Heparin Addition to Insulin in Implantable Pumps to Prevent Catheter Obstruction
J. Selam,Peter C. Lord,W. Antwerp,R. Drazin,J. Kandel,M. Charles
DOI: https://doi.org/10.2337/diacare.12.1.38
1989-01-01
Diabetes Care
Abstract:Now that stable insulin solutions are available (1), coagulation, fibrin, and tissue deposition are the major causes of implantable insulin pump failures because they cause catheter malfunction. (2). Heparin may prevent this phenomenon by blocking fibrin formation and/or other coagulation processes. We have reported that heparin may be mixed with insulin (Organon, Saint Denis, France) in pumps (Siemens, FRG) without altering the stability or other basic properties of either compound (3). This study expands on the above data by evaluating the miscibility of heparin with the only special insulin (Hoechst-Roussel, Somerville, NJ) and programmable implantable pump (MiniMed, Sylmar, CA) available and currently under clinical trial in the United States (C. Saudek, J.-L.S., C. Pitt, K. Waxman, M.A. Charles, and R.E. Fischell, unpublished observations). Four solutions were tested: U-400 human buffered neutral insulin stabilized by a surfactant (Genapol, Hoeschst-Roussel, Somerville, NJ), pure sodium heparinate (500 U/ml, Organon), a combination of insulin at 400 U/ml and heparin at 500 U/ml, and a control solution of phenol and water. Each solution was used to fill four implantable pumps (model MIP 2001, MiniMed). Pumps were set at 0.7 U/h and shaken horizontally at 83 Hz, 6.4 x g and placed at 37°C for 3 mo. Eluates were collected at the extremity of the pump catheter (polyethylene covered externally with Silastic; length 25 cm, internal diam 0.3 mm) every week. The turbidity of the solution was assessed visually. High-performance liquid chromatography (HPLC; Waters system, Axxiom columns, and absorbance detection at 220 nm) was used to analyze insulin solutions as previously described (1,3). To improve accuracy, each peak area was adjusted by the corresponding extinction coefficient, determined by ultraviolet spectroscopy at 280 and 220 nm. Zinc, phenol, and pH were measured by standard methods. Heparin activity was assessed by biological (clotting times with the Sigma panel of reagents) and chemical (colorimetry with orthotoluidine blue) methods. No significant changes were noted in clarity or pH. Over the 3-mo period, zinc was reduced by 10% and phenol by 12-15% from both insulin and heparinized insulin. HPLC analysis of insulin, expressed in percentages of total insulin at time 0, gave the following results. Pure insulin solution results at peak 1 (unmodified insulin) were 89.8 and 66.4% at 0 and 3 mo; at peak 2 (denatured insulin), 5.0 and 2.9%; and at peak 3 (other products), 5.2 and 10.2%. Heparinized insulin solution results at 3 mo were 65.3% (peak 1), 13.5% (peak 2), and 10.7% (peak 3). Heparin activity, expressed in percentage of initial heparin activity (500 U/ml = 100% activity) was 43 and 65% for the heparinized insulin solution, assessed biologically and chemically at 3 mo, and 46 and 76% for the pure heparin solution. In summary, a significant drop in pure insulin and heparin content was noted after 3 mo. However, further analyses (data not shown) indicated that most of the alterations were caused by shaking rather than by temperature, storage, or material contact; that the heparin drop is still very compatible with full anticoagulant effect; and, most important, that heparin addition did not further affect any of the parameters tested. We conclude that addition of heparin to insulin in conditions reproducing those of in vivo implanted devices did not alter the chemical stability of either com-