Emerging Treatments and Technologies ORIGINAL ARTICLE Pramlintide Reduces Postprandial Glucose Excursions When Added to Regular Insulin or Insulin Lispro in Subjects With Type 1 Diabetes A dose-timing study 1 3 CHRISTIAN WEYER, MD MARIA GUTIERREZ, MD espite important advances in the 1 1 ALAN GOTTLIEB, PA YAN WANG, PHD delivery and pharmacology of insu- 1 1 DENNIS D. KIM, MD JAMES A. RUGGLES, PHD 1 1 lin (1,2), most patients with type 1 AREN UTZ PHD RVILLE OLTERMAN MD D K L , O G. K , 2 1 diabetes are still unable to achieve glyce- SHERWYN SCHWARTZ, MD DAVID G. MAGGS, MD mic targets with insulin therapy alone (3,4). The limitations of current insulin replacement therapy are especially evi- dent during the postprandial period, when rapid and profound changes in glu- OBJECTIVE — To assess the postprandial glucose-lowering effect of the human amylin an- alog pramlintide when given with either regular insulin or insulin lispro in subjects with type 1 cose flux are required to accommodate diabetes, with an emphasis on the optimal dose timing relative to meals. the sudden appearance of meal-derived glucose into the circulation (5–7). RESEARCH DESIGN AND METHODS — In this randomized, single-blind, placebo- In healthy individuals, meal ingestion controlled, five-way crossover study, 19 subjects with type 1 diabetes using regular insulin and leads to the rapid release of two gluco- 21 subjects with type 1 diabetes using insulin lispro underwent five consecutive mixed meal regulatory ␤-cell hormones, insulin and tests. In randomized order, subjects received subcutaneous injections of placebo at Ϫ15 min or amylin (8). These hormones act in con- 60 ␮g pramlintide at Ϫ15, 0, ϩ15, or ϩ30 min relative to the meal after an overnight fast. cert to limit the postprandial glucose ex- Ϫ Regular insulin or insulin lispro was injected at 30 and 0 min, respectively, at doses that were cursion, with amylin reducing the rate of adjusted appropriately for both the content of the standardized meal and the anticipated effects glucose appearance into the circulation of pramlintide. Plasma glucose concentrations were measured before and during the 4-h post- meal period. and insulin stimulating the rate of glucose disappearance from the circulation (5–7). RESULTS — In both the regular insulin and insulin lispro groups, pramlintide injections at all In addition, the secretion of glucagon, the four time points lowered the postprandial glucose excursion (36 to Ͼ100% reduction in incre- main hormonal stimulus for hepatic glu- mental area under the concentration time curve from 0 to 4 h (AUC0–4 h) compared with cose production, is normally suppressed placebo. However, only preprandial injections of pramlintide (Ϫ15 and 0 min) were able to in response to a carbohydrate meal (9). prevent the initial postprandial surge in glucose. The optimal time for pramlintide injection was In people with type 1 diabetes, post- 0 min, which reduced the postprandial glucose excursion by Ͼ100% compared with regular prandial insulin and amylin responses are Ϫ Ϯ Ϯ ⅐ Ϫ1 ⅐ Ϫ1 Ͻ insulin plus placebo (incremental AUC0–4 h: 0.6 2.5 vs. 11.0 2.9 mmol h l , P Ϯ completely absent (5–8), and glucagon se- 0.0007) and by 75% compared with insulin lispro plus placebo (incremental AUC0–4 h: 2.5 Ϫ Ϫ cretion is abnormally increased, leading to 2.1 vs. 10.0 Ϯ 2.5 mmol ⅐ h 1 ⅐ l 1, P Ͻ 0.0098). No serious adverse events were reported. excessive postprandial glucose excursions CONCLUSIONS — Pramlintide, given at or just before a meal, reduces the postprandial (9–11). Mealtime insulin replacement re- glucose excursion in subjects with type 1 diabetes, regardless of whether added to regular insulin duces postprandial hyperglycemia (12). or a rapid-acting insulin analog. However, even with the use of rapid-acting insulin analogs, it is still not possible to rep- Diabetes Care 26:3074–3079, 2003 licate the rapid release of endogenous insu- lin into the portal vein or to correct the ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● abnormal postprandial rise in glucagon of- From the 1Amylin Pharmaceuticals, Inc., San Diego, California; the 2Diabetes and Glandular Disease Re- ten seen in patients with type 1 diabetes search Associates, San Antonio, Texas; and the 3Clinical Studies, Ltd., Fort Lauderdale, Florida. (6,7). Both of these abnormalities contrib- Address correspondence and reprint requests to Christian Weyer, MD, Amylin Pharmaceuticals, Inc., ute to the lack of appropriate suppression of 9360 Towne Centre Dr., San Diego, CA 92121. E-mail: [email protected]. Received for publication 2 April 2003 and accepted in revised form 24 July 2003. hepatic glucose production and, hence, to C.W., A.G., D.D.K., K.L., S.S., Y.W., J.A.R., O.G.K., and D.G.M. are employed by and hold stock in excessive postprandial glucose excursions Amylin Pharmaceuticals, Inc. S.S. is a member of an advisory panel for Amylin Pharmaceuticals, Inc. J.A.R. (9–11). also holds stock in Bristol-Myers Squibb and Schering Plough. Pramlintide is a synthetic analog of Abbreviations: AUC, area under the curve. A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion human amylin that is under development factors for many substances. as an adjunct to insulin therapy in people © 2003 by the American Diabetes Association. with types 1 and 2 diabetes (5–7,13). 3074 DIABETES CARE, VOLUME 26, NUMBER 11, NOVEMBER 2003 Weyer and Associates Short-term clinical studies in patients and all subjects provided written in- reduction to avoid postprandial hypogly- with type 1 diabetes have shown that formed consent before participation. This cemia if the premeal glucose was near- mealtime amylin replacement via pre- study was conducted in accordance with normal) but not for the purpose of prandial injections of pramlintide, in ad- principles described in the Declaration of glycemic control (dose increase to im- dition to regular insulin injections, Helsinki (1964), including all amend- prove postprandial hyperglycemia). The corrects postprandial hyperglucagonemia ments up to and including the South Af- basal insulin regimens, which were com- (14) and slows the rate of gastric empty- rica revision (1996). prised of continuous subcutaneous insu- ing (15,16). As a result, the appearance of lin infusion (n ϭ 8) or subcutaneous both endogenous (liver-derived) and ex- Study design injections of intermediate- or long-acting ogenous (meal-derived) glucose into the Within 14 days after screening, consenting insulin (n ϭ 32), were held constant circulation is controlled to better match subjects were admitted to the clinical re- throughout the study. the rate of insulin-mediated glucose dis- search center for at least 6 days. Between appearance, leading to a substantial re- screening and admission, subjects were Statistical analyses duction of postprandial glucose asked to record their daily food intake, in- Main pharmacodynamic parameters in- excursions (17–19). Long-term clinical sulin regimen, and self-monitored blood cluded the incremental plasma glucose studies have shown that the postprandial glucose results in a diary. Upon admission, area under the concentration time curve glucose lowering effect of pramlintide each subject underwent a mixed meal test (AUC) from 0 to 2 h (AUC0–2h), incre- translates into a clinically meaningful and on 5 consecutive days in a randomized, sin- mental AUC0–4h, and the incremental statistically significant reduction in A1C gle-blind, placebo-controlled, five-way plasma glucose concentrations at specific in patients with types 1 and 2 diabetes crossover design. The meal was a standard- sampling times. The mean Ϯ SEM incre- (20–23). ized breakfast, consisting of a bagel with mental plasma glucose concentration The objective of the present study was margarine and jam, cheese, yogurt, milk, profiles were calculated and plotted by to further characterize the effect of pram- and orange juice. The size of the meal was treatment and by study group. For each lintide on postprandial glucose concen- calculated individually to provide 30% of a study group, the pharmacodynamic pa- trations when used as an adjunct to either subject’s total daily caloric requirements rameter data were summarized descrip- regular insulin or insulin lispro, with an with a macronutrient composition accord- tively and were analyzed using mixed emphasis on the optimal dose timing rel- ing to the American Diabetes Association effect models. The mixed effect models ative to meals. nutritional recommendations (55%/15%/ included treatment, treatment sequence, 30% of kcal from carbohydrate/protein/fat, and period as fixed effects and subject- RESEARCH DESIGN AND respectively). The size of the standardized within-sequence as random effects. METHODS breakfast meal was the same on each study The P values for comparisons among day for each individual, and the meal was the least square means of the incremental Study population always consumed within 15 min. AUC0–2h, incremental AUC0–4h, and in- A total of 21 subjects with type 1 diabetes On each day, subjects received one of cremental glucose concentrations at vari- using insulin lispro and 19 subjects with five treatments (a subcutaneous injection of ous time points between dose timings type 1 diabetes using regular insulin un- placebo at Ϫ15minor60␮g pramlintide were provided. Because the study was not derwent a standardized mixed meal test [0.6 mg/ml, Amylin Pharmaceuticals, Inc., powered to perform multiple compari- on 5 consecutive days. Subjects were San Diego, CA] at Ϫ15, 0, ϩ15, or ϩ30 sons, the P values were not adjusted for 18–65 years of age and had the following min relative to the standardized breakfast) multiple comparisons. characteristics: a history of type 1 diabetes according to a randomized sequence, after Safety evaluations were based on re- for at least 1 year, a baseline A1C value of an overnight fast.