requires first determining the total daily dose of basal insulin needed. Half of this dose is administered as basal insulin and the other half as bolus insulin; 10% to 20% of the total bolus dose is administered at each meal, with the largest dose administered with the largest meal of the day. The basal insulin dose is then adjusted based on FPG, while the prelunch, predinner, and bedtime glucose levels are used to adjust the morning, lunch, and dinner bolus doses, respectively, using a titration schedule as shown in Table 3.

The third approach, which is especially suitable for individuals with a regular schedule, uses premixed insulin in place of basal-bolus therapy, with 1 dose initiated before dinner, while discontinuing the basal insulin. The initial dose of premixed insulin is 12 units in insulin-naïve patients, or 70% of the total daily insulin dose, and is increased by 1 to 3 units/day every 2 to 3 days. If FPG and/or PPG levels remain elevated, additional doses of premixed insulin before mealtime can be administered. The amount of the added dose should be in direct proportion to the total daily dose, with the other dosage(s) reduced accordingly.

A more exact approach to determining the dose of bolus insulin involves the 450 Rule, which estimates the amount of carbohydrate disposed of by 1 unit of bolus insulin. If this dose of bolus insulin is not sufficient to lower the PPG level to the target level, the amount of additional bolus insulin is determined using the 1500 Rule (for regular human insulin) or 1800 Rule (for a rapid-acting insulin analogue). When using the 450 Rule, the first step is to divide 450 by the total daily insulin dose. For example, if the total daily dose of insulin is 70 units (450 ÷ 70 = 6.4), then 6 g of carbohydrate are disposed of by 1 unit of bolus insulin (6 g/unit). The second step is to calculate the total amount (in grams) of carbohydrate to be consumed during a meal. The final step is to calculate the dose of bolus insulin needed for the meal planned. Therefore, if the meal to be consumed contains 60 g of carbohydrate, the amount of bolus insulin needed would be determined as follows: 60 g ÷ 6 g/unit, which equals 10 units of bolus insulin. This assumes that the patient’s preprandial glucose is at an acceptable level. If it is not, additional bolus insulin must be administered, with the amount calculated using the 1500 or 1800 rule. For example, for a patient taking 70 units of rapid-acting insulin: 1800 ÷ 70 units/day = 25.7 mg/dL; thus, 26 mg/dL is the amount by which blood glucose is lowered by 1 unit of rapid-acting insulin. If the patient’s premeal blood glucose is 182 mg/dL and the FPG target is 130 mg/dL: (182 mg/dL - 130 mg/dL) ÷ 26 mg/dL per 1 unit of insulin = 2 units. This is the amount of rapid-acting insulin that is needed in addition to the 10 units calculated by the 450 rule to dispose of the 60-g carbohydrate meal.


Insulin is the most effective agent available for lowering blood glucose levels, typically reducing HgbA1C by 1.5% to 2.5%. Recent clinical evidence and practice guidelines indicate insulin should be initiated early in management and built into a comprehensive treatment program that includes addressing weight, exercise, and blood pressure and cholesterol levels. Evidence-based guidelines developed by the ACE/AACE recommend an HgbA1C level of 6.5% or lower.15,16 Treatment with insulin (basal and/or prandial) is recommended if the HgbA1C level is 8% or higher. Assuring the patient that their treatment will be based on their needs and preferences is essential for successful diabetes management.

Initiating insulin should first focus on normalizing FPG using basal insulin. Because basal insulin is often ineffective in treating postprandial hyperglycemia, bolus insulin should be used to lower PPG when elevated, particularly if the HgbA1C remains at 7% or above. Postprandial hyperglycemia is an important therapeutic target in the management of type 2 diabetes, because it and glycemic variability are strongly associated with oxidative stress, which results in endothelial dysfunction and microvascular and macrovascular complications. Several approaches can be used to successfully initiate and titrate bolus insulin therapy, of which basal-bolus or premixed insulin therapies are the most frequently used.


This manuscript is based on a 2-day conference on intensifying diabetes management geared to family medicine chief residents. The conference and manuscript were supported by an educational grant from Novo Nordisk. Dr Brunton serves on advisory boards for Amylin, Novo Nordisk, and Pfizer. Dr Levy has no affiliations to disclose that might represent a conflict of interest with the content of this article.