Pramlintide Improved Glycemic Control and Reduced Weight in Patients with Type 2 Diabetes Using Basal Insulin

Clinical Care/Education/Nutrition/Psychosocial Research ORIGINAL ARTICLE

Pramlintide Improved Glycemic Control and Reduced Weight in Patients With Type 2 Diabetes Using Basal Insulin

1 2 MATTHEW RIDDLE, MD CARL BROWN, PHD achieving glycemic targets, intensification 2 2 JUAN FRIAS, MD KAREN LUTZ, PHD of therapy with the addition of mealtime 2 2 BEI ZHANG, MD ORVILLE KOLTERMAN, MD 2 insulin increases the risk of hypoglycemia HOLLY MAIER, PHD (5–7) and often results in undesirable weight gain (8–10). Pramlintide is a synthetic analog of OBJECTIVE — To assess the efficacy and safety of pramlintide in patients with type 2 diabetes human amylin, a naturally occurring neu- suboptimally controlled with basal insulin. roendocrine hormone cosecreted with insulin by pancreatic ␤-cells (11). Amylin RESEARCH DESIGN AND METHODS — In a 16-week, double-blind, placebo- regulates gastric emptying (12), sup- controlled study, 212 patients using insulin glargine with or without oral antidiabetes agents presses inappropriate postprandial gluca- (OAs) were randomized to addition of pramlintide (60 or 120 ␮g b.i.d./t.i.d.) or placebo. Insulin glargine was adjusted to target a fasting plasma glucose concentration of 70–100 mg/dl. One gon secretion (13), and reduces food coprimary end point was the change in A1C at week 16. The other coprimary end point was a intake (14,15). Through mechanisms composite measure of overall diabetes control comprising A1C Յ7.0% or reduction Ն0.5%, similar to those of amylin, pramlintide re- mean daily postprandial glucose (PPG) increments Յ40 mg/dl, no increase in body weight, and duces postprandial glucose (PPG), im- no severe hypoglycemia. Patients meeting all four conditions at week 16 achieved this end point. proving overall glycemic control (16,17), and increases satiety, resulting in reduced RESULTS — More pramlintide- than placebo-treated patients achieved the composite end food intake and weight loss (16–19). point (25 vs. 7%; P Ͻ 0.001). Reductions (means Ϯ SE) in A1C (Ϫ0.70 Ϯ 0.11% vs. Ϫ0.36 Ϯ Therapies that improve glycemic con- Ͻ Ϫ Ϯ Ϫ Ϯ Ͻ 0.08%; P 0.05) and PPG increments ( 24.4 3.6 mg/dl vs. 0.4 3.0 mg/dl; P 0.0001) trol without weight gain and its associated were greater in pramlintide- versus placebo-treated patients, respectively. Glycemic improve- long-term complications and do not in- ments were accompanied by progressive weight loss with pramlintide and weight gain with placebo (Ϫ1.6 Ϯ 0.3 kg vs. ϩ0.7 Ϯ 0.3 kg; P Ͻ 0.0001). No treatment-related severe hypogly- crease the risk of severe hypoglycemia cemia occurred. will significantly enhance treatment of patients with type 2 diabetes. This study in- CONCLUSIONS — Pramlintide improved multiple glycemic parameters and reduced vestigated the efficacy and safety of weight with no increase in hypoglycemia in patients with type 2 diabetes who were not achieving pramlintide therapy with basal insulin ti- glycemic targets with basal insulin with or without OAs. tration in patients with type 2 diabetes suboptimally controlled with basal insu- Diabetes Care 30:2794–2799, 2007 lin, with or without OAs.

ype 2 diabetes is characterized by ally a basal, long-acting preparation, is RESEARCH DESIGN AND insulin resistance and progressive eventually required to achieve adequate METHODS — Enrolled patients were T ␤-cell dysfunction resulting in defi- glycemic control. While basal insulin aged 25–75 years with type 2 diabetes and ciencies of insulin and amylin. Due to the therapy can result in adequate fasting glu- not achieving adequate glycemic control progressive nature of the disease, therapy cose control, it does not address post- with insulin glargine (no mealtime insulin), for most patients starts with medical nu- prandial hyperglycemia (1,2). Even with with or without OA therapy (metformin, trition therapy and exercise and is fol- rigorous basal insulin titration, ϳ30– sulfonylurea, and/or thiazolidinedione). lowed by the addition of one or more oral 40% of patients do not reach acceptable Inclusion criteria at screening included antidiabetes agents (OAs). Insulin, usu- A1C levels (Յ7.0%) (3,4). For those not A1C Ͼ7.0% and Յ10.5%, BMI 25–45 ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● kg/m2, insulin glargine treatment Ն3 From the 1Division of Endocrinology, Diabetes and Clinical Nutrition, Department of Medicine, Oregon months with a stable dose (Ϯ10%) for Ն1 Health and Science University, Portland, Oregon; and 2Amylin Pharmaceuticals, San Diego, California. month, and, if applicable, a stable dose of Address correspondence and reprint requests to Orville Kolterman, MD, Amylin Pharmaceuticals, 9360 OAs for Ն2 months. Female patients were Towne Centre Dr., San Diego, CA 92121. E-mail: [email protected]. Received for publication 25 March 2007 and accepted in revised form 3 August 2007. postmenopausal, surgically sterile, or Published ahead of print at http://care.diabetesjournals.org on 13 August 2007. DOI: 10.2337/dc07- used adequate contraception throughout 0589. Clinical trial reg. no. NCT00240253, clinicaltrials.gov. the study. Patients were excluded if they M.R. has received grant/research support from Amylin Pharmaceuticals, Eli Lilly, and sanofi-aventis; has had a history of hypoglycemia unaware- received consulting fees from Amylin Pharmaceuticals, ConjuChem, Emisphere, Eli Lilly, and sanofi-aventis; and has received honoraria from Amylin Pharmaceuticals, Eli Lilly, GlaxoSmithKline, Pfizer, and sanofi- ness or recurrent severe hypoglycemia aventis. during the preceding 6 months, were par- Abbreviations: FPG, fasting plasma glucose; OA, oral antidiabetes agent; PPG, postprandial glucose. ticipating in a weight loss program, were A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion using antiobesity agents, or had a con- factors for many substances. firmed diagnosis of gastroparesis or any © 2007 by the American Diabetes Association. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby other significant medical condition. marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The study protocol was approved by

2794 DIABETES CARE, VOLUME 30, NUMBER 11, NOVEMBER 2007 Riddle and Associates an institutional review board. All patients Study end points nary end points. A general linear model provided written informed consent be- Two coprimary end points were evalu- including treatment, baseline A1C stra- fore study initiation. The study was con- ated in this study. The first coprimary end tum (Յ8.0 or Ͼ8.0%), BMI stratum (Յ35 ducted in accordance with principles point was the change in A1C from base- or Ͼ35 kg/m2), and sulfonylurea use (yes/ outlined in the Declaration of Helsinki line to week 16. The second coprimary no) as covariates was used to compare the (1964), including all amendments end point was a dichotomous composite change in A1C at week 16. Parametric through the South Africa revision (1996). end point assessing the proportion of pa- analyses of secondary continuous end This was a 16-week, randomized, tients meeting all of the following pre- points were performed using general lin- double-blind, placebo-controlled, multi- specified criteria at week 16: 1)A1C ear models including treatment and base- center study conducted in the U.S. (41 Յ7.0% or an A1C reduction from baseline value as covariates. Descriptive sites) between October 2005 and June line Ն0.5%, 2) mean daily PPG incre- analyses and P values used the arithmetic 2006. After a screening visit, eligible pa- ments Յ40 mg/dl, 3) no weight gain, and and least squares means, respectively. tients made six visits to the study site 4) no severe hypoglycemia. Severe hypo- (baseline, 2, 4, 8, 12, and 16 weeks). At glycemia was defined as a hypoglycemic RESULTS the baseline visit, patients were random- event requiring assistance from another ized to receive pramlintide (Amylin Phar- individual and/or administration of glu- Patient disposition and baseline maceuticals, San Diego, CA) or placebo cagon or intravenous glucose. Secondary demographics (Amylin Pharmaceuticals). Randomiza- end points included components of the Of 212 patients randomized, 91 (85%) tion was stratified according to screening composite end point, the proportion of placebo-treated and 87 (83%) pramlint- Յ Յ visit A1C (Յ8orϾ8%), BMI (Յ35 or patients achieving A1C 7.0 or 6.5% ide-treated patients completed the study Ͼ35 kg/m2), and sulfonylurea use (yes/ and changes from baseline to each time (Table 1). One patient in the placebo- no). point in A1C, seven-point glucose pro- treated arm withdrew consent before in- Study medication (pramlintide or files, PPG increments, FPG, weight, and jection of study medication, resulting in placebo) was self-administered subcuta- insulin glargine dose. Similar ad hoc anal- an intent-to-treat population of 211 pa- neously immediately before major meals yses for secondary end points were per- tients. Baseline demographics were well depending on the patient’s typical meal formed on patients divided into matched between treatment arms (Table pattern (b.i.d. or t.i.d.). Patients initiated subgroups according to baseline A1C 1). Eighty-nine percent used at least one Յ Ͼ OA, and 50% used two or three OAs. study medication at a volume equivalent 8.5 or 8.5%. Within the pramlintide-treated popula- to 60 ␮g pramlintide per dose and esca- Statistical analyses tion, 98 (93%) patients escalated to the lated to a volume equivalent to 120 ␮g per A sample size of 90 patients per treatment 120-␮g dose. dose within 3–7 days if no clinically sig- arm was predicted to provide ϳ90% nificant nausea occurred. Once the main- power to detect a difference in the propor- Coprimary end points tenance dose was achieved, investigators tion of patients achieving the coprimary A1C. A1C values progressively de- were asked to make weekly adjustments composite end point and ϳ95% power to creased throughout the study. Pramlintide- in the insulin glargine dose to target a fast- demonstrate noninferiority of pramlint- treated patients achieved a significantly ing glucose concentration of Ն70 to Ͻ Ϯ Ͻ ide versus placebo for change in A1C (P 0.05) greater reduction (means SE) 100 mg/dl using an algorithm previ- from baseline. Noninferiority for change from baseline at week 16 (Ϫ0.70 Ϯ 0.11%) ously described by Riddle et al. (3). Pa- in A1C was concluded if the upper limit of than placebo-treated patients (Ϫ0.36 Ϯ tients self-monitored fasting glucose the two-sided 95% CI for the difference 0.08%), exceeding the noninferiority crite- concentrations daily and completed two between pramlintide and placebo was be- rion (upper limit of 95% CI ϭϪ0.04%) self-monitored, seven-point glucose pro- low the noninferiority margin of 0.4%. (Fig. 1A). Mean (ϮSE) A1C values at week files during the week before each visit The overall power for reaching both 16 were 7.8 Ϯ 0.1% (pramlintide) and consisting of measurements taken 15 min coprimary end points was expected to be 8.1 Ϯ 0.1% (placebo). The proportion of before and 1.5–2 h after the start of each ϳ85%. As both coprimary end points patients achieving an A1C Յ7.0 or Յ6.5% meal and at bedtime. Patients were re- were required to be met, no adjustment to was 23 and 11% with pramlintide and 13 quired to eat three meals on profile days. the significance level (␣ϭ0.05) was and 4% with placebo, respectively. Patients used study-provided Accu-Chek required. Composite end point. At week 16, sig- Aviva blood glucose monitors (Roche Di- Analyses were performed on patients nificantly more pramlintide-treated pa- agnostics, Indianapolis, IN), reporting within the intent-to-treat population, all tients achieved the composite end point plasma-referenced glucose concentra- of whom received at least one dose of than placebo-treated patients (25 vs. 7%; tions. At each visit, weight and vital signs study medication. Missing individual P Ͻ 0.001) (Fig. 1B). were measured and self-monitored blood data were imputed from the last sched- glucose values, insulin dose, and adverse uled visit using the last-observation- Secondary end points events reviewed. A1C was measured at carried-forward approach for all efficacy Components of the composite end screening, baseline, and every 4 weeks analyses, with the exception of FPG, in- point. The percentage of pramlintide- thereafter. Laboratory measurement of sulin dose, and the seven-point glucose versus placebo-treated patients achieving fasting plasma glucose (FPG) was per- profiles that were analyzed using the in- an A1C Յ7.0% or an A1C reduction formed at baseline and week 16. Patients tent-to-treat observed population. Fish- Ն0.5% was not significantly different were instructed to maintain their usual er’s exact test was used to compare the (Fig. 1C). Significantly more pramlintide- diet and exercise regimens throughout proportion of patients achieving the treated patients achieved mean PPG in- the study. coprimary composite and secondary bi- crements Յ40 mg/dl (P Ͻ 0.0001) and

DIABETES CARE, VOLUME 30, NUMBER 11, NOVEMBER 2007 2795 Pramlintide, basal insulin, and type 2 diabetes

Table 1—Patient disposition and baseline demographics

Placebo Pramlintide Total population baseline A1C baseline A1C Placebo Pramlintide Յ8.5% Ͼ8.5% Յ8.5% Ͼ8.5% Disposition (n) Randomized 107 105 59 48 63 42 Completed 91 87 51 40 52 35 Withdrew 16 18 8 8 11 7 Reason for withdrawal Withdrawal of consent 12 95763 Adverse event 141031 Investigator decision 302100 Protocol violation 020011 Lost to follow-up 030012 Baseline demographics Intent-to-treat population (n) 106 105 58 48 63 42 Sex (male/female) (n) 55/51 48/57 35/23 20/28 32/31 16/26 Race (Caucasian/other) (n) 77/29 77/28 40/18 37/11 47/16 30/12 Age (years) 55 Ϯ 10 55 Ϯ 955Ϯ 11 56 Ϯ 956Ϯ 853Ϯ 9 Weight (kg) 103 Ϯ 18 103 Ϯ 18 105 Ϯ 20 99 Ϯ 16 104 Ϯ 18 102 Ϯ 16 BMI (kg/m2) 35 Ϯ 635Ϯ 535Ϯ 635Ϯ 635Ϯ 536Ϯ 4 Diabetes duration (years) 10 Ϯ 611Ϯ 610Ϯ 611Ϯ 611Ϯ 611Ϯ 6 A1C (%) 8.5 Ϯ 0.9 8.5 Ϯ 0.9 7.7 Ϯ 0.4 9.3 Ϯ 0.6 7.9 Ϯ 0.4 9.4 Ϯ 0.7 FPG (mg/dl) 140 Ϯ 54 146 Ϯ 52 133 Ϯ 57 150 Ϯ 50 132 Ϯ 45 167 Ϯ 56 PPG increments (mg/dl) 57 Ϯ 27 59 Ϯ 29 58 Ϯ 29 56 Ϯ 26 58 Ϯ 28 62 Ϯ 32 OA use (n) 969254425933 Sulfonylurea use (no/yes) (n) 50/56 54/51 26/32 24/24 30/33 24/18 Daily insulin dose (units) 54 Ϯ 42 48 Ϯ 25 57 Ϯ 43 51 Ϯ 42 44 Ϯ 22 55 Ϯ 29 Data are means Ϯ SD, unless otherwise indicated. did not gain weight (P Ͻ 0.0001). Com- PPG increments. Mean (ϮSE) PPG in- 9.4%) and mean FPG (132 vs. 158 mg/ pared with placebo, more pramlintide- crements at week 16 were 34.8 Ϯ 2.7 dl). Insulin glargine dosage increased treated patients achieved both A1C and mg/dl (pramlintide) and 56.6 Ϯ 2.3 steadily from baseline to week 16 in both PPG components (P Ͻ 0.005), more mg/dl (placebo), reflecting significant de- subgroups. reached the A1C goal without weight gain creases in PPG increments from baseline Baseline A1C <8.5%. At week 16, (P Ͻ 0.0001), and more had well- to week 16 in pramlintide- versus place- pramlintide-treated patients exhibited re- controlled PPG without weight gain (P Ͻ bo-treated patients (Ϫ24.4 Ϯ 3.6 mg/dl ductions from baseline in mean (ϮSE) 0.0001) (Fig. 1D). One episode of severe [pramlintide] vs. Ϫ0.4 Ϯ 3.0 mg/dl [pla- A1C (Ϫ0.36 Ϯ 0.13%), FPG (Ϫ17.3 Ϯ hypoglycemia occurred in a pramlintide- cebo]) (P Ͻ 0.0001) (Fig. 2B). 7.1 mg/dl), PPG increments (Ϫ24.9 Ϯ treated patient but was deemed unrelated to Weight. Pramlintide treatment resulted 4.4 mg/dl), and weight (Ϫ2.0 Ϯ 0.4 kg). pramlintide treatment by the investigator. in progressive weight loss, while placebo- In contrast, placebo-treated patients ex- Insulin. Insulin glargine dosage in- treated patients gained weight (week 16: hibited a reduction from baseline in mean creased steadily throughout the study Ϫ1.6 Ϯ 0.3 kg vs. 0.7 Ϯ 0.3 kg, P Ͻ (ϮSE) FPG (Ϫ7.5 Ϯ 6.8 mg/dl) but did (Fig. 2A). Mean (ϮSE) week 16 dosage 0.0001) (Fig. 2C). At week 16, approxi- not exhibit changes from baseline in A1C was 61.4 Ϯ 3.4 units (pramlintide) and mately two-thirds (68%) of pramlintide- (Ϫ0.08 Ϯ 0.09%), PPG increments 69.5 Ϯ 5.3 units (placebo), reflecting in- treated patients had lost weight compared (Ϫ3.6 Ϯ 3.8 mg/dl), or weight (0.4 Ϯ 0.4 creases of 11.7 Ϯ 1.9 units and 13.1 Ϯ with approximately one-third (35%) of kg). 1.6 units, respectively. placebo-treated patients (P Ͻ 0.0001) Baseline A1C >8.5%. At week 16, Fasting plasma glucose. Mean (ϮSE) (Fig. 2D and E). pramlintide-treated patients exhibited re- FPG concentrations at week 16 were ductions from baseline in mean (ϮSE) 119.5 Ϯ 4.1 mg/dl (pramlintide) and Patient stratification according to A1C (Ϫ1.19 Ϯ 0.14%), FPG (Ϫ44.4 Ϯ 122.8 Ϯ 4.3 mg/dl (placebo), reflecting baseline A1C 12.7 mg/dl), PPG increments (Ϫ23.7 Ϯ an average change from baseline of To further explore the implications of 5.9 mg/dl), and weight (Ϫ1.0 Ϯ 0.3 kg). Ϫ28.3 Ϯ 6.8 mg/dl (pramlintide) and these results in clinical practice, we di- Placebo-treated patients exhibited reduc- Ϫ12.0 Ϯ 5.6 mg/dl (placebo). An FPG vided the study population into two sub- tions from baseline in mean (ϮSE) FPG concentration Ͻ100 mg/dl was achieved groups according to the mean baseline (Ϫ18.4 Ϯ 9.4 mg/dl) and A1C (Ϫ0.69 Ϯ by 28 of 105 (27%) pramlintide-treated A1C (Յ8.5 or Ͼ8.5%) (Table 1). These 0.13%) but did not exhibit a change in and 33 of 106 (31%) placebo-treated pa- subgroups were similar in baseline char- PPG increments (3.2 Ϯ 4.6 mg/dl), and tients at week 16. acteristics, except for mean A1C (7.8 vs. they gained weight (1.1 Ϯ 0.4 kg). The

2796 DIABETES CARE, VOLUME 30, NUMBER 11, NOVEMBER 2007 Riddle and Associates

ductions in fasting glucose resulting from basal insulin titration. Moreover, as in prior studies of pramlintide used in com- bination with mealtime insulin (16,17,19), this treatment regimen resulted in weight loss, while insulin titration alone caused weight gain. The coprimary composite study end point, comprising A1C, PPG, weight, and severe hypoglycemia components, was designed to measure the proportion of patients achieving a highly desirable clinical out- come. Significantly more pramlintide- treated patients achieved this end point (25%) than patients receiving insulin alone (7%), confirming the clinical ad- vantages of pramlintide plus basal insulin over basal insulin alone. Therapies that reduce PPG and body weight may provide long-term benefits to patients with type 2 diabetes. Postpran- dial hyperglycemia has been implicated in the development of micro- and macrovascular complications through mechanisms including increased oxidative stress and inflammation (20–22). Moreover, obesity is very common in patients with type 2 diabetes and contributes to an already- increased risk of cardiovascular disease. Ϯ Ͻ Whether the severity of A1C elevation Figure 1—A: Mean ( SE) change in A1C from baseline to each visit. *P 0.05 for pramlintide at baseline affects the benefits of adding vs. placebo. , placebo; f, pramlintide. B: Percentage of patients achieving the composite end point at week 16. C: Percentage of patients achieving each component within the composite end pramlintide is of clinical interest. There- point at week 16. , placebo; f, pramlintide. D: Percentage of patients achieving at least two fore, ad hoc analyses were performed on components, not including severe hypoglycemia, within the composite end point at week 16. , patient subgroups with baseline A1C placebo; f, pramlintide. B–D:*P Ͻ 0.005; **P Ͻ 0.0001 for pramlintide vs. placebo. Ͼ8.5 or Յ8.5%. In patients with higher baseline A1C, basal insulin titration alone reduced A1C at the price of weight gain, reduction in PPG increments in pramlint- deemed this event unrelated to pramlintide while pramlintide plus basal insulin titra- ide- but not placebo-treated patients in treatment. tion resulted in greater reductions in A1C both A1C subgroups is illustrated by sev- (via PPG reductions) and induced weight en-point glucose profiles performed at CONCLUSIONS — Patients with sub- loss. In patients with lower baseline A1C, baseline and week 16 (Fig. 3). optimal glycemic control on basal insulin basal insulin titration alone did not pro- therapy may further improve control by in- vide much benefit, indicating the need for Safety creasing the basal insulin dose and/or add- additional therapy. In contrast, pramlint- The most common adverse events were ing mealtime insulin, but at the expense of ide plus basal insulin titration reduced mild to moderate nausea (31% pramlint- additional weight gain and an increased risk both A1C and weight. Thus, pramlintide ide, 10% placebo) and mild to moderate of hypoglycemia (1,10). In addition to their provided benefits beyond those of basal hypoglycemia (44% pramlintide, 47% clinical significance, these side effects are insulin alone regardless of baseline A1C. placebo). Most nausea occurred within the disliked by patients and, thus, may deter This study had several limitations. first week of treatment and decreased over intensification of insulin therapy. First, the relatively short 16-week dura- time. Two pramlintide-treated patients This study demonstrated that the ad- tion was not long enough to allow insulin withdrew from the study due to mild or dition of pramlintide with continued dosage, A1C, and weight to plateau. Sec- moderate nausea. Other adverse events basal insulin titration allowed such pa- ond, many patients entering this study leading to withdrawal were treatment- tients to achieve improved glycemic con- had high A1C values despite substantial related pruritis at the injection site (one pa- trol and additional metabolic benefits not basal insulin doses (ϳ54 units daily for tient in each treatment arm) and alopecia, achieved with insulin titration alone. those with baseline A1C Ͼ8.5%), sug- which was not considered treatment- Pramlintide, as an adjunct to basal insu- gesting that endogenous insulin secretion related (one patient in the pramlintide arm). lin, allowed patients to achieve an A1C was low. Many of those patients will even- One event of severe hypoglycemia occurred lower than that achieved with basal insu- tually need mealtime insulin to reach an in a pramlintide-treated patient who acci- lin titration alone. This was accomplished A1C Յ7.0%. Studying the use of pram- dentally took a dose of rapid-acting insulin through pramlintide-dependent reduc- lintide with basal insulin earlier in the instead of insulin glargine. The investigator tions in PPG increments coupled with re- course of type 2 diabetes is therefore of

DIABETES CARE, VOLUME 30, NUMBER 11, NOVEMBER 2007 2797 Pramlintide, basal insulin, and type 2 diabetes

Figure 2—A: Mean (ϮSE) daily insulin glargine doses. , placebo; f, pramlintide. B: Mean (ϮSE) change in PPG increments from baseline during the study. , placebo; f, pramlintide. *P Ͻ 0.0001 for pramlintide vs. placebo. C: Mean (ϮSE) change in body weight from baseline during the study. *P Ͻ 0.0001 for pramlintide vs. placebo. , placebo; f, pramlintide. D and E: Individual weight changes from baseline for placebo-treated (D) and pramlintide-treated (E) patients. Percentages of patients that gained or lost weight are indicated. interest. Third, the seven-point glucose profiles demonstrated improved but per- sistently high postbreakfast glucose increments in pramlintide-treated patients. Some pramlintide-treated patients might have benefited from mealtime insulin at breakfast to achieve adequate glycemic control. Pramlintide added to basal insulin was generally well tolerated. Earlier studies of pramlintide indicated an increased risk of insulin-induced severe hypoglycemia, which occurred primarily in the more hypoglycemia-prone type 1 diabetic population (16,17). In contrast, no treatment-related severe hypoglycemia occurred in the present study. Also, the frequency of mild-to-moderate hypoglycemia was similar between the two treatment arms, despite the fact that pramlintide-treated patients achieved significantly better glycemic control. In summary, adding pramlintide to basal insulin improved multiple aspects of diabetes control, thereby addressing important challenges associated with in- Figure 3— Mean (ϮSE) seven-point glucose profiles in patients with baseline A1C Յ8.5% (A and tensifying insulin therapy. These findings B)orϾ8.5% (C and D).

2798 DIABETES CARE, VOLUME 30, NUMBER 11, NOVEMBER 2007 Riddle and Associates support pramlintide as a potential option Martinez Ravn G, Clauson P, Home P: A T, Scharrer E: The anorectic effect of a for the next therapeutic step when pa- 26-week, randomized, parallel, treat-to- chronic peripheral infusion of amylin is tients with type 2 diabetes are not achiev- target trial comparing insulin detemir abolished in area postrema/nucleus of the ing glycemic targets with basal insulin with NPH insulin as add-on therapy to solitary tract (AP/NTS) lesioned rats. Int J therapy. Further studies examining oral glucose-lowering drugs in insulin-na- Obes Relat Metab Disord 25:1005–1011, ive people with type 2 diabetes. Diabetes 2001 pramlintide as an alternative to mealtime Care 29:1269–1274, 2006 insulin are warranted. 15. Rushing PA, Lutz TA, Seeley RJ, Woods 5. Jaap AJ, Jones GC, McCrimmon RJ, Deary SC: Amylin and insulin interact to reduce IJ, Frier BM: Perceived symptoms of hy- food intake in rats. Horm Metab Res 32: poglycaemia in elderly type 2 diabetic pa- 62–65, 2000 Acknowledgments— Data from this study tients treated with insulin. Diabet Med 15: were presented at the 67th annual meeting of 16. Whitehouse F, Kruger DF, Fineman M, 398–401, 1998 Shen L, Ruggles JA, Maggs DG, Weyer C, the American Diabetes Association, Chicago, 6. Gurlek A, Erbas T, Gedik O: Frequency of Kolterman OG: A randomized study and Illinois, 22–26 June 2007, and at the 43rd an- severe hypoglycaemia in type 1 and type 2 open-label extension evaluating the long- nual meeting of the European Association for diabetes during conventional insulin the Study of Diabetes, Amsterdam, the Neth- therapy. Exp Clin Endocrinol Diabetes 107: term efﬁcacy of pramlintide as an adjunct erlands, 17–21 September 2007. 220–224, 1999 to insulin therapy in type 1 diabetes. Di- 7. Hepburn DA, MacLeod KM, Pell AC, abetes Care 25:724–730, 2002 Scougal IJ, Frier BM: Frequency and 17. Ratner RE, Dickey R, Fineman M, Maggs APPENDIX DG, Shen L, Strobel SA, Weyer C, Kolter- Participating investigators: A. Ahmann, S. symptoms of hypoglycaemia experienced by patients with type 2 diabetes treated man OG: Amylin replacement with pram- Aronoff, R. Bhushan, W. Cheatham, R. with insulin. Diabet Med 10:231–237, lintide as an adjunct to insulin therapy Cherlin, G. Collins, M. Feinglos, N. Fish- 1993 improves long-term glycaemic and man, G. Grunberger, R. Guthrie, P. Hol- 8. U.K. Prospective Diabetes Study (UKPDS) weight control in type 1 diabetes mellitus: lander, R. Kaplan, D. Karl, A. King, E. Group: Intensive blood-glucose control a 1-year randomized controlled trial. Dia- Klein, L. Levinson, P. Levy, M. May, M. with sulphonylureas or insulin compared bet Med 21:1204–1212, 2004 Magee, A. McCall, W. Miers, T. Moretto, with conventional treatment and risk of 18. Chapman I, Parker B, Doran S, Feinle- L. Olansky, W. Petit, A. Philis-Tsimikas, complications in patients with type 2 di- Bisset C, Wishart J, Strobel S, Wang Y, S. Plevin, J. Pullman, B. Ramlo-Halsted, abetes (UKPDS 33). Lancet 352:837–853, Burns C, Lush C, Weyer C, Horowitz M: M. Rendell, M. Riddle, D. Robertson, R. 1998 Effect of pramlintide on satiety and food Rood, J. Rosenstock, J. Shapiro, T. Wahl, 9. Henry RR, Gumbiner B, Ditzler T, Wal- intake in obese subjects and subjects with lace P, Lyon R, Glauber HS: Intensive con- R. Weinstein, F. Whitehouse, K. Wil- type 2 diabetes. Diabetologia 48:838–848, ventional insulin therapy for type II 2005 liams, A. Wynne, C. Wysham, and L. Ze- diabetes: metabolic effects during a 6-mo mel. 19. Hollander PA, Levy P, Fineman MS, outpatient trial. Diabetes Care 16:21–31, Maggs DG, Shen LZ, Strobel SA, Weyer C, 1993 Kolterman OG: Pramlintide as an adjunct 10. Purnell J, Weyer C: Weight effect of cur- to insulin therapy improves long-term References rent and experimental drugs for diabetes glycemic and weight control in patients 1. U.K. Prospective Diabetes Study Group: mellitus: from promotion to alleviation of with type 2 diabetes: a 1-year randomized U.K. Prospective Diabetes Study 16. obesity. Treat Endocrinol 2:33–47, 2003 controlled trial. Diabetes Care 26: Overview of 6 years’ therapy of type II 11. Young AA: Amylin’s physiology and its diabetes: a progressive disease. U.K. Pro- role in diabetes. Curr Opin Endocrinol Diab 784–790, 2003 spective Diabetes Study Group. Diabetes 4:282–290, 1997 20. Ceriello A: Oxidative stress and diabetes- 44:1249–1258, 1995 12. Young AA, Gedulin B, Vine W, Percy A, associated complications. Endocr Pract 12 2. DeFronzo RA: Pharmacologic therapy for Rink TJ: Gastric emptying is accelerated (Suppl. 1):60–62, 2006 type 2 diabetes mellitus. Ann Intern Med in diabetic BB rats and is slowed by sub- 21. Brownlee M: Biochemistry and molecular 131:281–303, 1999 cutaneous injections of amylin. Diabetolo- cell biology of diabetic complications. Na- 3. Riddle MC, Rosenstock J, Gerich J: The gia 38:642–648, 1995 ture 414:813–820, 2001 treat-to-target trial: randomized addition 13. Gedulin BR, Rink TJ, Young AA: Dose- 22. Home P: Contributions of basal and post- of glargine or human NPH insulin to oral response for glucagonostatic effect of prandial hyperglycaemia to micro- and therapy of type 2 diabetic patients. Diabe- amylin in rats. Metabolism 46:67–70, macrovascular complications in people tes Care 26:3080–3086, 2003 1997 with type 2 diabetes. Curr Med Res Opin 4. Hermansen K, Davies M, Derezinski T, 14. Lutz TA, Mollet A, Rushing PA, Riediger 21:989–998, 2005

DIABETES CARE, VOLUME 30, NUMBER 11, NOVEMBER 2007 2799