Feature Article / Mixing Glargine and Rapid-Acting Insulin Mixing Insulin Glargine With Rapid-Acting Insulin: A Review of the Literature Dana G. Carroll, PharmD, BCPS, CDE, and Lisa Meade, PharmD, CDE Abstract Purpose. To review the literature Two of the clinical outcomes examining the mixing of insulin studies did not report significant glargine with rapid-acting insulin differences in A1C levels or prepran- (RAI). dial, postprandial, or nocturnal Methods. A literature search blood glucose levels from mixing was conducted via PubMed and glargine and RAI. One of the clinical Medline (from 1948 to August 2012) outcome studies reported improved using the search terms “diabetes,” blood glucose control (A1C and fast- “insulin glargine,” “short acting ing blood glucose) with RAI mixed insulin,” “rapid acting insulin,” and with glargine compared to RAI “mixing.” Literature was limited to mixed with NPH insulin. There were no significant differences in hypogly- English-language articles reporting cemia in any of the clinical outcome on human studies. Studies with data trials at any time measured. describing mixing glargine with any Conclusion. Initial small clinical short-acting insulin or RAI were trials indicate that there are no sig- included. Four studies met nificant changes in clinical outcomes inclusion criteria. (blood glucose levels, A1C levels, and Results. Of the four studies hypoglycemia) when mixing glargine assessing mixing glargine, one was with RAI. Additional studies with a pharmacokinetic study. The other larger patient populations and longer three assessed clinical outcomes in trial durations are needed before “real-world” settings. All of these mixing glargine with RAI can be studies were conducted in pediatric recommended on a routine basis in patients with type 1 diabetes. clinical practice. Type 1 diabetes accounts for about FDA warns against mixing glargine 5–10% of patients with diabetes; with other insulins because of its pH however, 75% of patients diagnosed and solubility.2 Therefore, basal-bolus with type 1 diabetes are < 18 years of therapy with glargine increases the age.1 The benefits of intense insulin number of daily injections compared therapy to lower A1C and prevent to regimens using intermediate-acting diabetes-related complications are NPH insulin. However, glargine well documented. offers less hypoglycemia than NPH Insulin glargine, the first once- because it is relatively peakless.3 daily basal insulin approved by the Glargine is, however, more expensive U.S. Food and Drug Administration than NPH. (FDA), is used in regimens to treat The treatment of diabetes in chil- patients with type 1 or type 2 dia- dren presents a number of challenges, betes.2 The clear, colorless solution including pain at injection sites, of glargine is buffered to a pH of 4 frequent injections, injection-related Address correspondence to Dana G. and is completely soluble at this level. anxiety (pain or needle phobia), Carroll, PharmD, BCPS, CDE, 850 After injection, a microprecipitate fear of hypoglycemia, and concerns 5th Avenue East, Tuscaloosa, AL forms at the injection site that delays about weight gain. The insulin regi- 35401. and prolongs glargine’s action. The men must be acceptable to patients 112 Diabetes Spectrum Volume 26, Number 2, 2013 Feature Article / Carroll and Meade and their family, while achieving was significantly reduced in the the authors cautioned that titration glycemic control and minimizing mixing group (separate 7.1 ± 1 vs. of the evening dose of glargine may hypoglycemia. In patients 8–21 years mixed 3.9 ± 1 mg/kg/minute, P = be required to prevent hypoglycemia. of age, compliance decreases with 0.04). With regard to pharmaco- This study was conducted under frequency of injections.4 Some clini- kinetics, the peak concentration of real-life conditions, and there was no cians try to minimize the number of lispro was significantly less and the difference in pain or reported reac- injections by mixing glargine with time to peak action was significantly tions with the mixed regimen. The a rapid-acting insulin (RAI) despite delayed in the mixing group (Table study did document cloudiness when FDA warnings against this practice. 1). Mixing the two insulins flattens glargine was mixed with lispro or This review examines the data on and delays the lispro peak. These aspart, which is consistent with data mixing glargine with an RAI. results might lead to poorer control on file with the manufacturer regard- of postmeal glucose levels. ing glargine mixing.2 Limitations of Methodology One limitation of this study is this study include a lack of discussion A literature search was conducted via the use of a fixed 1:2 ratio of lispro about the actual insulin doses and PubMed and Medline (articles pub- to glargine. The insulin lispro dose the short study duration (10 days for lished from 1948 to January 2012) could be increased to overcome the each study group). using the search terms “diabetes,” delay and diminished action peak. Fiallo-Scharer et al.7 conducted “insulin glargine,” “short-acting The delay of action with the lispro a 24-week prospective, case-control insulin,” “rapid-acting insulin,” and can be > 5 hours. Clinically, there study in 110 pediatric patients (70 “mixing.” The search was limited to could be an increased risk of noc- male; mean age 13.4 ± 3.8 years in articles published in English describ- turnal hypoglycemia if lispro and the mixed group and 12.9 ± 4.0 years ing human studies. Studies with data glargine were dosed at supper. in the nonmixed group) with type describing mixing glargine with any The results in this study are con- 1 diabetes. This trial compared the short-acting insulin or RAI were sistent with the results observed in efficacy and safety of administering included. Four studies met these previous animal studies when mixing glargine mixed with either lispro or inclusion criteria. regular insulin with glargine (delayed aspart to that of glargine adminis- Literature Review onset of action and a delayed time to tered separately from an RAI. The 5 Four trials have been published maximum effect). patients mixing insulin (n = 55) were 6 regarding the mixing of insulin Kaplan et al. conducted a 30-day matched with controls (n = 55) based glargine and an RAI.5–8 Cengiz crossover study in 14 pediatric on age, sex, duration of diabetes, and et al.5 conducted a glucose clamp patients (6 male; mean age 13.5 use of identical types of insulin but study in 11 adolescents (six males ± 0.5 years) with type 1 diabetes. who were not mixing before admin- and five females, mean age 15.1 ± This trial compared the efficacy of istration. Patients were allowed to 3 years) with type 1 diabetes. This mixing glargine with either lispro or continue the doses and frequency of trial compared the pharmacody- aspart insulin and administering the insulin administration they had been namic and pharmacokinetic effects doses twice daily compared to once using at the time of enrollment for of administering glargine mixed daily. Continuous glucose monitor- both glargine and their RAI. Data with insulin lispro to administer- ing (CGM) was used to assess each were collected for 6 months before ing glargine separately. Patients had regimen. Patients used glargine at and 6 months after enrollment. The diabetes for at least 1 year and were baseline once daily with three to four primary outcome was change in A1C. on continuous subcutaneous insulin injections of an RAI. Patients were There was no significant differ- infusion for at least 3 months. The randomized to one of two groups ence between the two groups in A1C participants were admitted to the with a crossover design: 1) separate at 6 months (mixing 8.54 ± 1.14 vs. hospital the night before the study injections (glargine before breakfast nonmixing 8.61 ± 1.14%, P = 1.0). and randomized to receive 0.2 units/ and before supper and RAI with each There also were no significant differ- kg of lispro and 0.4 units/kg of meal, and 2) mixed (pre-breakfast ences in the incidence of nonsevere glargine, either mixed or not mixed. and pre-supper glargine mixed with hypoglycemia, severe hypoglycemia, The insulins were mixed at room an RAI and an RAI dose at lunch). or diabetic ketoacidosis at 3 or 6 temperature immediately before the Each patient underwent three studies months (Table 1). However, it should injection. The insulin pump was over a 4- to 6-week period. Patients be noted that four patients in the suspended before the dose was given. continued each regimen for 10 days, mixing group chose to discontinue Two clamp studies were performed, and CGM was performed during the mixing because of an increased fre- with patients receiving the opposite final 3 days of each regimen. quency of hypo- and hyperglycemic treatment regimen during the second There was no significant differ- episodes. The strengths of this trial study, which was within 4 weeks of ence in hypoglycemia, preprandial included a well-matched controlled the first clamp study. glucose, or postprandial glucose lev- group and a 6-month follow-up Both the pharmacodynamic els between baseline, separate, and period. Limitations included the and pharmacokinetic profiles were mixed groups (Table 1). However, study design, lack of blinding, lack altered with the mixing of glargine the mixed group experienced lower of standardized insulin dosing and and lispro. The peak effect of lispro nocturnal blood glucose levels, and frequency, and the fact that 15% of Diabetes Spectrum Volume 26, Number 2, 2013 113