Many strategies have been used for BG control in Type 1 Diabetes, ranging from attempts to prescribe both food and insulin quantities and timing to adaptive strategies that allow patients free reign in their lifestyles. Here's an outline of a range of control strategies: each step including additional information (and calculations) that enable better BG control.

One can use either Regular or a Fast Acting insulin; such as Lispro, Aspart, or Glulisine. Regular is cheaper than the others, and does not require a prescription in the USA, but the Fast Acting insulins are preferred because they enable dosing closer to the meal, and are used up sooner, mitigating possible "stacking" of insulin doses.

Two injections/day

For those who must be prescribed a basic, unchanging regimen and cannot handle more than 2 shots/day, the least effective strategy is pre-mixed 70/30 or 75/25 insulin, given before breakfast and dinner, where meal size is fixed. This regimen is discouraged (see this eTalk) but may be all that can be done for some patients.

A more flexible version involves a variable combination of bolus insulin with NPH, also taken twice a day before breakfast and dinner. This allows a different ratio of breakfast to lunch size, and better adaptation of dinner and overnight insulin quantities. This "Split/Mixed" regimen was once very common in the USA, but tends to lead to hypoglycemia prior to lunch and at 3am, due to the inappropriate use of NPH for lunch, and administration of overnight NPH at dinner, instead of bedtime.

Three injections/day

The addition of a third injection, NPH at bedtime, to either of the above (retaining the split/mixed or 70/30 shot at breakfast) allows a single dose of R or Fast Acting insulin for dinner. This 3 times per day regimen still requires prescribing a consistent meal size for breakfast and lunch (of fixed size ratio when using 70/30), but allows flexibility with dinner and moves the NPH peak closer to dawn.

Four plus Injections/day

Basal/Bolus is the term used for these regimens, because insulin is applied as a combination of a basal dose with a separate bolus dose for each meal/snack. This regimen mimics the way the pancreas supplies insulin; with a background insulin release adequate to keep BG steady absent food, and a bolus for food consisting of a first phase that is released at once to raise insulin levels quickly, followed by a second phase that continues until glucose returns to normal. Insulin formulations do not shut themselves off the way the pancreas second phase does, but overall, Fast Acting insulins emulate the combined effect of first and second phase pancreatic insulin release reasonably well.

Basal Insulin

Unfortunately, no study has been done to determine what the Basal insulin dose profile really is, or how it varies by age, weight, sex, etc. during the day. There is a tendency for glucose to rise in the early morning, known as the Dawn Phenomenon, which is stimulated by growth hormone. It is possible that this diminishes with age, but no studies have been done to substantiate it. Exercise is known to affect insulin sensitivity as well, with the lack of walking around 'exercise' overnight contributing to higher BG in the morning, and the resumption of daily activity to lower BG in the afternoon.

With basal insulin need as much as three times higher in the early morning as in the late afternoon, long acting insulins such as Glargine and Detemir are inappropriate for basal use in Type 1 patients, though they may have a place in Type 2 management. With basal need varying with exercise, and potentially changing from day to day with sports activity, the most flexible basal regimen requires an insulin pump for hour by hour management. For those whose exercise level is consistent, however, it is possible to adapt two or three NPH doses to daily basal need. Either way, it is important to establish and verify this while fasting (a description may be found in this eTalk).

Bolus Insulin

The following control strategies involve bolus doses calculated according to the patient's Insulin Sensitivity Factor (ISF), the food eaten, residual insulin remaining from an earlier dose, food yet unabsorbed from prior meals, and exercise. For these calculations to work as intended, it is imperative that the underlying basal dose be independent of any food intake, so getting that right is perhaps the most important thing a Type 1 patient can do for a true Basal/Bolus regimen. With that in place, calculation verified with experimentation can establish the ISF (how much 1 Unit of insulin lowers BG), the Carbohydrate Ratio (how much insulin balances 10g of carbohydrate), and their product (how much 10g of carbohydrate raises BG). These may then be used in the calculations needed below.

Bolus adapted to the Meal

Basic Basal/Bolus regimens add up the carbohydrate in a meal, apply the Carbohydrate Ratio, and inject that amount as a bolus dose. Meals high in protein can add to that with an insulin/Protein ratio. The effect of the bolus will have run out (with Fast Acting insulins, anyway) at the next meal, when you can see how close you came with the prior meal's bolus. Over time, and with log books to help, the attentive patient can learn how much to take for a familiar, well-known meal. Fast food and vending machine snacks often provide carbohydrate and protein quantities that ease bolus calculation using the above ratios.

Bolus adjusted for pre-prandial BG

If pre-prandial BG is not normal, the dose calculated for the meal may be supplemented by using the ISF, reducing the bolus for low BG and increasing it for high. This is the main reason for testing BG before eating, though seeing how you did with the prior meal's (or bedtime) bolus is enlightening.

Adjustments for on-board Insulin

Whenever insulin is given soon enough after the prior insulin dose that that prior dose has not been used up, it is necessary to take what is called the "on-board" insulin into account when making any of these bolus dose calculations. This is of particular concern at bedtime, but can be necessary before and after meals as well.

The percent of injected insulin remaining on-board as a function of time may be calculated as one minus the integral of insulin action up to that point in time, divided by its ultimate total action, expressed as a percent. This may be represented graphically by a curve of percent vs. time, starting at 100%, then falling ever faster up to the steepest point when insulin action peaks, then flattening out as the curve approaches 0% at about 5 hours (for Fast Acting insulin). An example may be found in this paper.

Using the time since the prior injection and this curve you can get the percent remaining, and multiplying this by the injected amount yields the unabsorbed insulin. Multiplying that by the ISF tells you how much BG will ultimately be lowered (from where it is now) because of it.

Bedtime Corrections

In the past, supplementing insulin at bedtime was discouraged for fear of overnight hypoglycemia. This was often due to "stacking" insulins (absent on-board adjustment) and from the too-early NPH dose common with split/mixed regimens of the past. With a carefully established 'true basal' dose, however, and adjustments for on-board insulin at bedtime, the ultimate, 3am, BG can be predicted. If low, a food supplement may be calculated. If high, an insulin supplement. If close to normal, nothing need be done (apart from a possible basal dose).

Pre-prandial on-board Adjustment

Any bolus adjusted for pre-prandial BG (as above) that occurs soon after (less than 5 hours for Fast Acting insulin) a prior bolus should be further adjusted for the on-board insulin as described above.

Post-prandial Corrections

There is a period after eating when the insulin bolus is not yet used up, but most of the meal has been absorbed. During this period impending hypoglycemia from an excessive bolus can be predicted while BG is yet high enough to deal with it. With Lispro and Aspart, for example, about 25-30% of the bolus remains on-board two hours after injection. This eTalk describes how to calculate how much further BG will fall (because of the on-board insulin) from what it is at that time. Should the ultimate BG be too low, you now know hypoglycemia is coming, and can use the predicted BG level to calculate how much carbohydrate need be eaten to preclude that and wind up normal instead.

Should post-prandial BG adjusted for on-board bolus insulin be too high, on the other hand, then an insulin supplement may be calculated (with the ISF) and taken immediately. Except for the effect of unabsorbed food, this should result in normal BG when used up, making the bolus for the next meal depend only on the meal.

Adjustments for on-board Food

The time course of action of food as it is absorbed depends on the contents of the meal. Fat delays gastric emptying, but otherwise does not add to BG, protein is absorbed slowly and does affect BG, while carbohydrate is absorbed relatively quickly. There are now at least three models of the effect of food on BG (Worthington, Hovorka, Cobelli). My model proposes its parameters as a comprehensive measure of the associated food, in lieu of the simplistic Glycemic Index, but to date these models and their parameters remain theoretical. With experience, however, patients can estimate on-board food from the meal size, content and time since eating. Whatever the effect, it will always be positive, so presuming food to be completely absorbed in any calculation will result in BG no lower than expected.

Adjustments for Exercise

Exercise is known to affect insulin sensitivity, and heavy exercise affects insulin sensitivity for some time. There has been little research done, however, and no model exists as there is for on-board insulin, or even on-board food. This is an area where patients will have to pay attention and learn what effect exercise has for them.