Case Study: Five Steps to Freedom: Dose Titration for Type 2 Diabetes Using Basal-Prandial-Correction Insulin Therapy

Mary, age 53, was seen as a new patient. She had a history of type 2 diabetes for 17 years. Her BMI was 31 kg/m². She had attempted to adhere to lifestyle measures for treatment of diabetes and had previously consulted with diabetes educators. She was fairly active at her job.

She was taking 90 units of peakless insulin (glargine) at bedtime in addition to pioglitazone, 45 mg daily; metformin, 1,000 mg twice daily; and glimperide, 8 mg daily. Her hemoglobin A_1c (A1C) was 8.5%. She had average prebreakfast glucose readings of 110 mg/dl. At other times of day,there were premeal glucose readings of 170-240 mg/dl and some peak postprandial readings > 300 mg/dl.

Under the supervision of her previous health care provider, over a 12-month period of combined glargine and triple oral therapy, the strategies of changing the time of peakless insulin administration to the morning and,later, splitting the peakless insulin dose into equal morning and bedtime components had failed to change her blood glucose pattern. A conceptual diagram of the average daily excursions of blood glucose at the time of her presentation is shown in Figure 1.

At the conclusion of the visit, she commented that she wanted to do whatever was necessary to get better results, and she agreed to replace the glimepiride with prandial use of insulin. The glimepiride was discontinued. The pioglitazone dose was reduced to 30 mg daily, with consideration of possible future discontinuation because of cost factors. The metformin was continued.

The combined total amount of the starting doses of insulin under a basal-prandial-correction regimen was assigned at 80% of the former peakless insulin dose. The insulin was divided into 36 units of peakless insulin at bedtime and 12 units of prandial insulin at each of the three major meals.

Recognition of single 15-g servings of carbohydrate was reinforced. As a temporary plan, a consistent carbohydrate diet including 60 g of carbohydrate at the major meals was recommended to the patient. She was issued with an insulin dose titration schedule, with change dates entered by the practitioner, such that an upward dose adjustment would be made every third day, alternating between a 2-unit increment of each of the three premeal lispro doses or a 6-unit increment of the bedtime glargine dose. The one-page handout given to the patient is shown in Figure 2.

Mary was instructed to revisit within 4 weeks or, if some readings approached target, to phone or visit sooner for pattern review. At that point,she was to discontinue the forced titration of insulin.

As Mary's blood glucose readings approached target, in order to achieve pattern adjustment, the practitioner adjusted specific components of the insulin therapy. The doses were revised as follows: 20 units of rapid-acting insulin analog at meals and 60 units of peakless insulin at bedtime. The patient revisited the diabetes educator with her logbook showing 1 week of recorded mealtime intake (still ∼ 60 g or four servings of carbohydrate at each meal); unusually heavy activity on two days accompanied by symptomatic lows; blood glucose readings; and insulin doses taken.

An insulin-to-carbohydrate ratio was identified (5 units per serving), and,in terms appropriate to her educational level, the patient received education on how to use the ratio. A simple correction rule was described for taking extra amounts of rapid-acting analog before meals if her blood glucose was high: for blood glucose of 140-199 mg/dl, add 4 units; for blood glucose >200 mg/dl, add 8 units.

Mary was cautioned to reduce the rapid-acting analog by half if unusual activity was expected to follow the dose of rapid-acting analog within 2-3 hours after injection. Otherwise, for unexpected exercise or for exercise occurring outside the timeframe of action of the rapid-acting analog, she was advised to take 15 g extra carbohydrate (one serving) for every 20 minutes of heavy exercise and not to cover the extra carbohydrate with extra insulin. The consistent carbohydrate requirement was eliminated.

Three months after initiating the basal-prandial-correction regimen, Mary's A1C had fallen 1%. Hypoglycemia was infrequent. Although she still experienced some fluctuation in blood glucose, her logbook showed premeal averages in the low 100-mg/dl range without a clear cut upward trend over the course of the day.

Occasionally, Mary performed postprandial testing 1-2 hours after meals to see the effect of her food choices and discovered that some postprandial readings exceeded 200 mg/dl. After a reappointment with the dietitian to review carbohydrate counting with special emphasis on portion size, her postprandial readings seldom exceeded 180 mg/dl and often were < 140 mg/dl.

1. What blood glucose pattern suggests the need for greater prandial insulin effect?

2. What are the potential advantages of basal-prandial-correction therapy(using rapid-acting analog at each meal and peakless basal insulin) over the use of premixed insulin?

3. How can basal-prandial-correction therapy be initiated and titrated in a primary care setting?

4. How can literacy and numeracy of patients be considered when providing instruction on advanced carbohydrate counting and on use of an insulin-to-carbohydrate ratio to determine the prandial insulin doses?

The case describes a time point in the course of type 2 diabetes at which peakless insulin, combined with daytime oral agents, has failed to achieve satisfactory control.

Patients having type 2 diabetes often fail combination oral therapy and require the addition of insulin. For each of the five major classes of oral agents, successful combination of the oral agent has been reported with use of peakless (glargine) or bedtime NPH insulin. During satisfactory control, the pre meal and postprandial blood glucose should be near target at all times of day, the A1C should be near target, and the patient should have only minimal trouble with hypoglycemia. If a patient has satisfactory control during combination oral therapy with peakless or bedtime NPH insulin, then prandial insulin is not required. Conversely, failure to achieve satisfactory control should suggest the need for intensification, to be achieved potentially by the addition of prandial insulin coverage.

The need for improved prandial coverage may be further suggested by a specific pattern of blood glucose selfmonitoring results: progressive rising of blood glucose during the day, with peaking after meals and gradual correction overnight. The first morning blood glucose is the lowest of the day. If a meal delay occurs, there may be hypoglycemia.

The foregoing blood glucose pattern may indicate that any prescribed basal insulin has been prescribed in doses that exceed true basal requirements. The basal insulin is being used to “play catch-up” every night for blood glucose elevations that occurred as a result of eating.

When daytime oral agents, combined with peakless or bedtime NPH insulin,fail to achieve daytime control, the solution often is to replace insulin secretogogue therapy with prandial insulin therapy. At the juncture in a patient's course at which prandial insulin therapy is added to basal insulin therapy, clincians must decide whether to continue or interrupt metformin and thiazolidinedione therapy, based on considerations of cost, patient wishes,and evidence for previous efficacy of the metformin or thiazolidinedione.

One way of delivering prandial insulin is to cover both prandial and basal requirements with intermediate-acting insulin or premixed insulin. Many patients having some endogenous insulin production do reasonably well for a while on such a regimen of premixed insulin.

The use of a rapid-acting analog in comparison to regular insulin,^1,2 peakless insulin in comparison to NPH,³  and advanced carbohydrate counting in comparison to standard dose therapy⁴  have all been demonstrated in type 1 diabetes to result in improvement of A1C results. Because patients with type 2 diabetes have some endogenous insulin production,the advantages of such therapies in improving A1C may be more difficult to demonstrate. Nevertheless, in type 2 diabetes, the use of peakless insulin in comparison to NPH during prandial use of regular insulin⁵  and the use of a rapid-acting analog in comparison to regular insulin during multiple daily injection therapy⁶  have been shown, respectively, to produce less hypoglycemia and to reduce postprandial hyperglycemia.

In practice, it is because of the problem of hypoglycemia that many patients with type 2 diabetes become frustrated with attempts at intensification. Yet these patients do want to achieve glycemic targets. When offered the option of multiple daily injections, they often say that they are willing to do what is necessary. Despite the requirement for four daily injections, the multiple daily dose treatment plan makes better sense to them than the use of premixed insulin because they understand the purpose of each component of therapy: to cover each meal with a rapid-acting analog, and to provide a peakless insulin to cover between meals and overnight. This is seen as “nature's way.”

In order to achieve the desired results, many patients with unsatisfactory glycemic control indicate ready willingness to move directly from zero or one to four daily injections of insulin. The use of pen injection devices for prandial coverage makes it easy to include all three meals. Patients are pleased to hear that, while making healthy choices, the end result is the freedom to eat what they like, when they like.⁷ 

Should clinicians require patients to fail twice-daily injections of premixed insulin before offering the preferred therapy,basal-prandial-correction therapy? To some extent, the answer is determined by patient preference. But to a large extent, unfortunately, the answer is determined by the perceived difficulty of implementing basal-prandial-correction therapy in a busy office practice. There are at least three great stumbling blocks:

1. Impatience with dose-finding. At baseline, before starting multiple daily injections, most patients with type 2 diabetes require substantial up-titration of their total daily dose of insulin. However, if there is concern about hypoglycemia, prudence usually requires an initial modest dose reduction from previous insulin therapy, if any was in place. To prevent premature abandonment of the treatment plan, patients need to be forewarned that a period of time will be required during dose titration of insulin before target-range results are seen.

2. Neglect of prandial insulin up-titration. Under time pressure during a busy office visit, the path of least resistance is to force the titration of the basal insulin. Neglecting the titration of the prandial insulin leads to lability of control and sometimes to abandonment of the concept of basal-prandial-correction therapy.

3. Repeated revision of the insulin-to-carbohydrate ratio. Part of the task of dose discovery is to be sure that the total daily amount of rapid-acting analog remains roughly equal to the amount of basal insulin. To do it right, balancing basal and prandial insulin appropriately, one must remain mindful of the average carbohydrate intake of the particular patient. Repeated revisions and restatements of the insulin-to-carbohydrate ratio will confound all but the most adept arithmeticians among our patients. Nevertheless, some practitioners achieve dose titration through repeated revisions of the insulin-to-carbohydrate ratio.

In a primary care practice, implementation of basal-prandial-correction therapy can be achieved easily in five steps(Table 1). The following description applies not only to the situation of transitioning from a single dose of peakless or bedtime NPH insulin, but also from premixed insulin or no insulin to multiple daily injections.

1. Determine the total daily starting dose of insulin. For insulin users, the total daily starting dose is 80-100% of the previous daily total insulin dose. For new insulin starts for patients having type 2 diabetes, the total daily starting dose is 0.3 units/kg body weight.

2. Divide the total daily dose of insulin between peakless insulin and rapid-acting analog. The insulin should be given half as peakless insulin and half as premeal rapid-acting analog, divided into three parts.

3. Assign a consistent carbohydrate diet. This strategy is a temporary measure to be used during insulin dose titration. Assign an equal amount of carbohydrate for each meal (60 or 75 g for most adults). Many patients prefer to count carbohydrates as servings, each containing ∼ 15 g(four or five servings per meal for most adults).

4. Provide an insulin titration schedule. Increase peakless insulin and rapid-acting analog in an alternating pattern. Have the patient make a change every third day, as long as all readings are high. For a total daily dose < 48 units, alternate between increasing the rapid-acting analog by 1 unit and increasing the peakless insulin by 3 units. For a total daily dose≤ 48 units, alternate between increasing the rapid-acting analog by 2 units and increasing the peakless insulin by 6 units.

5. Instruct on matching insulin to carbohydrate, and eliminate consistent-carbohydrate requirement. As readings approach target,discontinue forced titration of insulin. Adjust individual component doses of insulin until pattern adjustment is achieved. Evaluate insulin-to-carbohydrate ratios. Instruct the patient on advanced carbohydrate counting and the use of insulin-to-carbohydrate ratios at each meal to determine the dose of rapid-acting analog. Using the insulin-to-carbohydrate ratio and skills of advanced carbohydrate counting, the patient then has freedom to change the carbohydrate content of meals.

* * *

Some people have different insulin-to-carbohydrate ratios for each meal. For example, Mary, in our case above, may realize that her supper has more protein and fat than other meals, and that she is less active after supper. However, many patients are best served by a simple rule for a single insulin-to-carbohydrate ratio to be used at any meal. Often, it is best to state only one rule and to be reasonably sure what that rule is before articulating it to the patient.

Should patients with type 2 diabetes be taught to count carbohydrates by servings or by grams? One carbohydrate serving contains ∼ 15 g. Most people eat about three to five servings or 45-75 g of carbohydrate at each major meal and one to two servings or 15-30 g of carbohydrate at snacks. Some centers teach all patients to count by grams, other centers teach all patients to count by servings, and still other centers use either method based on patient characteristics and preference(Table 2).

For most patients, counting by servings is probably easier than counting by grams. Counting by servings requires multiplication or memorization of sequences of numbers together with finger-counting of the number of servings. In the simplest form, counting by servings might be expressed as a memorizable rule such as “15 units for three servings, 20 units for four servings,25 units for five servings.” On the other hand, the use of food labels is more difficult because prior conversion from grams to servings is required. The method of counting by servings is good for relatively high total insulin doses and therefore is good for type 2 diabetes.

When counting by grams, the demands on the quantitative skills of the patient probably are greater than when counting by servings. Counting by grams requires division (sometimes long division). Counting by grams may encourage the use of packaged foods. However, counting by grams is well-adapted for perfection of care and for use by patients with no endogenous insulin reserve and patients who are sensitive to small differences in insulin dose. Counting by grams allows calculation of fractions of units by users of insulin pumps. Smart pumps may do the math for the patient. Counting by grams often is preferred for insulin-deficient patients.

For unusual exercise planned to occur within 2 hours after administration of rapid-acting analog, patients may be instructed to reduce the rapid-acting dose by 25-50%. For unusual exercise beyond 2 hours after administration,patients may eat one extra 15-g carbohydrate choice for every 20 minutes of exercise (and should not cover the extra carbs with extra insulin).

Correction therapy may be assigned according to the “rule of 1800.” Patients may supplement the pre-meal dose of rapid-acting analog intended to cover the carbohydrate content of the meal with an extra amount intended to correct hyperglycemia. The number 1800 is divided by the approximate total daily dose of insulin to give the drop of blood glucose in mg/dl that 1 unit of rapid-acting analog would produce.

If a preprandial dose of rapid-acting analog already has been given to cover a meal and to correct any preprandial hyperglycemia, and if a correction dose then is added within 2-2.5 hours to correct postprandial hyperglycemia,the patient may experience hypoglycemis for at least two reasons: 1)the correction rule used by the patient may target a preprandial rather than a postprandial blood glucose goal, and 2) the action of the first dose of rapid-acting analog may overlap with the second dose (referred to as“stacking”). To avoid stacking, the use of correction doses should be restricted to premeal timing except on sick days.

It is a common mistake during establishment of basal-prandial-correction therapy to press the dose of the basal insulin to achieve first-morning normoglycemia while neglecting the up-titration of prandial therapy. The amounts should be about equal. When there is an excessive reliance on basal insulin, lability of control may occur. Upward trending of blood glucose during the day is followed by overnight correction. If meals are delayed,hypoglycemia occurs. Balance may be reestablished by “borrowing from Peter to pay Paul” (i.e., subtracting part of the basal insulin and adding the same number of units to the total daily prandial insulin dose). Depending on the comfort level of the patient, such rebalancing may be achieved all at once or gradually in several steps.

Why bother with such apparent complexity? In reality, multiple daily injection therapy is less complex than the alternative. The necessity to fend off repeated episodes of hyperglycemia and hypoglycemia during fixed dose therapy, without any clear comprehension of what each dose of insulin achieves, is not easy for patients. Basal-prandial-correction insulin therapy quickly becomes intuitive. Each component of insulin therapy has an obvious purpose. Freedom of lifestyle is achieved.

• A pattern of rising blood glucose during the day, with partial or complete correction overnight, suggests insufficient prandial insulin effect.

• In comparison with premixed insulin regimens, basal-prandial-correction therapy (multiple daily injections of insulin) with use of peakless insulin for basal coverage and rapid-acting analog for prandial coverage reduces hypoglycemia and allows greater freedom for patients to eat as they would like while making healthy food choices.

• During dose titration, the basal and total prandial components of insulin therapy should be kept roughly equal in amount.

• When instruction of patients on the use of an insulin-to-carbohydrate ratio occurs, the individual patient's needs must be considered.

• Initiation can be achieved with a fivestep program.