Current issues of ACP Journal Club are published in Annals of Internal Medicine


Premixed insulin resulted in fewer self-reported preparation errors in elderly patients with type II diabetes

ACP J Club. 1993 May-June;118:80. doi:10.7326/ACPJC-1993-118-3-080

Source Citation

Coscelli C, Calabrese G, Fedele D, et al. Use of premixed insulin among the elderly: reduction of errors in patient preparation of mixtures. Diabetes Care. 1992 Nov;15:1628-30.



To compare preferences for and the safety and effectiveness of premixed and self-mixed insulins used by elderly patients.


5-month, randomized, controlled, crossover trial.


Diabetes clinics in Italian university hospitals.


64 insulin-treated patients with non-insulin-dependent diabetes (mean age 67 y, 34 women, mean duration of diabetes 16 y, mean body mass index 27 kg/m2). Follow-up was complete.


All patients maintained their low-fat, high-carbohydrate diet, and levels of physical activity during the study. After a 4-week observation period, patients were randomized to either self-mixed short-plus intermediate-acting insulin from 2 vials or premixed insulin from 1 vial (short:intermediate ratio ranging from 10:90 to 40:60; Humulin, Lilly). After 8 weeks the groups were crossed over for another 8 weeks.

Main outcome measures

A blood glucose profile (monthly) and HbA1c levels. Patients recorded hypoglycemic events, filled out questionnaires to measure their skills and attitudes about insulin preparation, and did a test of skill in insulin preparation.

Main results

The daily insulin requirement remained constant for all patients on both regimens. The regimens did not differ in blood glucose profiles, HbA1c, serum lipid levels, and number of hypoglycemic episodes except for lower after-dinner glucose levels during self-mixed insulin treatment. Fewer self-reported insulin errors occurred while using premixed insulin (χ2 trend, P < 0.02). Patients preferred the premixed insulin: 42 (66%) patients described premixed insulin as "very easy" to use and 21 (32%) as "easy" compared with 11 (17%) and 43 (67%) for the self-mixed insulin. 94% of the patients wanted to use premixed insulin in the future. Results of the insulin preparation skills test indicated that insulin dosing was more accurate with 1 vial (premixed) than with 2 vials (self-mixed) (P < 0.001).


Elderly patients with non-insulin-dependent diabetes preferred premixed insulin over self-mixed insulin. Although the use of premixed insulin did not change blood glucose levels or the reported rate of hypoglycemic events compared with self-mixed insulin, fewer self-reported preparation errors occurred. On a skills test, premixed insulin was associated with more accurate insulin dosing than self-mixed insulin.

Source of funding: Not stated.

For article reprint: Dr. C. Coscelli, I Divisione Medica, Ospedale di Parma, Via Gramsci 14, 43100 Parma, Italy. FAX 39-521-293-296.


The study by Coscelli and colleagues addresses the accuracy of insulin dosing, focusing specifically on the relative benefits of premixed compared with self-mixed insulin in the elderly, nonvisually impaired patient with diabetes. The study consists of 2 separate investigations: 1) a one-time assessment of patients' skills in drawing up their insulin, and 2) a short-term clinical trial of glycemic control.

The clinical trial results should be interpreted with caution, however, because the limited Methods and Results sections make it difficult for the reader to personally assess the study design or examine the raw data. The results of the skills test are consistent with previous findings (1) in that errors occur with both methods, but to a lesser degree with premixed insulin (1 vial) compared with self-mixed insulin (2 vials). Interestingly, the apparent advantage of premixed insulin was not reflected in improved glycemic control when patients in the clinical trial received premixed insulin. This may have been caused by the difficulty of titrating fixed-dose combinations of short- and intermediate-acting insulin, despite the range of combinations now available.

The study by Coscelli and colleagues emphasized the need for both patients and health care providers to consider the potential pitfalls associated with insulin preparation and dosing. Their results support the intuitive hypothesis that premixed insulin is easier to use, therefore resulting in smaller dosing errors than self-mixed insulin in elderly patients with diabetes. Additional study is needed to determine whether the use of premixed insulin is associated with better glycemic control than equivalent amounts of self-mixed insulin.

Paul D. Levinson, MD
Brown University School of MedicineProvidence, Rhode Island, USA


1. Bell DS, Clements RS, Perentesis G, Roddam R, Wagenknecht L. Dosage accuracy of self-mixed vs premixed insulin. Arch Intern Med. 1991;151:2265-9.