Boluses in Insulin Therapy

Abstract

Giving a bolus is one major part in multiple dose insulin therapy (MDI) along with basal insulin substitution. To adjust the bolus optimally, different factors like carbohydrate content and composition of the meal, correction factors, and timing have to be considered. Advances in technologies like bolus advisors can assist the patients but still there a several open questions and technical challenges regarding boluses. This commentary provides an opportunity to address several of the above-mentioned factors influencing the result of bolusing. It shall draw attention to those factors and address the current opportunities, for example, continuous subcutaneous insulin infusion (CSII), as well as the need for further studies which can help to improve diabetes insulin therapy by means of the correct use of boluses.

Treatment with multiple daily injections (MDI) of prandial and basal insulin is the recommended insulin therapy for most individuals with type 1 diabetes and some with type 2 diabetes. ¹ However, safe and efficacious use of this therapeutic approach is often challenging.^2,3 The differences of injecting intermediate or long-acting insulin to cover basal insulin requirements in MDI therapy or the option to infuse short-acting insulin with low rates in insulin pump therapy (continuous subcutaneous insulin infusion, CSII) are obvious. One of the advantages of CSII is that in principle, by increasing the infusion rate adequately, any potential type of bolus can be infused to optimally cover the prandial insulin requirements. The required bolus might vary according to meal composition, for example, in amount or time of delivery. Modern insulin pumps offer different types of preprogrammed boluses to enable coverage of prandial insulin requirements.

The aim of this commentary is to discuss the different requirements, challenges and pitfalls regarding boluses in CSII in a more technical way and show possible clinical implications.

Importance of Boluses

Prandial insulin coverage and insulin for correction of high blood glucose is delivered as bolus. But what is the clinical relevance of selecting adequate and “right” boluses on the outcome of insulin therapy as a whole? In other words, do patients with an optimal coverage of prandial insulin requirements have a better glucose control in comparison to those that do not use the right bolus and sometimes do not select the optimal dose for various reasons? Especially challenging are the difficulties in estimating the right carbohydrate content of meals and the mathematical calculation of the insulin dose. As the recommended portion of the total daily dose (TDD) of bolus insulin should be in the range of 50-70%, one would assume that a deviation from this portion or an incorrect bolus application has a substantial influence on glucose control as measured by glycated hemoglobin (A1c) and frequency of hypoglycemia. ⁴ One has to acknowledge that the prandial insulin exerts its influence because of the length of insulin action during nearly the whole day with the exception of the night hours when no meals are taken. On the other hand the influence of bolus insulin, given for carbohydrate coverage or correction, can only be exactly accounted for when the basal insulin is correctly dosed, preferably checked by fasting tests.

However, for optimal coverage of prandial insulin requirements not only the insulin dose is of relevance but also the timing of the bolus and the speed of insulin absorption. Circulating insulin levels have to increase rapidly to suppress hepatic glucose production and enable high glucose uptake when there is the highest need. This applies, for example, after a meal when large amounts of glucose are absorbed in the gut and delivered into the blood stream. On the other side higher insulin levels have to be sustained for some time in situations of prolonged carbohydrate absorption.

In clinical practice it is also about the compliance of patients to give a bolus at each meal. In 1 study 38% of patients had missed >15% of the bolus doses the previous day; the question here is which are the reasons to not deliver a bolus?^5-7 Patients may skip bolusing or rely on empirical estimates, which can lead to severe negative clinical consequences. ⁸ Four missed meal boluses per week can result in a 1% increase in HbA1c. ⁶ One possible way to reduce missed boluses can be so-called reminder functions in medical devices such as blood glucose meters or insulin pumps.

So, there are different topics to discuss when it comes to optimization of prandial insulin delivery with insulin therapy:

1. Insulin

2. Speed of insulin resorption

3. Infusion site

Selecting the right insulin dose

Carbohydrate counting

Glucose monitoring

Calculated amount of insulin for a bolus by the patient

Use of a bolus calculator

Different types of boluses

Amount of boluses possible and increments of change

Standard, extended or multiwave boluses

Speed of delivery (“super bolus”) and changes of infusion rates over time

Timing of boluses around meals

Accuracy of the amount of insulin delivered (given by the insulin pen or insulin pump itself, but also the amount of insulin actually being delivered into the subcutaneous tissue via the needle or insulin infusion system)

It is of interest to note that only the influence of a few of these factors has been studied systematically. ⁹

Type of Insulin, Speed of Insulin Resorption, and Infusion Site

Today, rapid-acting insulin analogs have in some places already replaced regular insulin, especially in CSII therapy. ¹⁰ In the future, hopefully ultrafast insulins or measures to accelerate insulin absorption like local warming or the addition of, for example, hyaluronidase may improve insulin therapy, especially CSII therapy. Also different routes of insulin application, for example, intraperitoneal (Accu-Chek DiaPort, Roche), can change the profile of insulin action.

Another clinically relevant but often ignored aspect is the influence of the injection site on insulin absorption, that is, when a given skin area is used over and over again for inserting the needle and infusing the insulin, local changes in the subcutaneous tissue might develop (eg, lipohypertrophy or dystrophy and localized amyloidosis).^11-13 A recent clinical-experimental study with insulin injections into skin areas with and without such skin changes showed a considerable difference in the time-action profile of the applied insulin. ¹⁴

Selecting the Right Insulin Dose, Using Carbohydrate Counting and Glucose Monitoring

Giving an insulin bolus for meals is the principle of modern insulin therapy. But so-called fixed meal dose bolusing is still being used in people with type 2 diabetes and in some with type 1 diabetes, especially in regions of the world where carbohydrate counting is not common. Manual calculation of the appropriate insulin dose for a given meal or glucose value correction is a complex, difficult, time-consuming task and discourages adherence. In practice the correct calculation of the bolus dose requires both mathematical competency and an understanding of the factors that must be considered (eg, insulin-to-carbohydrate [CHO] ratios, insulin sensitivity factors [ISF]). Furthermore, the way how to adjust these numerical factors special situations (eg, sports, common infections) for the calculation needs to be known.^15,16

Correct calculation of the appropriate dose has 2 parts and depends primarily on 2 variables: glucose level and carbohydrate content of the meal (and possibly of the fat/protein content) on one side and the correction factor (ISF) and carbohydrate factor (CHO) on the other side. ¹⁷ Methods to find the right factors are not commonly accepted, although there are recommendations, such as in the United States, United Kingdom, and China.^18-20 In addition, manual calculation does not account for the effect of insulin still remaining metabolically active from the previous bolus, which can lead to inappropriate “stacking” of the insulin effects of different boluses, resulting in hypoglycemia. ²¹

Measuring the accurate blood glucose level for the bolus calculation is also more difficult than one might think, as numerous studies and the daily experience of anyone involved in patient-care of people with diabetes showed.^22,23 The right timing and frequency of self-monitoring of blood glucose (SMBG; or other modes of monitoring glucose levels) may vary extremely, intra- as well as interindividually. ²⁴ Most studies show that numerous patients do not use SMBG regularly or not at the advised times (eg, fasting or postprandial) and neither continuous glucose monitoring (CGM) has been shown to be used every day by the patients. Therefore these patients are not able to calculate correction boluses with the correct amount of insulin. The same holds true for estimating the amount and type of carbohydrate in food. This can be estimated by weighing the food and looking up the carbohydrate content in tables or using patient experience after specific training programs.^25,26 As shown in various studies patients tend to underestimate the carbohydrate content of meals to a great extent. As a consequence most patients select a too small prandial insulin dose.^27-30

Bolus Advisors

This explains why usage of bolus advisors (BA) has been shown in clinical trials to be of great help to improve the “quality” of boluses and subsequently also glycemic control.^31-36 Not only the mere calculation of the bolus according to the carbohydrate and correction factors is done by the BA, but also the action time of insulin and the influence of different events (sports, infections or others and even different ones in parallel) can be taken into account. However, calculation of the bolus represents only one part of optimizing prandial metabolic control. As mentioned, the part of estimating the carbohydrate (and fat/protein) content of a meal is still a great source of error, ³⁷ and this part can be supported only partly by the use of a BA.^34,38,39

Using a BA in daily routine by patients and health care professionals also relies to a great extent on the right settings for duration of insulin action, and so on. These settings, as well as the individual bolus factors for carbohydrates and insulin sensitivity, should be reviewed and adjusted regularly in clinical practice, for example, every 3 months during regular visits.^40,41 Taken together, it has been shown that consequent use of a BA over time in MDI and CSII therapy by adults as well as children and adolescents can reduce HbA1c and the frequency of hypoglycemic events.^34,42 One has to add in a critical way though, that to our knowledge, there are no studies until now which examined the different possible settings regarding insulin action in a systematic way. Studies to show this would be difficult to design as these parameters are highly individually different.

Usage of Different Types of Boluses

In a number of studies it was shown that different bolus types enable a better postprandial glycemic control, preventing postprandial hyperglycemia or late postprandial hypoglycemic events.^43-51 Clearly this depends heavily on the composition of the meal. Pankowska and others could show, that the delivery of insulin over longer time for meals with high fat/protein content is beneficial, but one has to acknowledge that the inclusion of fat/protein counting is somewhat elaborate.^37,47 This is the main reason why practically all modern insulin pumps offer different types of boluses but it is not well known if patients use different types of boluses in practice at all. ⁵²

An interesting question is how often and how consistently all the different bolus options are being used by the patients during their daily therapy. This will depend very much on the appropriate education of patients in adequate teaching programs. To our knowledge not too many surveys or systematic data analyses from pump downloads were published analyzing how often patients use the different types of boluses. Another issue is that each manufacturer is using his own terms to describe the different types of boluses their pump can deliver (eg, standard, extended, delayed, dual-wave, multiwave). The use of a “super bolus” (to include some of the following basal rate in a bolus for a high glycemic index meal) has been suggested, but this has to be calculated manually and no larger clinical study has been published to show a positive effect. ⁵³

In this context a critical question is, if the use of different modes of boluses, usage of a BA and other features of modern insulin pumps do make a difference in glycemic control in patients with diabetes on CSII. In other words, are all these features of clinical relevance and what is the evidence for using them? Different clinical endpoints are to be considered, in the short run the frequency of hypoglycemic events and postprandial hyperglycemia and over time the level of glucose control in general (= A1c) as well as patient therapy satisfaction, quality of life and autonomy.^54,55

As stated above, the main reason why patients with diabetes do not use them is probably that they are not appropriately taught in an adequate and consistent way by the diabetes teams to do so. Possibly these diabetes teams themselves lack solid information and experience about the advantage of different bolus types. ⁵⁶ Again, specific, well-planned, and well-executed studies are needed to show the effect and evaluate the importance of the use of different bolus types and their influence on glycemic control. Although it is clear that different bolus types may be only one factor of good glycemic control, today it should be possible to show the effect of different bolus types on different meal types with the use of CGM.

Timing of Boluses Relative to the Meal

It is well known for MDI that the time interval between insulin injection and start of the meal (the so called injection-meal interval) has a significant influence on postprandial glycemic excursions. ⁵⁷ Mostly an interval of 10 to 20 minutes is accompanied by a reduced glycemic excursion. Similar effects can be reached by accelerating insulin speed using a device for local warming of the skin after the bolus is delivered. However, not much is known about the influence of bolus timing regarding a meal with CSII.^24,58,59

Accuracy of Boluses

Interestingly, there are only a small number of published studies investigating the accuracy with which insulin is actually delivered as a bolus via an insulin pump. This is in clear contrast to numerous publications about the accuracy of insulin delivery via insulin pens.^60-66 From a clinical point of view it is of high relevance to know that the intended amount of insulin is truly being delivered to the patient using CSII. It is a special advantage of CSII that very small amounts of insulin can be delivered, but they must be released precisely. In pediatric diabetology, amounts as small as 0.02 IU/h of insulin are applied as basal rates and also boluses as small as 0.1 IU are often used in infants and small children, which is only possible with the use of CSII and not with insulin pens or syringes. In some countries, like Germany, the usage of CSII is very common in this age group (>80%) and accuracy of insulin delivery is therefore of utmost importance. ¹⁰ Insulin dilution could improve the accuracy, especially in children with low insulin need, due to the higher volume delivered per unit and insulin U10 has shown to reduce glycemic variability. Currently there is no diluted insulin commercially available in most countries and not approved for use in CSII and self-diluting is associated with many problems including inverting of the insulin concentrations.

How can the amount of insulin delivered by the pump be measured and what is the relevant site for estimating this? At the end of the catheter/tip of the needle in the subcutaneous tissue or at the tip of the insulin cartridge in the pump/beginning of the catheter? In principle one would prefer to measure the amount of insulin which is actually reaching the subcutaneous space, as this is the site where insulin is affecting postprandial glycemic excursions in reality; however, in practice this is difficult to perform. Again, one has to state that not much is known about the impact of the type of insulin infusion catheters, type of needles (steel or Teflon), length and kind/material of the tubing or kind of connectors that might have an impact on the amount of insulin delivered. ⁶⁷ It would be desirable that such studies are performed and published, even if they can only be accomplished in an experimental way in the laboratory setting for obvious reasons. Even if there will be answers to these questions, one has to look for batch-to-batch differences as it is known from test strips for SMBG that there might be a problem with some manufacturers. ⁶⁸

The impact of different infusion rates and types of boluses in regard to accuracy might also be important. Studies evaluating factors that influence accuracy of insulin infusion most often employ 1 or 2 different insulin rates (when it comes to basal rates) or 1 or 2 “typical” insulin doses with a standard bolus; however, it would be clinically relevant to evaluate if deviations from the target dose are linear over a range of doses or if there are greater deviations (percentage wise) at smaller doses. Along the same line of thinking it would be interesting to evaluate the accuracy of insulin delivery when a so-called multiwave bolus is given. Are the amounts of insulin infused within the different parts of the “wave” all applied with the same accuracy? What about prolonged boluses? Is the prolongation linear, that is, is the total amount of insulin delivered at the same level across the total time or is it given in “bursts”? ⁶⁹ To argue about the pharmacodynamic properties of a given rapid-acting insulin analog can be quite nonessential when an extended bolus is given in intervals compared to true continuous insulin application. ²¹

Speed of Delivery

Different pump types are delivering the insulin bolus with different speed. Differences from 1 IU in 2 seconds up to 40 seconds are possible. Therefore the amount of bolus insulin is an additional factor and this together can influence the time of delivery too. Also the speed of delivery can be adjusted by the patients themselves as for some of them a slower mode of delivery is more comfortable, especially with large amounts of insulin. But this in itself might alter the time of insulin-action for a given bolus. On the other hand this could lead to a better penetration of insulin into the subcutaneous space, avoiding the development of small insulin depots and therefore result in a more homogenous insulin absorption.

Summary

Only few clinicians and patients probably know about such open questions regarding boluses or even think about them at all, although boluses are a major part of insulin treatment in diabetes. Regular, accurate boluses with the right amount of insulin in a timely manner that take into account meal composition, absorption site, and time and other factors influencing insulin action are important. One way of utilizing different ways of giving boluses is the use of CSII with its already potential technical possibilities, for example BA, extended or dual-wave bolus, to name a few.

We do not exactly know how such different factors make a difference from a clinical point of view; however, one has to be aware of this topic when treating and advising patients with insulin treatment.

We hope that our commentary stimulates some discussion and subsequent systematic evaluation of the questions raised about insulin boluses. It might be worthwhile to discuss potential issues regarding basal rates in CSII in a similar manner. ⁷⁰ Only when all possibilities are being used, can optimal therapy of insulin-treated diabetes be achieved.