Programmable implantable insulin pumps have proven to be safe and effective for achieving good metabolic control(1) and decreasing the rate of severe hypoglycemia (2). In 1994, a change in the production of insulin resulted in insulin precipitation and recurrent events of underdelivery in the MIP 2001 models (Minimed, Sylmar,CA). A new insulin variant with improved stability in delivery systems was produced in 1997 by Hoechst Marion Roussel and evaluated by the Evaluation dans le Diabete du Traitement par Implants Actifs (EVADIAC) Study Group(3). Yet, persistent underdelivery was still observed and could not be resolved by flushing procedures through the side port catheter. Difficulties in adjusting insulin dosage remained problematic. This underdelivery was explained by the fact that the forward stroke of the piston pump occurred in 2 ms with a volume of 0.5μl, and the catheter lumen was only 0.2 mm in diameter. Thus, its compliance was too low to pass the stroke volume in 2 ms. Therefore, the manufacturer modified the catheter side port with the addition of three small titanium pillows that act as an accumulator, storing the initial impact of the hydraulic force pression of the stroke and adding compliance to the catheter system, resulting in an improved insulin delivery during bench tests.

The EVADIAC Group designed a study protocol to investigate the effect of this new side port catheter on the stability of insulin delivery and to assess the clinical improvement of the catheter compliance. After the approval of an ethical committee, 40 type 1 diabetic patients, currently implanted, were consecutively enrolled for a new implantation with the Minimed implantable pump MMT 2001 and the modified side port catheter MMT 4027 (Minimed). The patients were seen every 45 days to refill the pump reservoirs, to examine the accuracy of the insulin delivery system, to measure the HbA1clevels, and to check for adverse events. The main end point was the accuracy of insulin infusion (% error), calculated from the ratio of the difference between the programmed and the actual infused insulin volume on programmed insulin delivery.

Throughout the study, 40 patients were followed for 450 days, with clinical visits every 45 days (10 refill procedures). The HbA1c levels remained stable at 7.70 ± 0.98% (mean ± SD). The mean percent error was negative at the first 2 refills and then remained nonsignificantly modified between 8.31 ± 7.3% at the 6th refill and 10.4 ± 11.8%at the 10th refill (NS). Thus, during the 450 days of follow-up, the infusion accuracy remained acceptable under the percent error threshold of 15%,previously defined as a level of unacceptable dysfunction. During the same period, the number of insulin units actually infused per dose plateaued at 46± 20 U/kg at the 6th refill and 55 ± 21 U/kg at the 10th refill(NS). Thus, the indexes of good pump delivery throughout the catheter appeared constant and stable over the course of observation.

During this observation period, eight adverse events occurred between the 9th and the 14th month (16 of 100 patient-years). These events included one catheter encapsulation, one catheter obstruction, and six catheter/pump-related underdeliveries, all solved by pump NaOH rinse, catheter flushing procedures, or catheter surgical clearance (for the catheter encapsulation). Several of these adverse events were similar to those observed before 1994 (4).

After 450 days of follow-up in 40 patients, we concluded that modified side port catheters with titanium pillows restore the expected infusion accuracy of implantable pump systems without increasing device complications. The new system appears to be safe and effective. Due to the combined new insulin variant from Hoechst (3) and modified side port catheter by Minimed, the implantable insulin infusion system has returned to its state observed before the change in insulin production (4). Accordingly,our results need long-term confirmation, examined while continuing to refill pump reservoirs at 90-day intervals, as should be allowed by reservoir capacity.

APPENDIX — The members of the EVADIAC Study Group are as follows: Bogdan Catargi, MD, Henri Gin, MD, Bordeaux; Guillaume Charpentier,MD, Jean Pierre Riveline, MD, Corbeil; Jean-Marcel Brun, MD,Agnès Pacaud-Brun, MD, Dijon;Jean-François Martin, MD, Le Mans;François Gilly, MD, Lyon; Pauline Bélicar, MD,Véronique Lassmann-Vague, MD, Philippe Vague,MD, Marseille; Jacques Bringer, MD, Eric Renard, MD, Montpellier; Pierre Drouin, MD, Bruno Guerci, MD, Eric Benamou, MD, Nancy;Marie-Joëlle Haardt, MD, Jean-Louis Selam, MD,Paris; Hervé Grulet, MD, Corinne Leborgne, MD,Reims; Bruno Estour, MD, Luc Millot, MD, Saint Etienne; Sophie Boivin, MD,Nathalie Jeandidier, MD, Michel Pinget, MD, Strasbourg;Hélène Hanaire-Broutin, MD, Vincent Melki, MD, Jean-Pierre Tauber, MD, Toulouse,France.

Pinget M, Jeandidier N: Long-term safety and efficacy of intraperitoneal insulin infusion by means of implantable pumps.
Horm Met Res
Jeandidier N, Selam JL, Renard E, Guerci B, Lassmann-Vague V,Rocher L, Hanaire-Broutin H, EVADIAC Study Group: Decreased severe hypoglycemia frequency during intraperitoneal insulin infusion using programmable implantable pumps (Letter).
Diabetes Care
Boivin S, Belicar P, Melki V, EVADIAC Group: Assessment of in vivo stability of a new insulin preparation of implantable insulin pumps: a randomized multicenter prospective trial (Letter).
Diabetes Care
Hanaire-Broutin H, Broussole C, Jeandidier N, Renard E, Guerci B,Haardt MJ, Lassmann-Vague V, the EVADIAC Study Group: Feasibility of intraperitoneal insulin therapy with programmable implantable pumps in IDDM: a multicenter study.
Diabetes Care