Current issues of ACP Journal Club are published in Annals of Internal Medicine


Implantable insulin pump improved quality of life in NIDDM

ACP J Club. 1997 Mar-Apr;126:40. doi:10.7326/ACPJC-1997-126-2-040

Source Citation

Saudek CD, Duckworth WC, Giobbie-Hurder A, et al., for the Department of Veterans Affairs Implantable Insulin Pump Study Group. Implantable insulin pump vs multiple-dose insulin for non-insulin-dependent diabetes mellitus. A randomized clinical trial. JAMA. 1996 Oct 23/30;276:1322-7.



To assess whether an implantable insulin pump (IIP) and multiple daily insulin (MDI) injections could achieve equal glycemic control and whether the IIP offers any specific advantages over MDI in patients with non-insulin-dependent diabetes mellitus (NIDDM).


1-year randomized controlled trial.


7 Veterans Affairs medical centers (VAMCs) in the United States.


121 men (mean age 56 y) who were receiving care from a VAMC. Inclusion criteria were ages 40 to 69 years; NIDDM for 2 to 15 years; inadequate glycemic control (hemoglobin A1c [HbA1c] > 8%) despite insulin therapy; and no history of hypoglycemia, unawareness, life-threatening conditions, or use of medications that would interfere with the study drugs. Follow-up was > 85% in both study groups.


Patients were stratified by center and allocated to an IIP (n = 59) or MDI (n = 62). The pump (MiniMed Technologies, Sylmar, CA) was surgically implanted in the abdomen and delivered insulin into the abdominal cavity; patients used an external telemeter to control infusion rates and bolus doses. Patients allocated to MDI received self-administered subcutaneous insulin (mean number of injections/d was 2.3). Doses were adjusted according to clinical judgment, self-monitored glucose levels, meals, and activity levels.

Main outcome measures

Changes in fasting plasma glucose and HbA1c levels. Secondary outcomes were glycemic variation, hypoglycemic events, body weight, and quality of life.

Main results

The mean fasting plasma glucose level decreased from 9.1 to 8.2 mmol/L (P < 0.001) with the IIP and from 9.8 to 9.1 mmol/L (P < 0.001) with MDI, with no difference between groups (P = 0.57). Levels of HbA1c decreased for the first 3 months from 8.8% to 7.3% with the IIP (P < 0.001) and from 8.6% to 7.5% with MDI (P < 0.001), with no difference between groups (P = 0.15). Fluctuations in glucose were smaller with the IIP than with MDI (P < 0.001), and fewer occurrences of mild hypoglycemia were seen with the IIP (P ≤ 0.007). Patients in the MDI group gained a mean of 0.36 kg, and patients in the IIP group lost a mean of 1.0 kg (P = 0.01). Patients in the IIP group had higher scores on the total quality-of-life measure than did patients in the MDI group (P = 0.03).


Insulin therapy with an implantable insulin pump and multiple daily insulin injections achieved adequate glycemic control in patients with non-insulin-dependent diabetes mellitus. The implantable insulin pump was superior in reducing glycemic variability, hypoglycemic events, and weight gain and in improving quality of life.

Source of funding: Department of Veterans Affairs Medical Research Service.

For article reprint: Dr. C.D. Saudek, Osler Building, Room 576, Johns Hopkins Hospital, Baltimore, MD 21205, USA. FAX 410-955-0801.


The importance of the study by Saudek and colleagues lies less in its actual results than in the fact that it takes NIDDM seriously. This is a sign of a new era in which the medical community acknowledges the devastating complications of NIDDM and the need to find out how to prevent them.

This study shows that tight glycemic control in NIDDM can be achieved and sustained for a year with an IIP or MDI injections. The IIP was associated with less variation of glucose levels and thus with less hypoglycemia. Further, patients who received an IIP lost rather than gained weight and were more satisfied with their treatment than were patients who received MDI. The insulin reservoir for the pump had to be refilled only every 4 to 12 weeks.

The costs of the IIP were not completely delineated. Surgery was required, usually under general anesthesia, and procedures had to be repeated within the same year for 20% of patients because of catheter failure. Also, underdelivery occurred in 24 patients (41%), leading to 1.4% of treatment days off IIP therapy. The cost of the pumps was high: U.S. $10 000 each.

The acceptability of insulin therapy to most patients with NIDDM is unclear. In the trial, 8 of 59 patients who initially agreed to participate refused to have surgery after being allocated to an IIP. 8 of 62 patients allocated to MDI withdrew before the end of the 12-month trial.

An important lesson from the trial is the usefulness of pumps in decreasing glucose variability; this phenomenon has also been seen with external subcutaneous pumps and is often used as the rationale for switching patients with IDDM to pump therapy.

With the recent and future introduction of new classes of oral agents for NIDDM, multidrug therapy using combinations of classes (sulfonylureas, biguanides, α-glucosidase inhibitors, and thiazolidinediones) is likely to become increasingly popular and, it is hoped, effective in achieving glycemic control. This would leave the IIP or MDI for a select subgroup of patients in whom oral agents have failed.

Jacqueline A. Pugh, MD
Audie L. Murphy Memorial Veterans Administration HospitalSan Antonio, Texas, USA

Jacqueline A. Pugh, MD
Audie L. Murphy Memorial Veterans Administration Hospital
San Antonio, Texas, USA