Clinical Trial of Programmable Implantable Insulin Pump For Type I Diabetes

  1. Implantable Insulin Pump Trial Study Group
  1. Department of Medicine, University of California Irvine, California; San Raffaele Institute University of Milan Medical School Milan, Italy; Duke University Medical Center Durham, North Carolina; and the Diabetes Unit, Massachusetts General Hospital, Harvard Medical School Boston, Massachusetts
  1. Address Correspondence and Reprint Requests to David M. Nathan, MD, Diabetes Unit, Massachusetts General Hospital, 32 Fruit Street, Boston, MA 02114.


Objective The first step in the evolution of an artificial pancreas is the development of a reliable implantable pump for insulin delivery. Despite recent advances, significant issues remain, including small size of studies and frequent irreversible catheter obstructions. We report safety, feasibility, and efficacy results from 56 patients, representing 73 patient-yr of pump experience, entered into a multicenter trial with a new implantable programmable pump.

Research Design and Methods All patients had insulin-dependent (type 1) diabetes, were 38 ± 8 yr old, and were not prone to severe hypoglycemia. The pump (Infusaid 1000) has a pulsatile mechanism powered by freon-vapor pressure. Its rate is regulated by battery-powered valves, operated via a hand-held programmer. The pump is refilled transcutaneously with 25 ml U100 insulin (Hoechst 21PH) on a monthly basis and has a second septum (side port) proximal to the catheter, which allows flushing the catheter or lavaging the pump unit. The pumps were implanted after 3 mo intensive subcutaneous insulin therapy and catheters were positioned either in the peritoneum (i.p., n = 38) or the superior vena cava (i.v., n = 18).

Results All implanted pumps have functioned safely with no instance of overdelivery or stoppage. The most frequent complications were flow slow downs, presumably due to insulin precipitation within the pump, which occurred in 86% of pumps and were resolved in all but one case by lavaging the pump in situ with alkaline solution. Flow slow downs due to catheter obstruction occurred in 52% of the intravenous catheters but only 21% of the intraperitoneal catheters (P < 0.05) and were resolved in all but two cases by diluent flushing through the sideport. Incidence of severe hypoglycemia decreased from 0.47 before implant to 0.05 episodes/patient-yr after pump implantation (P < 0.001). Mean HbA1c fell from 7.4 ± 1.2% after intensive subcutaneous therapy to 7.1 ± 1.0% 12 mo after implantation. Only 2 patients withdrew from study after recurrent catheter problems, and quality-of-life questionnaires showed improvement in satisfaction with diabetes-specific quality of life when on implantable pump therapy.

Conclusions Insulin therapy with implantable pumps is effective and safe for periods up to 1.7 yr with a decreased risk of severe hypoglycemia than with intensive subcutaneous insulin therapy.

  • Received May 22, 1991.
  • Revision received September 18, 1991.
  • Accepted September 18, 1991.
| Table of Contents