Using insulins from three manufacturers, we examined the recovery by radioimmunoassay of short-acting soluble insulin when mixed with long-acting insulin as a function of the ratio of the mixture and the time of premixing. In ratios of 1:2, 1:3, and 1:5 (short- to long-acting insulin), all Novo, Nordisk, and Lilly short-acting insulins tested showed a significant loss of solubility when mixed with the respective company's long-acting insulin either for less than 75 s or for 20 min before centrifugation.
In ratios of 1:1, Novo's Actrapid (regular) with Monotard (lente) and Lilly's regular with lente showed no significant loss of solubility when pre-mixed for less than 75 s, and the regular insulin also showed no significant loss when pre-mixed for 20 min. However, when Lilly's regular was mixed with either NPH or ultralente in a 1:1 ratio, a significant loss of solubility of the short-acting insulin occurred regardless of time [as was also found with Nordisk's Velosulin (regular) with Insulatard (NPH)].
When Lilly regular was incubated with Lilly lente in ratios of 1:3 for <75 s, 20 min, 4 h, and 24 h before centrifugation, there was a progressive loss of solubility. In contrast, with the same ratios and times of premixing, Lilly regular when mixed with Lilly NPH showed a rapid initial loss of solubility that plateaued by 20 min before centrifugation.
Supernatants from the centrifugation of Lilly's long-acting insulins were assayed and found to contain only 0.67–1.5% of the total insulin content. When Lilly regular was mixed in a 1:3 ratio with these supernatants, there was a small but significant loss of solubility.
These in vitro data indicate that short-acting insulins may lose solubility when mixed with long-acting insulins as the proportion of the latter increases and the time Of pre-mixing iS prolonged.