Comparison of an Insulin Analog, Insulin Aspart, and Regular Human Insulin with No Insulin in Gestational Diabetes Mellitus

Pathophysiology/Complications ORIGINAL ARTICLE

Comparison of an Insulin Analog, Insulin Aspart, and Regular Human Insulin With No Insulin in Gestational Diabetes Mellitus

1 2 DAVID J. PETTITT, MD JERZEY W. KOLACZYNSKI, MD, PHD tion measured after eating has been 1 1 PAULINA OSPINA, MHS LOIS JOVANOVIC, MD recommended because of fear of hypoglycemia due to the peak action of human regular insulin. Because the peak postprandial glucose concentration is 1 h after eating (11), a rapid-acting insulin analog OBJECTIVE — To assess the short-term efficacy of insulin aspart in comparison with regular that reaches peak concentration at about human insulin in women with gestational diabetes mellitus (GDM) during standardized meal 1 h after injection would be ideal. Data on tests. nonpregnant diabetic patients document that the rapid-acting insulin analog, insu- RESEARCH DESIGN AND METHODS — The study included 15 women with GDM who had inadequate diabetes control with diet alone. On 3 consecutive days, breakfast meal tests lin aspart, peaks at 40 to 60 min after in- were performed—the first with no exogenous insulin and the other two after the injection of jection and thus can be used to blunt the either regular insulin or insulin aspart. peak postprandial glucose concentrations (15). Before insulin aspart can be pre- RESULTS — The peak insulin concentration was higher and the peak glucose and C-peptide scribed in pregnancy, the safety and effi- concentrations were lower with both insulin preparations than with no exogenous insulin. cacy of this insulin analog will need to be Glucose areas under the curve above baseline were significantly lower with insulin aspart (180- studied. The aim of this article is to doc- –1 ϭ –1 ϭ min area, 7.1 mg h dl ; P 0.018), but not with regular insulin (30.2 mg h dl ; P 0.997), ument the short-term efficacy of insulin –1 than with no insulin (29.4 mg h dl ). aspart in pregnant women during a standardized meal. CONCLUSIONS — This study demonstrates that effective postprandial glycemic control in women with GDM who required insulin was brought about by insulin aspart through higher insulin peak and lower demand on endogenous insulin secretion. RESEARCH DESIGN AND METHODS — Women with GDM Diabetes Care 26:183–186, 2003 who had been unable to control their glucose concentrations with diet and exer- cise, and were therefore candidates for estational diabetes mellitus (GDM), cation clearly associated with GDM is insulin therapy, were recruited to partic- defined as “carbohydrate intoler- macrosomia (5), which occurs at rates as ipate in a randomized trial of either insu- ance of varying degrees of severity high as 40% of neonates in untreated lin aspart or regular human insulin (Table G 1). Before insulinization, three 4-h break- with onset or first recognition during GDM (6). In addition, neonatal macroso- pregnancy” (1,2), occurs in nearly 4% of mia is associated with the metabolic syn- fast meal tests were administered, when- all pregnancies in the U.S., but the actual drome of hyperinsulinemia and ever possible, on 3 consecutive days. The prevalence differs with ethnicity and ma- deposition of fat in the visceral cavity (7). Table 1—Characteristics of study partici- ؍ ternal age (3). Women with high blood The literature suggests that the risk of glucose concentrations experience greater macrosomia rises as maternal glycemia in- pants (n 15) risk of adverse maternal and fetal out- creases (8–10). Specifically, the risk of comes, including preeclampsia, cesarean macrosomia appears to increase with in- Mean Ϯ SE Range delivery, macrosomia, congenital anoma- creasing postprandial glucose concentra- Ϯ lies, and increased risk for future devel- tions (11–13). Traditionally, in the Age at enrollment 31.9 1.9 20.4–46.5 (years) opment of type 2 diabetes (4). The most guidelines for care of nonpregnant dia- Ϯ common and significant neonatal compli- betic patients (14), a glucose concentra- Height (cm) 156.1 1.6 149.0–173.5 Prepregnancy 65.5 Ϯ 3.9 51.8–102.3 ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● weight (kg)* 2 Ϯ From the 1Sansum Medical Research Institute, Santa Barbara, California; and 2Novo Nordisk Pharmaceuti- BMI (kg/m ) 26.8 1.5 21.4–41.2 cals, Inc., Princeton, New Jersey. Fasting plasma 93.5 Ϯ 3.4 76.0–120.0 Address correspondence and reprint requests to David J. Pettitt, MD, Sansum Medical Research Institute, glucose† 2219 Bath St., Santa Barbara, CA 93105. E-mail: [email protected]. Gravity 2.8 Ϯ 0.4 1–7 Received for publication 17 May 2002 and accepted in revised form 26 September 2002. Ϯ Abbreviations: AUC, area under the curve; GDM, gestational diabetes mellitus. Parity 1.7 0.4 0–6 A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion *Data missing for one participant. †At baseline meal factors for many substances. test.

DIABETES CARE, VOLUME 26, NUMBER 1, JANUARY 2003 183 Insulin aspart in GDM

Table 2 —Glucose, insulin, and C-peptide during three meal tests

No exogenous insulin Regular insulin Insulin aspart Time Glucose Insulin C-peptide Glucose Insulin C-peptide Glucose Insulin C-peptide (min) (mg/dl) (␮U/ml) (ng/ml) (mg/dl) (␮U/ml) (ng/ml) (mg/dl) (␮U/ml) (ng/ml) Ϫ30 93.7 Ϯ 3.1 14.4 Ϯ 3.6 2.5 Ϯ 0.4 93.0 Ϯ 2.5 16.7 Ϯ 3.1 2.3 Ϯ 0.2 94.3 Ϯ 2.7 15.2 Ϯ 2.9 2.6 Ϯ 0.4 0 93.5 Ϯ 3.4 14.6 Ϯ 3.3 2.5 Ϯ 0.4 89.6 Ϯ 2.5 18.2 Ϯ 3.8* 2.5 Ϯ 0.4 93.0 Ϯ 2.6 14.6 Ϯ 3.1† 2.6 Ϯ 0.4 15 91.7 Ϯ 2.7 16.0 Ϯ 2.8 2.5 Ϯ 0.3 87.3 Ϯ 3.0‡ 28.1 Ϯ 6.9‡ 2.5 Ϯ 0.4 90.5 Ϯ 2.6 20.0 Ϯ 2.9‡ 2.6 Ϯ 0.3 30 101.7 Ϯ 2.9 37.4 Ϯ 3.9 4.0 Ϯ 0.4 101.8 Ϯ 2.9 51.5 Ϯ 7.4‡ 4.2 Ϯ 0.4 102.2 Ϯ 3.8 56.1 Ϯ 6.5* 4.2 Ϯ 0.4 60 122.8 Ϯ 3.0 72.3 Ϯ 9.7 6.8 Ϯ 0.7 115.8 Ϯ 2.9‡ 84.7 Ϯ 10.8‡ 5.8 Ϯ 0.4 111.6 Ϯ 3.9* 95.9 Ϯ 10.9* 5.5 Ϯ 0.3‡ 90 112.8 Ϯ 2.8 53.9 Ϯ 8.6 6.2 Ϯ 0.6 108.4 Ϯ 2.6‡ 68.2 Ϯ 10.4‡ 5.8 Ϯ 0.5 103.3 Ϯ 4.3‡ 77.4 Ϯ 8.8* 5.4 Ϯ 0.5 120 102.5 Ϯ 2.5 35.6 Ϯ 6.2 5.4 Ϯ 0.7 100.6 Ϯ 2.5 55.2 Ϯ 9.4* 5.4 Ϯ 0.6 93.1 Ϯ 3.5* 51.7 Ϯ 9.8* 4.8 Ϯ 0.7 180 85.1 Ϯ 2.1 20.0 Ϯ 4.1 3.9 Ϯ 0.5 79.9 Ϯ 2.6‡ 26.9 Ϯ 4.5‡ 3.1 Ϯ 0.4 72.5 Ϯ 3.4*† 24.6 Ϯ 4.6 2.7 Ϯ 0.4*† 240 74.7 Ϯ 1.7 11.2 Ϯ 3.0 2.6 Ϯ 0.5 66.7 Ϯ 2.0* 18.8 Ϯ 3.3* 1.8 Ϯ 0.4* 65.5 Ϯ 2.0* 15.4 Ϯ 1.9‡ 1.5 Ϯ 0.2*† Data are means Ϯ SE. *P ϭ 0.01; †P Ͻ 0.05 vs. meal using regular human insulin; ‡P Ͻ 0.05 vs. meal with no exogenous insulin. meal, which was plain yogurt (Alta Dena, lated at 120, 180, and 240 min. Paired t ing each of the three meal tests and was City of Industry, CA), contained 11 g fat, tests were used to compare the insulin significantly lower during the meal in 19 g carbohydrate, and 12 g protein per and glucose areas and the glucose, insu- which no exogenous insulin was admin- cup (227 g) and was topped with one-half lin, and C-peptide concentrations at each istered (72.6 Ϯ 9.7 ␮U/ml [mean Ϯ SE]) sliced banana. This meal comprised time point. than with either regular insulin (84.7 Ϯ ϳ40% carbohydrate, 20% protein, and Fifteen women who met the criteria 10.8 ␮U/ml; t ϭ 2.35, P ϭ 0.034 com- 40% fat and was calculated to provide for inclusion in the study participated in pared with no insulin) or insulin aspart 20% of the woman’s daily caloric require- all three meal tests. During one meal test, (95.9 Ϯ 10.9 ␮U/ml; t ϭ 3.6, P ϭ 0.009). ment. The same volume of yogurt was the intravenous catheter ceased to func- The insulin concentrations are plotted in given for each of the three meal tests. On tion after the 60-min sample was col- Fig. 1. In contrast, serum C-peptide peak day 1, a baseline meal was consumed lected, and subsequent phlebotomy by and overall response were lower after in- without any exogenous insulin. On day 2, venipuncture was unsuccessful. Conse- jection of insulin aspart (5.5 ng/ml) than each woman was randomly assigned to quently, blood was not obtained from one after no insulin (6.8 ng/ml; t ϭ 2.40, P ϭ take either regular human insulin or insu- woman for glucose measurements after 0.031). The peak C-peptide after the in- lin aspart, and on day 3, she took the the 60-min time point during the meal at jection of regular insulin (5.8 ng/ml) was other insulin preparation. Regular insulin which she took regular insulin. With the only slightly lower than after no insulin, was administered 30 min before and in- exception of one skipped sample due to and the difference was not statistically sig- sulin aspart 5 min before the meal. The difficulty with a catheter, the other 14 nificant (t ϭ 1.93, P ϭ 0.074). After the same dose of insulin was given on both women completed all three of the 4-h peak at 60 min, the C-peptide tended to days and was two-ninths of the woman’s meal tests. be lower after the insulin aspart injection calculated daily requirement (0.8 units than after the human regular injection, kg–1 day–1 for gestational age Ͻ26 RESULTS — Table 2 shows the mean and this reached stastical significance at weeks; 0.9 units kg–1 day–1 for gesta- glucose, insulin, and C-peptide concen- 180 and 240 min (Table 2). tional age Ն26 weeks). An indwelling in- trations during each of the three meal Similarly, the peak glucose was at the travenous catheter was placed and blood tests. The peak insulin was at 60 min dur- 60-min time point during all three meal was drawn for plasma glucose, serum insulin, and C-peptide concentrations at –30, 0, 15, 30, 60, 90, 120, 180, and 240 min. No other food was consumed during this time, but symptomatic hypoglycemia was treated with oral glucose tablets. Plasma glucose concentration was determined by the glucose oxidase method. Serum insulin was determined with en- zyme-linked immunosorbent assay (ELISA) (Roche Diagnostics, Indianapo- lis, IN) and serum C-peptide by radioim- munoassay (Diagnostic Products, Los Angeles, CA). The serum insulin and plasma glucose areas under the curve (AUCs) above Figure 1—Mean insulin concentrations during 4-h meal tests with no exogenous insulin, regular the baseline glucose (time 0) were calcu- human insulin, or insulin aspart.

184 DIABETES CARE, VOLUME 26, NUMBER 1, JANUARY 2003 Pettitt and Associates

lower after the insulin aspart injection than after the human regular insulin. This difference, which was statistically significant after 180 min, indicated that after the insulin aspart injection there was a lower demand for endogenous insulin than after human regular insulin. In a review summarizing a Medline database search performed to examine the pharmacology, therapeutics, pharmaco- kinetics, dosing guidelines, adverse effects, and drug interactions of insulin aspart in nonpregnant healthy volunteers Figure 2—Mean glucose concentrations during 4-h meal tests with no exogenous insulin, regular and nonpregnant type 1 and type 2 dia- human insulin, or insulin aspart. The horizontal lines are the glucose concentrations at time 0 for betic patients (16), it was reported that the meal at which no exogenous insulin was administered (upper line) and for the meal at which when insulin aspart is administered im- regular insulin was administered (lower line). mediately before a meal, it is at least as effective as regular human insulin in con- tests and was significantly lower with in- respectively). The AUC for regular insulin trol of postprandial blood glucose con- sulin aspart (111.6 Ϯ 3.9 mg/dl; t ϭ 4.2, was similar to that for no insulin through- centrations. In all published reports of P Ͻ 0.001) and with regular insulin out the meal test. nonpregnant persons, insulin aspart (115.8 Ϯ 2.9 mg/dl; t ϭ 2.4, P ϭ 0.034) Seven women had glucose concentra- achieved higher peak insulin concentra- than without exogenous insulin (122.8 Ϯ tions Ͻ60 mg/dl during one or both of the tions in less time and with a shorter dura- 3.0 mg/dl). Fifteen minutes into the meal, meal tests at which insulin was adminis- tion of action than regular human insulin. the women had a lower glucose concen- tered, but no hypoglycemic episode oc- The authors of the review concluded that, tration after taking regular insulin (87.3 curred before 3 h after the meal. There based on these data, insulin aspart is a mg/dl [95% CI 81.4–93.1]) than they did were no glucose concentrations Ͻ50 mg/ convenient premeal insulin for use by pa- during the baseline meal test (91.7[95% dl, the hypoglycemia cut point for this tients requiring mealtime insulin, and fur- CI 86.4–97.1]; t ϭ 2.5, P ϭ 0.025), and study. thermore, due to favorable pharmaco- the glucose was lower at 60, 90, 120, and kinetics, insulin aspart blunts the post- 240 min. With insulin aspart, the glucose prandial blood glucose concentrations at was lower at all time points from 60 min CONCLUSIONS — The major find- least as well as regular human insulin and onward. ing of this study is that insulin aspart was contributes to improved quality of life. The rise in glucose concentration very effective in reducing the peak post- Our present report of the use of insulin from the nadir at 15 min to the peak at 60 prandial glucose concentration, resulting aspart in women with gestational diabetes min was similar during the meals in in a significantly lower glucose area than is the first clinical trial to study the use of which regular insulin (28.1 Ϯ 2.9 mg/dl) without insulin. Regular human insulin insulin aspart in comparison with regular and no insulin (31.1 Ϯ 2.9 mg/dl) were failed to make a significant impact in low- human insulin in pregnant women. Al- used (t ϭ 0.8, P ϭ 0.432). The rise during ering the postprandial glucose concentra- though there are reports of the use of an- the meal in which aspart was used was tion. Because the insulin assay used in this other short-acting insulin analog, insulin lower (21.1 Ϯ 3.7 mg/dl; t ϭ 3.3, P ϭ study measured total insulin (i.e., both lispro, in women with gestational diabe- 0.006 compared with no insulin). exogenous and endogenous insulin), the tes, those reports do not compare the The glucose concentrations are plot- contribution of the endogenous insulin to rapid analog with no insulin (17). ted in Fig. 2, and the glucose areas are the postprandial glycemic profile can be In this study, insulin aspart was effec- shown in Table 3. The 120-, 180-, and ascertained by comparison of the C- tive in decreasing the postprandial glu- 240-min glucose AUCs were lower for in- peptide response. From the 60-min time cose concentration when administered 5 ϭ sulin aspart (P 0.05, 0.02, and 0.005, point onward, the C-peptide tended to be min before a meal to women with GDM. Because elevated postprandial glucose Table 3—Glucose AUC (mg h dl؊1) at 120, 180, and 240 min during three meal tests concentrations have been associated with an increased risk of macrosomia (11–13), decreased postprandial glucose concen- No exogenous insulin Regular insulin Insulin aspart trations with insulin aspart in women 120 min 29.1 (4.9) 29.8 (4.2) 17.3 (5.4) with insulin-requiring GDM should re- P ϭ 0.960 P ϭ 0.052 sult in an improvement in birth weight. 180 min 29.4 (7.5) 30.2 (6.1) 7.1 (7.9) Follow-up of a sufficient number of P ϭ 0.997 P ϭ 0.018 women requiring insulin during preg- 240 min 15.8 (10.1) 11.0 (8.2) Ϫ16.9 (10.4) nancy will be needed to verify this finding P ϭ 0.888 P ϭ 0.005 and to ensure the safety of this insulin P values are relative to no exogenous insulin. analog in pregnant women with diabetes.

DIABETES CARE, VOLUME 26, NUMBER 1, JANUARY 2003 185 Insulin aspart in GDM

and glycemic control in women with ges- 12. Combs CA, Gunderson E, Kitzmiller JL, Acknowledgments— This study was sup- tational diabetes. Am J Obstet Gynecol 159: Gavin LA, Main EK: Relationship of fetal ported in part through a contract with Novo 1478–1483, 1988 macrosomia to maternal postprandial Nordisk Pharmaceuticals, Inc. 6. Persson B, Hanson U: Neonatal morbidi- glucose control during pregnancy. Diabe- ties in gestational diabetes mellitus. Dia- tes Care 15:1251–1257, 1992 betes Care 21 (Suppl. 2):B79–B84, 1998 13. deVeciana M, Major CA, Morgan MA: References 7. Jovanovic L, Crues J, Durak E, Peterson Postprandial versus preprandial blood 1. Metzger BE, Coustan DR, The Organizing CM: Magnetic resonance imaging in preg- glucose monitoring in women with gesta- Committee: Summary and Recommenda- nancies complicated by diabetes predicts tional diabetes mellitus requiring insulin tions of the Fourth International Work- infant birthweight ratio and neonatal mor- therapy. N Engl J Med 333:1237–1241, shop-Conference on Gestational Diabetes bidity. Am J Perinatol 10:432–437, 1993 1995 Mellitus. Diabetes Care 21 (Suppl. 2): 8. Hod M, Rabinerson D, Peled Y: Gesta- 14. Schade DS, Santiago JV, Skyler JS, Rizza B161–B167, 1998 tional diabetes mellitus: is it a clinical en- RA: Self-monitoring of blood glucose. In 2. Jovanovic L (Ed.): Medical Management of tity? Diabetes Rev 3:603–613, 1995 Intensive Insulin Therapy. Amsterdam, Ex- Pregnancy Complicated by Diabetes. Alex- 9. Ogata ES: Perinatal morbidity in offspring cerpta Medica, 1983, p. 176–193 andria, VA, American Diabetes Associa- of diabetic mothers. Diabetes Rev 3:652– 15. Moses RG, Lucas EM, Knights S: Gesta- tion, 2000 657, 1995 tional diabetes mellitus: at what time 3. Engelgau M, German R, Herman W, Aub- 10. Langer O: Is normoglycemia the correct should the postprandial glucose level be ert R, Smith J: The epidemiology of diabe- threshold to prevent complications in the monitored? Aust NZ J Obstet Gynecol 39: tes and pregnancy in the U.S., 1988. pregnant diabetic patient? Diabetes Rev 457–460, 1999 Diabetes Care 18:1029–1033, 1995 4:2–10, 1995 16. Setter SM, Corbett CF, Campbell RK, 4. McCance DR, Pettitt DJ, Hanson RL, Ja- 11. Jovanovic L, Reed GF, Metzger BE, Mills White JR: Insulin aspart: a new rapid-act- cobsson LTH, Knowler WC, Bennett PH: JL, Knopp RH, Aarons JH: Maternal post- ing insulin analog. Ann Pharmacother 34: Birth weight and non-insulin-dependent prandial glucose levels and infant birth 1423–1431, 2000 diabetes: thrifty genotype, thrifty pheno- weight: the Diabetes in Early Pregnancy 17. Jovanovic L, Ilic S, Pettitt DJ, Hugo K, type, or surviving small baby genotype? Study. The National Institute of Child Gutierrez M, Bowsher RR, Bastyr EJ: Met- BMJ 398:942–945, 1994 Health and Human Development–Diabe- abolic and immunologic effects of insulin 5. Langer O, Mazze R: The relationship betes in Early Pregnancy Study. Am J Obstet lispro in gestational diabetes. Diabetes tween large-for-gestational-age infants Gynecol 164:103–111, 1991 Care 22:1422–1427, 1999

186 DIABETES CARE, VOLUME 26, NUMBER 1, JANUARY 2003