Continuous Subcutaneous Insulin Infusion and Hypoglycemia

One of the major limiting factors in achieving glycemic control is hypoglycemia.¹ The causes of hypoglycemia are often multifactorial, but it is commonly caused by excess insulin administration due to either an overestimation in the dosing of insulin or active insulin stacking in patients with diabetes who require insulin.² Other causes can include delayed gastric emptying due to gastroparesis, medication, exercise, or alcohol. Hypoglycemic episodes are common for patients with type 1 diabetes and those with advanced type 2 insulin-dependent diabetes.

One of the major limiting factors in achieving glycemic control is hypoglycemia.¹ The causes of hypoglycemia are often multifactorial, but it is commonly caused by excess insulin administration due to either an overestimation in the dosing of insulin or active insulin stacking in patients with diabetes who require insulin.² Other causes can include delayed gastric emptying due to gastroparesis, medication, exercise, or alcohol. Hypoglycemic episodes are common for patients with type 1 diabetes and those with advanced type 2 insulin-dependent diabetes. In type 1 diabetes, plasma glucose concentrations may be <50mg/dl for as much as 10% of the time.² It has also been estimated that the average patient with diabetes can suffer up to two episodes of symptomatic hypoglycemia per week and one or more episodes of severe disabling hypoglycemia per year.²

Participants in the Diabetes Control and Complications Trial (DCCT) who followed a regimen of intensive insulin therapy reduced their risk for microvascular complications such as retinopathy. Patients in the DCCT study who had a glycated hemoglobin (HbA_1c) level <7.5% experienced a threefold increased incidence in severe hypoglycemia.³ Therefore, optimal glucose control is often the result of a tenuous balancing act. Euglycemia and its maintenance are achieved by attempting to minimize hyperglycemia while simultaneously trying to avoid hypoglycemia.² There is also a high economic cost associated with hypoglycemia. The relative cost of caring for one patient with hypoglycemia over a 12-month period has been estimated to be approximately $9,000 more expensive than caring for a patient without hypoglycemia.⁴

Continuous subcutaneous insulin infusion (CSII) provides a closer approximation to the physiological secretion of insulin by the pancreatic islet cells than is possible with multiple daily injections (MDIs).⁵ Most insulin pump models have multiple ‘basal rates.’ Instead of using one or two long-acting insulin analogs to supply basal insulin requirements, the pump can be programmed to deliver multiple (up to 48 different basal rates) pre-set rates of insulin delivery. The patient is able to change these basal rate settings at different times of the day and adjust to accommodate for exercise and illness. Insulin can be delivered in increments as small as 0.05–0.25 units an hour. A cross-sectional analysis showed that the average number of basal rates used by a patient (after being properly adjusted) was between four and five different settings over a 24-hour period.⁶ Another pharmacodynamic advantage of CSII is the relatively small inter-individual variation in the absorption of the quickacting insulin analogs (less than 3% daily) compared with MDIs.⁷ In contrast, intermediate- or long-acting insulin (neutral protamine Hagedorn [NPH], lente, ultralente, or glargine), which usually supplies the basal component of a patient’s insulin needs, can have tremendous variations in absorption (absorption rates can range from 19 to 55% in the same individual).⁸ Utilizing a different injection site three to four times a day further exacerbates the variability in insulin absorption based on the different regions employed rather than the same single injection site for two to three days that CSII uses.⁵
Hypoglycemia and Continuous Subcutaneous Insulin Infusion
Several prospective comparative studies between MDIs and CSII have reported an improvement in overall glucose variability manifested as a lowering of average glycemia with concurrent reduction or maintenance of hypoglycaemia incidence.^9–13 Conversely, if the glucose control is maintained in insulin pump therapy versus MDI, there is often an associated decrease in the overall incidence of hypoglycemia.¹⁰ This reduction in glucose variability has been observed in both pediatric and adult patient populations.^9–13 Boland et al. prospectively observed a large group of adolescents with established type 1 diabetes on intensive insulin therapy during a 12-month period with either MDIs or CSII. One-third of 75 youths aged 12–20 years utilized CSII as their mode of treatment. Although both MDI- and CSII-treated adolescents initially exhibited improved metabolic control, this level of control did not continue to improve in the MDI group (at six months HbA_1c was 8.1 compared with 8.3 at 12 months), whereas the CSII group continued to show improvements in their glucose control during the 12 months of treatment (at six months HbA_1c was 7.7 compared with 7.5 at 12 months). Most importantly, the lower HbA_1c levels in CSII- versus MDI-treated patients were associated with a lower rate of severe hypoglycemic events—almost a 50% reduction in the CSII group (p=0.01).¹¹

A parallel observational study by Hissa et al. compared the 29 patients with type 1 diabetes in Brazil (17 patients on CSII therapy and 12 patients on MDI therapy). They prospectively cross-compared HbA_1c values with repeated measured analysis determined at baseline and every three months over a period of 18 months. The incidence of severe hypoglycemia was also noted. After three months of treatment, patients in the CSII group had a significantly lower HbA_1c level than the MDI group (p<0.05). Of the 17 patients treated with CSII, all had HbA_1c values below 7.5% at 18 months, and 13 (76%) had HbA_1c levels below 7%. No episodes of severe hypoglycemia were noted in either group. The achievement of lower target HbA_1c levels of less than 7% at 18 months was also not associated with an increase in hypoglycemic episodes.¹⁴

In a well-known non-parallel observational study by Bode and colleagues, the incidence of severe hypoglycemia in patients who crossed over from MDIs to insulin pump therapy was studied. From a population of 225 patients, the subjects met the following selection criteria: a minimum of one year on intensive therapy with MDIs and a minimum of one year on CSII after cross-over. They found that the incidence of severe hypoglycemia during MDI therapy declined from 138 to 22 events per 100 patient-years during the first year of CSII treatment (p<0.0001), and remained significantly lower in years two, three, and four on CSII (26, 39, and 36, respectively). Insulin pump therapy was associated with a significant and sustained reduction in severe hypoglycemia without a worsening in the level of glycemic control.¹⁰

Hypoglycemia-associated Autonomic Failure
The concept of hypoglycemia-associated autonomic failure (HAAF) in type 1 diabetes and advanced type 2 diabetes has also been well described by Cryer et al. In a thorough review on the mechanism of HAAF, Cryer et al. explain that repeated iatrogenic hypoglycemia causes both defective glucose counterregulation (by reducing the epinephrine response in the absence of an appropriate glucagon response) and hypoglycemia unawareness (by reducing the sympathoadrenal response and the resultant attenuated response of neurogenic symptoms). Thus, hypoglycemia can beget hypoglycemia through an alteration of the counter-regulatory system. The end result is a vicious cycle of recurrent hypoglycemia.¹⁵

Hypoglycemic unawareness is often seen in patients with diabetes who have experienced repeated hypoglycemia.¹⁶ Hypoglycemic unawareness is often a strong prognostic factor of an increased risk for severe hypoglycaemia and subsequent loss of consciousness.¹⁶ Many patients with diabetes and hypoglycemia unawareness do not experience appropriate autonomic warning symptoms before the development of neuroglycopenia and/or have a reduced counter-regulatory hormone response. It occurs particularly in patients with long-term type I diabetes or in those undergoing intensive insulin therapy.¹⁶ After a reduction of hypoglycemia, a restoration (be it partial or complete) of the counter-regulatory hormone response can be seen.¹⁶ Fanelli et al. looked at the improvement of hypoglycemia unawareness through careful intensive insulin therapy in patients with long and short insulin-dependent diabetes mellitus (IDDM) durations. They studied 21 patients with hypoglycaemia unawareness while on conventional insulin therapy and 20 non-diabetic control subjects. Neuroendocrine cells with cognitive and symptom responses were assessed in a stepped hypoglycemia clamp at baseline and then at two weeks, three months, and one year. In the experimental group, 16 patients remained on insulin therapy with the goal of meticulous prevention of hypoglycemia, and the remaining five individuals were placed in a control group with conventional therapy.

After a year, the individuals in the experimental group who experienced less hypoglycemia (the frequency of hypoglycemia decreased from 0.5±0.05 to 0.045±0.02 episodes/patient-day) and higher HbA_1c levels (5.83±0.18 to 6.94±0.13%) showed improvements in all counter-regulatory hormone and symptom responses to hypoglycemia. In fact, after two weeks of meticulous prevention of hypoglycaemia, the responses of plasma adrenaline increased (1.56±0.15 versus 1.09±0.12nmol/l; p<0.05) after hypoglycemic induction. At three months, further improvement in the plasma adrenaline was seen in its maximal plateau to 2.22±0.19nmol/l (p<0.05 versus resut at basal and two weeks). At one year, the adrenaline responses were the same as at three months. The responses of plasma glucagon to hypoglycemia after one year increased by 20% at the lowest hypoglycemic plateau (from 95±15 to 116±15pg/ml; p<0.05). Similarly, the prevention of hypoglycemia resulted in an improvement in autonomic symptom scores (an increase from 2.2±0.2 to 6.9±1) and of neuroglycopenic symptoms (from 4.1±0.9 to 9.7±1.1; p<0.05). After three months, both the autonomic and neuroglycopenic symptoms normalized to that of their non-diabetic controls.¹⁶
Hypoglycemia Awareness and Continuous Subcutaneous Insulin Infusion
In a small but noteworthy acute feasibility study by Kanc et al., there was an improvement in the counter-regulatory hormone response when nocturnal hypoglycemia was avoided in 14 patients on CSII. CSII with short-acting insulin utilized only at night was compared with a multiple daily injection regimen in patients using the more variable bedtime NPH insulin. During a stepwise hypoglycemic clamp, they studied the effect of this regimen on counterregulatory hormonal responses, warning symptoms, and cognitive function. CSII was associated with a lower frequency of hypoglycemia episodes (mean ± standard error of the mean [SEM]: 16.1±3.1 versus 23.6±3.3; p=0.03) during the six weeks of treatment with maintenance of good glycemic control (HbA_1c 7.2±0.2 versus 7.1±0.2%; p=0.2). Hypoglycemic thresholds for the growth hormone response and for autonomic symptoms were lower for the MDI group on NPH than for those receiving nocturnal CSII treatment. Nocturnal CSII utilization also improved warning symptoms and counter-regulatory hormonal responses to hypoglycemia.¹⁷

Nocturnal Hypoglycemia and Continuous Subcutaneous Insulin Infusion
Sleep still represents one of the most dangerous times for the phenomenon of hypoglycemia-associated autonomic failure to occur.¹⁵ It has been estimated that about 50% of nocturnal hypoglycemia often goes unrecognized. This inability to appropriately treat the hypoglycemia due to reduced consciousness can result in severe hypoglycemia.¹⁸ Persons with type 1 diabetes can have substantially reduced sympathoadrenal responses to a given level of hypoglycemia, and these responses are often further reduced during sleep. Moreover, probably because of their markedly reduced sympathoadrenal responses, persons with type 1 diabetes are much less likely to be awakened by hypoglycemia than non-diabetic persons. Thus, while sleeping, patients with type 1 diabetes have both defective glucose counter-regulation and a form of hypoglycemia unawareness (reduced arousal from sleep).¹⁹

Nocturnal glucose control has been assessed with CSII in children. In a one-year prospective analysis of 51 children (age 10.7±3.1 years, mean ± standard deviation [SD]) HbA_1c was observed before and after introducing CSII; the incidence of hypoglycemia was also measured. After insulin pump initiation, HbA_1c fell to 7.7±0.2% (p<0.001) within three months and remained decreased (7.9±0.1%) at 12 months (p<0.01). Although severe hypoglycemia (<50mg/dl) was reduced in the entire cohort, CSII was effective in lowering HbA_1c without an increase in the occurrence of severe nocturnal hypoglycemia.²⁰ As a popular long-acting insulin analog, glargine is typically thought to reduce the incidence of hypoglycemia versus intermediate-acting human insulin. Armstrong looked at CSII versus insulin glargine with CGMS in order to compare glycemic control. Eight subjects were treated with glargine and 11 subjects were treated with CSII. Subjects had matching HbA1c levels (7.1 versus 7.3%; p=not significant), frequency of reported hypoglycemia and self-monitoring blood glucose (SMBG), age, gender, and body mass index. All subjects were monitored for up to three days by CGMS, including a specific period of four hours following their last meal of the day and overnight until the first meal of the next day. CGMS profiles demonstrated that subjects treated with glargine spent a significantly greater amount of time with low blood glucose (<70mg/dl) than subjects treated with CSII (p=0.03).²¹

Sensor-augmented Pump and Hypoglycemia
Sensor-augmented insulin pump therapy is a convergence of two technologies: continuous insulin infusion therapy and realtime continuous interstitial glucose monitoring (RT-CGM). There have been few published randomized studies on the integration of these technologies when initiated in patients with type 1 patients who are pump-naïve.²² More recently, continuous glucose monitoring has given patients the ability to view their glucose levels in realtime, as well as to review graphs of recent trends in their glycemic control.²³ The application of realtime alarms warns users of impending hypo- and/or hyperglycemia, thereby potentially allowing for either preventive or, if need be, corrective action.^24,25

In a multicenter, randomized, prospective study looking at continuous glucose monitoring, Garg and colleagues observed 91 insulin-requiring patients with type 1 (n=75) and type 2 (n=16) diabetes. Subjects wore a transcutaneous, three-day, continuous glucose-sensing system for three consecutive 72-hour periods. Subjects were randomly assigned to either a control group (continuous glucose data not provided) or a display group (continuous glucose data not provided during period one but displayed during periods two and three). During periods two and three, patients in the display group had realtime access to sensor glucose values and were provided with high (200mg/dl) and low (80mg/dl) alerts, and a low (55mg/dl) alarm. Compared with control subjects, the display group spent 21% less time being hypoglycemic (55mg/dl), 23% less time being hyperglycemic (240mg/dl), and 26% more time within the target glucose range (81–140mg/dl; p=0.001 for each comparison). Nocturnal (10:00pm to 6:00am) hypoglycemia, as assessed at two thresholds, was also reduced by 38% (55mg/dl; p=0.001) and 33% (55–80mg/dl; p=0.001) in the display group compared with control subjects.²⁶

Data from long-term implantable sensors have demonstrated reduced glucose excursions when realtime continuous glucose values were available to patients with type 1 diabetes.²⁷ Given the established evidence in the medical literature on the efficacy of glucose control with insulin pump therapy and the emerging positive balance of data on RT-CGM, it was not surprising that patients who were pump-naïve and on multiple daily injections (MDIs) benefited when they transitioned to the combined use of realtime glucose monitoring integrated with insulin pump therapy (Paradigm® REAL-Time 722 System).²² Lee et al. reported their site-specific results in a pilot study, the Sensor Augmented Insulin Pump therapy (STAR 2) trial. The Paradigm REALTime System enabled participants with type 1 diabetes to achieve greater reductions in HbA_1c levels than those achieved by patients maintained on MDI. Most importantly, these significant reductions occurred without an increased incidence of severe hypoglycemia. While the magnitude of the HbA_1c improvement from baseline was greater in the study arm (-2.05) compared with the control arm (-1.08), it is important to note that the study arm had a higher initial HbA_1c average and had a greater number of clinician visits as necessitated by initiation of the pump and RT-CGM therapy (the study arm included eight visits and the control arm included five visits).²²
Conclusion
Large, multicenter, randomized trials are needed to assess the role of sensor-augmented insulin pump therapy in improving glucose control and reducing the incidence of hypoglycemia. Nevertheless, a solid body of evidence has been reported in the literature on the association between reduced hypoglycemia incidence and CSII in the last decade. The reversibility of hypoglycemic awareness by reducing hypoglycemia incidence is an obvious important clinical point in the discussion of insulin pump therapy. Breaking the vicious cycle of HAFF by reducing hypoglycemia incidence is critical for the literal survival of our patients predisposed to severe hypoglycemia with intensive insulin therapy goals. Therefore, proper management of glucose in our patients with diabetes should factor not only in the importance of improving average glycemia as measured by the HbA_1c, but also should aim at limiting repeated hypoglycemia. One important observation made by Fanelli et al. was that patients who had insulin-requiring diabetes for less than 15 years had a greater improvement in their counter-regulatory recovery than patients with insulin-requiring diabetes for more than 15 years.¹⁶ Such data should encourage us as providers to look more closely at the initiation of therapies such as CSII in our patients earlier in order to help prevent not only the longterm complications of hyperglycemia, but also the long-term complications (i.e. HAAF) of repeated hypoglycemia.■