The Alteration of Aspart Insulin Pharmacodynamics When Mixed

Clinical Care/Education/Nutrition/Psychosocial Research ORIGINAL ARTICLE

TheAlterationofAspartInsulin Pharmacodynamics When Mixed With Detemir Insulin

EDA CENGIZ, MD, FAAP JENNIFER L. SHERR, MD clamp technique to demonstrate that the KARENA L. SWAN, MD MELODY MARTIN, CRA pharmacodynamic (PD) time-action pro- WILLIAM V. TAMBORLANE, MD STUART A. WEINZIMER, MD file of lispro insulin is markedly blunted and delayed when it is mixed with glargine insulin prior to injection (6). This study OBJECTIVEd Mixing rapid acting insulin analogs with detemir insulin to minimize daily was undertaken to examine whether mixing injections has been adopted as a common regimen, especially for some children with type 1 ’ of insulin aspart with insulin detemir diabetes despite the manufacturing company s caution against mixing these analogs in the same causes a similar alteration in aspart’sPD syringe. The effect of this practice on the pharmacodynamics (PD) of rapid-acting insulin has not been widely studied. This crossover, randomized study was undertaken to determine whether action. mixing aspart with detemir insulin has an adverse effect on the early glucodynamic action of RESEARCH DESIGN AND rapid-acting insulin analog in humans. METHODS RESEARCH DESIGN AND METHODSdEight adolescents with type 1 diabetes (17.3 6 0.6 years and A1C of 7.3 6 0.3%) had two euglycemic glucose clamps during which 0.2 units/kg Subjects aspart and 0.4 units/kg detemir insulin were injected either as a separate or single mixed injection Eight subjects with type 1 diabetes (four in random order. female) who attended the Yale Children’s RESULTSdMixing the two insulins diminished the peak and overall early aspart insulin action Type 1 Diabetes Clinic were studied. Sub- jects with a clinical diagnosis of type 1 with significantly lower maximum glucose infusion rate (GIRmax separate 6.1 6 0.7 mg/kg/min vs. mix 4.5 6 0.5 mg/kg/min; P = 0.03) values and the area under curve for GIR during the first 3 diabetes for at least 1 year’s duration, h of the insulin action study (separate 757 6 105 mg/kg vs. mix 491 6 66 mg/kg; P = 0.04). age ranging from 11–21 years, continu- ous subcutaneous insulin infusion ther- CONCLUSIONSdThese data demonstrate that mixing aspart with detemir insulin markedly , fi apy for at least 3 months, A1C 9.0%, lowers the early PD action of aspart and prolongs its time-action pro le as compared with the BMI ,95% for age and sex, and the ability separate injection of these analogs. These changes in insulin PD should be weighed against the to comprehend written and spoken En- added convenience of mixing when considering such unlicensed use of these insulins in youth with type 1 diabetes. glish were eligible for enrollment. Exclu- sion criteria included any other medical disease aside from type 1 diabetes or treated hypothyroidism, use of medications that might affect glycemic control, pregnancy asal-bolus insulin therapy given as sanofi-aventis, available from http://www. or breast-feeding, not consistently using either multiple daily injections or by lantus.com/hcp/closing.aspx; insulin dete- B barrier methods or abstinence as contra- an insulin pump is a mainstay of mir rDNA origin, brand name Levemir ception, or any other condition that in the diabetes treatment for achieving optimal drug insert; Novo Nordisk, Bagsvaerd, judgment of the investigators would inter- glycemic control in type 1 diabetes. The Denmark). Despite these warnings, some fere with the subject’s or parent’s ability to relatively flat and prolonged duration of patients and pediatric practitioners con- provide informed consent or the investi- action of insulin detemir and glargine (1,2) tinue to mix both types of insulin analogs gator’s ability to perform the study. The make them better options for basal insulin in order to improve compliance by reduc- Yale University Human Investigation replacement in multiple daily injection reg- ing the number of daily injections. This Committee approved the study. imens in comparison with NPH insulin. In practice has been supported by clinical tri- Study personnel explained the risks children and adolescents, a negative aspect als that failed to demonstrate an adverse and benefits of the study to the subjects of both long-acting insulin analogs is that impact of mixing insulins on A1C levels and parents, obtained informed consent they increase the number of daily insulin or on continuous glucose monitoring from the parents and informed assent injections due to warnings against mixing profiles (3–5). from the subjects, performed history tak- long- and rapid-acting insulins together In keeping with labeled warnings ing and physical examinations, and mea- (insulin glargine, Lantus; sanofi-aventis; against mixing, we recently used the glucose sured HbA1c during the enrollment visit. Each subject was randomized to the ccccccccccccccccccccccccccccccccccccccccccccccccc mixed versus separate injection during From the Division of Pediatric Endocrinology and Diabetes, Yale School of Medicine, New Haven, the first versus second admission by the Connecticut. flip of a coin. Corresponding author: Eda Cengiz, [email protected]. Received 18 April 2011 and accepted 25 December 2011. Procedures DOI: 10.2337/dc11-0732. Clinical trial reg. no. NCT00652288, clinicaltrials.gov. © 2012 by the American Diabetes Association. Readers may use this article as long as the work is properly Subjects were admitted to the Yale-New cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ Haven Hospital Research Unit on the licenses/by-nc-nd/3.0/ for details. evening prior to the euglycemic clamp to care.diabetesjournals.org DIABETES CARE 1 Diabetes Care Publish Ahead of Print, published online February 28, 2012 Pharmacodynamics of mixing aspart and detemir monitor blood glucose levels. An intrave- Statistical analyses nous catheter was placed to measure blood Data are expressed as means 6 SEM. Rate glucose levels hourly overnight, and insulin of exogenous glucose infusion (GIR) ana- dose was adjusted via insulin pump to lyzedevery10minwasadjustedfor achieve glucose levels between 80 and 120 changes in the glucose space, as previ- mg/dL on the morning of the euglycemic ously described (9). The PD parameters clamp. A second intravenous catheter was that were calculated for each clamp study placed in the contralateral arm in the included area under the curve of the glu- morning for infusion of 20% dextrose. cose infusion rate (AUCGIR), maximum All subjects underwent two euglycemic glucose infusion rate (GIRmax), and clamp procedures on separate days in time to reach maximum glucose infu- random order within 4 weeks of one sion rate (TGIRmax). Statistical compari- another, as previously described (7,8). sons between the two study conditions Subjects received 0.2 units/kg aspart and (mixed versus separate injections) were 0.4 units/kg detemir as a single mixed in- performed using GraphPad Prism ver- jection or two separate injections. Subjects sion 5.0 (GraphPad Software Inc., La Figure 1dMean glucose infusion rate with who received the mixed injection before Jolla, CA). Paired t tests were used to SEM for separate (circles with solid line) and fi the rst clamp were given separate injec- compare GIRmax,GIR300min,TGIR max, mixed (squares with dashed line) injections of tions prior to the second euglycemic AUCGIR 0–300min,andAUCGIR 0–180min aspart and detemir insulins. clamp performed within 4 weeks of the between control and mixed injection first clamp and vice versa. Insulin aspart days. Each subject acted as its own con- is very similar to that we observed in and detemir were mixed in the same sy- trol for the paired t test analysis of plasma previous studies following the same 0.2 ringe (BD insulin syringe with ultra-fine glucose and glucose clamp data. units/kg dose of insulin aspart alone (7,8). needle, 8 mm, 31 gauge; BD Biosciences, Specifically, in the 5 h following a bolus of RESULTSd Franklin Lakes, NJ) at room temperature Eight subjects (four fe- aspart alone, the GIRmax and TGIRmax were immediately before the injection into the male) with type 1 diabetes, age 17.3 6 similar to corresponding values in this deep subcutaneous tissue of the left arm 0.6 years and A1c of 7.3 6 0.3%, were study following the separate injection. Be- through a two-finger pinch of skin at a 45– enrolled and completed both clamp stud- cause previous studies have shown that 908 angle. Subjects were given detemir in- ies. Plasma glucose levels were similar for detemir does not reach 50% maximum sulin in the left and aspart insulin in the the mixed and separate injection studies action until 4–14 h after injection (10), right arm on the day that they were ran- during the 5 h of the clamp (100 6 5vs. it is unlikely that it made a major contri- domized to receive insulins separately. 100 6 5 mg/dL, respectively; P =0.5) bution to the early stimulation of glucose Neither the subject nor the investigator with an intraindividual plasma glucose metabolism following the separate injec- was blinded to mixing versus separate in- concentration coefficient of variation of tion in this study. sulin injection. The infusion of insulin via 5.3 6 0.2% vs. 4.7 6 0.1% (mean 6 In contrast to the results following the insulin pump was suspended just SD) for separate and mixed injections, separate injections, our data demonstrate prior to the administration of aspart and respectively. that mixing insulin aspart with insulin detemir. The overall time-action profiles and detemir diminishes the peak action of Plasma glucose levels were measured PD parameters following the separate and aspart insulin and shifts the time-action every 5 min, and a 20% dextrose infusion mixed injections are shown in Fig. 1 and curve to the right, as evidenced by the was adjusted to clamp plasma glucose con- Table 1. As can be seen in Fig. 1, when lower GIRmax and greater GIR300min in the centrations between 80 and 100 mg/dL aspart and detemir were given as separate mixed compared with the separate injec- during 5 h of the study, as previously injections, there was a rapid and sharp in- tions. Overall insulin action reflected ; described (7,8). Blood for measurement of crease in GIR, which decreased to 25% of by AUCGIR 0–300 min over the whole 5 h of plasma insulin levels was collected every peak values at the end of the study at 300 the study was only slightly reduced with 10 min for the first 90 min, every 15 min min. In contrast, the early rise in GIR was mixing because the decreased AUC during for the next 90 min, and every 30 min for blunted, but late insulin action was in- the first 3 h (AUC GIR 0–180 min) was offset the last 120 min. creased following the mixed compared by the increase in AUC during the follow- with separate injections. As shown in ing hours. The number of subjects in our Biochemical methods Table 1, mixing aspart with detemir signif- study was small but sufficient to demon- A1C was measured by the DCA Vantage icantly reduced the GIRmax but increased strate statistically significant differences Analyzer (Siemens Medical Equipment, the GIR300min as compared with separate between the PD measures of insulin for Malvern, PA) and plasma glucose by the injections. Although there were no signif- mixed and separate injection groups. YSI Glucose Analyzer (YSI Life Sciences, icant differences between TGIRmax and However, failure to detect statistically sig- fi Yellow Springs, OH). The very high plasma AUCGIR 0–300min,theAUCGIR 0–180min ni cant differences, as in AUCGIR 0–300 min, levels of albumin-bound insulin that are was decreased with mixing. must be interpreted with caution in view achieved after detemir injection precluded of the small sample size in this study. measurements of plasma aspart insulin lev- CONCLUSIONSdOur data demon- These changes are qualitatively similar to els due to cross-reactivity with the Mercodia strate that when aspart and detemir in- what we have previously reported with iso-insulin ELISA assay (Mercodia Iso- sulinsaregivenatthesametimeas mixing of insulin lispro and glargine (6). Insulin ELISA Technical Note and insert, separate injections, the time-action pro- In a recent clinical trial, Nguyen and 2007; Mercodia Inc., Uppsala, Sweden). file in the 300 min following the injection colleagues (5) examined the effects of

2 DIABETES CARE care.diabetesjournals.org Cengiz and Associates

Table 1dPD summary measures after subcutaneous injection of aspart insulin of human insulin. Pharm Res 2004;21: and detemir insulin in separate or mixed injections 1498–1504 2. Heinemann L, Linkeschova R, Rave K, Hompesch B, Sedlak M, Heise T. Time- P PD parameters Separate injection Mix injection value action profile of the long-acting insulin analog insulin glargine (HOE901) in GIRmax (mg/kg/min) 6.1 6 0.7 4.5 6 0.5 0.03 6 6 comparison with those of NPH insulin GIR300min (mg/kg/min) 1.6 0.4 2.9 0.4 0.01 6 6 and placebo. Diabetes Care 2000;23: TGIRmax (min) 110 18 130 23 0.3 644–649 6 6 AUCGIR 0–300min (mg/kg) 1137 149 888 112 0.2 3. Fiallo-Scharer R, Horner B, McFann K, 6 6 AUCGIR 0–180min (mg/kg) 757 105 491 66 0.04 Walravens P, Chase HP. Mixing rapid- Data are mean 6 SEMs. P values in boldface refer to the significance of differences between separate and acting insulin analogues with insulin mixed injections. glargine in children with type 1 diabetes mellitus. J Pediatr 2006;148:481–484 mixing aspart and detemir during outpa- due, at least in part, to the mixed dose 4. Hassan K, Rodriguez LM, Johnson SE, tient treatments of 14 pediatric patients of aspart and detemir that was given at TadlockS,HeptullaRA.Arandomized, with type 1 diabetes. No significant differ- dinnertime. controlled trial comparing twice-a-day insulin glargine mixed with rapid-acting ences in overall sensor glucose AUC or insulin analogs versus standard neutral mean amplitude of glycemic excursions d protamine Hagedorn (NPH) therapy in were observed on separate and mixed in- Acknowledgments This study was made newly diagnosed type 1 diabetes. Pediat- jection days during this randomized, possible by Clinical and Translational Science rics 2008;121:e466–e472 crossover, open-label study that used Award 5 KL2 RR024138 from the National 5. Nguyen TM, Renukuntla VS, Heptulla RA. 48-h continuous glucose monitoring as Institutes of Health. Mixing insulin aspart with detemir does W.V.T. is a consultant for Medtronic, Novo not affect glucose excursion in children its primary outcome measure. Although Nordisk, Eli Lilly, Bayer, and Unomedical these results might appear to be inconsis- with type 1 diabetes. Diabetes Care 2010; and a speaker for Novo Nordisk and Eli Lilly. – tent with our findings, they are not in- 33:1750 1752 S.A.W. is a consultant for Medtronic, Animas, 6. Cengiz E, Tamborlane WV, Martin- compatible with our results. We only and BD Medical and a speaker for Eli Lilly. fi fl Fredericksen M, Dziura J, Weinzimer SA. studied a xed aspart to detemir unit No other potential con icts of interest relevant Early pharmacokinetic and pharmacody- doseratioof1:2(equivalenttoamolar to this article were reported. namic effects of mixing lispro with glargine ratio of 1:8), whereas in clinical practice, E.C. researched data, reviewed data, con- insulin: results of glucose clamp studies in the fraction of aspart can be increased to tributed to the discussion, and wrote/revised youth with type 1 diabetes. Diabetes Care compensate for the delayed and diminished the manuscript. K.L.S. researched data. W.V.T. 2010;33:1009–1012 peak action of aspart when it is mixed with reviewed data, contributed to the discussion, 7. Swan KL, Dziura JD, Steil GM, et al. Effect and reviewed/edited the manuscript. J.L.S. of age of infusion site and type of rapid- detemir. Moreover, our study did not assess contributed to the discussion. M.M. researched the impact of other factors such as previ- acting analog on pharmacodynamic pa- data. S.A.W. reviewed data, contributed to the rameters of insulin boluses in youth with ously administered albumin-bound detemir discussion, and reviewed/edited manuscript. that could influence the effect of mixing on type 1 diabetes receiving insulin pump E.C. is the guarantor of this work and, as such, therapy. Diabetes Care 2009;32:240– insulin action in the clinical setting, because had full access to all the data in the study and 244 all of our subjects were receiving insulin takes responsibility for the integrity of the data 8. Swan KL, Weinzimer SA, Dziura JD, et al. pump therapy prior to this study. and the accuracy of the data analysis. Effect of puberty on the pharmacody- Parts of this study were presented in abstract It is noteworthy that after mixing, fi namic and pharmacokinetic properties of the peak postdinner sensor glucose val- form at the 68th Scienti cSessionsofthe insulin pump therapy in youth with type 1 American Diabetes Association, San Francisco, – ues exceeded 200 mg/dL in almost all of – diabetes. Diabetes Care 2008;31:44 46 ’ California, 6 10 June 2008. 9. DeFronzo RA, Tobin JD, Andres R. Glucose Nguyen s subjects, and the nighttime The authors thank Yale Pediatric Diabetes sensor glucose values in the borderline clamp technique: a method for quantifying Clinic and Research Center Team members, insulin secretion and resistance. Am J to frankly hypoglycemic range were ob- patients, and families. – served in 9 of the 14 subjects (5). Given Physiol 1979;237:E214 E223 10. Plank J, Bodenlenz M, Sinner F, et al. A the alterations in the time-action pro- fi double-blind, randomized, dose-response les observed in our study, we speculate References study investigating the pharmacodynamic that postdinner hyperglycemia and in- 1. Havelund S, Plum A, Ribel U, et al. The and pharmacokinetic properties of the long- creased risk of nocturnal hypoglycemia mechanism of protraction of insulin acting insulin analog detemir. Diabetes Care in the study by Nguyen et al. (5) were detemir, a long-acting, acylated analog 2005;28:1107–1112

care.diabetesjournals.org DIABETES CARE 3