In practice with the use of insulin analogues, it is common to find many patients requiring as little as 40% of their insulin requirements as basal insulin. Perhaps the most frequent mistake in the use of insulin therapy is providing the majority of insulin as basal insulin. The reasons for this are multiple, but this evolves over time as patients and clinicians observe hyperglycemia and assume the problem requires more basal insulin.
Most patients with type 1 diabetes will require 0.4 to 0.8 units of insulin/kg/day. This may vary based on numerous factors, including family history of type 2 diabetes, lifestyle (sedentary vs. active), adiposity, gender (males usually require more than females), concomitant illness, and any remaining endogenous insulin secretion. For a typical 75 kg man, a usual total daily insulin dose would be between 30 and 60 units/day [calculated as: 0.4 units/kg x 75 kg = 30 units/day to 0.8 units/kg x 75 kg = 60 units/day]. Let’s say that this person is using 0.7 units/kg, this would be 0.7 units/kg x 75 kg = 52 units/day, and thus the usual basal insulin dose would be 40 to 50% of that, or 21 to 26 units.
Both long-acting insulin analogues – insulin glargine and insulin detemir – can be administered once or twice daily. In general, the lower the dose of insulin, the shorter the duration of action. If it is decided to start with once daily basal insulin, it is usually administered at bedtime. That way, basal insulin deficiency late the following day can be compensated by the use of prandial insulin at dinner time. Insulin detemir will often require larger doses of insulin compared to NPH insulin.
Titration of basal insulin is generally accomplished by observing glucose trends during periods of fasting. The most common way this is done is observing glucose swings overnight when glycemia is not contaminated by food or prandial insulin. Ideally, glucose levels within target (90-130 mg/dL) at bedtime remain stable until the next morning. To confirm stability, it may be beneficial to measure a middle-of-the night glucose reading. Skipping breakfast to confirm stability through the morning will further allow the ability to review basal insulin effectiveness. However, problems with morning insulin resistance (the dawn phenomenon) or insulin waning, particularly into the afternoon with low doses of basal insulin, convince many patients and their physicians to add a second injection of basal insulin in the morning (or proceed to insulin pump therapy). Most often, if two injections of basal insulin are administered, it is given in two equal doses, but this detail has not been studied. One school of thought (championed by IBH) is that if the basal insulin is given once or twice daily, the timing of the injections should be consistent from day-to-day to minimize variable insulinemia. Another school of thought (championed by JSS) is that with a peakless basal insulin (e.g., insulin glargine) the timing of the two injections is less important, and that having the patient administer half of the basal dose before retiring and the other half upon arising offers substantial convenience and flexibility in lifestyle without either sacrificing glycemic control or creating substantial variability in insulinemia.
Practical aspects of prandial insulin include lag times as noted above, and the incorporation of either carbohydrate counting or some mechanism to best match an appropriate dose of insulin with the calorie content of that meal. Typical doses of rapid-acting analogue for type 1 diabetes would be 1 unit per 10 to 15 grams of carbohydrate, but there are many exceptions to this. Obese patients with type 1 diabetes may require as much as 1 unit per 5 grams of carbohydrate while thin insulin sensitive patients may require doses no higher than 1 unit per 20 grams of carbohydrate.
Similarly, patients require pre-determined algorithms for premeal hyperglycemia. These “correction doses”, or supplements, are added to the prandial dose. Typical doses of a rapid-acting analogue for the correction dose are 1 unit per 30 to 50 mg/dL, with targets that need to be individualized for each patient. For example, a common scenario would be to have a correction dose of 1 unit per 40 mg/dL above 130 mg/dL. The other term for correction dose is “insulin sensitivity factor”. Thus, if this individual had a premeal glucose of 250 mg/dL, the correction dose would be (250-130)/40 = 3 units. In addition, for this degree of hyperglycemia a greater lag time, typically 20 minutes would be recommended.
Perhaps the most important practical advice is the use of between-meal SMBG, particularly if there is premeal hyperglycemia or if the carbohydrate content of the meal is unknown, particularly common when meals are not prepared by the patient or family member. Between-meal hyperglycemia can be treated with additional correction dose insulin. While the ADA suggests a 1 to 2-hour postprandial target of less than 180 mg/dL and the American Association of Clinical Endocrinologists recommends even a more stringent target of 140 mg/dL, in actuality these targets are difficult to achieve unless the carbohydrate content of the meal is quite low and the insulin is matched perfectly. Therefore, to understand how to safely correct between-meal hyperglycemia, the patient needs to appreciate insulin-action times, i.e., the pharmacodynamics of insulin after injection (figure 11). Modern-day insulin pumps do these calculations for the patient, but the MDI patient needs to understand this concept without any assistance from computer software. After injection of a rapid-acting insulin analogue, insulin activity will typically last 5 to 6 hours in adults. Additional insulin injected within this timeframe without taking into consideration how much “insulin-on-board” is available risks “insulin stacking” and hypoglycemia. Therefore, the insulin dose for between-meal hyperglycemia needs to be calculated by the patient by the formula:
Total Correction Dose – Insulin-On-Board = Suggested Correction Dose
Patients monitoring frequent between-meal glucose levels often learn quickly that if this concept is not appreciated hypoglycemia will result from overly aggressive insulin dosing from the stacking of the insulin. The fundamental problem with this entire strategy of between-meal adjustments from SMBG data is that even if one knows the amount of insulin-on-board, the direction of the glucose (glycemic trend) is not necessarily appreciated. As an example, a patient with a target of 100 mg/dL, an insulin sensitivity factor of 30 (i.e.. 1 unit for every mg/dL above target [in this case 100 mg/dL] and a blood glucose of 250 mg/dL gave 6 units of insulin aspart 2 hours previously. The patient needs to appreciate that well over 3 units of insulin are still active. However, if the meal happened to be a high carbohydrate, low protein/low fat meal, the glycemic trend two hours later is moving quickly downward, and giving the normal correction dose [(250-100)/30]- 3 = 2 units will likely be too much insulin since the glucose is trending down anyway. Often, the actual glycemic trend is not known, making any correction dose insulin difficult to estimate. This is the main advantage of the use of continuous glucose monitors (CGM). By knowing the glycemic trend, it is possible to much more safely add appropriate doses of correction dose insulin between meals.