Diabetes Care

Volume 22 Supplement 2
Improving Prognosis in Type 1 Diabetes
Proceedings from an Official Satellite Symposium
of the 16th International Diabetes Federation Congress

These pages are best viewed with Netscape version 3.0 or higher or Internet Explorer version 3.0 or higher. When viewed with other browsers, some characters or attributes may not be rendered correctly.


How to Ameliorate the Problem of Hypoglycemia in Intensive As Well As Nonintensive Treatment of Type 1 Diabetes

Geremia B. Bolli, MD

Maintenance of long-term near-normoglycemia by intensive therapy largely, if not fully, prevents the onset of microangiopathic complications and delays progression of complications in type 1 diabetic patients. However, intensive therapy has been reported to increase the frequency of severe hypoglycemia. In addition, a number of experimental studies have shown that a few episodes of mild, recurrent hypoglycemia blunt the symptom and hormonal responses to hypoglycemia over the next few days. At present, the critical "post-DCCT" (Diabetes Control and Complications Trial) questions are: is it possible to maintain long-term HbA1c<7.0 %, first, without increasing the frequency of severe hypoglycemia, and second, without increasing the frequency of mild, recurrent hypoglycemia? The answer is yes. The key factors are use of a physiological model of insulin replacement and the education of patients to appropriate the decision of insulin dose based on blood glucose monitoring and eating patterns. Hypoglycemia unawareness should be suspected whenever HbA1c is <6.0 (upper normal limit 5.5%) and the patient does not report autonomic symptoms when their blood glucose level is <3.0 mmol/l. The unaware patients should be treated with a short-term program of meticulous prevention of hypoglycemia, which reverses the abnormalities of responses of symptoms, hormonal counterregulation, and brain cognitive function. In turn, reversal of these abnormalities decreases the risk for severe hypoglycemia. Importantly, a program of meticulous prevention of hypoglycemia does not result in loss of long-term near-normoglycemia, i.e., it is compatible with the glycemic targets of intensive therapy.

Diabetes Care 22 (Suppl. 2):B43–B52, 1999

The year 1993 is a milestone in the modern history of type 1 diabetes. Two large intervention studies independently proved that long-term maintenance of near-normoglycemia, as indicated by the percentage of HbA1c, strongly protects against onset and/or progression of microangiopathic complications in type 1 diabetes (1,2). Thus, there is no longer a scientific excuse not to intensively treat patients with type 1 diabetes, especially young patients with a recent onset or short duration of the disease. These are the patients who likely benefit most from maintenance of long-term near-normoglycemia because of their long life expectancy.

Unfortunately, several barriers still prevent the majority of type 1 diabetic patients all over the world from being on intensive treatment. The reasons are multiple. Perhaps the most common are economical, cultural, and organizational. There is no doubt that intensive therapy is expensive, especially the cost of self-monitoring of blood glucose. However, besides money, a new culture of treatment of type 1 diabetes by diabetologists and an appropriate structure of diabetes clinic are required.

An additional barrier to the generalization of intensive treatment of type 1 diabetes is inherent in the results obtained in the above-mentioned studies (1,2). Severe hypoglycemia (defined as an episode severe enough to require help from a third party) has been reported to be more frequent in intensively treated type 1 diabetic patients with an HbA1c of ~7.0 % than in those "conventionally" treated with an HbA1c of ~9.0 % (300% more frequent in the Diabetes Control and Complications Trial [DCCT]) (2). Hence, there is diffuse concern that intensive therapy, though beneficial on one hand in terms of protection against long-term complications, may be detrimental and risky for type 1 diabetic patients on the other hand because of the elevated risk of severe hypoglycemia.

In this review, the Manichean view that intensive therapy is "good on one hand, but bad on the other"—that it necessarily carries the risk for more frequent severe hypoglycemia and that it is in principle not applicable to the vast majority of type 1 diabetic patients—will be confuted. In addition, it will be made clear that prevention of hypoglycemia is an essential part, but also the natural consequence, of a rational plan of insulin therapy.

GLUCOSE AND BRAIN METABOLISM— Tissues such as muscle and liver may easily switch from oxidation of glucose to other nonglucose fuels, i.e., nonesterified free fatty acids, ketones, and lactate. Thus, these tissues are not energy-deprived during hypoglycemia. In contrast, the brain can utilize only glucose as source of energy. In fact, although in theory the brain can oxidize ketones (3,4) and lactate (4,5), this occurs in humans only under experimental conditions where supraphysiological concentrations of these substrates are produced in plasma during an exogenous infusion. During insulin-induced hypoglycemia of type 1 diabetic patients, plasma ketone concentration decreases after hyperinsulinemia, and lactate concentration does not increase substantially (6). Thus, these substrates cannot compensate for the condition of neuroglycopenia after hypoglycemia. Because the brain cannot perform gluconeogenesis nor store considerable amounts of glucose in the form of glycogen, the brain is strictly dependent on continuous glucose delivery from the circulation for its metabolism and function.

DEFINITION OF HYPOGLYCEMIA— In classic textbooks, hypoglycemia is usually defined as plasma glucose concentration <50 mg/dl (~2.7 mmol/l) (7,8). However, in nondiabetic subjects, the suppression of endogenous insulin secretion (the very first counterregulatory response to hypoglycemia) already occurs with a marginal decrease in plasma glucose below normal postabsorptive values (~10 mg/dl, ~0.5 mmol/l), and the responses of counterregulatory hormones occur when plasma glucose concentrations decrease to ~65 mg/dl (~3.5 mmol/l) (9). Thus, a more conservative, physiological definition of hypoglycemia would be plasma glucose concentration below ~70 mg/dl (~4.0 mmol/l). This concept is important, not only in physiology but also in the practical treatment of type 1 diabetes, to defining safe glycemic targets for intensive therapy (10,11).

006004a.gif (20927 bytes)
Figure 1—After subcutaneous injection of a therapeutic dose of insulin resulting in similar hyperinsulinemia, the resulting hypoglycemia is more severe in type 1 diabetic patients because of impaired counterregulation. Data from Bolli et al. (12).

WHY HYPOGLYCEMIA OCCURS IN TYPE 1 DIABETES— It is true that in type 1 diabetic patients, hypoglycemia develops because (too much) insulin is injected. However, this is not the sole explanation. When type 1 diabetic patients and nondiabetic subjects are normalized for hyperinsulinemia, hypoglycemia is still more severe and prolonged in the former (Fig. 1) (12). Because type 1 diabetes is a condition of insulin resistance (13), the more severe hypoglycemia of type 1 diabetic patients as compared with nondiabetic subjects cannot be explained by greater insulin sensitivity, but by compromised counterregulatory defenses (11).

The most common and nearly universal defect in glucose counterregulation in type 1 diabetes is loss of glucagon response to hypoglycemia (11). This defect occurs early in the natural history of type 1 diabetes and is irreversible. Unfortunately, many type 1 diabetic patients also suffer from reduced responses of the second key counterregulatory hormone, adrenaline, especially in long-term diabetes (11). Clearly, diabetic patients who have combined defects of glucagon and adrenaline are at the greatest risk for severe hypoglycemia (11). Blunted responses of adrenaline to hypoglycemia are largely due to antecedent, recurrent iatrogenic hypoglycemia, especially in patients with only a few years of diabetes duration (14,15). However, this mechanism appears less operative in long-term diabetes (15) and in clinically overt autonomic neuropathy (16).

WHY IT IS IMPORTANT TO PREVENT HYPOGLYCEMIA IN TYPE 1 DIABETES— The most well-known consequence of hypoglycemia is severe episodes where brain dysfunction may ultimately result in unconsciousness if oral or intravenous glucose or parenteral glucagon are not given at the appropriate time by a third party.

However, in recent years, it has become clear that even mild hypoglycemia, i.e., self-treated episodes of recurrent hypoglycemia, play a major role in the pathogenesis of hypoglycemia unawareness in type 1 diabetes (17). Under these conditions, patients lose most of the autonomic symptoms that normally occur in nondiabetic individuals at a blood glucose threshold of ~55 mg/dl (~3.0 mmol/l). Therefore, not only do unaware patients experience less symptoms, but they realize the symptoms at a lower than normal blood glucose concentration (high thresholds). In addition, in hypoglycemia unawareness, the release of counterregulatory hormones, especially adrenaline, is impaired and again occurs at a lower than normal blood glucose concentration (see above definition of hypoglycemia). Thus, the glycemic thresholds for release of counterregulatory hormones and initiation of symptoms are both shifted downward (high thresholds). In practical terms, this means that type 1 diabetic patients cannot correct hypoglycemia (by endogenous counterregulation and/or oral glucose) as they would otherwise do in the presence of normal thresholds for release of counterregulatory hormones and symptoms. Under these circumstances, whenever blood glucose decreases to below ~50 mg/dl (~2.7 mmol/l), neuroglycopenia and cognitive dysfunction not preceded by symptoms occur. In fact, although the threshold for onset of cognitive dysfunction (which normally occurs at plasma glucose of ~50 mg/dl, or ~2.7 mmol/l) is also increased in hypoglycemia unawareness (i.e., impairment of brain function occurs at lower than normal plasma glucose) (18), the failure to counterregulate with endogenous/exogenous glucose causes a rapid decrease in blood glucose, ultimately resulting in sudden, severe impairment in brain function.

The mechanisms by which mild, recurrent hypoglycemia induces unawareness in type 1 diabetes are unclear. One possibility, derived from experiments in chronically hypoglycemic rats (17), is an increase in the fractional extraction of glucose from blood by the brain (19). Under these circumstances, the paradox of blood glucose being low in the blood (hypoglycemia) but not in the brain (absent neuroglycopenia) occurs. This would explain why the brain does not need to activate either counterregulation or the symptoms, and apparently works appropriately (absent cognitive dysfunction) at a lower than normal blood glucose concentration (18). This hypothesis has been indirectly supported by the observation that in type 1 diabetic patients with hypoglycemia unawareness and recent, antecedent hypoglycemia, glucose utilization by the brain does not decrease during experimental hypoglycemia, in contrast to aware type 1 diabetic patients and normal, nondiabetic subjects, in whom it does decrease (19). An alternative possibility to explain unawareness after recurrent, recent episodes of hypoglycemia is that the response of cortisol to antecedent hypoglycemia blunts the autonomic hormone responses to subsequent hypoglycemia (20,21).

006005a.gif (20927 bytes)
Figure 2—Vicious circle where recurrent hypoglycemia during intensive treatment of type 1 diabetes causes hypoglycemia unawareness and impaired counterregulation, ultimately increasing the risk for severe hypoglycemia.

Regardless of the mechanism, it is intuitive how risky hypoglycemia unawareness is for type 1 diabetic patients. By impairing counterregulation and blunting the warning symptoms, this condition predisposes to further decreases in blood glucose while the patient is unaware of such an emergency. Although at a plasma glucose concentration of ~50–40 mg/dl (~2.7–2.1 mmol/l), the brain of unaware type 1 diabetic patients is "protected,"—i.e., brain function is normal (14,15,18)—if there is a further fall in plasma glucose, severe neuroglycopenia may suddenly occur. At this point, even the appearance of symptoms cannot be helpful to the patient because of the brain dysfunction (18). Therefore, prevention of severe hypoglycemia is only possible if hypoglycemia unawareness is prevented, i.e., if mild, self-treated episodes of hypoglycemia are prevented. Failure to do so creates a vicious circle where repeated episodes of hypoglycemia ultimately result in unawareness, which in turn increases the risk for severe hypoglycemia (Fig. 2). Therefore, there are vital reasons to reduce the frequency of mild, recurrent hypoglycemia in type 1 diabetes.

006006a.gif (20927 bytes)
Figure 3—Normal glucose homeostasis. Data from Schade et al. (22).

HOW TO PREVENT HYPOGLYCEMIA IN INTENSIVE THERAPY— Intensive therapy of type 1 diabetes aims at HbA1c<7.0% to prevent onset and/or progression of long-term microangiopathic complications (1,2). The following steps are of primary importance in prevention of recurrent hypoglycemia while aiming at the goal of long-term near-normoglycemia.

Physiological models of insulin replacement
It is much easier to prevent hypoglycemia if insulin delivery mimics the physiology of endogenous insulin secretion of normal, nondiabetic subjects (Fig. 3) (22). This can be best achieved with the technique of continuous subcutaneous insulin infusion (CSII) by means of a minipump. However, for several reasons, including cost, pumps are unfortunately not very popular among type 1 diabetic patients. A good alternative to pumps is the model of multiple daily insulin injections (MDI) (Fig. 4). It is important to emphasize that this model of treatment may be used to successfully prevent hypoglycemia during both day and night.

052.T1.JPG (19176 bytes)

Blood glucose control during the day. Regular (soluble, fast-acting) insulin is injected 20–30 min before each meal in the area of faster subcutaneous absorption, i.e., the abdomen. The dose depends on the size of carbohydrate meal, but it is interesting that on average, type 1 diabetic patients on a nutritional schedule of three meals per day require relatively low units of regular insulin (Table 1).

006007a.gif (20927 bytes)
Figure 4—Model of insulin therapy based on MDI.

For example, the insulin requirements are low at breakfast because of the carry-over effect of NPH insulin given at bedtime the night before. In fact, one important practical recommendation to patients is not to overestimate the fasting blood glucose concentration to decide the morning dose of regular insulin. Because the "dawn phenomenon" (23) is transient, fasting hyperglycemia quite rapidly wanes during the late morning hours. In other words, type 1 diabetic patients should be educated not to worry too much if they occasionally get up in the morning with fasting blood glucose of 250–300 mg/dl (~14–16 mmol/l) and to be cautious in increasing the insulin dose at breakfast (no more than 1–2 units). Importantly, additive measures under these circumstances are a longer interval between injection and breakfast (45 min) and, if possible, a lighter meal. Of course, if the pattern of marked fasting hyperglycemia recurs over time, the nighttime insulin dose of NPH should be increased. When seeing patients, it is important that the diabetologist look at the recorded pre-lunch blood glucose values and discuss with the patients the algorithm used to decide the breakfast insulin dose. Because mild decreases in blood glucose before lunch are common, patients should be warned against this risk, especially if they are going to exercise in the morning hours.

006008a.gif (20927 bytes)
Figure 5—Importance of the time interval between subcutaneous insulin injection and meal ingestion to improve postprandial blood glucose control in type 1 diabetes. Data from Dimitriadis and Gerich (24).

At lunch and supper, type 1 diabetic patients exhibit similar insulin requirements. One good question is how to calculate the insulin dose for lunch and dinner. Because regular insulin has a somewhat slow absorption, the 2-h postmeal blood glucose cannot be taken into account as a measure of appropriate meal dose. It is the 4-h postmeal blood glucose that indicates if the antecedent insulin dose was appropriate. Thus, patients should not worry about hyperglycemia 90–120 min after lunch, because this is expected with the pharmacokinetics of human regular insulin (Fig. 5) (24). What is important is that blood glucose decreases later on to the target values. This strategy changes substantially with the use of the short-acting insulin analog lispro (see below). Several type 1 diabetic patients, especially those who are totally C-peptide negative and have a long time interval between lunch and dinner (e.g., in southern Europe), may develop quite elevated blood glucose before dinner despite an appropriate dose of regular insulin at lunch (i.e., 4-h postlunch blood glucose of 140–160 mg/dl, or ~8–9 mmol/l). An increase in the prandial dose of regular insulin would increase the risk for postprandial hypoglycemia instead of decreasing the pre-dinner blood glucose (Fig. 4). In this case, it is useful to add 5–8 units of NPH to the prandial dose of regular insulin. This not only improves the early and late afternoon blood glucose (25) but also smoothes the postdinner and overall 24-h blood glucose profile. In fact, by giving NPH twice daily, at lunch and bedtime, the replacement of basal insulin improves as compared with a single bedtime dose of NPH. When few units of NPH are added to the regular insulin at lunch, the dinner insulin dose should be decreased by 1–2 units to avoid after-dinner hypoglycemia.

006009a.gif (20927 bytes)
Figure 6—Risk for preprandial hypoglycemia after subcutaneous injection of a 30/70 insulin mixture of regular/NPH insulin before breakfast in type 1 diabetic patients because of inappropriate hyperinsulinemia in the late morning hours as compared with normal, nondiabetic subjects. Data from Dimitriadis and Gerich (26).

Figure 6 is an example of what one should not do in treating patients (26). If the morning insulin dose is a mixture of regular (30%) and NPH (70%), blood glucose decreases markedly in the late morning hours and results in hypoglycemia despite a midmorning snack. One would easily imagine what happens if lunch were postponed for whatever reason in the experiment of Fig. 6. Thus, to prevent hypoglycemia in the morning, it is important that NPH is not added to the short-acting insulin at breakfast. Unfortunately, most type 1 diabetic patients are still given a large morning dose of NPH at breakfast, and this is an important factor of risk for hypoglycemia. This is also the case with a morning injection of "premixed" formulations (10/90, 20/80, etc.), which should be abandoned. Exceptions to this recommendation are small children who do need a midmorning snack. However, in this case no more than 2–4 units of NPH should be added to the regular insulin for a final ratio 80/20 or 90/10 (regular/NPH).

Blood glucose control during the night. At least 50% of hypoglycemic episodes occur at night, and most of them are not recognized. Unrecognized nocturnal hypoglycemia not only results in patients not feeling well the next morning, but also causes unawareness of hypoglycemia (27). Thus, if prevention of hypoglycemia is always important, prevention of nocturnal hypoglycemia is even more important.

Type 1 diabetic patients have variable insulin requirements at night—i.e., they require less insulin between midnight and 4:00 a.m. as compared with 5:00–7:00 a.m. (23). Unfortunately, both NPH (as well as lente and ultralente) insulin exhibit an early peak between 2 and 5 h after subcutaneous injection, with subsequent waning of the insulin effect (28). This means that if NPH is injected at dinner, it peaks at approximately midnight and increases the risk for nocturnal hypoglycemia around that time. The best approach to this problem is the use of CSII at a variable rate (23). However, for the vast majority of patients who are on MSI, it is crucial to split the evening insulin dose into regular insulin at dinner and NPH at bedtime to minimize the risk for nocturnal hypoglycemia. This applies to all type 1 diabetic patients, whether they be children, adults, or elderly people, because prevention of hypoglycemia is a key issue for everyone. Unfortunately, the vast majority of type 1 diabetic patients are still on insulin mixtures before dinner with sometimes a snack to prevent nocturnal hypoglycemia. But of course the snack is just a tentative remedy that attempts to fix wrong insulin pharmacokinetics. Rather, it makes much more sense to rationalize the insulin pharmacokinetics first. This is not to say that the split of the evening insulin dose is the solution to the problem of nocturnal hypoglycemia, but it is a measure that helps a lot (29). With the split of the evening insulin dose, the bedtime snack is no longer necessary unless blood glucose is <120 mg/dl (~7 mmol/l). In this case, the night dose of NPH should not be reduced to prevent marked hyperglycemia next morning.

Specific questions
To snack or not to snack? The prescription of snacks by diabetologists dates back to the time when type 1 diabetic patients were treated with twice-daily NPH injections or the so-called split-mixed insulin regimen (mixture of regular + NPH twice daily). The rationale of using snacks was to prevent hypoglycemia at the time of inappropriate hyperinsulinemia, i.e., at 10:00–11:00 a.m. (Fig. 6) (26) and bedtime. With the present regimen of MDI of regular insulin, the snacks are not only not necessary, but actually detrimental to blood glucose control because they increase blood glucose before the next meal. Common clinical experience indicates that patients are in fact happy not to be obliged to snack any longer. Usually patients like only three meals a day. Abolition of snacks is also important to minimize the weight gain associated with intensive therapy.

The obvious exception are preschool- and school-aged children. The midmorning snack cannot be forbidden to a little child or young student, who should socialize with his/her friends during the breaks at school. The negative consequences of the midmorning snack on pre-lunch blood glucose can be minimized by adding 2–4 units of NPH to the breakfast dose of regular insulin. In any case, a less than optimal blood glucose control under these situations has be accepted, because the lifestyle and satisfaction of little children is more important than ideal blood glucose control. However, even under these circumstances, children have to follow the MDI regimen and manage the insulin dose to make the best effort to achieve the best (possible) control. As time goes by and children approach the adult age, it is possible to ask them to stop snacking. This results in an immediate improvement in blood glucose control when they continue the MDI treatment they are already on.

NPH or ultralente? Both insulin preparations are poor surrogates of basal insulin and far less than optimal compared with replacement of basal insulin delivered by CSII. NPH and ultralente demonstrate poor reproducibility from day to day in terms of the kinetics and dynamics of their effects (30). Consequently, these preparations may expose type 1 diabetic patients to sudden, wide fluctuations in blood glucose. This can be easily appreciated in the fasting state, when blood glucose concentration of patients is entirely dependent on these insulin preparations given the night before. New analogs of long-acting insulin preparations (acylated and soluble long-acting insulins) are presently under development (31,32) or investigation (33), but it will take some time before they will be available to patients. It is hoped that these new preparations possess better pharmacokinetics (reproducibility) than do NPH and ultralente.

The longer the duration of action of an insulin preparation, the greater is its variability in subcutaneous absorption (30). Therefore, the real advantage of CSII is the use of regular insulin as basal insulin, because absorption of infused regular insulin is highly reproducible (30). The duration of action of NPH is less than that of ultralente, and therefore its absorption is less variable (30). However, this also depends on the dose of insulin injected, because a large subcutaneous depot in itself increases variability in absorption (30). NPH, in small doses twice daily (Table 1) is preferable to ultralente because of less variability in absorption and better mixture with regular, should this be needed (34).

It is recommended that NPH be injected into the inner part of the thigh, where absorption is slower as compared to other sites.

Is the short-acting insulin analog lispro of help in preventing hypoglycemia? Yes. The benefit is small (~10 % less risk) but well documented, at least for severe hypoglycemia (35). However, one should note that those results (35) were obtained in a meta-analysis of different studies including the most variable insulin regimens, usually not aiming at intensive treatment. Also, because in those studies HbA1c did not decrease with lispro as compared with human regular insulin, despite lower 2-h postmeal blood glucose with lispro, one can conclude that the protective effect of lispro on the risk for severe hypoglycemia was perhaps due to greater increases in blood glucose before meals and in the early part of the night (36).

At present, three studies have examined the effect of lispro on the risk for hypoglycemia in intensive treatment of type 1 diabetes (3739). With both CSII (37) and MDI (38,39), it is possible to decrease HbA1c by ~0.3–0.4% with lispro as compared with human regular insulin, with no increase in the risk for mild and/or severe hypoglycemia. Taken together, these observations indicate that for an identical HbA1c level, lispro decreases the risk for hypoglycemia as compared with human regular insulin, thus confirming previous reports (35). In addition, if lispro is used in programs of intensive therapy that decrease HbA1c, the risk for hypoglycemia does not increase in contrast to human regular insulin (1,2).

Is the short-acting insulin analog lispro useful in intensive treatment of type 1 diabetes? One word of comment is needed about the use of lispro in intensive therapy. This short-acting insulin analog is quite appealing because it can be injected at mealtime or right after the meal and still improve the postprandial blood glucose. For this reason, the short-acting insulin analog is becoming very popular among patients undergoing intensive therapy. However, the duration of action of lispro is shorter than that of human regular insulin, and simply substituting lispro "unit by unit" for human regular insulin results in deficiency of basal insulin and therefore greater hyperglycemia before meals and in the fasting state. This is the reason why in the majority of studies, the better postprandial blood glucose control achieved with lispro as compared with human regular insulin has not resulted in lower HbA1c (40).

In this regard, one teaching example is the recent study of Jacobs et al. (41), who transferred intensively treated type 1 diabetic patients from human regular to lispro insulin before each meal while maintaining the basal insulin as a once-daily injection at bedtime. Under these conditions, mean daily blood glucose did not improve with lispro, despite lower 2-h postprandial blood glucose, and HbA1c did not change because preprandial blood glucose was greater with lispro, especially before dinner (41). This confirms that when lispro is used as meal insulin instead of human regular, more appropriate replacement of basal insulin is required to improve mean daily blood glucose and HbA1c (25,38,39).

052.T2.JPG (19176 bytes)

The problem of substituting for basal insulin when lispro is used at meals in intensive management of type 1 diabetes may be easily solved with CSII (37), but it is more complicated with MDI (38,39) because NPH must then be administered 2–4 times daily (Table 2).

Blood glucose monitoring
Its is suggested that type 1 diabetic patients regularly check their blood glucose concentration before each insulin injection. As stated before, because of the slow absorption of human regular insulin, usually the 2-h postmeal blood glucose is remarkably elevated, and there is no reason to further frustrate patients with repeated demonstration of poor glycemic results at that time of day. The titration of the meal insulin dose is based on the 4-h postmeal blood glucose, and the titration of bedtime NPH is based on the fasting blood glucose.

The above suggested schedule for home blood glucose monitoring should be different with the use of lispro as the mealtime insulin, because the short-acting insulin analog improves the 2-h postmeal blood glucose to a greater extent than human regular insulin. Thus, at least when starting therapy, patients should really check the preprandial and the 2-h postmeal blood glucose to titrate the dose of lispro. With the combination of lispro + NPH indicated on Table 2, the preprandial blood glucose is used to titrate the dose of NPH given with the antecedent injection.

Unfortunately, blood glucose monitoring is the most expensive part of intensive therapy. This is a problem for many patients, both in Western countries such as the U.S. and in developing countries.

052.T3.JPG (19176 bytes)

Blood glucose targets
Table 3 indicates the desired "realistic" values of blood glucose to aim at before meals. It is clear that these values do not strictly aim at normoglycemia but at moderate hyperglycemia. In other words, because our current treatment of type 1 diabetes with subcutaneous delivery of insulin is so imperfect, we should make prevention of hypoglycemia a priority, not make absolute normoglycemia the primary goal. The good part of the DCCT results is that with an HbA1c value of ~7.0 %, i.e., a value slightly greater than normal, the risk for onset and/or progression of complications appears to be largely if not fully prevented. Thus, there seems to be a "window" of values of HbA1c just above the upper limit of the nondiabetic range where patients are protected simultaneously from the onset/progression of complications (1,2) and the risk for severe hypoglycemia (10,42). Thus, until new, safer means of therapy are available, there is no reason to decrease the HbA1c to the normal range, where the risk for hypoglycemia is greater (10).

Time spent with the patient is the most valuable way to decrease the HbA1c and frequency of hypoglycemia. Unfortunately, patients are many, and real diabetologists are few. Diabetologists too busy in academics are not the ideal professional people to take care of patients. Nurses, although very important in their job, cannot be fully delegated to assist most of the patients' requirements.

The most important job that diabetologists should do is to transmit to their patients the motivation and the enthusiasm to carry the burden of intensive therapy (nearly) throughout their lives. Patients should be reassured that they can have a normal lifestyle and can be nearly free of hypoglycemia and, based on the DCCT data, long-term complications. Only when patients develop a positive attitude toward their diabetes and consider type 1 diabetes not a as disease, but as a manageable, though sometimes tedious condition, does the diabetologist succeed in his/her job. Understandably, this is particularly important with the youngest patients.

LONG-TERM RESULTS WITH INTENSIVE THERAPY IN NEWLY DIAGNOSED TYPE 1 DIABETIC PATIENTS— The results of the above strategy to maintain long-term near-normoglycemia, minimize the risk for severe and mild, recurrent hypoglycemia, and prevent hypoglycemia unawareness have recently been described in detail (10). Notably, the frequency of severe hypoglycemia in that report (10) was ~60 times less than in the experimental group of the DCCT (2), despite similar HbA1c levels. We believe that the result can be explained primarily on the basis of long-term experience with intensive therapy, continuing education, and close contact with patients. Over the last few years, we have found it important to assist the newly treated type 1 diabetic patient on an intensive regimen with the cellular phone. Over the initial weeks of the new treatment, the patient is allowed to call (free) each time he/she performs blood glucose monitoring to discuss the insulin dose, the meal plan, etc. It is our belief that such a link with the patient helps him/her to accept the uncertainties of a new treatment, such as intensive insulin therapy, under the most different and sometimes unexpected circumstances. Most important, this approach avoids hospitalization and its artificial daily schedule and nutrition plan.

Results similar to ours have just been published by Bott et al. (42) in a larger group of type 1 diabetic patients. What Bott et al. have shown is that it is possible to decrease the HbA1c with intensive therapy and at the same time decrease the risk for severe hypoglycemia. These results, combined with ours (10), are reassuring and should prompt us to implement intensive therapy in nearly every type 1 diabetic patient without the intellectual limitation derived from the DCCT (2) that the risk for severe hypoglycemia necessarily increases when glucose control improves.

WHAT TO DO IN TYPE 1 DIABETIC PATIENTS WITH RECURRENT HYPOGLYCEMIA, HYPOGLYCEMIA UNAWARENESS, AND HIGH RISK FOR SEVERE HYPOGLYCEMIA— Because unawareness of hypoglycemia in type 1 diabetes is largely, if not fully, the result of antecedent, recurrent hypoglycemia, it is expected that meticulous prevention of hypoglycemia reverses most of the symptom abnormalities. In fact, this is exactly what has been demonstrated as early as after few days of meticulous prevention of hypoglycemia in a group of type 1 diabetic patients (14). Most important, these results can be maintained long term (15). Also, the responses of adrenaline to hypoglycemia improve after meticulous prevention of hypoglycemia in patients with short-term type 1 diabetes. This indicates that most, if not all, of the abnormal adrenaline responses to hypoglycemia are secondary to antecedent hypoglycemia. In contrast, in patients with long-term type 1 diabetes, responses of adrenaline to hypoglycemia improve but do not normalize (16). This indicates that (unknown) factors other than antecedent hypoglycemia account for reduced responses of adrenaline to hypoglycemia in long-term type 1 diabetes. Taken together, these pioneering observations, which have been subsequently confirmed (43,44), indicate that hypoglycemia unawareness is a "functional" syndrome that may come and go depending on the frequency of antecedent hypoglycemia. It is important to mention that the results of the above-mentioned studies (14,15,43,44) have also been reproduced in type 1 diabetic patients with autonomic neuropathy (16). In conclusion, a therapeutic program aiming at prevention of hypoglycemia like the one described above can be implemented not only in patients with short-term but also in those with long-standing type 1 diabetes and/or autonomic neuropathy, and positive results in recovery of awareness and counterregulation should be expected as long as hypoglycemia is successfully prevented.

What is surprising about the program of prevention of hypoglycemia is that the HbA1c increases no more than 0.5–0.7%, which indicates that the protection against onset/progression of long-term complications is not lost (1,2). Thus, it is possible to combine prevention of long-term complications and prevention of hypoglycemia.

HOW TO PREVENT HYPOGLYCEMIA IN TYPE 1 DIABETIC PATIENTS ON NONINTENSIVE INSULIN THERAPY— Nonintensive therapy is defined as a model of insulin delivery based on MDI aiming at mean blood glucose concentration and HbA1c well above the values indicated by the DCCT to prevent long-term complications (Table 3). For different reasons, special populations, such as young children, adults with devastating advanced complications, and type 1 diabetic patients above ~65 years of age, should be treated with nonintensive, rather than intensive, therapy. For example, in small children it is often not possible to achieve the glycemic goals of intensive therapy, whereas in adults with major complications and elderly patients, the irreversibility of complications and/or the reduced life expectancy makes it useless to aim at near-normoglycemia.

One concept should be clear, though. The difference between intensive and nonintensive therapy as defined in this article is strictly limited to the glycemic targets outlined in Table 3. Everything else described above in terms of therapy for type 1 diabetes (MDI model of insulin treatment, diet, blood glucose monitoring, education, etc.) should be identical in nonintensive and intensive therapy. Thus, there is a fundamental difference between the nonintensive therapy proposed in this article and the "conventional" treatment of the Stockholm (1) and DCCT studies (2), where "conventional" is one or two daily insulin injections of mixtures of insulins. The "conventional" insulin treatment proposed by those studies (1,2) is no longer acceptable, neither for the patients in whom it is not possible nor for those in whom it is not convenient to aim at long-term near-normoglycemia, because of the high risk of hypoglycemia that the "conventional" treatment carries over. For example, in the DCCT study, the "conventional" treatment resulted in a frequency of severe hypoglycemia of ~0.20 episodes/100 patient-year, despite markedly elevated HbA1c (up to ~9.0 %). In other words, on average, each individual patient experienced severe hypoglycemia approximately once every 5 years, despite long-term elevated blood glucose. Understandably, elevated HbA1c and a high risk for severe hypoglycemia is the worst therapeutic combination for the type 1 diabetic patient. On the contrary, the priority of nonintensive therapy is prevention of hypoglycemia, which can be solely achieved with the above-described MDI treatment and the other therapeutic measures.

Unfortunately, for cultural, economical, and organizational reasons, the "conventional" treatment (1,2) is still the most popular treatment of type 1 diabetes around the world. Unless this changes to either intensive or nonintensive treatment, where appropriate, over the next few years, it is difficult to expect a global decrease in the frequency of hypoglycemia in insulin-treated type 1 diabetes in the near future.

CONCLUSIONS— As of today, the intensive treatment of type 1 diabetes can be considered successful when it fulfills the following endpoints. First, HbA1c should be maintained <7.0% to prevent long-term complications (1,2), but >6.0% to prevent hypoglycemia unawareness (10). Second, type 1 diabetic patients should be free of the risk for severe hypoglycemia. Third, type 1 diabetic patients should maintain intact over the years the early autonomic symptoms of hypoglycemia that inform them about dangerous decreases in blood glucose and the need to eat to prevent severe hypoglycemia and reverse it in an early phase. Last but not least, intensively treated type 1 diabetic patients are well treated when they enjoy a normal lifestyle, similar to that of nondiabetic subjects.

These goals are feasible nowadays. For young type 1 diabetic patients, the initiation of insulin therapy is like the start of a lifelong journey. They should be given a good car (i.e., CSII or MDI model combined with the "instructions"; see above) from the very first clinical onset of diabetes. At a very young age, they should initially drive slowly (i.e., accept the risk of elevated HbA1c) to gain familiarity with the technology of the car (i.e., learn how insulin works). Then, they should speed up (to decrease HbA1c to the realistic goals of intensive therapy) and either maintain a constant speed or vary it depending on what happens in life. As decades go by and the journey gets close to the end, there is no reason to continue to speed up and risk accidents; better to slow down (i.e., switch to nonintensive therapy) and enjoy the last part of life relaxed, with no risks for hypoglycemia. But to achieve this, pediatric and elderly type 1 diabetic patients have to keep the good car (i.e., MDI treatment) and not switch to "conventional" treatment (1,2) just because they are very young or elderly.

Acknowledgments— I wish to acknowledge the many patients of our diabetes center on intensive and nonintensive therapy.

1.  Reichard P, Nilsson BY, Rosenquist U: The effect of long-term intensified insulin treatment on the development of microvascular complications of diabetes mellitus. N Engl J Med 329:304–309, 1993

2.  The Diabetes Control and Complications Trial Research Group: The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N Engl J Med 329:977–986, 1993

3.  Amiel S, Archibald H, Chusney G, Williams A, Gale E: Ketone infusion lowers hormonal responses to hypoglycaemia: evidence for acute cerebral utilization of a non-glucose fuel. Clin Sci 81:189–194, 1991

4. Veneman T, Mitrakou A, Mokan M, Cryer P, Gerich J: Effect of hyperketonemia and hyperlactacidemia on symptoms, cognitive dysfunction and counterregulatory hormone responses during hypoglycemia in normal humans. Diabetes 43:1311–1317, 1994

5. Maran A, Cranston I, Lomas J, Macdonald I, Amiel S: Protection by lactate of cerebral function during hypoglycaemia. Lancet 343:16–20, 1994

6. Fanelli C, Di Vincenzo A, Modarelli F, Lepore M, Ciofetta M, Epifano L, Pampanelli S, Brunetti P, Bolli GB: Post-hypoglycaemic hyperketonaemia does not contribute to brain metabolism during insulin-induced neuroglycopenia in humans. Diabetologia 36:1191–1197, 1993

7.  Foster DW, Rubenstein AH: Hypoglycemia. In Harrison's Principles of Internal Medicine. Wilson JD, Braunwald E, Isselbacher KJ, Petersdorf RG, Martin JB, Fauci AS, Root RK, Eds. New York, McGraw-Hill, 1991, p. 1759–1762

8. Young CW, Karam JH: Hypoglycemic disorders. In Basic and Clinical Endocrinology. Greenspan FS, Ed. East Norwalk, CT, Appleton & Lange, 1991, p. 651–654

9.  Fanelli C, Pampanelli S, Epifano L, Rambotti AM, Ciofetta M, Modarelli F, Di Vincenzo A, Annibale B, Lepore M, Lalli C, Del Sindaco P, Brunetti P, Bolli GB: Relative roles of insulin and hypoglycaemia on induction of neuroendocrine responses to, symptoms of, and deterioration of cognitive function in hypoglycaemia in male and female humans. Diabetologia 37:797–807, 1994

10. Pampanelli S, Fanelli C, Lalli C, Ciofetta M, Del Sindaco P, Lepore M, Modarelli F, Rambotti AM, Epifano L, Di Vincenzo A, Bartocci L, Annibale B, Brunetti P, Bolli GB: Long-term intensive insulin therapy: effects of HbA1c, risk for severe and mild hypoglycaemia, status of counterregulation and unawareness of hypoglycaemia. Diabetologia 39:677–686, 1996

11. Bolli GB: From physiology of glucose counterregulation to prevention of hypoglycaemia in type 1 diabetes mellitus. Diabetes Nutr Metab 3:333–349, 1990

12. Bolli GB, Dimitriadis GD, Pehling GB, Baker BA, Haymond MW, Cryer PE, Gerich JE: Abnormal glucose counterregulation after subcutaneous insulin in insulin-dependent diabetes mellitus. N Engl J Med 310:1706–1711, 1984

13. Yki-Jarvinen H, Helve E, Koivisto VA: Hyperglycemia decreases glucose uptake in type 1 diabetes. Diabetes 36:892–896, 1987

14. Fanelli C, Epifano L, Rambotti AM, Pampanelli S, Di Vincenzo A, Modarelli F, Lepore M, Annibale B, Ciofetta M, Bottini P, Porcellati F, Scionti L, Santeusanio F, Brunetti P, Bolli GB: Meticulous prevention of hypoglycemia (near-)normalizes magnitude and glycemic thresholds of neuroendocrine responses to, symptoms of, and cognitive function during hypoglycemia in intensively treated patients with IDDM of short duration. Diabetes 42:1683–1689, 1993

15. Fanelli C, Pampanelli S, Epifano L, Rambotti AM, Di Vincenzo A, Modarelli F, Ciofetta M, Lepore M, Annibale B, Torlone E, Perriello G, De Feo P, Santeusanio F, Brunetti P, Bolli GB: Long-term recovery from unawareness, deficient counterregulation and lack of cognitive dysfunction during hypoglycaemia, following institution of a rationale, intensive insulin therapy in IDDM. Diabetologia 37:1265–1276, 1994

16. Fanelli C, Pampanelli S, Lalli C, Del Sindaco P, Ciofetta M, Lepore M, Porcellati F, Bottini P, Di Vincenzo A, Bottini P, Brunetti P, Bolli GB: Long-term intensive therapy of IDDM diabetic patients with clinically overt autonomic neuropathy: effects on awareness of, and counterregulation to hypoglycemia. Diabetes 46:1172–1181, 1997

17. Bolli GB: Hypoglycemia unawareness. Diabetes Metab 23 (Suppl. 3):29–35, 1997

18. Fanelli CG, Pampanelli S, Porcellati F, Bolli GB: Shift of glycaemic thresholds for cognitive function in hypoglycaemia unawareness in humans. Diabetologia 41:720–723, 1998

19. Boyle P, Kempers S, O'Connor AM, Nagy RJ: Brain glucose uptake and hypoglycemia unawareness in patients with insulin-dependent diabetes mellitus. N Engl J Med 333:1726–1731, 1995

20.  Davis SN, Shavers C, Costa F, Mosqueda-Garcia R: Role of cortisol in the pathogenesis of deficient counterregulation after antecedent hypoglycemia in normal humans. J Clin Invest 98:680–691, 1996

21. Pampanelli S, Lalli C, Del Sindaco P, Lepore M, Ciofetta M, Fanelli C, Brunetti P, Bolli GB: Effect of recent, antecedent hypoglycemia and responses of cortisol per sť, on responses to subsequent hypoglycemia in humans. Diabetologia 40 (Suppl. 1):A27, 1997

22. Schade S, Santiago JV, Skyler JS, Rizza RA (Eds): Intensive Insulin Therapy. Amsterdam, Excerpta Medica, 1983

23. Bolli GB, Perriello G, Fanelli C, De Feo: Nocturnal blood glucose control in type 1 diabetes mellitus. Diabetes Care 16 (Suppl. 3):71–89, 1993

24.  Dimitriadis G, Gerich J: Importance of timing of preprandial subcutaneous insulin administration in the management of diabetes mellitus. Diabetes Care 6:374–377, 1983

25.  Torlone E, Pampanelli S, Lalli C, Del Sindaco P, Di Vincenzo A, Rambotti AM, Modarelli F, Epifano L, Kassi G, Perriello G, Brunetti P, Bolli GB: Effects of the short-acting insulin analog [LYS(B28),-PRO(B29)] on postprandial blood glucose control in IDDM. Diabetes Care 19:945–952, 1996

26. Dimitriadis G, Gerich J: Effect of twice daily subcutaneous administration of a long-acting somatostatin analog on 24-hour plasma glucose profiles in patients with insulin-dependent diabetes mellitus. Horm Metab Res 10:510–511, 1985

27. Veneman T, Mitrakou A, Mokan M, Cryer P, Gerich J: Induction of hypoglycemia unawareness by asymptomatic nocturnal hypoglycemia. Diabetes 42:1233–1237, 1993

28. Bolli GB: The pharmacokinetic basis of insulin therapy in diabetes mellitus. Diabetes Res Clin Pract 6:S3–S16, 1989

29. Fanelli C, Pampanelli S, Lepore M, Porcellati F, Bartocci L, Brunetti P, Bolli GB: Prevention of nocturnal hypoglycemia in intensive therapy of IDDM (Abstract). Diabetes 47 (Suppl. 1). A109, 1998

30. Binder C, Lauritzen T, Faber O, Pramming S: Insulin pharmacokinetics. Diabetes Care 7:188–199, 1984

31. Markussen J, Havelund S, Kurtzhals P: Soluble, fatty acid acylated insulins bind to albumin and show protracted action in pigs. Diabetologia 39:281–288, 1996

32. Radziuk J, Pye S, Bradley B, Braaten J, Vignati L, Roach P, Bowsher R, Di Marchi R, Chance R: Basal activity profiles of NPH and [N+-palmitoyl Lys(B29)] human insulins in subjects with IDDM. Diabetologia 41:116–120, 1998

33. Seipke G, Geisen K, Neubauer HP, Pittius C, Rosskamp R, Schwabe D: New insulin preparations with prolonged profiles: A21-modified argine insulins. Diabetologia 35 (Suppl. 1):A4, 1992

34. Heine RJ, Bilo HJG, Sikkenk AC, van der Veen EA: Mixing short and intermediate acting insulins in the syringe: effect on postprandial blood glucose concentrations in type 1 diabetics. BMJ 290:204–205, 1985

35. Brunelle RL, Llewelyn J, Anderson VA, Gale EAM, Koivisto VA: Meta-analysis of the effect of insulin lispro on severe hypoglycemia in patients with type 1 diabetes. Diabetes Care 21:1726–1731, 1988

36. Ahmed AME, Home PD: The effect of insulin analog lispro on nighttime blood glucose control in type 1 diabetic patients. Diabetes Care 21:32–37, 1998

37. Zinman B, Tildesley H, Chiasson JL, Tsui E, Strack TR: Insulin lispro in CSII: results of a double-blind crossover study. Diabetes 46:440–443, 1997

38. Jansson P-A, Ebeling P, Smith U, Conget I, Coves MJ, Gomis R, Lalli C, Del Sindaco P, Bolli GB, Koivisto VA: Improved glycemic control can be better maintained with insulin lispro than with human regular insulin. Diab Nutr Metab 11:194-199, 1998

39. Del Sindaco P, Ciofetta M, Lalli C, Perriello G, Pampanelli S, Torlone E, Brunetti P, Bolli GB: Use of the short-acting insulin analogue lispro in intensive treatment of IDDM: importance of appropriate replacement of basal insulin and time-interval injection-meal. Diabet Med 15:592–600, 1998

40. Anderson JH, Brunelle R, Koivisto VA, Pfuetzner A, Trautman M, Vignati L, Di Marchi R: Insulin analogue treatment reduces postprandial hyperglycemia and frequency of hypoglycemia in IDDM patients. Diabetes 46:265–270, 1997

41. Jacobs MAJM, Keulen ETP, Kanc K, Casteleijn S, Scheffer P, Deville W, Heine RJ: Metabolic efficacy of preprandial administration of Lys(B28),Pro(B29) human insulin analog in IDDM patients. Diabetes Care 20:1279–1292, 1997

42. Bott S, Bott U, Berger M, Muehlhauser I: Intensified insulin therapy and the risk for severe hypoglycaemia. Diabetologia 40:926– 932, 1997

43. Cranston I, Lomas J, Maran A, Macdonald I, Amiel SA: Restoration of hypoglycaemia awareness in patients with long-duration insulin-dependent diabetes. Lancet 344:283– 287, 1994

44. Dagogo-Jack S, Rattarasarn C, Cryer PE: Reversal of hypoglycemia unawareness, but not defective glucose counterregulation, in IDDM. Diabetes 43:1426–1434, 1994

From the Department of Internal Medicine, Endocrinology and Metabolism, University of Perugia, Perugia, Italy.

Address correspondence and reprint requests to Geremia B. Bolli, MD, University of Perugia, Di.M.I.SEM, Via E.Dal Pozzo, 06126 Perugia, Italy. E-mail: gbolli@dimisem.med.unipg.it.

Received for publication 27 May 1998 and accepted in revised form 20 August 1998.

Abbreviations: CSII, continuous subcutaneous insulin infusion; DCCT, Diabetes Control and Complications Trial; MDI, multiple daily insulin injections.

This article is based on a presentation at a satellite symposium of the 16th International Diabetes Federation Congress. The symposium and the publication of this article were made possible by educational grants from Hoechst Marion Roussel AG.

Copyright © 1999 American Diabetes Association
Last updated: 3/99
For ADA Related Issues contact

For Technical Issues contact webmaster@diabetes.org