Intermediate-Acting Insulin Formulations
There are two widely used types of intermediate-acting insulin preparations: Neutral Protamine Hagedorn (NPH) and Lente. Both formulations achieve extended time-action by necessitating the dissolution of a precipitated and/or crystalline form of insulin. This dissolution is presumed to be the rate-limiting step in the absorption of intermediate- and long-acting insulin. Consequently, the time-action of the formulation is prolonged by further delaying the dissociation of the hexamer into dimers and monomers.
NPH, named after its inventor H.C. Hagedorn (Hagedorn et al., 1936), is a neutral crystalline suspension that is prepared by the cocrystallization of insulin with protamine. Protamine consists of a closely related group of very basic proteins that are isolated from fish sperm. Protamine is heterogeneous in composition; however, four primary components have been identified and show a high degree of sequence homology (Hoffmann et al., 1990). In general, protamine is ~30 amino acids in length and has an amino acid composition that is primarily composed of arginine, 65–70%. Using crystallization conditions identified by Krayenbu¨hl and Rosenberg (1946), oblong tetragonal NPH insulin crystals with volumes between 1 and 20 mm3 can be consistently prepared from protamine and insulin (Deckert, 1980). These formulations, by design, have very minimal levels of soluble insulin or protamine in solution. The condition at which no measurable protamine or insulin exists in solution after crystallization is referred to as the isophane point.
NPH has an onset of action from 1 to 2 hr, peak activity from 2 to 8 hr, and duration of activity from 14 to 24 hr (Table 2). As with other formulations, the variations in time-action are due to factors such as dose, site of injection, temperature, and the patient’s physical activity.
NPH can be readily mixed with Regular insulin either extemporaneously by the patient or as obtained from the manufacturer in a premixed formulation (Table 2). Premixed insulin, e.g., 70/30 or 50/50 NPH/Regular, has been shown to provide the patient with improved dose accuracy and consequently improved glycemic control (Bell et al., 1991). In these preparations, a portion of the soluble Regular insulin will reversibly adsorb to the surface of the NPH crystals through an electrostatically mediated interac-tion under formulation conditions (Dodd et al., 1995); however, this adsorption is reversible under physio-logical conditions and consequently has no clinical significance (Galloway et al., 1982; Hamaguchi et al., 1990; Davis et al., 1991). Due, in part, to the reversibility of the adsorption process, NPH/ Regular mixtures are uniquely stable and have a two-year shelf life.
The rapid-acting insulin analog, insulin lispro, can be extemporaneously mixed with NPH; however, such mixtures must be injected immediately upon preparation due to the potential for exchange between the soluble and suspension components upon long-term storage. Exchange refers to the release of human insulin from the NPH crystals into the solution phase and concomitant loss of the analog into the crystalline phase. The presence of human insulin in solution could diminish the rapid time-action effect of the analog. One way to overcome the problem of exchange is to prepare mixtures containing the same insulin species in both the suspension and the solution phases, analogous to human insulin Regular/NPH preparations. However, this approach requires an NPH-like preparation of the rapid-acting analog.
An NPH-like suspension of LysB28, ProB29-human insulin has been prepared and its physicochemical properties relative to human insulin NPH have been described (DeFelippis et al., 1998). In order to prepare the appropriate crystalline form of the analog, significant modifications to the NPH crystallization procedure are required. The differences between the crystallization conditions have been proposed to result from the reduced self-association properties of LysB28, ProB29-human insulin.
Pharmacological studies reported for the LysB28, ProB29-human insulin NPH-like suspension, com-monly referred to as neutral protamine lispro (NPL) (Janssen et al., 1997; DeFelippis et al., 1998) indicate that the pharmacokinetic and pharmacodynamic properties of this analog suspension are analogous to human insulin NPH. The availability of NPL suspension allows for the preparation of homoge-neous, biphasic mixture preparations containing intermediate-acting NPL and rapid-acting solutions of LysB28, ProB29-human insulin that are not impacted by exchange between solution and crystalline forms. As with LysB28, ProB29-human insulin, premixed formulations of the AspB28-human insulin have been prepared in which rapid-acting soluble insulin aspart has been combined with a protamine-retarded crystalline preparation of insulin aspart (Balschmidt, 1996). Clinical data on LysB28, ProB29-human insulin mix-tures and those composed of AspB28-insulin have been reported in the literature (Weyer et al., 1997; Heise et al., 1998). The pharmacological properties of the rapid-acting analogs are preserved in these stable mixtures (Table 3). Premixed formulations of both rapid-acting analogs are now commercially available in many countries.
Immunogenicity issues with protamine have been documented in a small percentage of diabeticpatients (Kurtz et al., 1983; Nell and Thomas, 1988). Individuals who show sensitivity to the protamine in NPH formulations are routinely switched to Lente or Ultralente preparations to control their basal glucose levels.
Lente insulin is a zinc insulin suspension that was designed for single daily injection (Hallas-Møller et al., 1952). Lente insulin is a mixture of two insoluble forms of insulin, 70% rhombohedral zinc insulin crystals (Ultralente component) and 30% amorphous insulin particles (Semilente component). The pH of the formulation is neutral and contains acetate and excess zinc. The surplus zinc in the formulation presumably binds to weak-binding metal sites on the insulin hexamer surface, reducing the solubility of the insulin, and thus slowing the time-action of the insulin (Deckert, 1980). The crystal volume of the Ultralente component is routinely between 200 and 1200 mm3 (Deckert, 1980).
Lente has an onset of action of 1–2 hr, peak activity from 3 to 10 hr, and duration of action from 20 to 24 hr (Table 2). As with other formulations, the variations in observed time-action are due to factors such as dose, site of injection, temperature, and the patient’s physical activity.
The mixability of Lente with Regular insulin is restricted to extemporaneous mixtures that are used immediately upon preparation (Deckert, 1980; Galloway et al., 1982). Prolonged storage of Lente/ Regular insulin mixtures leads to a change in the course of effect due to precipitation of the insulin from the Regular section (Deckert, 1980). The precipitation of Regular insulin is presumably due to binding of the surplus zinc found in the Lente formulation to weak-binding sites on the insulin hexamer. Use of Lente insulin among persons with diabetes continues to decline and its continued commercial viability is in doubt.