Diabetes Technology During the Past 30 Years: A Lot of Changes and Mostly for the Better

I finished my pediatric endocrinology fellowship just over 30 years ago. I can truly say I have participated in the revolution in diabetes management. In 1977, insulin consisted of animal-sourced, relatively impure regular insulin and NPH (neutral protamine Hagedorn), Lente, or PZI (protamine zinc insulin),names many younger readers will not even recognize. Insulin treatment was considered advanced if more than one injection was used for type 1.

Glucose monitoring was with urine, mostly Clinitest tablets that boiled urine, along with some glucose monitoring tapes. Ketones were measured with tablets that changed colors. Dipsticks for urine were just being introduced.

That said, right at the end of my fellowship at the University of Minnesota, Dr. Don Etzwiler sent me to Wyoming with a dozen teens to hike above the tree line. My mantra then was, “Keep it a trace of glucose.” I have no idea about the true glucose values, but I expect, if measured, the hemoglobin A_1c (A1C) values would have been > 9%. We all came home; the greatest danger was the raging thunderstorm in South Dakota on the way back.

Those were simple times, even though we thought we were practicing“state-of-the-art” diabetes management. Technology was really nonexistent when compared to 2008. In fact, compared to the past 30 years, the 30 years from 1947 to 1977 really didn't bring any major changes in management. Urine monitoring and a fixed dose of insulin usually one or two times a day was the norm. What, then, did we have to do in 1977, and how is it different in 2008? What are the true costs and the true benefits of 30 years of diabetes management and the changes experienced?

Perhaps the best way to look at the true cost of technology is through the lives of real patients. Their experiences may give clues to issues facing diabetes professionals as we try to improve the lives of people with diabetes. When appropriate, I will illustrate a point of diabetes management with a graph of glucose data taken from a patient in my practice. After nearly 30 years of pediatric diabetes care, examples of most things that can go wrong and some that have gone right are easy to find. The examples will illustrate a point more completely based on published evidence. I would expect many readers have similar examples in their own diabetes practices.

Type 1 diabetes remains a disease that is all encompassing of a patient's life.^1–4 Diabetes continues to need attention multiple times a day, whether it is to check glycemia by whatever means used or to calculate each meal and snack. Patients must remain attentive at all times to feelings that might indicate hypoglycemia. Few other conditions require this level of diligence. This fact has not changed in the history of modern diabetes management. At this moment in 2008, we have not been able to substantially reduce the diligence necessary to and inherent burden of managing diabetes.

Health care professionals who do not live with diabetes likely can never know the depth of this burden. Although summer camp provides a moment and, in epidemiological terms, more than a year of patient-days for me, I know at the end of the camp, my time with diabetes is over for the year. Such is not the case for family members of children affected by diabetes. Perhaps the most important statement of all is that in spite of all that has changed and become better during the past 30 years, the human cost to adequately care for diabetes has not changed. And the burden might actually have increased.

Compliance literature suggests that if patients actually take medication 80% of the time, they are considered to have been compliant with their treatment.⁵  In diabetes, doing that will get you an A1C of 9%, and if the 20% of the time that you don't take medication is a few days in a row, it will get you a hospital stay. Perhaps as we professionals sit with patients, we should acknowledge that all we offer today does not obviate for 1 minute the diligence expected in the care for diabetes. It was so 60 years ago and will continue to be so into the foreseeable future.

As a pediatric endocrinologist, I will focus mainly on type 1 diabetes. This discussion will review the major milestones in technology for diabetes:the insulins we use, the way we measure glycemia, and our methods of delivering insulin.

Before continuing this survey of the effects of technology, some discussion of nutrition therapy is indicated. One might even say that nutrition therapy also has been affected by the technological revolution in diabetes.⁶  It, too,comes with many controversies. I, certainly, am not an expert here, but I have participated in these changes in management. There are basically two modes of thinking about dietary advice and its relationship to insulin therapy. I will not discuss other issues, such as the types of fat to eat, the glycemic index,the Food Pyramid, the Mediterranean diet, or low-carbohydrate diets. The real issue seems to be whether we make the insulin match the food or the food match the insulin. The thinking is really fundamentally different from when I started diabetes practice. I recognize that individuals may have approached dietary advice differently, but I speak for mainstream pediatric diabetes medicine as practiced by most of my peers.

For much of my career, I prescribed diets (or my nutrition team did above my signature). Patients met with a dietitian, where a daily regimen was prescribed. Typically, patients were given American Diabetes Association (ADA)exchange lists or books with food lists and carbohydrate values. Meals were prescribed, and the insulin dose was fixed. Phone discussions with patients included measuring glucose and then adjusting insulin based on the glucose value, believing that food intake was a constant. Scores of stories of parents begging children to eat or children begging for more are in all of our histories. Worrying about the potential hypo- or hyperglycemic event from mismatched insulin and food was a constant part of diabetes care. The power struggles between parents and children were a part of everyday diabetes care. Clinic visits were full of questions about how to get a child to eat. While this discussion is still a frequent part of diabetes care, it is no longer about “making” someone eat.

Today this has changed, not too radically I hope, but fundamentally changed. One error physicians and other health professionals make is to say that today with diabetes, “you can eat whatever you want and take insulin to match it.” This is not true, but today the approach is to try to match the insulin to what is actually eaten. I frequently tell patients that the dose of rapid-acting insulin never needs to be the same meal-to-meal or day-to-day. The ability to give insulin after meals makes this approach much more attractive and possible.^7,8 

But this expectation comes at a cost that must be realized. Patients are now asked to spend time analyzing meals as to their content, type, and amount and then making judgments as to how much insulin to administer. Expectations of skill sets are obvious, and not all patients or their families are capable of such analysis. Patients or their parents are expected to possess the abstract math skills needed to calculate an insulin dose. Stressing this fact is very important because every patient cannot accomplish the complex management expected today. Careful evaluation of ability is crucial to success. Today, phone calls are not about how much insulin to take, but are instead about changing the carbohydrate ratio for a certain meal or snack or changing the insulin sensitivity factor for a particular time of day.

Because I attend diabetes camp every summer, I have seen this in action. In decades past, the children received insulin before a meal, came to the cafeteria, picked up the trays with the prescribed amount of food (not really,but we tried), ate the food (again, not really), and went about the events of the day. Today, the children come to the cafeteria first and select their own food choices. After the meal, the counselor and the child add up the food eaten and, taking into account the glucose level before the meal, calculate an insulin dose. Having done it both ways, I can say that it takes longer today to eat and give insulin.

Today, however, the act of sitting and urging a child to finish something he or she dislikes (adults do this, too) is history. We now try to match insulin to the food eaten. This is a fundamental shift in our approach to management. Understanding this is crucial to understanding patient management in 2008.

Today, patients are expected to take long-acting insulin in addition to rapid-acting mealtime insulin. These are both analog insulin formulations. Blood glucose is tested at least four times a day, and in addition, there are other opportunities to test and to give insulin because of changes in activity and the eating of snacks. The differences from a decade ago include the type of insulin used as well as the new concept of insulin for food. Many more patients take multiple insulin doses than in the past. Lunchtime insulin doses are no longer avoided “to save the patient from another injection.” The need to manage diabetes not only for short-term benefit but also to avoid long-term complications is universal.

This thinking has led to more work for patients, as well as for their support personnel. I use this term because now individuals other than parents or family members must assume a role in diabetes management. For children,schools have become involved. Children with diabetes expect to be safe at school.⁹  The ADA has, through its advocacy programs, instituted a “Safe at School”initiative, through which school personnel are trained to help with blood glucose testing and insulin administration at school and during school-sponsored events.

Although considered essential, these efforts do come with a cost to families and society. Parents must engage the school system and frequently spend time teaching individuals about diabetes and proper care. School systems must appoint qualified individuals and train them in diabetes care. Today, I expect most parents to be within cell phone contact range to discuss management with their child and school personnel when necessary. Collaborative management among children with diabetes, their family members, and school faculty and staff leads to successful experiences for children with diabetes every day. Gone are the days when people with diabetes were excluded from many of life's events.

Typical patients with type 1 diabetes in the late 1970s were asked to use two types of insulin (regular insulin and a longer-acting insulin preparation), to buy syringes, and to test urine. Typical patients tested their urine one to four times a day and gave insulin one or two times a day. The mindset at the time was, “the less the intrusion into daily life,the better.” Physicians and patients tried to avoid adding to the burden of managing diabetes.

Typical patients in the late 1980s still used regular and long-acting insulin, but it likely was human insulin rather than animal-sourced insulin in the United States. Arguments continue over the relative value of human versus animal insulin; however, the marketplace has driven animal insulins out of diabetes care in the United States.¹⁰ 

The major change in that decade was the introduction of self-monitoring of blood glucose (SMBG). The proper application of this new technology was controversial. I remember discussions early in the decade with the chief medical officer of Blue Cross of Florida, during which we argued over the need to test blood glucose at home and, if necessary at all, why every day, let alone multiple times a day? In the late 1980s, the relationship of control to long-term outcomes was still being debated and awaited the outcome of the Diabetes Control and Complications Trial (DCCT) in 1993.^11,12 

One outcome that is clear today is that the number of SMBG checks performed during a day is directly correlated to outcome of overall blood glucose control, as measured by A1C.^13–16 Indeed, in long-term outcome studies, SMBG intensity may be the most important variable that predicts success. This need does come with a distinct increase in the cost of diabetes care. Blood glucose strips retail at nearly $1 per test. This is clearly one true cost of technology during these past 30 years that carries a benefit for those with type 1 diabetes. I will write prescriptions for patients to test as often as they desire.

Likewise during the 1980s, A1C tests became widely used. Many clinics tested every visit, and some offered point-of-care A1C testing to save laboratory visits and venipuncture.¹⁷ 

One interesting outcome in pediatric diabetes has been the ongoing Hvidore results. The study, in mostly European pediatric diabetes clinics, has enrolled clinics and patients for a decade. Diabetes outcome, as measured by A1C, is clinic-dependent and does not seem to change over time.^18,19 This is in spite of continued participation in the study and knowledge of the clinic's results. Continuing research as to why this is the case may help us understand the effect of individual clinical management styles and outcomes. Perhaps the Hvidore study provides even more interesting data. These international centers practiced a variety of diabetes management. Some use regular and NPH insulin twice a day. Others use multiple daily injections and,of course, some use insulin pumps. There is speculation that diabetes control involves more than just how you manage glycemia. There seems to be some combination of the patients, professionals, clinic management, philosophy, and other factors that together achieves an outcome in diabetes care. I insert this here because many of the advances in technology described below have been available to all the participants in the Hvidore study; yet the outcome as evidenced by A1C is so different among the centers in the study.

By the late 1990s, insulin analogs entered the U.S. market. Today,fast-acting analogs have a major place in diabetes management, especially in North America and Europe. Companies are considering withdrawal of regular insulin in some western countries. Likewise, long-acting analogs were introduced around the turn of the millennium and have enjoyed widespread use.

Unlike glucose testing, the use of analogs has not been shown to improve outcomes as measured by A1C.^20,21 This has led some to suggest that the increased cost of analogs is unjustified.^22,23 This must be tempered by the clear reduction in hypoglycemia with analog use and the lifestyle benefit of timing injections either just before or just after a meal.^24–26 The above-mentioned benefits have been demonstrated for both short- and long-acting insulin analogs. Although overall glucose control, as measured by A1C, is an admirable outcome in diabetes care and clearly important for long-term success, improvements that help ease the day-to-day burden of diabetes management should not be undervalued.

Analog insulins have changed my practice. My first example is of a 5-year-old girl who came to my practice on mixed rapid analog and NPH. She had visits to the Emergency Department for hypoglycemic seizures three times in the past 2 years. The glucose pattern(Figure 1) demonstrates the classic problem with two daily injections of NPH. She had intervals of considerable hyperglycemia and many nights where she was nearly if not actually hypoglycemic. Figure 2shows the same child when switched to a long-acting analog. Technology has helped this child and her mother.

The evolution of diabetes management has led to the introduction of sophisticated continuous subcutaneous insulin pumps. Although the technology is nearly 30 years old, the widespread use of these devices is a fact of the past decade.^27–30 The use of insulin pumps is truly in the mainstream, with articles on their use now in primary care journals. It is no longer the case that pump therapy is restricted to specialized centers with dedicated staff, including nurses,dietitians, and psychosocial support teams.

Patients see friends and colleagues using pumps and request one. The literature is clear. Articles from sophisticated centers with dedicated pump staff have achieved outstanding results even in very young children. There is no doubt that patients whose diabetes is well managed in an insulin pump have excellent results. However, better long-term diabetes control, as measured over years by A1C, does not always result from insulin pump use. Most likely,as with anything new, the novelty wears off and the intense work of diabetes management must continue to achieve good results.³¹ 

Two attributes of insulin pumps bear mentioning because they are true pitfalls in the use of the technology. First, the risk of diabetic ketoacidosis (DKA) cannot be ignored. Too many pump users forget this risk while all is going well. The pump delivers a small amount of insulin on a constant basis (the basal rates). Pump failure, such as a dislodged catheter or other mechanical mishap, leads to a halt in insulin delivery. Lacking any insulin, the patient's glucose rises dramatically. I tell my patients that they are really only 6 hours away from the Emergency Department at any moment. Waking up in DKA is a real threat every morning.

Knowing the pump is delivering insulin at bedtime is crucial. Hyperglycemia at bedtime must be corrected and knowledge that the pump is functioning and the site is adequate is a crucial skill that cannot be overstressed. Changing the pump site and going straight to bed is a mistake, although it makes sense to change it right after an evening bath. Patients must be sure the site will deliver insulin. This is an example where the cost of technology is the time needed to be sure of insulin delivery. Unless the insulin has lost potency,which does happen, an injection with a syringe or pen is a sure way of knowing insulin has been delivered. This must be stressed and restressed at clinic visits. I have seen patients who have successfully used a pump for years confront this situation and not even have syringes in their home. The directions from every manufacturer are clear, but must be re-taught regularly.

Figure 3 shows data from a 12-year-old and demonstrates a very high glucose level. The child received an injection, the site was changed, and the glucose was corrected very quickly. A potential crisis was avoided. The attention to detail with pump management also includes site management.

The second pitfall is what I call “going on autopilot” with the pump. This is essentially forgetting to bolus insulin for food or“living on basal.” This will lead to an A1C of ∼ 9% and poor control, but no immediate symptoms. Regrettably, this occurs far too often. The pump requires more, rather than less, attention to diabetes management. The software included in modern pumps that helps calculate insulin doses based on carbohydrates eaten aids success and improves blood glucose levels. We must assume that the documented skipped boluses with pumps are not unique. We just couldn't know that a traditional shot was skipped because we had no real way to document it. Some clinicians have actually increased the basal rate at meals to help compensate for skipped boluses. This is just 1970s management with a pump. A delayed meal will have the same effect it did then:hypoglycemia.

Choosing patients who are good candidates for pumps is sometimes difficult. There are guidelines, but in practice, knowledge of individuals and their medical and psychosocial situation are crucial variables for successful pump use. Inability or unwillingness to take the necessary time and schedule the needed visits for training and follow-up are a sign of potential difficulty. Part of the cost of the pump, over and above the considerable capital investment by either the patient or the insurance company, is the time needed to properly learn to use the pump and become aware of the potential pitfalls of pump use. A casual or cavalier attitude does not bode well for success.

Some clinicians, including myself, require a trial of pump use before the actual decision to purchase a pump. Often, previous experience with a pump consists of no more than seeing someone else use a pump or reading newspaper or Internet descriptions of a pump and its features. The cost of a pump and its needed supplies comes to several thousand dollars a year over the life of the pump. Reasonably careful selection of pump candidates and then thorough teaching really is the best we can do at present. Regrettably, no treatment works for everyone, and it seems a bigger shame when an expensive pump is prescribed to the wrong patient. Even with this caveat, facilitating pump therapy to patients who request it has proven to be a benefit in most cases.

Although patients' personal knowledge of their blood glucose has evolved quite dramatically, there are significant gaps in that knowledge every day. Even testing 8–10 times a day does not prevent episodes of hyper- or hypoglycemia. The introduction of continuous glucose monitoring (CGM) systems has allowed us to clearly document glucose values to see where either hyper-or hypoglycemia occurs at intervals between glucose tests.

CGM has evolved during the past 5 years. Today, three systems are available for patient use. The questions about this new technology are not all answered. Does it add sufficient data to be worthy of giving over a second site (for pump users) on the body for its placement? Should it be used all the time or just to gather trends? How much does it improve overall glucose control? Does it prevent glucose highs and lows? Is it accurate?

These questions are the subject of articles reporting investigations of the devices.^32–34 One outcome that has been noted is that, within a few days, CGM has allowed patients with very high glucose levels to reduce their glucose and reduce glycemic excursions. A lesser effect has been seen in patients who already have low A1C levels and therefore have less to gain from the new technology. Reductions in the amount of time spent in hyper- or hypoglycemia have been demonstrated. Some have questioned the accuracy of the CGM systems, however,especially when glucose levels are falling.³⁵ 

Two examples of the potential use of CGM are offered for consideration. One of the major issues in pump use has been that glucose tends to rise to unacceptable hyperglycemic ranges after meals even when the next premeal glucose level is in target. Likewise, corrections cause hypoglycemia. The patient illustrated in Figure 4demonstrates a common pattern of postmeal hyperglycemia and occasional hypoglycemia. Perhaps a more clear indication for CGM can be seen in Figure 5, from a college student living alone for the first time. She shares an apartment with three friends, but they do not “supervise” her diabetes. They know about diabetes and could help with an emergency. The glucose values this past fall are telling. While at home for Thanksgiving and winter holidays, the values fell, but back at school, she was unwilling to “go low.” She was willing to consider a CGM system, however.

The unanswered questions enumerated above about CGM systems might have been written in the early 1980s about SMBG. Like CGM systems today, SMBG was technology trying to fit into a diabetes world little changed from the 1920s. I suspect it really took the DCCT—a very expensive National Institutes of Health trial—to confirm the need to monitor and control diabetes. The diabetes community is fortunate that the participants in the DCCT were willing to continue their contributions to diabetes research in the Epidemiology of Diabetes Interventions and Complications follow-up studies, which have extended the DCCT patient data set to > 20 years. Long-term control of blood glucose is crucial. On this point, there is no argument. The arguments are about the best way to get there.

I would never have predicted some of the changes in diabetes management seen during these past 30 years. Diabetes care, although better, is still far from ideal. Hopefully, the horizon is even brighter. One thing is certain:dedicated patients and their physicians will continue to try to use emerging technologies to the best advantage to improve diabetes care.