History of Pump Technology

Dr. Arnold Kadish of Los Angeles, California, devised the first insulin pump in the early 1960s. It was worn on the back and was roughly the size of a Marine backpack (Figure 1). Rigorous clinical testing for CSII began in the late 1970s,[3-5] and by the early 1980s, CSII was being considered as a possible alternative form of insulin delivery for patients with type 1 diabetes.[6]

Figure 1. The first insulin pump. The AutoSyringe model, also known as the "Big Blue Brick," was the first commercial pump. [7] Upon its 1978 introduction, excitement spread throughout the medical community, and several companies began to promote the development of insulin pumps. However, many of these pumps lacked the controls necessary to ensure safe insulin delivery. They were not very user-friendly, and some models even required the use of a screwdriver for dosage adjustment.[8] In addition, the idea of wearing a large, heavy pump and being hooked to a machine understandably met with resistance among patients.[7] In the early 1980s, pump therapy was reserved only for the most difficult-tomanage cases, and the results were often unsatisfactory.[8] By the late 1980s, insulin pump therapy was still used in only a minority of patients. The 1990s brought major advances in the field of medical device technology, which allowed dramatic reductions in the size of the pump, brought enhanced safety, and allowed greater ease of use for patients. Currently, most pumps are about the size of a hand-held pager (Figures 2, 3, 4, 5, and 6) and have features such as programmable memory, multiple basal rates, several bolus options, safety lockout features, and remote control. [1]

Figures 2-6. Today's insulin pumps are the size of a hand-held pager. (Figure 6 courtesy Pauline Devney.)

Methods of Insulin Delivery in CSII

CSII is able to mimic the physiologic secretion of insulin by the islet cells more closely than is possible with multiple daily injections (MDI). Current insulin pump models allow a preset rate of insulin delivery to be programmed over a 24-hour period, also known as the "basal rate." The patient can adjust the basal rate to match subtle changes in insulin requirements that take place during exercise, weight change, menstruation, illness, and other circumstances. Precise delivery of insulin can be delivered in increments as small as 0.05-0.25 U an hour. A cross-sectional analysis showed that the average number of basal rates (after being properly adjusted) was between 4 and 5 different settings over a 24-hour period.[9] Another advantage of CSII compared with MDI is the predictable absorption of the soluble, short-acting forms of insulin used in CSII. Absorption of these quick-acting forms of insulin varies by less than 3% daily. [10] By contrast, MDI usually requires 1 or more daily injections of intermediate or long-acting insulin (NPH, lente, ultralente, or glargine) as the basal component. Compared with absorption of the short-acting forms, absorption of these insulins has been shown to vary between 19% and 55% in the same individual and can increase the risk of glucose excursions. [11] Furthermore, intraregional variability in glucose absorption is decreased with CSII, since it employs a single injection site for 2 to 3 days rather than a new injection site 3 to 4 times per day, as with MDI. [12] The pump also delivers insulin on demand; that is, the user can initiate the immediate delivery of insulin. This method of insulin delivery is referred to as a "meal bolus," and it is usually performed before a meal to address anticipated postprandial needs. The amount of insulin delivered is typically calculated from the amount of carbohydrate

consumed. An insulin bolus can also be delivered if the random glucose level is high and there is a need for additional insulin -- "the correction bolus." Insulin pump therapy also can deliver insulin in a more physiologic way than as a single premeal bolus of insulin. Several insulin pump models have 2 additional bolus insulin delivery functions, that is, the square-wave bolus and the dual-wave bolus. The square-wave bolus insulin is evenly infused over a duration determined by the user at the time of administration. The delivery of insulin by the dual-wave bolus is similar to the release of first- and second-phase insulin secretion by the pancreas. The dual-wave bolus combines standard and square-wave bolus delivery. The standard bolus portion of the dual-wave bolus is administered over a period of minutes to address the rapid postprandial rise following carbohydrate intake. The square-wave portion of the dual-wave bolus addresses the postprandial glycemic excursion that occurs following ingestions of protein or fat. [13] Comparisons between standard and dual-wave bolus insulin delivery demonstrate that the standard bolus portion of the dual-wave bolus effectively controls the initial rise in postprandial glucose observed following a high-fat meal, and that the dual-wave bolus provides the best method for managing sustained postprandial glucose excursions after a high-fat meal. Chase and colleagues[14] demonstrated significantly lower postprandial glucose 4 hours following a carbohydrate-rich, high-fat meal and dual-wave bolus combination vs a single bolus combination (P = .04). It has also been demonstrated that dual-wave insulin delivery effectively controls both the initial and sustained postprandial glycemic excursion (over a 14-hour period) observed after a high-fat meal.[13]

Intensive Insulin Therapy

The Diabetes Control and Complications Trial (DCCT)[15] provided unequivocal data that tight glucose control in persons with type 1 diabetes mellitus helps prevent long-term microvascular complications. The landmark 9-year trial of 1441 patients, published in 1993, demonstrated that intensive insulin therapy (insulin pump or at least 3 daily injections and frequent self-monitoring of blood glucose) significantly delays the progression of diabetic retinopathy, nephropathy, and neuropathy compared with conventional therapy (1 or 2 daily insulin injections). The results of the DCCT trial prompted the American Diabetes Association (ADA) to recommend that blood glucose control equivalent to that achieved with intensive insulin therapy (< 7%) should be the goal for most patients with type 1 diabetes.[16] In 2002, the American Association of Clinical Endocrinologists (AACE) recommended the following goals for glucose control: HbA1c level of 6.5% or less; preprandial glucose of 6 mmol/L (110 mg/dL) or less; and postprandial glucose of 8 mmol/L (144 mg/dL) or less.[17]

Current Experience With CSII

The ADA and AACE guidelines for glucose control can be difficult to attain for many diabetes patients on intensive insulin therapy. The incidence of hypoglycemic reactions was 3-fold higher in the DCCT as patients with type 1 diabetes approached an HbA1c of 7.2%,[15] and a major limiting factor to reaching glucose goals for most diabetes patients on insulin is hypoglycemia. Utilization of CSII in several studies has shown substantial reductions in the incidence of both mild and severe hypoglycemia.[8,18,19] Combined outcomes from the pediatric population, pregnant population, and adults reveal improved glucose control with CSII vs MDI. A recent comprehensive meta-analysis of 52 studies supported the improvement of glycemic control with CSII compared with conventional therapy or MDI.[20] A randomized trial of CSII in patients with long-standing poor glycemic control found that CSII improved glycemic control as well as some aspects of health-related quality of life.[21] In this study, 79 patients were randomized to 16 weeks of CSII followed by 16 weeks of MDI, or the reverse order. At 16 weeks, mean HbA1c was 0.84% (95% CI 1.31 to -0.36) lower in the CSII group compared with the MDI group (P = .002). In addition, scores on the "general health" and "mental health" subscales of the Short-Form 36-Item Health Survey improved in the CSII group compared with the injection group (P = .048 and .050, respectively).[21] In the DCCT trial, 124 subjects who used CSII for more

than 90% of the mean 6.5-year study duration had HbA1c levels that were on average 0.2% to 0.4% lower than levels of patients using MDI (P < .001).[15] Perhaps the greatest advantage of CSII from the patient's perspective is that it allows a more normal lifestyle. Pump therapy simplifies irregular meal schedules and allows for flexibility in mealtimes and other aspects of a patient's life. Indeed, this flexibility is the most frequent reason patients give for choosing and remaining on CSII. [22] Patients are able to modify insulin availability by the hour, which makes possible activities that would otherwise be risky, such as skipping or delaying meals, sleeping late on weekends, and engaging in vigorous exercise.[23] Studies demonstrate high patient satisfaction and low discontinuation rates for CSII therapy (< 10%). [24-26] Although some would argue that pump therapy is an expensive alternative to MDI, uncontrolled glucose is the major driving factor in escalating healthcare costs for diabetes.[27] Regression analysis was used to estimate the relationship between glycemic control and medical care charges for 3017 adults with diabetes in a health maintenance organization.[27] Medical care charges increased significantly for every 1% increase above an HbA1c of 7%. A recent analysis conducted in the United Kingdom compared the cost-effectiveness of CSII with that of MDI and found that CSII was most cost-effective in patients who had more than 2 severe hypoglycemic events per year and who required admission to hospital at least once every year.[28]

Precautions for CSII Therapy

Historically, there have been concerns about an increased risk of diabetic ketoacidosis (DKA) with CSII. Because CSII does not make use of long-acting insulin, DKA can potentially develop if the flow of regular short-acting insulin is interrupted. However, the current literature does not support an increased risk of DKA with CSII,[29-35]and the studies that did report significant increases in DKA with CSII date from 1984-1986. A recent meta-analysis of 52 studies showed no reported increase in incidence of DKA in clinical trials since 1993.[20] Bode and colleagues[34] found rates of DKA to be actually less with CSII than MDI, although the difference was not significant (7.2 vs 14.5 events per 100 patient-years). Similarly, Zeigler and associates[35] found an incidence of DKA of 9.7 events per 100 years in patients assigned to CSII and 8.1 events per 100 years in patients assigned to MDI. Patient education about the management of severe hyperglycemia on the insulin pump is important to avoid DKA. One study that examined the incidence of DKA in patients once they started CSII found reduced rates in both adolescents and adults compared with when they were on MDI. [36] In this study, patients were instructed to give themselves a manual subcutaneous insulin injection if blood glucose level was greater than 14 mmol/L (250 mg/dL) on 2 consecutive readings, along with a change in infusion set. A current clinical approach is to have patients give up to 2 correction boluses for glucose levels above 14 mmol/L (250 mg/dL). If glucose remains elevated, patients are then instructed to give a manual injection and change the infusion set. In the past, the so-called "pump runaway" was reported. However, this occurrence was documented only once. [37]This results when the pump delivers an excess amount of insulin as the result of a malfunction. Enhanced safety mechanisms on current models have allowed for no reports of this for over a decade. [23] Infections and inflammation at the infusion site can be prevented by good hygiene and catheter site changes every 2 to 3 days. Although rare, patients should still be counseled about the risk of contact dermatitis.[38,39] Applying a topical antibiotic ointment to mild skin infections will usually address them. Creams with aloe, vitamin E, or corticosteroids may help prevent contact dermatitis in prone individuals.[23]

CSII in Children and Adolescents

Intensive insulin management in the pediatric population can be challenging. Factors related to growth, puberty, and peer pressure make diabetes management in this group especially difficult. Nonetheless, several clinical trials of CSII in young people have shown consistently positive outcomes. [18,40,41] One recent study[40] of 95 children and adolescents, mean age 12 years (range 4-18), with type 1 diabetes showed improvements in glucose control with corresponding reductions in hypoglycemic events (12 vs 17, rate ratio 0.46, 95% CI = 0.21-1.01). HbA1c decreased significantly 3-6 months after pump initiation (P = .03). After 1 year, with adjustment for duration and age, the

difference in mean HbA1c remained significantly lower than before pump initiation (P < .001). There were no significant increases in admissions to the emergency room or DKA after pump initiation. Another study of 74 youths aged 12 to 20 years found that CSII was more effective than MDI in lowering HbA1c levels and reducing the risk of hypoglycemia (P = .01).[24] In addition, adolescents using CSII found coping with diabetes to be less difficult than did adolescents using MDI. [18] Positive outcomes with CSII in adolescents were also reported at the recent 63rd Scientific Sessions of the ADA. In an Italian study[42] enrolling 15 patients with an average age of 12.7 years, HbA1c was reduced from 9.2% to 7.7% at 6 months and 7.9% at 12 months with pump therapy. Clear improvement in quality of life was also reported. Similar benefits of CSII have been observed in toddlers and preschoolers. Compared with older children and adolescents, younger children with type 1 diabetes are more likely to experience DKA, impaired consciousness, and higher blood glucose levels at the time of diagnosis.[43] A study by Litton and colleagues[44] analyzed the effects of pump therapy in 9 toddlers who developed type 1 diabetes between the ages of 10 and 40 months. After a mean of 13.7 months of therapy with MDI, patients were treated with insulin pumps for 7 to 19 months. In these children, HbA1c levels declined from a mean of 9.5% 0.4% at baseline to 7.9% 0.3% (P < .001). In addition, the incidence of severe hypoglycemia decreased from 0.52 episodes per month to 0.09 episodes per month (P < .05). In a study reported at the 63rd Scientific Sessions of the ADA of 65 children with a mean age of 4.5 years at pump initiation (range, 1.4-6.9 years), mean HbA1c decreased significantly from 7.4% to 6.9% at 12 months, and severe hypoglycemia was reduced by 53%.[45]

CSII Use During Pregnancy

Maintaining tight glycemic control to bring glucose into a near-normal range is especially critical during pregnancy. There is an overwhelming body of scientific evidence demonstrating that hyperglycemia during pregnancy is strongly associated with congenital abnormalities, including intrauterine growth delay, macrosomia, neonatal hypoglycemia, respiratory distress syndrome, and neonatal jaundice.[46,47] Furthermore, the risk of hypoglycemia significantly increases during the first trimester. One study reported rates of severe hypoglycemia, defined as coma, seizure, or incapacitation requiring the help of others, to be as high as 41% in pregnant women with diabetes. [48] Several reports suggest that CSII is superior to conventional therapy and at least comparable to MDI during pregnancy.[49-51] A study has demonstrated the benefits of starting CSII in pregnant women with type 1 diabetes mellitus.[51] One subgroup began insulin pump therapy during pregnancy (n = 24), another was treated with MDI (n = 24), and the third was already using an insulin pump before pregnancy (n = 12). When CSII was initiated during pregnancy, glycemic control was well maintained, and maternal and perinatal outcomes and healthcare costs were comparable to those in the other 2 groups. Furthermore, women who began pump therapy in pregnancy were highly likely to continue using the pump after delivery. Thus, CSII appears to represent an attractive choice for women who are or intend to become pregnant.

General Guidelines for Insulin Pump Therapy

The ADA recommends insulin pumps as an effective therapy in type 1 diabetes which should be covered by usual payment mechanisms.[52] The ADA emphasizes that the use of CSII "requires care by skilled professionals, careful selection of patients, meticulous patient monitoring, and thorough patient education." Physicians prescribing CSII should be familiar with CSII and capable of supporting patients. Furthermore, the ADA recommends that candidates for CSII must be "strongly motivated to improve glucose control and willing to work with their healthcare provider in assuming substantial responsibility for their day-to-day care."[52]

Indications for Pump Therapy

All candidates for CSII should meet certain prerequisites. Above all, patients considering CSII must be willing and able to meet the demands of pump therapy, such as the need to change the infusion set regularly and to monitor blood glucose. CSII may be especially useful in patients who are pregnant, experience hypoglycemia unawareness, demonstrate insulin sensitivity (require < 20 U of insulin per day), have severe early morning insulin resistance, or have an unpredictable meal schedule. Bode and colleagues[12] have suggested the following criteria for the selection of pump candidates:

Inadequate glycemic control, defined as HbA1c above target (> 7%). "Dawn phenomenon," with glucose levels greater than 8-9 mmol/L (> 144-162 mg/dL) in the morning. Marked daily variations in glucose levels. History of hypoglycemia unawareness or of hypoglycemic events requiring assistance. Need for flexibility in lifestyle. Pregnancy or intention to become pregnant. Low insulin requirements (< 20 U/day).

Getting Started on Insulin Pump Therapy

Implementing pump therapy requires that the initial insulin dosage, both basal and bolus requirements, be calculated prior to pump initiation. Both the basal rate and the amount of insulin administered as a bolus should be fine-tuned after initiation of the pump based on the response of the patient.

Dose Calculation
Bode and colleagues[12] recommend that the total daily insulin requirements be reduced by 25% to 30%; half of this dose should be administered in the basal infusion, and the other half should be given as a meal or correction bolus. The remaining portion of the insulin calculated for meals should be divided according to the patient's meal content. It is recommended that the patient adjust meal bolus doses on the basis of glucose response to intake by adjusting his or her carbohydrate-to-insulin ratio. Snacks should be avoided during the first 2 to 4 weeks until basal rates have been established.[12] For children, the total insulin requirement per 24 hours might stay the same once therapy is initiated [1] or it might need to be decreased by about 15% to 20%.[19,53] Approximately 40% to 50% of the daily insulin requirement in children is given as a basal rate, but some may need up to 60% given as the basal rate. Compared with adults and older children, children younger than 12 years of age may show a "reverse dawn" effect, in which the basal rate is higher from 9 PM to 3 AM and lower from 3 AM to 6 PM.[1] Use of lispro or aspart is especially suitable for infants and toddlers receiving CSII because of their unpredictable appetites and eating patterns.[1] A partial "priming" dose can be administered before a meal and the remainder can be given during or after the meal, depending on the amount of carbohydrates consumed. [1]

Initial Period Following CSII Initiation

Once CSII has been initiated, basal and bolus rates must be adjusted on the basis of ongoing monitoring of the patient's blood glucose levels before meals, 2 hours after eating, at bedtime, midnight, and 3 AM. [12] Basal rates can be programmed to counteract nocturnal hypoglycemia that can occur between 2 AM and 4 AM and the increase that occurs on waking. The basal rate can be adjusted to maintain overnight glucose level goals of 100 mg/L 30 mg/dL. If glucose levels rise beyond this limit from 3 AM to pre-breakfast, a second basal rate approximately 1.5 times the first basal rate can be added for 4 to 6 hours, starting between 3 and 4 AM. [12] The basal rate should be adjusted during the day when significant glucose excursions are occurring due to delayed or skipped meals. To guide the daytime basal rate adjustment, patients should be instructed to monitor blood glucose in a daytime fasting state every 2 hours.[12] In contrast to the basal dose, bolus insulin doses should be adjusted according to the 2-hour postmeal blood glucose levels. The patient can adjust his or her insulin boluses or carbohydrate-to-insulin ratio every 2 days to keep glucose levels within reasonable limits (< 10 mmol/L [180 mg/dL] 2 hours postmeal). After the bolus dose has been fine-tuned to work with a stable meal plan, the carbohydrate-to-insulin ratio can be estimated so that the patient can modify the bolus dose according to the carbohydrate content of each meal. Providing a ratio allows the patient to estimate adequate levels of insulin delivery rather than reacting to a prior mismatch of insulin delivery and carbohydrate intake and helps stabilize glucose levels. The optimal ratio ranges from 1 U/5 g carbohydrate to 1 U/25 g carbohydrate. [12] Control of blood glucose levels in the pregnant patient warrants special attention. Pregnancy is a time of accelerated growth, during which insulin demands and the production of insulin-opposing hormones are increased. These conditions increase the risk of ketoacidosis, which can result in fetal mortality. In one practice, the rate of ketoacidosis was reduced to less than 0.5% and fetal mortality to 0% by the administration of lente or NPH insulin (0.2 U/kg) at bedtime given in addition to the normal basal amount delivered by insulin pump. [54]

Troubleshooting
After the initiation of CSII therapy, the patient should continue to monitor blood glucose at least 4 times a day, and, once stabilized, should follow up with a physician on a quarterly basis.[12] At each visit, the physician reviews with the patient ways to prevent and troubleshoot occurrences of hyperglycemia and hypoglycemia. If HbA1c is above target, the patient's methods of glucose self-monitoring, recording frequency, diet, knowledge of food intake, and bolus adjustment practices should be analyzed.[12] Questions that can be asked when CSII results in nonoptimal blood glucose control include the following[1]:

Is testing of blood glucose levels being performed 4-6 times a day? Are the basal insulin rates set properly (does hypoglycemia or hyperglycemia occur when meals are skipped)? Are carbohydrate counting or exchange lists being performed accurately and with the proper insulin boluses? Is the proper correction bolus factor to treat elevated blood glucose levels occurring? Is the correct treatment of low blood glucose levels with the appropriate amount of carbohydrates occurring? Is bolus memory on the pump reflected correctly in the logbook records?

Exercise and the Pump

Clear benefits can be achieved from regular physical activity in patients with type 1 diabetes. However, exercise can result in both initial hyperglycemia during exercise and significant hypoglycemia post exercise. The ability of CSII to incrementally alter the amount of insulin delivered during and after exercise can help

maintain glycemic control and avoid dangerous hypoglycemia. [55] Adequate patient training is essential to allow safe changes in the basal infusion rate and to avoid inadvertent disruptions of the catheter site during brisk physical activity.[56]Healthcare providers should counsel patients about how to maintain glycemic control in these situations.[57] A clinical approach to managing appropriate glucose control on the insulin pump is to have a patient check the glucose level at the time of activity, 45 minutes into activity (when hyperglycemia can result), and 2 hours after the completion of exercise (when risk of hypoglycemia increases). Basal rates can then be adjusted on the basis of these glucose levels. An insulin pump can be worn during most sports activities with certain precautions, depending on the activity involved.[57] For patients engaged in contact sports, providers should discuss the possibility of damage to the pump or infusion set. Those who regularly participate in contact sports may need to plan for pump changes or temporary interruptions in insulin infusion during their sport.

Conclusion
Only a decade after the DCCT confirmed the benefits of intensive insulin therapy in achieving tight glucose control, the use of CSII continues to increase. One estimate suggests that more than 40% of people with type 1 diabetes will eventually use CSII.[1] CSII therapy has proven effective in treating many patients with type 1 diabetes, decreases the frequency and severity of hypoglycemic events, and enhances freedom and flexibility in lifestyle. Not surprisingly, more diabetes specialists with type 1 diabetes prefer the use of insulin pump therapy for themselves.[58] Participation and motivation on the part of patients, namely in monitoring glucose and proper use of pump equipment, can help greatly minimize any disadvantages and risks. At one time, the notion of an implanted closed loop system with a glucose sensor and insulin pump seemed only a lofty idea. The "artificial pancreas" is fast becoming a reality. Current investigation offers encouraging data on a closed loop system, which could dramatically improve the glycemic control in individuals with diabetes.