- Email: [email protected]
Therapy after single oral agent failure: adding a second oral agent or an insulin mixture?
Diabetes Research and Clinical Practice 62 (2003) 187 /195 www.elsevier.com/locate/diabres
Therapy after single oral agent failure: adding a second oral agent or an insulin mixture? James K. Malone a,*, Scott D. Beattie a, Barbara N. Campaigne a, Patricia A. Johnson a, Andrew S. Howard a, Zvonko Milicevic b a
Lilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285, USA b Eli Lilly and Company, Regional Operations GES, M.B.H., Vienna, Austria Received 2 May 2002; received in revised form 19 June 2003; accepted 4 August 2003
Abstract Aim: to compare the glycemic response to an insulin lispro mixture (25% insulin lispro and 75% NPL) twice daily plus metformin (Mix25/M) with glibenclamide plus metformin (G/M), in patients with type 2 diabetes inadequately controlled with a single oral agent. Methods: 597 patients treated in a randomized, open-label, 16-week parallel study. Variables evaluated: hemoglobin A1C (A1C), patient symptoms, hypoglycemia rate (episodes/patient/30 days), and incidence (% patients experiencing ]/1 episode). For a subset of patients (N/120), fasting, 1-h, and 2-h postprandial plasma glucose (FPG, 1-h ppPG, 2-h ppPG) in response to a standardized test meal (STM) and self-monitored blood glucose (BG) profiles were measured. Results: improved A1C at endpoint for both groups, and A1C changes from baseline to endpoint were not significantly different between treatments (Mix25/M, /1.879/1.35% vs. G/M, / 1.989/1.28%; p /0.288). Among patients completing STM; endpoint 2-h ppPG was significantly lower with Mix25/M (9.059/3.32 mmol/l vs. 12.319/3.65 mmol/l; p B/0.001), as was 2-h ppPG excursion (2-h ppPGex)(0.389/3.23 mmol/l vs. 2.889/1.98 mmol/l; p B/0.001). Percentage of patients achieving postprandial BG targets (B/10 mmol/l) at endpoint was significantly greater with Mix25/M (80% vs. 48%; p B/0.001). Although, overall hypoglycemia rates were similar, percentage of patients experiencing and rate of nocturnal hypoglycemia was less with Mix25/M (1% vs. 5%; p B/0.01, and 0.01 vs. 0.08 episodes/pt/30 d; p/0.007). Patients reported less polyuria with Mix25/M (p B/0.001). Conclusion: in patients with type 2 diabetes failing on metformin or a sulfonylurea, Mix25/M provided similar overall glycemic control, lower ppPG, reduced nocturnal hypoglycemia, and fewer hyperglycemic symptoms compared to G/M. # 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Type 2 diabetes; Oral agents; Insulin mixtures
1. Introduction
* Corresponding author. Tel.: /1-317-277-6472; fax: /1317-277-5458. E-mail address: [email protected] (J.K. Malone).
The management of type 2 diabetes begins with education, diet, physical activity, and the attempt to reduce body weight in overweight patients. When these measures fail to achieve adequate
0168-8227/03/$ - see front matter # 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.diabres.2003.08.003
188
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
glycemic control, oral antidiabetic agents are typically used. However, patients with type 2 diabetes treated solely with single oral agent therapy */for example, second-generation sulfonylurea or metformin */often find their glycemic control is inadequately maintained even at maximum dosages. An effective approach to sulfonylurea failure is to add metformin [1]; the addition of insulin therapy is generally accepted as a treatment option only after acceptable glycemic control cannot be achieved or maintained with multiple oral agents. However, sulfonylurea plus metformin therapy does not address postprandial hyperglycemia, an early and prominent metabolic defect in type 2 diabetes. According to the European Diabetes Policy Group, the criteria for good metabolic control are hemoglobin A1C B/ 6.5% and 2-h postprandial blood glucose less than 7.5 mmol/l (135mg/dl) [2]. A major goal of current diabetes management is to prevent or slow the progression of diabetes-related complications. Clinical studies in patients with type 1 or type 2 diabetes have demonstrated that the development of long-term complications of diabetes and the level of metabolic control are closely correlated [3 /5]. However, the association of postprandial blood glucose (BG) control with diabetes complications remains controversial [6]. Insulin lispro, a rapid-acting insulin analog, has been shown to result in lower postprandial BG and a lower risk of nocturnal hypoglycemia in comparison to regular human insulin [7,8]. A pre-mixed insulin containing 25% insulin lispro and 75% insulin lispro protamine suspension (Humalog† Mix25TM, Humalog† Mix75/25TM in the US) (Mix25) was approved for clinical use in 1998 and is now available in many countries. Compared to human insulin mixtures, twice daily administration of insulin lispro mixtures results in improved postprandial blood glucose, similar overall glycemic control, less risk of nocturnal hypoglycemia, and the convenience of dosing immediately before meals [9,10]. Introducing a rapid-acting insulin mixture in combination with metformin after a single oral agent fails to maintain adequate glycemic control could improve postprandial blood glucose levels, and improve overall glycemic control with a low
risk of hypoglycemia. In the current study, patients with type 2 diabetes who were inadequately controlled on a sulfonylurea or metformin were transferred to Mix25 twice daily plus metformin or glibenclamide plus metformin, and the effects of these two regimens on glycemic control were compared.
2. Materials and methods 2.1. Patient population This study included 597 patients with type 2 diabetes, as defined by the World Health Organization [11], ages 30 /75 years, with body mass index (BMI) B/40 kg/m2. Each patient’s diabetes was inadequately controlled, as determined by a hemoglobin A1C /125% of the upper limit of normal by the local laboratory within 4 weeks prior to or at study entry. For at least 3 months prior to entering the study, patients were required to be using a single oral-antidiabetic agent, specifically either metformin or a second-generation sulfonylurea (e.g., glibenclamide, glipizide, gliclazide, glimepiride) at a maximally clinically effective dose within the last 30 days or more as determined by the investigator. The study was approved by the ethical review boards of the participating centers. Patients were required to complete and sign an informed consent document to participate in the study, and the study was conducted in accordance with the Declaration of Helsinki and good clinical practice guidelines. 2.2. Study design This was a randomized, open-label, two-arm, parallel prospective study in patients with type 2 diabetes and inadequate glycemic control using either a second-generation sulfonylurea or metformin, and was conducted in 56 centers in 14 countries. After a 2-week lead-in period, during which patients were treated with either glibenclamide or metformin (depending upon pre-study oral agent therapy), patients were randomly assigned to one of the following two treatment groups: twice daily Mix25 plus metformin
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
(Mix25/M) or twice daily glibenclamide plus metformin (G/M). 2.3. Lead-in period The lead-in period consisted of 2 (9/1) weeks between visit 1 (study entry) and visit 2 (the baseline visit and randomization). At visit 1, after giving informed consent, the patient’s height, weight, history, and pre-existing conditions were obtained, and a physical examination was performed. Patients were given diaries for recording hypoglycemic episodes and study drug dosages. Patients who used a sulfonylurea prior to the study were given glibenclamide during the lead-in period at a dose equivalent to their previous sulfonylurea dosage. Patients using metformin prior to the study continued with the same dose of metformin during the lead-in period. 2.4. Treatment period The treatment period began at visit 2 (week 0) and ended at visit 6 (week 169/1). This allowed for 4 weeks of dose adjustment and 12 weeks for stabilization of glycemic control. At visit 2 patients were randomized to receive either Mix25/M or G/M for 16 weeks. Patients assigned to dual oral agent therapy added the second oral agent (G or M) to their existing oral agent. For the group randomized to receive Mix25/ M, the investigators titrated insulin doses to achieve BG targets for fasting and pre-meal BG B/7 mmol/l (126 mg/dl) and 2-h postprandial BG B/10 mmol/l (180 mg/dl), without increasing the frequency of hypoglycemia. Patients taking metformin during the lead-in period maintained their metformin dose and added Mix25 twice daily, immediately before the morning and evening meals. The metformin dose was titrated so that by visit 3 (4 weeks after the baseline visit 2) patients were taking at least 1500 mg per day (mg/day) but not more than 2550 mg/day, in two to three divided doses with meals. The dose of metformin was held constant after Visit 3. Patients taking glibenclamide during the lead-in period were switched to metformin and Mix25 was added twice-daily.
189
In the dual oral agent group, the metformin dose was titrated in the same way, while the glibenclamide dose was adjusted to reach the fasting and pre-meal blood glucose target throughout the remainder of the study. All patients returned to the study center once every 4 (9/1) weeks for routine assessments. At each visit, completed diaries were collected and new ones were distributed. Body weight, adverse events, episodes of hypoglycemia, study drug dose, and concomitant medication were also obtained. At visits 2 and 6, blood was drawn for measurement of A1C. The A1C assay was performed by ion-exchange high performance liquid chromatography (HPLC) on a Bio-Rad variant analyzer at a central laboratory (Covance, Indianapolis, IN; reference range, 4.3 /6.1%). The Variant system is certified by the National Glycohemoglobin Standardization Program (NGSP) to be traceable to the Diabetes Control and Complications Trial (DCCT) [3] reference method and values. At visits 2 and 6, all patients were asked to complete a treatment satisfaction questionnaire, and at visit 6 patients receiving Mix25/M were asked to complete an insulin delivery device questionnaire. All patients at sites selected to represent the specific regions of the world (sites in Romania, Australia, Canada, and Brazil) were asked to perform three 8-point BG profiles (premeal and 2-h postprandial for morning, mid-day and evening meals, as well as at bedtime and 03:00 h) on separate days during the 2-week period prior to visit 2 and visit 6. These patients also participated in an on-site standardized test meal (STM) on the mornings of visits 2 and 6. Upon arriving at the study center, the patients had fasting plasma glucose (FPG) measured, consumed the STM, and then had 1-h and 2-h postprandial plasma glucose (1 and 2-h ppPG) levels measured. The composition of the STM was approximately 369 kcal (57% carbohydrate, 11% protein and 32% fat) and consisted of Glucerna† (Ross Products Division, Abbott Laboratories, Abbott Park, IL) (237 ml) and a fruit cup (234 g). Patients arrived at the study site after an 8-h overnight fast having omitted their morning diabetes therapy. The study drugs were administered five minutes before the STM.
190
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
2.5. Study drugs Humalog† Mix25TM (Eli Lilly and Company, Indianapolis, IN) was provided in 3-ml cartridges, or 3-ml pre-filled pens*/100 units/ml. The oralantidiabetic agents were glibenclamide in 5 mg tablets (bioequivalent to 3.5 mg micronized form) and metformin hydrochloride in 500 and 850 mg tablets. 2.6. Statistical methods Randomization to treatment was performed, stratified by investigative site and by prestudy oral therapy (metformin or sulfonylurea), such that an approximately equal number of patients were assigned to the two treatments within each stratum. The sample size for the study was chosen to provide adequate power to determine treatment differences in A1C improvement (change from baseline to endpoint) within each prestudy oral therapy stratum. For this reason, enrollment was designed to randomize at least 240 patients from each prestudy therapy. This was planned to provide 63/80% power for a 0.4% treatment difference within each stratum and 78% power for a 0.3% difference overall. Statistical analyses were performed with SAS version 6.09 using two-sided tests at a 0.05 level of significance. Primary analyses utilized the endpoint data from each patient (last observation carried forward), and comparisons were made according to treatment received. Hemoglobin A1C and body weight were analyzed by Type III sums of squares from an analysis of variance (ANOVA) model with effects for treatment, country, and treatment /country interaction. Patient data from countries with fewer than 10 patients were pooled with geographic neighbors. Glucose values from the test meals and from self-monitored BG profiles were analyzed by a similar ANOVA model using investigator instead of country. Hypoglycemia was defined as any time a patient felt, or another person observed, that the patient was experiencing a sign/symptom which they associated with hypoglycemia or the patient recorded a BG B/3.5 mmol/l. Hypoglycemia data were collected via patient home BG monitoring
and patient records. Nocturnal hypoglycemia was defined as any time one of these episodes occurred after bedtime and before awakening. Episodes where the patient was unable to treat himself/ herself or that were associated with a BG 5/2.0 mmol/l were defined as severe. Incidence (percentage of patients experiencing) of treatment-emergent adverse events and episodes of hypoglycemia were analyzed using Fisher’s exact test. Rates of hypoglycemia (episodes/patient/30 days) were analyzed using a ranked ANOVA model with treatment and country effects. Answers to treatment satisfaction questions, measured on a five-point Likert scale, were analyzed using a cumulativelogit model that included treatment effect only. All continuous variables are reported as unadjusted mean (9/S.D.).
3. Results Of the 597 randomized patients, 296 were assigned to treatment with Mix25/M, and 301 were assigned to G/M. Prior to the study, 381 patients had used sulfonylurea therapy and 216 used metformin. A total of 543 patients completed the study; 25 patients dropped out while receiving Mix25/M for the following reasons: 1 */adverse event, 1*/death (not related to study drug), 1 */ lost to follow-up, 1 */moved away, 7*/patient decision, 3*/entry criteria not met, 2*/sponsor decision, 3 */physician decision, 3 */protocol violation, and 3*/inability to tolerate metformin. Twenty-nine patients dropped out while receiving G/M for the following reasons: 2 */adverse event, 1 */lack of efficacy, 2*/lost to follow-up, 1*/patient decision, 3 */entry criteria not met, 4*/physician decision, 6 */protocol violation, 10*/inability to tolerate metformin. At baseline, the patients’ mean age was 58 (9/9) years and their mean BMI was 29.7 (9/4.5) kg/m2 (Table 1). Patients in both treatment groups experienced a mild increase in body weight (Mix25/M, 0.89/3.4 kg; G/M, 0.39/2.8 kg) during the study, although there was no significant difference between treatments (Table 2). The mean insulin doses achieved at endpoint were 0.19 U/kg in the morning and 0.14 U/kg in
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
Table 1 Patients’ baseline characteristics by treatment group (mean9/ S.D.) Variable
Mix25/M (n/296)
G/M (n/301)
Age (years) 589/8.8 599/9.3 Males 169 (57%) 146 (49%) Females 127 (43%) 155 (51%) Total 296 301 Caucasian 263 (88.9%) 268 (89.0%) African 2 (0.7%) 3 (1.0%) Hispanic 22 (7.4%) 18 (6.0%) Other 9 (3.0%) 12 (4.0%) Weight (kg) 83.09/15.2 81.79/15.7 BMI (kg/m2) 29.89/4.4 29.69/4.5 11.509/3.08 FBG (mmol/l) 11.619/2.97 A1C (%) 9.179/1.50 9.279/1.55 Diabetes duration (years) 8.09/5.8 7.49/5.4
pvalue 0.124 0.036 0.793
0.292 0.589 0.759 0.181 0.160
the evening. Average total daily dose of metformin for the Mix25/M group was 1813 mg and for the G/M group was 1968 mg. The mean glibenclamide dose for the G/M group was 14.2 mg. 3.1. Hemoglobin A1C Both treatment groups experienced significant improvement in A1C as measured by change from baseline to endpoint (p B/0.001). However, there was no significant difference between treatments in A1C improvement (Mix25/M, /1.879/1.35% vs. G/M, /1.989/1.28%; p /0.288). Additionally, there was no significant difference between treatments in endpoint A1C (Mix25/M, 7.299/
191
1.00% vs. G/M, 7.339/1.14%; p/0.661) (Table 2) or in the percentage of patients achieving A1C values B/7.0% (Mix25/M, 40% vs. G/M, 41%), or the percentage of patients with A1C values B/ 6.5% (Mix25/M, 18% vs. G/M, 19%). In the subgroup of patients completing the test meals, the A1C levels were similar to the overall cohort at baseline and endpoint (Mix25/M, 9.649/1.60% decreased to 7.299/1.12% vs. G/M, 9.789/1.83% decreased to 7.539/1.27%; p/0.192).
3.2. Glycemic parameters on a subset of patients at selected study sites As previously described, self-monitored BG profiles were collected, and standardized test meals were administered to 120 patients. For self-monitored BG, Mix25/M patients reported significantly lower morning and evening 2-h ppBG values, lower BG at bedtime, and smaller 2-h ppBG excursions following the morning meal (Fig. 1). The daily average BG was significantly lower for Mix25/M patients at endpoint (7.969/1.78 mmol/l vs. 9.169/2.31 mmol/l; p /0.002). The percentage of patients achieving predefined BG targets for fasting and pre-meal BG (5/7 mmol/l) was not significantly different with Mix25/M (41% and 45%) and G/M (32% and 30%). However, the percentage of patients achieving the postprandial BG target ( 5/10 mmol/l) was significantly greater with Mix25/M compared with G/M (80% vs. 48%; p B/0.001).
Table 2 Select clinical characteristics of patients by treatment group at baseline and endpoint Variable
n
A1C (%) */all randomized patients A1C (%) */all test meal patients FPG (mmol/l) Overall hypoglycemia rate (episodes/patient/30 days) Nocturnal hypoglycemia incidence (% of patients) Nocturnal hypoglycemia rate (episodes/patient/30 days) Severe hypoglycemia incidence (% of patients) Severe hypoglycemia rate (episodes/patient/30 days) Body weight (kg)
589 9.179/1.50/7.299/1.00 118 9.649/1.60/7.299/1.12 117 13.309/3.79/8.679/3.36 597 0.089/0.59/0.319/1.07 597 1.4/1.0 597 0.039/0.23/0.019/0.11 597 0.3/1.0 597 0.019/0.09/0.019/0.11 597 83.09/15.2/84.09/15.1
Mean9/S.D. values at baseline (B)/endpoint (E).
Mix25/M B/E
G/M B/E
Endpoint p -value
9.279/1.55/7.339/1.14 9.789/1.83/7.539/1.27 12.999/3.78/9.439/3.39 0.079/0.57/0.489/1.17 0.0/5.0 0.009/0.00/0.089/0.40 0.0/1.3 0.009/0.00/0.029/0.15 81.79/15.7/82.29/15.4
0.661 0.192 0.173 0.067 B/0.01 0.007 1.000 0.755 0.330
192
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
Fig. 1. Mean (9/S.E.M.) eight-point blood glucose (BG) profiles for the test-meal subset (n/120).
For patients completing the on-site test meals, FPG values were not different between treatments at endpoint. The 1-h ppPG and 2-h ppPG at endpoint were significantly lower with Mix25/M (p/0.019, p B/0.001), as was the 2-h ppPG excursion (p B/0.001) (Fig. 2). The decrease in 2-
h ppPG from baseline to endpoint was significantly greater for Mix25/M, compared with G/ M (/6.899/4.69 mmol/l vs. /3.839/4.72 mmol/l; p/0.007), as was the 2-h ppPG excursion (/2.279/3.87 mmol/l vs. /0.229/1.99 mmol/l; p/0.009).
Fig. 2. Mean (9/S.E.M.) fasting, 1-h and 2-h postprandial plasma glucose (FPG, 1-h and 2-h ppPG) in the test-meal subset (n/120).
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
193
3.3. Hypoglycemia
3.5. Treatment-emergent adverse events
There were 352 reported episodes of hypoglycemia in the Mix25/M group (28 nocturnal, 19 severe) and 486 reported episodes in the G/M group (37 nocturnal, 12 severe). Of all hypoglycemic episodes 78% were confirmed to be B/3.5 mmol/l by self-monitored BG, while 14% were ]/ 3.5 mmol/l, and in 8% of episodes no BG was measured. At endpoint patients treated with Mix25/M reported a lower incidence and rate of nocturnal hypoglycemia than patients receiving G/M (incidence: 1% vs. 5%, p B/0.01; rate: 0.019/ 0.11 episodes/patient/30 days vs. 0.089/0.40 episodes/patient/30 days, p /0.007) and a trend in reduction in the overall rate of hypoglycemia (p/ 0.067), but not in the incidence or rate of severe hypoglycemia (p/1.000 and 0.755) (Table 2). There was no statistically significant difference between treatments in the incidence (p /0.087) or the rate (p /0.123) of daytime hypoglycemia.
Of the 597 randomized patients, 232 (39%) reported a least one treatment-emergent adverse event. There was no significant difference between treatments in the overall incidence of treatmentemergent events (Mix25/M, 106 (36%) patients; G/M, 126 (42%) patients). The most frequently reported adverse events were diarrhea (6%), headache (5%), abdominal pain (5%), rhinitis (4%), and flu syndrome (4%). There was no significant difference between treatments in any treatmentemergent adverse event. There were seven serious adverse events in the Mix25/M treatment group, five serious adverse events in the G/M treatment group, and none were related to the study drugs.
3.4. Patient satisfaction Responses to the treatment satisfaction questionnaire revealed the majority of patients in both treatment groups were satisfied with their treatment; more patients (97% vs. 92%) receiving G/ M reported a willingness to continue this treatment at study endpoint compared with Mix25/M (p/0.016). Patients on Mix25/M reported significantly fewer symptoms of hyperglycemia (e.g. increased energy level, p/0.002; greater wellbeing, p /0.003; lower thirst, p/0.003; fewer trips to the bathroom during the night, p B/ 0.001) than patients on G/M. Of the Mix25/M patients who responded to the questionnaire, the majority used the HumaPen Ergo device (HumaPen Ergo, 257 [92%]; Humalog Mix25 Pen, 24 [9%]). The majority of patients (99%) indicated they were satisfied with the devices (very satisfied, 232 patients [83%]; somewhat satisfied, 45 patients [16%]). A similar percentage of patients (98%) also found the devices easy to use overall (very easy, 222 patients [79%]; somewhat easy, 54 patients [19%]). When asked if they would continue to use the same device, the majority of patients (93%) indicated they would.
4. Discussion In patients failing on single oral agent therapy, 16 weeks of treatment with Humalog† Mix25TM plus metformin resulted in improved postprandial blood glucose, lower incidence and rate of nocturnal hypoglycemia, and reduced symptoms of hyperglycemia compared to patients treated with two oral agents. The two treatment groups experienced similar improvements in overall glycemic control, as measured by A1C. The improvement in postprandial blood glucose in the Mix25/M group is likely a direct effect of the rapid-acting insulin component of Mix25. The lack of change in the postprandial blood glucose excursion from baseline to endpoint in the dual oral agent therapy group indicates the lack of efficacy of this regimen on postprandial blood glucose control. This may be important in light of increasing evidence for the possible clinical consequences of postprandial hyperglycemia. Insulin lispro has been shown to result in lower postprandial BG, less nocturnal hypoglycemia, and greater patient satisfaction than regular human insulin in patients failing oral agent therapy [12]. Bastyr et al. [13] compared the effects of combination therapies (insulin lispro plus glyburide, metformin plus glyburide, or NPH plus glyburide) on fasting and postprandial BG and the impact of these measures on overall glycemic
194
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
control in patients inadequately controlled on a sulfonylurea. The mean FBG was significantly lower with NPH plus glyburide and the mean 2-h ppBG was lower with insulin lispro plus glyburide compared with the other therapies, while the overall rates of hypoglycemia were similar for all three treatment groups. Insulin lispro plus glyburide resulted in significantly lower A1C at endpoint compared to the other therapy groups. The authors concluded that when insulin lispro was used in combination therapy to focus on postprandial BG, there was a greater impact on overall metabolic control [13]. Regardless of the potential independent effect of postprandial BG on the development of complications, in patients who achieve their pre-meal BG targets, but whose overall glycemic control as evidenced by A1C is unsatisfactory, therapy to minimize postprandial BG excursions may be beneficial [6]. This becomes increasingly important in light of the fact that many patients newly diagnosed with type 2 diabetes already have evidence of the long-term complications of diabetes [14,15]. Introducing a rapid-acting insulin before two oral agents have failed would provide better postprandial BG control and possibly slow the progression of long-term complications associated with diabetes. Although A1C values did not differ between the treatment groups in the present study, both groups realized significant improvements in A1C levels (approximately 2%). Had more time (/4 months) been allotted to optimize insulin dose, a greater number of patients using insulin may have reached target fasting and pre-meal BG levels, in addition to reaching postprandial BG targets. In the smaller subset of patients who performed self-BG monitoring, Mix25/M provided a lower BG profile with less glucose fluctuations than G/M. A longer term study may be required to demonstrate a meaningful reduction of A1C with Mix25/M treatment.
5. Conclusion In patients with type 2 diabetes experiencing inadequate glycemic control on either metformin
or a sulfonylurea, treatment with Mix25/M provided similar overall glycemic control while lowering ppPG, lowering the risk of nocturnal hypoglycemia, and lessening symptoms of hyperglycemia compared to treatment with G/M. Twice daily Humalog Mix25 plus metformin was well accepted by patients and can be considered a valuable treatment option for patients failing single oral agent therapy.
Acknowledgements Appreciation is expressed to Peggy Campbell for editorial assistance with the manuscript. This work was sponsored by Eli Lilly and Company.
References [1] R.A. DeFronzo, A.M. Goodman, and the Multicenter Metformin Study Group, Efficacy of metformin in patients with non-insulin-dependent diabetes mellitus, N. Engl. J. Med. 333 (1995) 541 /549. [2] European Diabetes Policy Group, A desktop guide to type 2 diabetes mellitus, Ann. Intern. Med. 16 (1999) 716 /730. [3] The Diabetes Control and Complications Trial Research Group, The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus, N. Engl. J. Med. 329 (1993) 977 /986. [4] Y. Ohkubo, H. Kishikawa, E. Araki, T. Miyata, S. Isami, S. Motoyoshi, Y. Kojima, N. Furoyoshi, M. Shichiri, Intensive insulin therapy prevents the progression of diabetic microvascular complications in Japanese patients with non-insulin-dependent diabetes mellitus. A randomized prospective 6-year study, Diab. Res. Clin. Pract. 28 (1995) 103 /117. [5] UK Prospective Diabetes Study Group, UK Prospective Diabetes Study 33: Intensive blood /glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes, Lancet 352 (1998) 837 /853. [6] American Diabetes Association, Postprandial blood glucose, Diabetes Care 24 (2001) 775 /778. [7] J.H. Anderson, Jr, R.L. Brunelle, V.A. Koivisto, A. Pfutzner, M.E. Trautmann, L. Vignati, R. DiMarchi, Reduction of postprandial hyperglycemia and frequency of hypoglycemia in IDDM patients on insulin-analog treatment. Multicenter Insulin Lispro Study Group, Diabetes 46 (1997) 265 /270. [8] J.H. Anderson Jr, R.L. Brunelle, P. Keohane, V.A. Koivisto, M.E. Trautmann, L. Vignati, R. Di Marchi,
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195 Mealtime treatment with insulin analog improves postprandial hyperglycemia and hypoglycemia in patients with non-insulin-dependent diabetes mellitus. Multicenter Insulin Lispro Study Group, Arch. Intern. Med. 157 (2000) 1249 /1255. [9] P. Roach, L. Yue, V. Arora, Improved postprandial glycemic control during treatment with Humalog Mix25, a novel protamine-based insulin lispro formulation, Diabetes Care 22 (1999) 1258 /1261. [10] P. Roach, M. Trautmann, V. Arora, B. Sun, J.H. Anderson, Jr, Improved postprandial blood glucose control and reduced nocturnal hypoglycemia during treatment with two novel insulin lispro-protamine formulations, Insulin Lispro Mix25 and Insulin Lispro Mix50, Clin. Ther. 21 (1999) 523 /534. [11] J.C. Pickup, G. Williams (Eds.), Classification and diagnosis of diabetes mellitus and impaired glucose tolerance, in: Textbook of Diabetes, vol. 1, Blackwell Scientific Publications, Oxford, UK, 1991, p. 39.
195
[12] S.A. Ross, B. Zinman, R.V. Campos, T. Strack, Canadian Lispro Study Group, A comparative study of insulin lispro and human regular insulin in patients with type 2 diabetes mellitus and secondary failure of oral hypoglycemic agents, Clin. Invest. Med. 24 (2001) 292 /298. [13] E.J. Bastyr, III, C.A. Stuart, R.G. Brodows, S. Schwartz, C.J. Graf, A. Zagar, K.E. Robertson, Therapy focused on lowering postprandial glucose, not fasting glucose, may be superior for lowering HbA1c. IOEZ Study Group, Diabetes Care 23 (2000) 1236 /1241. [14] I. Kowalska, J. Prokop, H. Bachorzewska-Gajewska, et al., Disturbances of glucose metabolism in men referred for coronary arteriography: postload glycemia as a predictor of coronary atheriosclerosis, Diabetes Care 24 (2001) 897 / 901. [15] T. Temelkova-Kurktschiev, E. Henkel, F. Schaper, et al., Prevalence and atherosclerosis risk in different types of non-diabetic hyperglycemia. Is mild hyperglycemia an underestimated evil?, Exp. Clin. Endocrinol. Diabetes 108 (2000) 93 /99.
Therapy after single oral agent failure: adding a second oral agent or an insulin mixture? James K. Malone a,*, Scott D. Beattie a, Barbara N. Campaigne a, Patricia A. Johnson a, Andrew S. Howard a, Zvonko Milicevic b a
Lilly Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN 46285, USA b Eli Lilly and Company, Regional Operations GES, M.B.H., Vienna, Austria Received 2 May 2002; received in revised form 19 June 2003; accepted 4 August 2003
Abstract Aim: to compare the glycemic response to an insulin lispro mixture (25% insulin lispro and 75% NPL) twice daily plus metformin (Mix25/M) with glibenclamide plus metformin (G/M), in patients with type 2 diabetes inadequately controlled with a single oral agent. Methods: 597 patients treated in a randomized, open-label, 16-week parallel study. Variables evaluated: hemoglobin A1C (A1C), patient symptoms, hypoglycemia rate (episodes/patient/30 days), and incidence (% patients experiencing ]/1 episode). For a subset of patients (N/120), fasting, 1-h, and 2-h postprandial plasma glucose (FPG, 1-h ppPG, 2-h ppPG) in response to a standardized test meal (STM) and self-monitored blood glucose (BG) profiles were measured. Results: improved A1C at endpoint for both groups, and A1C changes from baseline to endpoint were not significantly different between treatments (Mix25/M, /1.879/1.35% vs. G/M, / 1.989/1.28%; p /0.288). Among patients completing STM; endpoint 2-h ppPG was significantly lower with Mix25/M (9.059/3.32 mmol/l vs. 12.319/3.65 mmol/l; p B/0.001), as was 2-h ppPG excursion (2-h ppPGex)(0.389/3.23 mmol/l vs. 2.889/1.98 mmol/l; p B/0.001). Percentage of patients achieving postprandial BG targets (B/10 mmol/l) at endpoint was significantly greater with Mix25/M (80% vs. 48%; p B/0.001). Although, overall hypoglycemia rates were similar, percentage of patients experiencing and rate of nocturnal hypoglycemia was less with Mix25/M (1% vs. 5%; p B/0.01, and 0.01 vs. 0.08 episodes/pt/30 d; p/0.007). Patients reported less polyuria with Mix25/M (p B/0.001). Conclusion: in patients with type 2 diabetes failing on metformin or a sulfonylurea, Mix25/M provided similar overall glycemic control, lower ppPG, reduced nocturnal hypoglycemia, and fewer hyperglycemic symptoms compared to G/M. # 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: Type 2 diabetes; Oral agents; Insulin mixtures
1. Introduction
* Corresponding author. Tel.: /1-317-277-6472; fax: /1317-277-5458. E-mail address: [email protected] (J.K. Malone).
The management of type 2 diabetes begins with education, diet, physical activity, and the attempt to reduce body weight in overweight patients. When these measures fail to achieve adequate
0168-8227/03/$ - see front matter # 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.diabres.2003.08.003
188
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
glycemic control, oral antidiabetic agents are typically used. However, patients with type 2 diabetes treated solely with single oral agent therapy */for example, second-generation sulfonylurea or metformin */often find their glycemic control is inadequately maintained even at maximum dosages. An effective approach to sulfonylurea failure is to add metformin [1]; the addition of insulin therapy is generally accepted as a treatment option only after acceptable glycemic control cannot be achieved or maintained with multiple oral agents. However, sulfonylurea plus metformin therapy does not address postprandial hyperglycemia, an early and prominent metabolic defect in type 2 diabetes. According to the European Diabetes Policy Group, the criteria for good metabolic control are hemoglobin A1C B/ 6.5% and 2-h postprandial blood glucose less than 7.5 mmol/l (135mg/dl) [2]. A major goal of current diabetes management is to prevent or slow the progression of diabetes-related complications. Clinical studies in patients with type 1 or type 2 diabetes have demonstrated that the development of long-term complications of diabetes and the level of metabolic control are closely correlated [3 /5]. However, the association of postprandial blood glucose (BG) control with diabetes complications remains controversial [6]. Insulin lispro, a rapid-acting insulin analog, has been shown to result in lower postprandial BG and a lower risk of nocturnal hypoglycemia in comparison to regular human insulin [7,8]. A pre-mixed insulin containing 25% insulin lispro and 75% insulin lispro protamine suspension (Humalog† Mix25TM, Humalog† Mix75/25TM in the US) (Mix25) was approved for clinical use in 1998 and is now available in many countries. Compared to human insulin mixtures, twice daily administration of insulin lispro mixtures results in improved postprandial blood glucose, similar overall glycemic control, less risk of nocturnal hypoglycemia, and the convenience of dosing immediately before meals [9,10]. Introducing a rapid-acting insulin mixture in combination with metformin after a single oral agent fails to maintain adequate glycemic control could improve postprandial blood glucose levels, and improve overall glycemic control with a low
risk of hypoglycemia. In the current study, patients with type 2 diabetes who were inadequately controlled on a sulfonylurea or metformin were transferred to Mix25 twice daily plus metformin or glibenclamide plus metformin, and the effects of these two regimens on glycemic control were compared.
2. Materials and methods 2.1. Patient population This study included 597 patients with type 2 diabetes, as defined by the World Health Organization [11], ages 30 /75 years, with body mass index (BMI) B/40 kg/m2. Each patient’s diabetes was inadequately controlled, as determined by a hemoglobin A1C /125% of the upper limit of normal by the local laboratory within 4 weeks prior to or at study entry. For at least 3 months prior to entering the study, patients were required to be using a single oral-antidiabetic agent, specifically either metformin or a second-generation sulfonylurea (e.g., glibenclamide, glipizide, gliclazide, glimepiride) at a maximally clinically effective dose within the last 30 days or more as determined by the investigator. The study was approved by the ethical review boards of the participating centers. Patients were required to complete and sign an informed consent document to participate in the study, and the study was conducted in accordance with the Declaration of Helsinki and good clinical practice guidelines. 2.2. Study design This was a randomized, open-label, two-arm, parallel prospective study in patients with type 2 diabetes and inadequate glycemic control using either a second-generation sulfonylurea or metformin, and was conducted in 56 centers in 14 countries. After a 2-week lead-in period, during which patients were treated with either glibenclamide or metformin (depending upon pre-study oral agent therapy), patients were randomly assigned to one of the following two treatment groups: twice daily Mix25 plus metformin
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
(Mix25/M) or twice daily glibenclamide plus metformin (G/M). 2.3. Lead-in period The lead-in period consisted of 2 (9/1) weeks between visit 1 (study entry) and visit 2 (the baseline visit and randomization). At visit 1, after giving informed consent, the patient’s height, weight, history, and pre-existing conditions were obtained, and a physical examination was performed. Patients were given diaries for recording hypoglycemic episodes and study drug dosages. Patients who used a sulfonylurea prior to the study were given glibenclamide during the lead-in period at a dose equivalent to their previous sulfonylurea dosage. Patients using metformin prior to the study continued with the same dose of metformin during the lead-in period. 2.4. Treatment period The treatment period began at visit 2 (week 0) and ended at visit 6 (week 169/1). This allowed for 4 weeks of dose adjustment and 12 weeks for stabilization of glycemic control. At visit 2 patients were randomized to receive either Mix25/M or G/M for 16 weeks. Patients assigned to dual oral agent therapy added the second oral agent (G or M) to their existing oral agent. For the group randomized to receive Mix25/ M, the investigators titrated insulin doses to achieve BG targets for fasting and pre-meal BG B/7 mmol/l (126 mg/dl) and 2-h postprandial BG B/10 mmol/l (180 mg/dl), without increasing the frequency of hypoglycemia. Patients taking metformin during the lead-in period maintained their metformin dose and added Mix25 twice daily, immediately before the morning and evening meals. The metformin dose was titrated so that by visit 3 (4 weeks after the baseline visit 2) patients were taking at least 1500 mg per day (mg/day) but not more than 2550 mg/day, in two to three divided doses with meals. The dose of metformin was held constant after Visit 3. Patients taking glibenclamide during the lead-in period were switched to metformin and Mix25 was added twice-daily.
189
In the dual oral agent group, the metformin dose was titrated in the same way, while the glibenclamide dose was adjusted to reach the fasting and pre-meal blood glucose target throughout the remainder of the study. All patients returned to the study center once every 4 (9/1) weeks for routine assessments. At each visit, completed diaries were collected and new ones were distributed. Body weight, adverse events, episodes of hypoglycemia, study drug dose, and concomitant medication were also obtained. At visits 2 and 6, blood was drawn for measurement of A1C. The A1C assay was performed by ion-exchange high performance liquid chromatography (HPLC) on a Bio-Rad variant analyzer at a central laboratory (Covance, Indianapolis, IN; reference range, 4.3 /6.1%). The Variant system is certified by the National Glycohemoglobin Standardization Program (NGSP) to be traceable to the Diabetes Control and Complications Trial (DCCT) [3] reference method and values. At visits 2 and 6, all patients were asked to complete a treatment satisfaction questionnaire, and at visit 6 patients receiving Mix25/M were asked to complete an insulin delivery device questionnaire. All patients at sites selected to represent the specific regions of the world (sites in Romania, Australia, Canada, and Brazil) were asked to perform three 8-point BG profiles (premeal and 2-h postprandial for morning, mid-day and evening meals, as well as at bedtime and 03:00 h) on separate days during the 2-week period prior to visit 2 and visit 6. These patients also participated in an on-site standardized test meal (STM) on the mornings of visits 2 and 6. Upon arriving at the study center, the patients had fasting plasma glucose (FPG) measured, consumed the STM, and then had 1-h and 2-h postprandial plasma glucose (1 and 2-h ppPG) levels measured. The composition of the STM was approximately 369 kcal (57% carbohydrate, 11% protein and 32% fat) and consisted of Glucerna† (Ross Products Division, Abbott Laboratories, Abbott Park, IL) (237 ml) and a fruit cup (234 g). Patients arrived at the study site after an 8-h overnight fast having omitted their morning diabetes therapy. The study drugs were administered five minutes before the STM.
190
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
2.5. Study drugs Humalog† Mix25TM (Eli Lilly and Company, Indianapolis, IN) was provided in 3-ml cartridges, or 3-ml pre-filled pens*/100 units/ml. The oralantidiabetic agents were glibenclamide in 5 mg tablets (bioequivalent to 3.5 mg micronized form) and metformin hydrochloride in 500 and 850 mg tablets. 2.6. Statistical methods Randomization to treatment was performed, stratified by investigative site and by prestudy oral therapy (metformin or sulfonylurea), such that an approximately equal number of patients were assigned to the two treatments within each stratum. The sample size for the study was chosen to provide adequate power to determine treatment differences in A1C improvement (change from baseline to endpoint) within each prestudy oral therapy stratum. For this reason, enrollment was designed to randomize at least 240 patients from each prestudy therapy. This was planned to provide 63/80% power for a 0.4% treatment difference within each stratum and 78% power for a 0.3% difference overall. Statistical analyses were performed with SAS version 6.09 using two-sided tests at a 0.05 level of significance. Primary analyses utilized the endpoint data from each patient (last observation carried forward), and comparisons were made according to treatment received. Hemoglobin A1C and body weight were analyzed by Type III sums of squares from an analysis of variance (ANOVA) model with effects for treatment, country, and treatment /country interaction. Patient data from countries with fewer than 10 patients were pooled with geographic neighbors. Glucose values from the test meals and from self-monitored BG profiles were analyzed by a similar ANOVA model using investigator instead of country. Hypoglycemia was defined as any time a patient felt, or another person observed, that the patient was experiencing a sign/symptom which they associated with hypoglycemia or the patient recorded a BG B/3.5 mmol/l. Hypoglycemia data were collected via patient home BG monitoring
and patient records. Nocturnal hypoglycemia was defined as any time one of these episodes occurred after bedtime and before awakening. Episodes where the patient was unable to treat himself/ herself or that were associated with a BG 5/2.0 mmol/l were defined as severe. Incidence (percentage of patients experiencing) of treatment-emergent adverse events and episodes of hypoglycemia were analyzed using Fisher’s exact test. Rates of hypoglycemia (episodes/patient/30 days) were analyzed using a ranked ANOVA model with treatment and country effects. Answers to treatment satisfaction questions, measured on a five-point Likert scale, were analyzed using a cumulativelogit model that included treatment effect only. All continuous variables are reported as unadjusted mean (9/S.D.).
3. Results Of the 597 randomized patients, 296 were assigned to treatment with Mix25/M, and 301 were assigned to G/M. Prior to the study, 381 patients had used sulfonylurea therapy and 216 used metformin. A total of 543 patients completed the study; 25 patients dropped out while receiving Mix25/M for the following reasons: 1 */adverse event, 1*/death (not related to study drug), 1 */ lost to follow-up, 1 */moved away, 7*/patient decision, 3*/entry criteria not met, 2*/sponsor decision, 3 */physician decision, 3 */protocol violation, and 3*/inability to tolerate metformin. Twenty-nine patients dropped out while receiving G/M for the following reasons: 2 */adverse event, 1 */lack of efficacy, 2*/lost to follow-up, 1*/patient decision, 3 */entry criteria not met, 4*/physician decision, 6 */protocol violation, 10*/inability to tolerate metformin. At baseline, the patients’ mean age was 58 (9/9) years and their mean BMI was 29.7 (9/4.5) kg/m2 (Table 1). Patients in both treatment groups experienced a mild increase in body weight (Mix25/M, 0.89/3.4 kg; G/M, 0.39/2.8 kg) during the study, although there was no significant difference between treatments (Table 2). The mean insulin doses achieved at endpoint were 0.19 U/kg in the morning and 0.14 U/kg in
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
Table 1 Patients’ baseline characteristics by treatment group (mean9/ S.D.) Variable
Mix25/M (n/296)
G/M (n/301)
Age (years) 589/8.8 599/9.3 Males 169 (57%) 146 (49%) Females 127 (43%) 155 (51%) Total 296 301 Caucasian 263 (88.9%) 268 (89.0%) African 2 (0.7%) 3 (1.0%) Hispanic 22 (7.4%) 18 (6.0%) Other 9 (3.0%) 12 (4.0%) Weight (kg) 83.09/15.2 81.79/15.7 BMI (kg/m2) 29.89/4.4 29.69/4.5 11.509/3.08 FBG (mmol/l) 11.619/2.97 A1C (%) 9.179/1.50 9.279/1.55 Diabetes duration (years) 8.09/5.8 7.49/5.4
pvalue 0.124 0.036 0.793
0.292 0.589 0.759 0.181 0.160
the evening. Average total daily dose of metformin for the Mix25/M group was 1813 mg and for the G/M group was 1968 mg. The mean glibenclamide dose for the G/M group was 14.2 mg. 3.1. Hemoglobin A1C Both treatment groups experienced significant improvement in A1C as measured by change from baseline to endpoint (p B/0.001). However, there was no significant difference between treatments in A1C improvement (Mix25/M, /1.879/1.35% vs. G/M, /1.989/1.28%; p /0.288). Additionally, there was no significant difference between treatments in endpoint A1C (Mix25/M, 7.299/
191
1.00% vs. G/M, 7.339/1.14%; p/0.661) (Table 2) or in the percentage of patients achieving A1C values B/7.0% (Mix25/M, 40% vs. G/M, 41%), or the percentage of patients with A1C values B/ 6.5% (Mix25/M, 18% vs. G/M, 19%). In the subgroup of patients completing the test meals, the A1C levels were similar to the overall cohort at baseline and endpoint (Mix25/M, 9.649/1.60% decreased to 7.299/1.12% vs. G/M, 9.789/1.83% decreased to 7.539/1.27%; p/0.192).
3.2. Glycemic parameters on a subset of patients at selected study sites As previously described, self-monitored BG profiles were collected, and standardized test meals were administered to 120 patients. For self-monitored BG, Mix25/M patients reported significantly lower morning and evening 2-h ppBG values, lower BG at bedtime, and smaller 2-h ppBG excursions following the morning meal (Fig. 1). The daily average BG was significantly lower for Mix25/M patients at endpoint (7.969/1.78 mmol/l vs. 9.169/2.31 mmol/l; p /0.002). The percentage of patients achieving predefined BG targets for fasting and pre-meal BG (5/7 mmol/l) was not significantly different with Mix25/M (41% and 45%) and G/M (32% and 30%). However, the percentage of patients achieving the postprandial BG target ( 5/10 mmol/l) was significantly greater with Mix25/M compared with G/M (80% vs. 48%; p B/0.001).
Table 2 Select clinical characteristics of patients by treatment group at baseline and endpoint Variable
n
A1C (%) */all randomized patients A1C (%) */all test meal patients FPG (mmol/l) Overall hypoglycemia rate (episodes/patient/30 days) Nocturnal hypoglycemia incidence (% of patients) Nocturnal hypoglycemia rate (episodes/patient/30 days) Severe hypoglycemia incidence (% of patients) Severe hypoglycemia rate (episodes/patient/30 days) Body weight (kg)
589 9.179/1.50/7.299/1.00 118 9.649/1.60/7.299/1.12 117 13.309/3.79/8.679/3.36 597 0.089/0.59/0.319/1.07 597 1.4/1.0 597 0.039/0.23/0.019/0.11 597 0.3/1.0 597 0.019/0.09/0.019/0.11 597 83.09/15.2/84.09/15.1
Mean9/S.D. values at baseline (B)/endpoint (E).
Mix25/M B/E
G/M B/E
Endpoint p -value
9.279/1.55/7.339/1.14 9.789/1.83/7.539/1.27 12.999/3.78/9.439/3.39 0.079/0.57/0.489/1.17 0.0/5.0 0.009/0.00/0.089/0.40 0.0/1.3 0.009/0.00/0.029/0.15 81.79/15.7/82.29/15.4
0.661 0.192 0.173 0.067 B/0.01 0.007 1.000 0.755 0.330
192
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
Fig. 1. Mean (9/S.E.M.) eight-point blood glucose (BG) profiles for the test-meal subset (n/120).
For patients completing the on-site test meals, FPG values were not different between treatments at endpoint. The 1-h ppPG and 2-h ppPG at endpoint were significantly lower with Mix25/M (p/0.019, p B/0.001), as was the 2-h ppPG excursion (p B/0.001) (Fig. 2). The decrease in 2-
h ppPG from baseline to endpoint was significantly greater for Mix25/M, compared with G/ M (/6.899/4.69 mmol/l vs. /3.839/4.72 mmol/l; p/0.007), as was the 2-h ppPG excursion (/2.279/3.87 mmol/l vs. /0.229/1.99 mmol/l; p/0.009).
Fig. 2. Mean (9/S.E.M.) fasting, 1-h and 2-h postprandial plasma glucose (FPG, 1-h and 2-h ppPG) in the test-meal subset (n/120).
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
193
3.3. Hypoglycemia
3.5. Treatment-emergent adverse events
There were 352 reported episodes of hypoglycemia in the Mix25/M group (28 nocturnal, 19 severe) and 486 reported episodes in the G/M group (37 nocturnal, 12 severe). Of all hypoglycemic episodes 78% were confirmed to be B/3.5 mmol/l by self-monitored BG, while 14% were ]/ 3.5 mmol/l, and in 8% of episodes no BG was measured. At endpoint patients treated with Mix25/M reported a lower incidence and rate of nocturnal hypoglycemia than patients receiving G/M (incidence: 1% vs. 5%, p B/0.01; rate: 0.019/ 0.11 episodes/patient/30 days vs. 0.089/0.40 episodes/patient/30 days, p /0.007) and a trend in reduction in the overall rate of hypoglycemia (p/ 0.067), but not in the incidence or rate of severe hypoglycemia (p/1.000 and 0.755) (Table 2). There was no statistically significant difference between treatments in the incidence (p /0.087) or the rate (p /0.123) of daytime hypoglycemia.
Of the 597 randomized patients, 232 (39%) reported a least one treatment-emergent adverse event. There was no significant difference between treatments in the overall incidence of treatmentemergent events (Mix25/M, 106 (36%) patients; G/M, 126 (42%) patients). The most frequently reported adverse events were diarrhea (6%), headache (5%), abdominal pain (5%), rhinitis (4%), and flu syndrome (4%). There was no significant difference between treatments in any treatmentemergent adverse event. There were seven serious adverse events in the Mix25/M treatment group, five serious adverse events in the G/M treatment group, and none were related to the study drugs.
3.4. Patient satisfaction Responses to the treatment satisfaction questionnaire revealed the majority of patients in both treatment groups were satisfied with their treatment; more patients (97% vs. 92%) receiving G/ M reported a willingness to continue this treatment at study endpoint compared with Mix25/M (p/0.016). Patients on Mix25/M reported significantly fewer symptoms of hyperglycemia (e.g. increased energy level, p/0.002; greater wellbeing, p /0.003; lower thirst, p/0.003; fewer trips to the bathroom during the night, p B/ 0.001) than patients on G/M. Of the Mix25/M patients who responded to the questionnaire, the majority used the HumaPen Ergo device (HumaPen Ergo, 257 [92%]; Humalog Mix25 Pen, 24 [9%]). The majority of patients (99%) indicated they were satisfied with the devices (very satisfied, 232 patients [83%]; somewhat satisfied, 45 patients [16%]). A similar percentage of patients (98%) also found the devices easy to use overall (very easy, 222 patients [79%]; somewhat easy, 54 patients [19%]). When asked if they would continue to use the same device, the majority of patients (93%) indicated they would.
4. Discussion In patients failing on single oral agent therapy, 16 weeks of treatment with Humalog† Mix25TM plus metformin resulted in improved postprandial blood glucose, lower incidence and rate of nocturnal hypoglycemia, and reduced symptoms of hyperglycemia compared to patients treated with two oral agents. The two treatment groups experienced similar improvements in overall glycemic control, as measured by A1C. The improvement in postprandial blood glucose in the Mix25/M group is likely a direct effect of the rapid-acting insulin component of Mix25. The lack of change in the postprandial blood glucose excursion from baseline to endpoint in the dual oral agent therapy group indicates the lack of efficacy of this regimen on postprandial blood glucose control. This may be important in light of increasing evidence for the possible clinical consequences of postprandial hyperglycemia. Insulin lispro has been shown to result in lower postprandial BG, less nocturnal hypoglycemia, and greater patient satisfaction than regular human insulin in patients failing oral agent therapy [12]. Bastyr et al. [13] compared the effects of combination therapies (insulin lispro plus glyburide, metformin plus glyburide, or NPH plus glyburide) on fasting and postprandial BG and the impact of these measures on overall glycemic
194
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195
control in patients inadequately controlled on a sulfonylurea. The mean FBG was significantly lower with NPH plus glyburide and the mean 2-h ppBG was lower with insulin lispro plus glyburide compared with the other therapies, while the overall rates of hypoglycemia were similar for all three treatment groups. Insulin lispro plus glyburide resulted in significantly lower A1C at endpoint compared to the other therapy groups. The authors concluded that when insulin lispro was used in combination therapy to focus on postprandial BG, there was a greater impact on overall metabolic control [13]. Regardless of the potential independent effect of postprandial BG on the development of complications, in patients who achieve their pre-meal BG targets, but whose overall glycemic control as evidenced by A1C is unsatisfactory, therapy to minimize postprandial BG excursions may be beneficial [6]. This becomes increasingly important in light of the fact that many patients newly diagnosed with type 2 diabetes already have evidence of the long-term complications of diabetes [14,15]. Introducing a rapid-acting insulin before two oral agents have failed would provide better postprandial BG control and possibly slow the progression of long-term complications associated with diabetes. Although A1C values did not differ between the treatment groups in the present study, both groups realized significant improvements in A1C levels (approximately 2%). Had more time (/4 months) been allotted to optimize insulin dose, a greater number of patients using insulin may have reached target fasting and pre-meal BG levels, in addition to reaching postprandial BG targets. In the smaller subset of patients who performed self-BG monitoring, Mix25/M provided a lower BG profile with less glucose fluctuations than G/M. A longer term study may be required to demonstrate a meaningful reduction of A1C with Mix25/M treatment.
5. Conclusion In patients with type 2 diabetes experiencing inadequate glycemic control on either metformin
or a sulfonylurea, treatment with Mix25/M provided similar overall glycemic control while lowering ppPG, lowering the risk of nocturnal hypoglycemia, and lessening symptoms of hyperglycemia compared to treatment with G/M. Twice daily Humalog Mix25 plus metformin was well accepted by patients and can be considered a valuable treatment option for patients failing single oral agent therapy.
Acknowledgements Appreciation is expressed to Peggy Campbell for editorial assistance with the manuscript. This work was sponsored by Eli Lilly and Company.
References [1] R.A. DeFronzo, A.M. Goodman, and the Multicenter Metformin Study Group, Efficacy of metformin in patients with non-insulin-dependent diabetes mellitus, N. Engl. J. Med. 333 (1995) 541 /549. [2] European Diabetes Policy Group, A desktop guide to type 2 diabetes mellitus, Ann. Intern. Med. 16 (1999) 716 /730. [3] The Diabetes Control and Complications Trial Research Group, The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus, N. Engl. J. Med. 329 (1993) 977 /986. [4] Y. Ohkubo, H. Kishikawa, E. Araki, T. Miyata, S. Isami, S. Motoyoshi, Y. Kojima, N. Furoyoshi, M. Shichiri, Intensive insulin therapy prevents the progression of diabetic microvascular complications in Japanese patients with non-insulin-dependent diabetes mellitus. A randomized prospective 6-year study, Diab. Res. Clin. Pract. 28 (1995) 103 /117. [5] UK Prospective Diabetes Study Group, UK Prospective Diabetes Study 33: Intensive blood /glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes, Lancet 352 (1998) 837 /853. [6] American Diabetes Association, Postprandial blood glucose, Diabetes Care 24 (2001) 775 /778. [7] J.H. Anderson, Jr, R.L. Brunelle, V.A. Koivisto, A. Pfutzner, M.E. Trautmann, L. Vignati, R. DiMarchi, Reduction of postprandial hyperglycemia and frequency of hypoglycemia in IDDM patients on insulin-analog treatment. Multicenter Insulin Lispro Study Group, Diabetes 46 (1997) 265 /270. [8] J.H. Anderson Jr, R.L. Brunelle, P. Keohane, V.A. Koivisto, M.E. Trautmann, L. Vignati, R. Di Marchi,
J.K. Malone et al. / Diabetes Research and Clinical Practice 62 (2003) 187 /195 Mealtime treatment with insulin analog improves postprandial hyperglycemia and hypoglycemia in patients with non-insulin-dependent diabetes mellitus. Multicenter Insulin Lispro Study Group, Arch. Intern. Med. 157 (2000) 1249 /1255. [9] P. Roach, L. Yue, V. Arora, Improved postprandial glycemic control during treatment with Humalog Mix25, a novel protamine-based insulin lispro formulation, Diabetes Care 22 (1999) 1258 /1261. [10] P. Roach, M. Trautmann, V. Arora, B. Sun, J.H. Anderson, Jr, Improved postprandial blood glucose control and reduced nocturnal hypoglycemia during treatment with two novel insulin lispro-protamine formulations, Insulin Lispro Mix25 and Insulin Lispro Mix50, Clin. Ther. 21 (1999) 523 /534. [11] J.C. Pickup, G. Williams (Eds.), Classification and diagnosis of diabetes mellitus and impaired glucose tolerance, in: Textbook of Diabetes, vol. 1, Blackwell Scientific Publications, Oxford, UK, 1991, p. 39.
195
[12] S.A. Ross, B. Zinman, R.V. Campos, T. Strack, Canadian Lispro Study Group, A comparative study of insulin lispro and human regular insulin in patients with type 2 diabetes mellitus and secondary failure of oral hypoglycemic agents, Clin. Invest. Med. 24 (2001) 292 /298. [13] E.J. Bastyr, III, C.A. Stuart, R.G. Brodows, S. Schwartz, C.J. Graf, A. Zagar, K.E. Robertson, Therapy focused on lowering postprandial glucose, not fasting glucose, may be superior for lowering HbA1c. IOEZ Study Group, Diabetes Care 23 (2000) 1236 /1241. [14] I. Kowalska, J. Prokop, H. Bachorzewska-Gajewska, et al., Disturbances of glucose metabolism in men referred for coronary arteriography: postload glycemia as a predictor of coronary atheriosclerosis, Diabetes Care 24 (2001) 897 / 901. [15] T. Temelkova-Kurktschiev, E. Henkel, F. Schaper, et al., Prevalence and atherosclerosis risk in different types of non-diabetic hyperglycemia. Is mild hyperglycemia an underestimated evil?, Exp. Clin. Endocrinol. Diabetes 108 (2000) 93 /99.