To evaluate the accuracy of elderly patients in their mixing of regular and intermediate insulins, to assess the safety and efficacy of premixed insulins compared with extemporarily mixed insulins, and to determine patients' preferences.


We conducted a crossover multicenter study of 5 mo duration. Premixed insulins and patient-mixed, human biosynthetic (rDNA) insulins were used among 64 insulin-treated patients with NIDDM. After a 4-wk run-in period, eligible patients were randomly assigned to treatment 1 (extemporarily mixed insulins) or treatment 2 (premixed insulins) for 8 wk. After that period, the two treatments were crossed for an additional 8-wk period. A blood glucose profile was recorded monthly and HbA1c was measured at the beginning and at the end of each treatment period. An in vitro skills test was performed to assess the accuracy and reproducibility of the patient preparation of insulin doses, and a questionnaire was used to determine their personal preferences for premixed versus extemporarily mixed insulin.


In our study, the quality of the metabolic control was the same whether patients used self-mixed or premixed insulin. The differences in blood glucose profiles and HbA1c were negligible between type and periods of treatment. The overall number of hypoglycemic episodes increased during the trial in both groups, but the difference between treatments was not significant.

The in vitro skills test, however, indicated that the accuracy in the preparation of insulin doses was significantly higher when patients aspirated from one vial compared with preparation from two vials (P < 0.001). The CVs were 3.7% when drawing up a single dose and 5.0% when preparing a mixture, but the ranges were rather elevated (0.1–20.7 and 0.6–35.8%, respectively). Forty-two patients described the preparation of their daily insulin dose as very easy and 21 described it as easy when using premixed insulins versus 11 and 43, respectively, when using extemporarily mixed insulins (P < 0.001).


While the quality of the metabolic control was the same whether patients used self-mixed or premixed insulin, the in vitro skills test indicated that insulin preparation by elderly patients is highly inaccurate. In some patients, a modification of the contents of the insulin is likely to occur in a few days. The use of premixed insulins should lessen the errors that occur in mixing insulins and from the contamination of the second insulin vial. Draw-up errors could partially account for the lack of improvement of glucose control during the period when patients received premixed insulins. A longer observation period probably is needed to assess appreciable changes in the quality of diabetes control.

This content is only available via PDF.