A, Examples of 3 insulin infusion pumps; B and C, 2 subcutaneous catheter infusion sets that may be used for continuous subcutaneous insulin infusion.
A typical profile of basal insulin infusion rates used in continuous subcutaneous insulin infusion. It is very common for people with diabetes to require a higher basal infusion of insulin in the predawn hours. Many people are more active in the late afternoon and more sedentary after dinner, requiring downward and upward adjustments, respectively.
Lenhard MJ, Reeves GD. Continuous Subcutaneous Insulin InfusionA Comprehensive Review of Insulin Pump Therapy. Arch Intern Med. 2001;161(19):2293-2300. doi:10.1001/archinte.161.19.2293
A tremendous amount of data suggest that near-normal glycemic control prevents or delays complications of diabetes, which has led to a dramatic increase in continuous subcutaneous insulin infusion (CSII) or insulin pump use. In this article, the data supporting CSII in type 1 diabetes is reviewed, and the advantages and disadvantages of CSII are analyzed. In addition, CSII use in specific situations is examined, including during childhood and pregnancy and while exercising. The published articles suggest that CSII provides better glycemic control than does conventional therapy and comparable to or slightly better control than multiple daily injections. The use of CSII may be especially indicated during pregnancy or for preconception care and for diabetes presenting in childhood or adolescence.
The Diabetes Control and Complications Trial (DCCT) demonstrated a dramatic reduction in the frequency and severity of complications of diabetes mellitus type 1 in adolescents and young adults by achieving and maintaining glucose control in the near-normal range.1 Comparable but less dramatic results were demonstrated in the UK Prospective Diabetes Study for people with type 2 diabetes.2 These studies established intensive management of diabetes as the standard of care, which has rekindled an interest in external insulin infusion pumps, referred to as continuous subcutaneous insulin infusion (CSII). The patients treated with CSII in the DCCT demonstrated slightly better glycemic control than those treated with multiple daily injections, and CSII therapy was well tolerated.3
Consequently, the popularity of CSII has increased dramatically. Approximately 50 000 to 70 000 adults in the United States are thought to be using CSII, representing about 8% of the adult population with type 1 diabetes.4,5 In addition, the number of insulin pump users is estimated to have doubled over the last 5 years.4 Despite this surge in popularity, a comprehensive review of CSII has not appeared in the literature for many years, although brief reviews have been published in textbooks.6,7 Continuous subcutaneous insulin infusion seems to be the most physiologic method of delivering insulin subcutaneously to achieve near-normal glycemic control. With patients and their families demanding to use CSII, this review is an attempt to describe the best uses, special considerations, and problems associated with CSII therapy.
All articles found in MEDLINE from 1979 to 2000 that contained the words "CSII" or "insulin pump" were read. There were 991 articles listed, although some articles appeared on both lists. Review articles were included and differences among conclusions explored. Both reviewers abstracted data independently.
Insulin pumps were introduced in the late 1970s. There was initial excitement over this new technology,8,9 but within a few years their popularity waned because their size, safety, and efficacy became troublesome issues.10,11 Insulin pumps had a resurgence in popularity in late 1993 after the DCCT results were published. The new pumps are smaller, more efficacious, and easier to use. There are presently 3 manufacturers of insulin pumps in the United States: Disetronic Medical Systems Inc (St Paul, Minn); MiniMed Technologies (Sylmar, Calif); and Animas (Frasier, Pa). The new pumps are small, weighing around 400 g, and they all operate similarly. The insulin pump contains an insulin-filled cartridge or a syringe connected to a catheter that is inserted into the subcutaneous tissue (Figure 1). The pump continuously delivers predetermined basal rates to meet nonprandial insulin requirements. The devices allow programming of many different basal infusion rates, although the average patient requires only 4 to 6 different rates (Figure 2). It also infuses a bolus to cover mealtime or snack time insulin requirements.
Therapy using CSII is not well understood by the public, patients with diabetes, or even some health care providers. There is no surgery involved. The subcutaneous catheter is manually inserted, typically in less than 5 minutes. The pump is not an artificial pancreas. While patients who use CSII may sleep later in the morning because their levels are better controlled for longer duration, appropriate adjustments in the insulin infusion rates must be made. They cannot ignore their calorie- and carbohydrate-restricted diets. The patient needs to self-monitor blood glucose levels as much as if not more than patients who rely on multiple daily injections (MDI). While there is good evidence that CSII will provide better glycemic and metabolic control than MDI, with fewer dangerous glycemic excursions, there is often a misconception that it will completely eliminate episodes of severe hypoglycemia or hyperglycemia.
Intensive diabetes management with CSII provides better glycemic control than does conventional management, which is usually defined as 2 or fewer injections per day and 2 self-monitored blood glucose checks.12,13 It also provides as good and often better glycemic control than intensive diabetes management with MDI.14- 30 In addition to lowering the levels of glycosylated hemoglobin, CSII has been shown to decrease glycemic variability31,32 and lower fasting glucose values.16,29,33- 35
Hypoglycemia is a serious risk associated with intensive therapy and occurs with both CSII and MDI. Early studies suggested that the risk of hypoglycemia with CSII was greater or similar to that of conventional diabetes management 36,37 and MDI.38,39 More recently, however, reports have suggested that severe hypoglycemia may be reduced by CSII as much as 4-fold compared with MDI treatment40,41 with no reduction in glycemic control. This decrease in hypoglycemic events has been accompanied by an increase in self-reported warning symptoms of hypoglycemia, as well as by an increase in counterregulatory hormonal responses to hypoglycemia.42,43 Severe hypoglycemia has now become an accepted indication for initiation of CSII therapy.
Intensive diabetes management with CSII improves glycemic control. The improved control is associated with fewer diabetic and metabolic complications. Treatment with CSII also improves or slows the progression of diabetic nephropathy,44- 46 peripheral and autonomic neuropathy,47- 49 retinopathy,50- 54 hypertriglyceridemia and hypoalphalipoproteinemia,55,56 and diabetic changes in transplanted kidneys.57
The improvement in lifestyle may be the most important reason to the patient who chooses CSII. The ability to increase flexibility in moment-to-moment living is the reason most frequently cited by individuals who have chosen CSII.58 It allows the patient to modify insulin availability hour by hour, which makes possible the performance of activities that would otherwise be risky: skipping or delaying meals, sleeping late on weekends, or engaging in vigorous exercise.59 This increased flexibility may be fueling the upsurge in patient demand for CSII more than any other factor.
As recently as 1990, some authorities asserted that "the use of CSII is discouraged in routine clinical practice," suggesting instead that it be limited to specific subsets of patients with type 1 diabetes.60 While earlier reports of increased and unexplained mortality among CSII users61 have largely been explained,11 there are several risks associated with CSII.
There is no subcutaneous depot of long-acting insulin with CSII. If the flow of the regular, short-acting insulin is interrupted, ketonemia and diabetic ketoacidosis can develop more rapidly and more frequently with CSII than with other treatments.21,36,62 The interruption of insulin may be intentional, to allow patients to participate in certain activities, or unintentional, caused by catheter occlusion, catheter disinsertion, battery failure, depleted insulin supply, and other causes.63- 65 Many times, the interruption of insulin is a result of patient error and inadequate training, particularly when patients do not take the emergency steps necessary in the event of unexplained hyperglycemia.6 In experimental settings where insulin delivery was intentionally interrupted, plasma-free insulin values fell to very low levels (4-8 mU/L) within a 6-hour period, accompanied by large increases in blood glucose, plasma 3-β-hydroxybutyrate, and free fatty acid values.66
While hypoglycemia generally occurs less frequently with CSII than with MDI, concern has been expressed about hypoglycemia resulting from unintentional insulin delivery, or "pump runaway." While this event has occurred,67 it is exceedingly rare and, to our knowledge, has not been found in the United States in the last 10 years. Technological advances in the microprocessor components and insulin delivery alarms of the currently marketed insulin pumps now make the occurrence of such an event extremely unlikely.
The most common complication associated with CSII is infection at the infusion site36,67,68; this is one of the most common causes listed for discontinuation of CSII.69 Most cases of infection have been bacterial, usually Staphylococcus or Streptococcus species,70 although Rhizomucor cellulitis has been found.71 Presumably, almost any pathogen could be a cause. There are conflicting reports as to whether CSII users are chronic carriers of Staphylococcus.70,72 Occasionally, the infection may lead to cellulitis or abscess formation requiring surgical debridement. The annual rate of catheter site infection has been estimated at 7.3 to 11.3 events per 100 years of patient follow-up.3 Additionally, occasional cases of contact dermatitis attributed to the components of the infusion sets and tape have been described.73,74 On very rare occasions, the contact dermatitis persists despite changing the type of tape and/or catheter, and is so severe that CSII must be discontinued.
The most common metabolic adverse effect of improved glycemic control is weight gain, largely attributable to reducing glycosuria. Participants in the DCCT who used intensive management gained about 10 pounds (4.5 kg) more than the conventional treatment group,1 although there was no difference in the weight gained between patients using CSII and those using MDI.3
The insulin pump and the supplies needed to begin therapy average about $5000. The infusion set and catheters must be purchased regularly for as long as CSII is used, at a yearly cost of approximately $1500. Most insurance companies, including Medicare and Medicaid, cover the cost of CSII treatment after medical approval. We are not aware of any detailed studies of the cost-benefit analysis of CSII.
To minimize the risk of ketoacidosis, patients must check their blood glucose levels at least 4 times a day to prevent the development of severe diabetic ketoacidosis. Frequent self-monitoring of blood glucose levels will also allow for early recognition of hypoglycemia. A change of catheter site every 2 to 3 days will minimize the risk for developing skin infections. The application of a local antibiotic ointment to mild skin infections will usually cure them, and creams with aloe, vitamin E, or corticosteroids may be helpful for contact dermatitis. Weight gain does not have to occur with CSII. Exercise and close attention to caloric intake can result in weight maintenance and, if necessary, weight reduction.
Phosphate-buffered insulin demonstrates a decreased incidence of catheter obstruction, and therefore is the preferred insulin for CSII.75 There are currently 2 insulin preparations available that contain a phosphate buffer. Velosulin (Novo Nordisk, Princeton, NJ) is buffered regular insulin, and lispro insulin (Lilly, Indianapolis, Ind) is an insulin analogue modified to provide very fast action. While only Velosulin has a Food and Drug Administration indication for CSII, lispro insulin has some clear advantages: lispro CSII has resulted in less severe and fewer cases of hypoglycemia and better glycemic control than Velosulin CSII.75- 78 However, since lispro is more rapidly absorbed, there is a concern that the interruption of lispro insulin infusion may lead to more rapid metabolic deterioration than an interrupted infusion with buffered regular insulin. Consensus is lacking. One study showed no temporal difference.79 Another study showed that with lispro metabolic deterioration occurred 1.5 to 2 hours earlier and was associated with a larger decrease in blood pH.80 Thus, CSII with lispro insulin seems to provide better glycemic control and fewer and less severe instances of hypoglycemia than CSII with buffered regular insulin, although diabetic ketoacidosis may develop more rapidly.
In contrast to the adult population with type 1 diabetes, there is scarce data on the use of CSII in the adolescent and childhood populations (especially childhood). The literature includes few studies, all very limited in scope; none are randomized, and most involve adolescents. Most of the studies are small, usually 25 or fewer subjects, and short in duration, usually 12 months or less.81- 90 In addition to nonrandomization, a limited number of subjects, and a relatively short duration, the 3 CSII studies in the childhood diabetic population have other major limitations. Two involve only toddlers,91,92 and the others involve only children with newly diagnosed diabetes.93,94
The results of the few studies in the adolescent and childhood diabetes population are contradictory. Most of the studies, especially the more recent ones, demonstrate that insulin pump therapy provides as good or better metabolic and glycemic control than MDI and that it is as well or better tolerated.3,81- 86 Earlier studies, however, demonstrated that it did not provide better control and was not well tolerated.87,88 Recent studies demonstrate that CSII is associated with lower or comparable rates of complications such as hypoglycemia, ketoacidosis, and weight gain than are MDI,85,89,90,95 while earlier studies demonstrated more complications.87,88 Notwithstanding the scarcity of information and some variance in the findings on the safety and efficacy of CSII in the pediatric diabetes population, children and adolescents with diabetes and their parents are demanding to use the pump because they perceive it to be the safest, easiest, and the most physiologic method of delivering insulin subcutaneously to achieve near-normal glycemic control.
The association between excessive maternal hyperglycemia in women with type 1 diabetes and the risk of fetal anomalies is well accepted.96,97 In addition to striving for preconception euglycemia, euglycemic or near euglycemic control should be the goal throughout pregnancy. Hypoglycemia increases during the first trimester of pregnancy. This may be due in part to attempts at improved metabolic control, the passive diffusion of glucose across the placenta, and alterations in the counterregulatory responses of epinephrine, growth hormone, and glucagon.98 Kimmerle et al99 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. With the advantages of CSII in decreasing hypoglycemia and improving glycemic variability, it is logical to assume that CSII would be beneficial for pregnant women with diabetes. As with early studies of CSII efficacy for glycemic control in nonpregnant patients, some of the earlier studies of CSII during pregnancy demonstrated no significant improvement in glycemic control, perinatal morbidity, or length of hospitalization.100 Subsequent studies, however, suggest that CSII is superior to conventional therapy97,101 and at least comparable to MDI102- 104 in achieving metabolic goals in pregnancy. Several aspects of CSII suggest that even if glycemic control is only comparable to MDI, pregnant women may prefer CSII because of its several advantages. These include increased ease of treating morning sickness and hyperemesis gravidarum, reductions in glycemic excursions and hypoglycemia, ease of treating the dawn phenomenon that increases during pregnancy, and improved management in the postpartum period when insulin requirements may fluctuate.105 Three studies have found comparable results for pregnant women with type 2 or gestational diabetes,104- 106 although this area has not been well studied.
The data on CSII therapy in type 2 diabetes is rather scarce compared with data for treating type 1 diabetes, although the enthusiasm for implementing CSII with type 2 diabetes has increased.107 Several short-term trials have demonstrated improvement in glycemic control at 6 months,108,109 improved β-cell function and first-phase insulin secretion,110,111 and improvement in the altered metabolic milieu associated with type 2 diabetes.112,113 There is only 1 randomized trial comparing CSII with conventional insulin therapy.114 In this 4-month trial involving 20 subjects, CSII achieved superior glycemic control, with comparable insulin doses and similar amounts of weight gain. Ultimately, CSII may prove to be of significant benefit to patients with type 2 diabetes, but the paucity of data makes it impossible to draw any conclusions at present. The many publications detailing the successful use of the experimental implantable insulin pump in type 2 diabetes gives hope that these results can be extended to CSII.115
Habitual physical activity has significant benefit to the patient with diabetes, and CSII may make it easier for the patient to maintain glycemic parameters acutely during exercise because of its ability to readily alter the rate of insulin delivery.116 Although CSII has not always led to significant improvement in postexercise hyperglycemia,117 insulin sensitivity and aerobic exercise capacity have been shown to improve with CSII.118,119 Several studies have suggested that CSII may be superior to conventional insulin therapy in improving cardiac function in subjects with cardiac autonomic neuropathy.119,120 And while regular exercise is certainly to be encouraged, adequate training of the CSII user is essential to allow safe changes in the basal infusion rate and to avoid disrupting the catheter site during brisk physical activity.121
The use of CSII may evoke psychological issues to a greater degree than conventional diabetes therapy.122 An emotional and psychosocial assessment is necessary prior to initiating CSII to prevent discontinuation and ensure proper use of the pump.
Several aspects of CSII are unique, and therefore may present barriers to successful glycemic control. Since the pump is visible to others, some patients express difficulty with body image and their self-perceived attractiveness.123 Other people feel inconvenienced explaining CSII to others and consider it an invasion of their privacy.124 Although CSII may help provide exceptional metabolic control, this mechanical dependency may invoke feelings of vulnerability and fear of device failure.122 Patient and family education prior to initiation of CSII helps to minimize this.122 Patient selection is vital to patient tolerance of CSII. It is not always possible to predict whether CSII will be successful.95 While some studies have shown CSII to be associated with significantly less anxiety and depression and improved family social dynamics,125 other studies have suggested that these benefits are most notable in patients who asked to remain on CSII after completion of the study.126,127 In effect, these studies self-selected those patients who were likely to express the greatest satisfaction and least psychological discomfort with CSII. There are few data on the psychological effect of CSII in children and adolescents, with studies showing both an improved sense of control over their life for some patients128 and an increase in anxiety for others.129
Several excellent texts and articles on the strategies and procedures of initiating CSII have been published and provide a detailed "how to" guide for the health care practitioner.7,130- 135 Beginning CSII involves more than merely changing the method of insulin delivery. Patients should participate in a formal educational program. A dietitian should instruct the patient in proper meal planning and carbohydrate counting, whereby the patient calculates the quantity of insulin to take at each meal based on the amount of carbohydrates to be consumed. A diabetes educator should instruct the patient in insulin pump management. And a psychologist and/or social worker should assess motivation, cognitive and problem-solving skills, maturity level, financial stability, and social support systems. It may be difficult to successfully implement CSII without the assistance of ancillary health care providers and a team approach. The successful implementation of CSII may be as dependent on a motivated, flexible, and skilled health care team as it is on the patient.
There are no universally applicable criteria for determining which patients will do well on CSII. Table 1 lists some characteristics that may help determine which patients will be successful. The current model is that the motivated, relatively stable patient is the best candidate for CSII. A different school of thought, however, holds that the patient with poorly controlled diabetes who may, in addition, not possess the highest level of motivation, cognitive skills, or social support may benefit to an even greater degree than the more enfranchised patient. There is currently no data to support this opinion, but relevant research is being conducted.
Continuous subcutaneous insulin infusion has proven to be extremely effective in treating type 1 diabetes. It provides glycemic control superior to that of conventional therapy and comparable or slightly superior to MDI. It also decreases the frequency and/or severity of hypoglycemic reactions and increases lifestyle flexibility. With good compliance and proper attention to details, the disadvantages and risks of CSII can be minimized.
Some specific patient situations may be especially attractive for the use of CSII, including pregnancy or preconception care and diabetes presenting in childhood or adolescence. For the successful use of CSII, a skilled and motivated health care delivery team is required, and a thorough evaluation and training of the CSII candidate is necessary prior to implementation. There is also a need for ongoing close contact between the pump user and the health care team.
Accepted for publication March 29, 2001.
We thank Zoe Meyers of Disetronic Medical Systems Inc (St Paul, Minn), and the librarians at the Christiana Care Medical Library, Wilmington, Del, especially Joan Smith, for assistance with the literature search.
The photographs were expertly taken at the Medical Photography section at Christiana Hospital.
Corresponding author and reprints: M. James Lenhard, MD, Diabetes and Metabolic Diseases Center, Section of Endocrinology, Christiana Care Health Services, 700 Lea Blvd, Suite 300, Wilmington, DE 19802 (e-mail: JLenhard@ChristianaCare.org).