Insulin Detemir Versus Glargine in the Hospital: Do Rates Differ? Michelle A. Crisher,1 Christopher A. Giuliano,1,2 and Carrie L. Hartner1 FEATURE ARTICLE ■ IN BRIEF Several studies have compared the safety and efficacy of insulin detemir and ; however, most have been conducted in the ambulatory care setting. This retrospective cohort study compared hypoglycemia rates between the two basal insulin analogs in hospitalized patients with . No difference was found between the two insulin cohorts in the proportion of patients who experienced hypoglycemic events. n 2015, it was estimated that ~30.3 Detemir is also derived using recom- million Americans were living binant technology and, when injected, Iwith type 1 or (1). forms a soluble depot. In addi- Common long-term complications tion, acylation and self-association of uncontrolled diabetes include properties of detemir allow for revers- end-stage renal disease, diabetic ret- ible binding to albumin, which results inopathy, amputations, and increased in its prolonged duration of action. hospitalizations (2,3). In the inpatient Detemir’s duration of action on blood setting, hyperglycemia, hypoglycemia, varies from 6 to 23 hours and blood glucose variability can lead and is largely dose dependent, with to increased costs, lengths of stay, and longer duration observed as dosage mortality (4,5). Long-acting insulin increases (6,9). Glargine and detemir analogs remain one of the mainstays may be administered once or twice of therapy in the management of daily depending on medication and blood glucose in patients with diabe- patient considerations (8,9). These can tes (5). Two commonly used basal in- include basal insulin dose, prandial sulin analogs are insulin glargine and insulin use, glycemic target, patient insulin detemir. preference, and endogenous insulin Detemir and glargine exhibit reserve (10–12). similar pharmacokinetic and phar- Studies comparing the safety and macodynamic profiles, which make efficacy of detemir and glargine have them appropriate for basal therapy. been done mostly in the ambulatory 1Ascension St. John Hospital, Detroit, MI Both demonstrate a peakless time- care setting and have demonstrated action profile and a long duration of similar glycemic control when added 2Wayne State University Eugene Applebaum College of Pharmacy and Health Sciences, action, although they achieve this to mealtime insulin or oral medica- Detroit, MI through different absorption mecha- tions in patients with type 2 diabetes Corresponding author: Christopher A. nisms (6,7). Glargine is derived using (13,14). In contrast, a subgroup of Giuliano, [email protected] recombinant technology and forms the PREDICTIVE study (15) found https://doi.org/10.2337/cd18-0065 a precipitate that dissolves slowly at improvements in glycemic control and physiologic pH, thus delaying absorp- a reduction of hypoglycemic events ©2018 by the American Diabetes Association. Readers may use this article as long as the work tion. This delayed absorption allows when switching patients with type 1 is properly cited, the use is educational and not for a duration of action that ranges or type 2 diabetes from glargine to for profit, and the work is not altered. See http:// creativecommons.org/licenses/by-nc-nd/3.0 from 10 to 24 hours, reaching up detemir at routinely scheduled clinic for details. to ~30 hours in a few studies (6,8). visits (15).

VOLUME 37, NUMBER 2, SPRING 2019 167 FEATURE ARTICLE

Despite the breadth of comparative lary agent (from glargine to detemir) Outcomes literature surrounding basal insulin occurred during this time. Patients The primary outcome was the pro- therapy in the ambulatory care envi- were identified via active orders for portion of patients who experienced ronment, there is a paucity of evidence either detemir or glargine within the hypoglycemia within 72 hours comparing the safety and efficacy electronic medical record (EMR) sys- of initiating detemir or glargine. of glargine and detemir in hospi- tem. EMRs of adult patients with a Hypoglycemia was defined as a blood talized patients. Galindo et al. (16) diagnosis of diabetes and an order glucose value <70 mg/dL. Secondary performed a retrospective database for detemir or glargine were reviewed outcomes included the proportion analysis of glargine and detemir use in for inclusion. Patients were excluded of patients who experienced hyper- glycemic events, the total number of hospitalized patients who had either if they were pregnant, had received type 1 or type 2 diabetes. No differ- hypoglycemic events, and total in- a continuous insulin infusion, had ence in average blood glucose values sulin requirements within 72 hours received detemir or glargine for <72 was noted between the two insulin of detemir or glargine initiation. cohorts throughout the hospital stay. hours, were admitted to an intensive Hyperglycemia was defined as a blood Similarly, there were no differences in care unit (ICU), or had missing data glucose value >180 mg/dL. needed for analysis (missing both maximum blood glucose or number Data Analysis of patients with severe hyperglycemia. height and weight, having no blood Our sample size was based on an over- Interestingly, there was an increase in glucose measurements, or missing se- all rate of hypoglycemia in the detemir hypoglycemic events in the detemir rum creatinine [SCr] values) within 72 group of 33.5% compared to 29% for cohort. Patients who received detemir hours of initiation of either detemir the glargine group, with an alpha error also received higher total daily doses or glargine. Institutional review board rate of 0.05 and 80% power (16). To (TDDs) of insulin, as well as more approval was obtained before the study find this difference, a sample size of daily injections. In contrast, Zhang commenced. 1,659 patients per group was need- et al. (17) performed a prospective Data Collection ed, for a total of 3,318 patients. Data randomized crossover trial comparing were analyzed using SPSS version glargine and detemir in hospital- Data were electronically abstract- 25.0 software (IBM Corp., Armonk, ized patients with type 2 diabetes. ed from the EMR system using N.Y.). Descriptive data were expressed The authors concluded that the two Access 2016 software (Microsoft as mean ± SD, median ± interquartile demonstrated similar time Corp., Redmond, Wash.) to ex- range, or frequency and percentage. to fasting blood glucose target and tract International Classification of Univariate analysis was performed us- similar TDDs. Hypoglycemic events Diseases, 9th or 10th revision, codes ing a Student t test, a Mann-Whitney occurred in three of the patients and Excel 2016 software (Microsoft U test, or a χ2 test for continuous, switched to detemir. However, this Corp., Redmond, Wash.) to extract ordinal, and categorical data, respec- analysis was limited to 42 patients, all other data. Authors M.A.C. and tively. Multivariate logistic regression thus restricting its power to detect C.A.G. organized the extracted data. was performed using hypoglycemia as safety events of statistical significance. Data collected within 72 hours of de- the dependent variable. Variables were More evidence comparing glargine temir or glargine initiation included initially considered for inclusion into and detemir in the inpatient environ- highest and lowest creatinine clear- the model if there was a difference be- ment is necessary. The purpose of this ance, all blood glucose values, short- tween the detemir and glargine groups study was to compare the efficacy and acting insulin, sliding scale type, insu- and there was an association with hy- safety profiles of glargine and detemir lin TDD, nothing-by-mouth status, poglycemia (P <0.1). by exploring the incidence of both acute kidney injury, chronic kidney Results hypoglycemia and hyperglycemia in disease, corticosteroids, beta-blockers, hospitalized patients with diabetes. Of the 5,200 patients initially total parenteral nutrition, enteral Design and Methods screened, 3,726 were identified for nutrition, and specific infections, in- inclusion. A total of 3,318 patients Sample Selection cluding urinary tract infection, pneu- (1,659 in each cohort) were random- This retrospective cohort study was monia, bacteremia, and diabetic foot ly selected for the final data analysis. conducted at a 772-bed community infection. Data collected outside of 72 The most common reason for exclu- teaching hospital in patients with a hours included baseline demograph- sion was no diagnosis of diabetes. diagnosis of type 1 or type 2 diabetes ics, length of stay, long-acting insu- Figure 1 shows an inclusion and ex- between January 2010 to November lin taken before admission, and most clusion flowchart. Baseline character- 2017. Data from the month of recent A1C. Electronically abstracted istics are summarized in Table 1 and December 2013 were not collected be- data were manually validated by ex- were mostly similar between the two cause a transition of preferred formu- amining 50 patients per cohort. cohorts. The detemir cohort includ-

168 CLINICAL.DIABETESJOURNALS.ORG c r i s h e r e t a l .

potential confounders, multivariate regression analysis was performed (Table 2). Age, baseline SCr, high-in- tensity insulin scale, and concomitant corticosteroid use all met criteria to be included in the final model. When controlling for these factors, there was no difference in hypoglycemia occurrence with detemir or glargine. No difference was seen in the propor- tion of patients with hyperglycemic events in the glargine versus detemir group (92.7 vs. 91.1%, P = 0.086).

The median number of hypoglyce- FEATURE ARTICLE mic events was 0 (range 0–0) for the glargine group versus 0 (range 0–1) in ■ FIGURE 1. Flowchart of inclusion and exclusion criteria. BG, blood glucose; DKA, the detemir group (P = 0.257). No dif- diabetic ketoacidosis; ICD, International Classification of Diseases. ference in insulin TDD was observed between the two cohorts: 34 ± 32 units for glargine versus 35 ± 33 ed more patients with a diagnosis of No difference was found in the units for detemir (P = 0.606). bacteremia and included more pa- proportion of patients with hypo- tients who received the high-intensity glycemic events between the two Discussion insulin scale, beta-blockers, and cohorts (24.2% glargine vs. 25.5% Data comparing the use of glargine and corticosteroids. detemir, P = 0.356). To assess for detemir are mostly derived from the ambulatory care setting, with very few TABLE 1. Baseline Characteristics studies dedicated to hospitalized pa- Glargine Group Detemir Group P tients. In our study, we found no differ- (n = 1,659) (n = 1,659) ence between the detemir and glargine Age, years* 64 ± 14.7 63 ± 14.3 0.019 cohorts in the proportions of patients who experienced hypoglycemic events. Male sex, n (%) 774 (46.7) 779 (47.0) 0.862 Additionally, there were no differences Race, n (%) 0.076 found in the proportion of patients African American 870 (52.4) 852 (51.4) with hyperglycemic events, overall Asian 4 (0.2) 9 (0.5) number of hypoglycemic events, and Caucasian 649 (39.1) 691 (41.7) insulin TDD. These results suggest that either agent is appropriate for the Other 136 (8.2) 107 (6.4) management of diabetes in hospital- BMI, kg/m2* 31.2 ± 11.2 31.7 ± 12.2 0.214 ized patients. However, our results also Baseline SCr, mg/dL* 1.4 ± 1.4 1.5 ± 1.7 0.081 highlight that blood glucose is difficult A1C, %* 8.5 ± 2.3 8.6 ± 2.2 0.326 to manage; ~25% of the patients in this Initial blood glucose, mg/dL* 213.0 ± 93.2 211.1 ± 91.7 0.555 study experienced hypoglycemia, and >90% experienced hyperglycemia. Bacteremia, n (%) 10 (0.6) 51 (3.1) <0.001 Our results differ from those of Insulin scale, n (%) Galindo et al. (16), who found an High intensity 2 92 (17.6) 357 (21.5) 0.004 increase in both hypoglycemic events Medium intensity 747 (45.0) 1101 (66.4) <0.001 and insulin TDD in patients receiv- Low intensity 418 (25.2) 529 (31.9) <0.001 ing detemir, which may be a result of different methods used for data col- ICU 10 (0.6) 19 (1.1) 0.093 lection. Data in the study by Galindo Custom 221 (13.3) 121 (7.3) <0.001 et al. relating to potential confound- Beta-blocker, n (%) 968 (58.3) 112 2 (67.6) <0.001 ers were collected throughout the Steroid, n (%) 263 (15.9) 324 (19.5) 0.006 entire hospital stay. These potential *Mean ± SD. confounders may not have been temporally related to the occurrence

VOLUME 37, NUMBER 2, SPRING 2019 169 FEATURE ARTICLE

the integrity of the data and the accuracy of TABLE 2. Multiple Logistic Regression for Long-Acting Insulin the data analysis. as a Predictor of Hypoglycemia Variable Odds Ratio (95% CI) P References 1. American Diabetes Association. Statistics Glargine versus detemir 0.911 (0.777–1.069) 0.255 about diabetes. Available from www.diabetes. Age, years 1.011 (1.005–1.016) <0.001 org/diabetes-basics/statistics. Accessed 8 June 2018 Baseline SCr, mg/dL 1.152 (1.099–1.207) <0.001 2. Valla V. Therapeutics of diabetes mellitus: Insulin TDD, units 1.005 (1.003–1.008) <0.001 focus on insulin analogues and insulin pumps. Exp Diabetes Res 2010;2010:178372 High-intensity insulin scale 0.715 (0.572–0.893) 0.003 3. World Health Organization. Diabetes fact Concomitant steroid 0.709 (0.565–0.889) 0.003 sheet. Available from www.who.int/en/news- room/fact-sheets/detail/diabetes. Accessed 8 of hypoglycemia or to the initia- mation bias, we performed manual June 2018 tion of long-acting insulin therapy. validation of electronically collected 4. Draznin B, Gilden J, Golden SH, Inzucchi SE. Pathways to quality inpatient manage- Additionally, confounding bias may data. Second, practice changes may ment of hyperglycemia and diabetes: a call have been introduced by a longer have occurred during the timeframes to action. Diabetes Care 2013;36:1807–1814 length of stay, which would have been in which data were collected, although 5. Finfer S, Chittock DR, Su SY-S, et al. difficult to control for. We collected insulin TDDs were similar between Intensive versus conventional glucose con- data within 72 hours of initiation the groups. Third, this study excluded trol in critically ill patients. N Engl J Med of long-acting insulin to attempt to patients admitted to the ICU, and 2009;360:1283–1297 address these issues. Collecting data results cannot be directly applied 6. Poon K, King AB. Glargine and detemir: safety and efficacy profiles of the long-acting for the first 72 hours is similar to to this patient population. Finally, basal insulin analogs. Drug Healthc Patient the strategy employed prospectively we excluded patients who received Saf 2010;2:213–223 in the study by Zhang et al. (17), in long-acting insulin but did not have a 7. Porcellati F, Rossetti P, Busciantella NR, which hypoglycemic outcomes were diagnosis of diabetes. et al. Comparison of and dynamics of the long-acting insulin analogs followed for 3 days after long-acting In conclusion, this study found glargine and detemir at steady state in type 1 insulin initiation. similar rates of hypoglycemia in diabetes: a double-blind, randomized, cross- The results of this study suggest patients receiving insulin glargine over study. Diabetes Care 2007;30:2447–2452 that a therapeutic interchange between and those receiving insulin detemir. 8. Sanofi. Insulin glargine [package insert]. glargine and detemir could be appro- Both insulins appear to be appropriate Bridgewater, N.J., Sanofi-Aventis, 2015 priate. Although these agents are options in hospitalized patients with 9. . Insulin detemir [package similar in price per vial, institutions diabetes. Further research is war- insert]. Plainsboro, N.J., Novo Nordisk, 2015 could select the insulin with the small- ranted to pursue a similar comparison 10. DeVries JH, Nattrass M, Pieber TR. Refining basal insulin therapy: what have est negotiated price. However, there are in critically ill patients, as well as to we learned in the age of analogues? Diabetes a few unanswered questions that could evaluate overall glycemic variability Metab Res Rev 2007;23:441–454 be explored in future studies. First, we using these two long-acting insulin 11. Wallace JP, Wallace JL, McFarland did not compare patients who came analogs. MS. Comparing dosing of basal insulin in on detemir or glargine and were analogues detemir and glargine: is it really unit-per-unit and dose-per-dose? Ann switched to the alternate agent versus Acknowledgments Pharmacother 2014;48:361–368 those who stayed on the same agent. The authors acknowledge the contributions 12. Nelson SE. Detemir as a once-daily basal Second, we did not evaluate glycemic of Susan M. Szpunar, PhD, to the statistical insulin in type 2 diabetes. Clin Pharmacol variability with glargine or detemir. analysis and Michael S. Palmer, RPh, to the 2011;3:27–37 electronic data abstraction. Additionally, we did not perform a 13. Hollander P, Cooper J, Bregnhøj J, regression on hyperglycemia to evalu- Pedersen CB. A 52-week, multinational, Duality of Interest open-label, parallel-group, noninferior- ate the effect of potential confounding No potential conflicts of interest relevant to ity, treat-to-target trial comparing insulin factors because our primary outcome this article were reported. detemir with insulin glargine in a basal- focused on hypoglycemia. We chose bolus regimen with mealtime hypoglycemia as our primary outcome Author Contributions in patients with type 2 diabetes. Clin Ther 2008;30:1976–1987 because of its association with mortal- M.A.C., C.A.G., and C.L.H. developed the project protocol. M.A.C. and C.A.G. 14. Rosenstock J, Davies M, Home PD, ity in hospitalized patients (5). collected and organized data. M.A.C. wrote Larsen J, Koenen C, Schernthaner G. A This study has limitations. First, it the first draft of the manuscript. C.A.G. and randomised, 52-week, treat-to-target trial was observational in design, and data C.L.H. reviewed and edited the manuscript. comparing insulin detemir with insulin were acquired electronically, which All authors reviewed the final version of the glargine when administered as add-on to manuscript. C.A.G. is the guarantor of this glucose-lowering drugs in insulin-naive could have led to information bias. work and, as such, had full access to all the people with type 2 diabetes. Diabetologia To decrease the likelihood of infor- data in the study and takes responsibility for 2008;51:408–416

170 CLINICAL.DIABETESJOURNALS.ORG c r i s h e r e t a l .

15. Yenigun M, Honka M. Switching patients 16. Galindo RJ, Davis GM, Fayfman M, 17. Zhang T, Lin M, Li W, et al. Comparison from insulin glargine-based basal-bolus reg- et al. Comparison of efficacy and safety of of the efficacy and safety of insulin detemir imens to a once daily insulin detemir-based glargine and detemir insulin in the man- and insulin glargine in hospitalized patients basal-bolus regimen: results from a subgroup agement of inpatient hyperglycemia and with type 2 diabetes: a randomized crossover of the PREDICTIVE study. Int J Clin Pract diabetes. Endocr Pract 2017;23:1059–1066 trial. Adv Ther 2016;33:178–185 2009;63:425–432 FEATURE ARTICLE

VOLUME 37, NUMBER 2, SPRING 2019 171