- Email: [email protected]
The rationale and design of the Glycemic Effects in Diabetes Mellitus
Journal of Diabetes and Its Complications 19 (2005) 74 – 79
The rationale and design of the Glycemic Effects in Diabetes Mellitus Carvedilol–Metoprolol Comparison in Hypertensives (GEMINI) trial George L. Bakrisa,*, David S.H. Bellb, Vivian Fonsecac, Richard Katholid, Janet McGille, Robert Phillipsf, Philip Rasking, Jackson T. Wright Jr.h, Malini Iyengari, Terry Holeslawi, Karen M. Andersoni, for the GEMINI Investigators a
Rush Hypertension Center, Rush Medical Center, Suite 470, 1700 W. Van Buren, Chicago, IL 60612, USA b University of Alabama at Birmingham, Birmingham, AL, USA c Tulane University, New Orleans, LA, USA d St. John’s Hospital, Springfield, IL, USA e Washington University School of Medicine, St. Louis, MO, USA f New York University, New York, NY, USA g University of Texas at Southwestern Medical Center, Dallas, TX, USA h University Hospitals of Cleveland, Cleveland, OH, USA i GlaxoSmithKline, Philadelphia, PA, USA Received 5 March 2004; received in revised form 6 July 2004; accepted 9 July 2004
Abstract h-blockers utilized in the Type 2 diabetic patient result in an even greater decrease in cardiac events than in the nondiabetic patient. Unfortunately, first-and second-generation h-blockers are associated with the worsening of insulin resistance, deterioration of glycemic control, peripheral vasoconstriction, potentially worsening peripheral vascular disease, and more frequent and severe hypoglycemia. The third-generation h-blockers have unique properties, including a1-blockade, and have been shown to lower insulin resistance, improve glycemic control, and vasodilate resistance arterioles. The Glycemic Effects in Diabetes Mellitus: Carvedilol–Metoprolol Comparison in Hypertensives (GEMINI) trial has been designed to compare a third-generation (cardevilol) with a second-generation h-blocker (metoprolol) in a cohort of participants with hypertension and Type 2 diabetes. The primary outcome measure of the study is change in the HbA1c. The study is powered to detect a difference in HbA1c of 0.3 units (%) between the groups. Secondary endpoints include changes in insulin resistance, fasting glucose, and the lipid profile. Differences in the side-effect profile (cold extremities, fatigue, impotence, and hypoglycemia) will also be assessed. The GEMINI trial, therefore, is the first large randomized trial to assess whether utilizing a thirdgeneration h-blocker yields a favorable metabolic profile in the patient with Type 2 diabetes and hypertension. D 2005 Elsevier Inc. All rights reserved. Keywords: Hypertension; Type 2 diabetes; h-blockers; Clinical trial; Carvedilol; Metoprolol
1. Introduction Type 2 diabetes is a cardiac disease equivalent—diabetic patients have a 20% chance of a cardiac event within 10 years
* Corresponding author. Tel.: +1 312 942 3140; fax: +1 312 942 8119. E-mail address: [email protected] (G.L. Bakris). 1056-8727/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jdiacomp.2004.07.003
(Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel, 2001). In patients with diabetes, h-blockade decreases mortality in post-MI patients and in patients with established coronary disease, even superimposed on current care standards of angiotensin converting enzyme inhibitors (ACEI) and fibrinolytics (Jona et al., 1996; Kjekshus et al., 1990). This may result from decreased myocardial workload, prevention and reversal of myocardial remodelling, prevention of
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
congestive heart failure, and antiatherogenic and antiinflammatory properties (Bell, 2003). The treatment of hypertension in diabetes, with the American Diabetes Association goal blood pressure of 130/80 mm Hg or less, often requires the use of several antihypertensive agents (Arauz-Pancheco, Parot, & Raskin, 2002; Bakris, 1999). After an ACEI or angiotensin receptor blocker (ARB), with their effects amplified by a thiazide diuretic, h-blockers should be the next choice rather than a calcium channel blocker (CCB) or vasodilator. Unfortunately, traditional h-blockers increase insulin resistance, lower HDL, and raise triglyceride and glucose levels (Bell, 1999). In nondiabetic patients, there is over a 20% increase in the development of diabetes with h-blockers but not with thiazide diuretics (Gress et al., 2000). In diabetic patients, increases in HbA1c of as much as 1% have been reported with a second-generation h-blocker (Janka, Holzgreve, & Nakor, 2000). The third-generation h-blocker, carvedilol, through its combined h- and a1adrenoceptor blockade, has been shown to lower insulin resistance and glucose and triglyceride levels and to increase HDL in both diabetic and nondiabetic subjects (Giugliano et al., 1997; Jacob et al., 1996). Peripheral vasoconstriction, also a potentially unacceptable side effect of first- and second-generation h-blockers in diabetic patients, does not occur because of carvedilol’s a1-blocking, vasodilating effect (Tomlinson, 1988). Because these encouraging preliminary data suggested potential advantages to using a third-generation h-blocker in diabetic hypertensive patients, the Glycemic Effects in Diabetes Mellitus: Carvedilol–Metoprolol Comparison in Hypertensives (GEMINI) trial is being performed.
75
in patients with controlled mild-to-moderate hypertension and Type 2 diabetes mellitus. The major secondary objectives are to determine whether carvedilol improves or maintains glycemic control and whether metoprolol worsens glycemic control, as measured by the change from baseline HbA1c. Additional secondary objectives are to assess effects on insulin sensitivity (HOMA analysis), fasting plasma glucose, serum insulin, triglycerides, total cholesterol, LDL cholesterol, HDL cholesterol, aldosterone, and body weight. The relationship between blood pressure and glycemic control will be examined, and the effects on urinary albumin excretion [albumin/creatinine ratio (ACR)] and the incidence of side effects (e.g., fatigue, cold extremities, and sexual dysfunction) are described. Safety parameters include the incidence of clinically defined symptomatic and asymptomatic hypoglycemia and the percentage of patients withdrawn from the trial. 2.2. Study population Participants recruited are males or females, 30 to 80 years of age, with Type 2 diabetes and a documented history of mild-to-moderate hypertension (systolic BP z140 to V179 mm Hg; diastolic BP z90 to V109 mm Hg). Patients must be taking an ACEI or ARB. Target enrollment is 1210 patients at approximately 250 sites. Specific inclusion and exclusion criteria are summarized in Table 1. 2.3. Study timelines and participant assessments The GEMINI trial includes six phases (Fig. 1): screening/ washout, baseline/randomization, uptitration, maintenance, downtitration, and follow-up. Maximum study length per participant is 35 weeks.
2. GEMINI trial design GEMINI is a randomized, double-blind, parallel-group, multicenter (United States) trial comparing the effects of carvedilol and metoprolol on glycemic control in hypertensive diabetic patients. Patients are titrated with blinded study medication to aggressive target blood pressures consistent with recently published recommendations (ArauzPancheco et al., 2002; The seventh report of the Joint National Committee on Prevention, Detection, and Treatment of High Blood Pressure, 2003) Following Institutional Review Board approval, all participants were consented according to the Declaration of Helsinki Recommendations Guiding Physicians in Biomedical Research Involving Human Subjects (1996). 2.1. Study objectives The primary objective of the GEMINI trial is to determine if the effect of carvedilol on glycemic control, assessed by differences in the change from baseline hemoglobin (Hb) A1c, is superior to the effect of metoprolol
2.3.1. Screening/washout A mandatory 4-week interval exists between screening and a prerandomization visit to verify the stability of glycemic control; HbA1c could not increase N0.5% prior to randomization. This interval comprises a 2-week screening interval and a 2-week washout period. During screening, after providing informed written consent, medical history, physical exam (including vital signs), electrocardiogram (ECG), and chest X-ray are done. Patients provide blood and urine samples for clinical chemistries to rule out any exclusionary disease and to assess whether C-peptide, HbA1c, and urinary ACR meet entry criteria. Once fulfilling the screening requirements, patients begin the 2-week washout phase. Patients on ACEI or ARB monotherapy are eligible for enrollment if they fulfill all other inclusion/ exclusion criteria. All patients are to remain on their ACEI or ARB, but additional antihypertensives are stopped. Patients are allowed to stay on other medications with antihypertensive properties only if prescribed for another indication (e.g., a diuretic for edema, an a-blocker for benign prostatic hypertrophy). Diabetes control must remain stable
76
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
Table 1 Inclusion and exclusion criteria to determine eligibility for enrollment into the GEMINI trial Inclusion criteria 1. z30 and V80 years of age. 2. A documented past history of mild to moderate hypertension (systolic BP z140 to V179 mm Hg; diastolic BP z90 to V109 mm Hg). 3. Taking an ACE inhibitor or an ARB either alone or as part of polypharmacy for control of blood pressure. 4. Stable antihypertensive regimen (no changes to medications and/or dosage) during the month prior to screening. 5. A documented history of Type II diabetes mellitus. Patients must be treating their diabetes with diet alone or with antidiabetic medications (may include stable insulina). 6. Stable antidiabetic regimen (no changes to diet, medication, dosage, or formulation) during the 3 months prior to screening. 7. Patients receiving antidiabetic medications must have an HbA1c of z6.5% and V8.5%; patients managing their diabetes by diet alone must have an HbA1c of z6.5% and V7.5%. 8. Body mass index of 22– 45 kg/m2. 9. Endogenous insulin production as manifested by a fasting C-peptide of z0.6 ng/ml at screening. 10. Evidence that diabetes is well controlled: HbA1c may not be increased by greater than 0.5 units (%) over the 4-week interval between screening and randomization. 11. Following washout of all antihypertensive medications, except ACE inhibitor or ARBs, the patient must have elevated blood pressure (systolic BP N130 to V179 mm Hg and/or diastolic BP N80 to V109 mm Hg). 12. If female of child-bearing potential, using adequate contraceptive measures. Exclusion criteria 1. Uncontrollable or symptomatic arrhythmias 2. Unstable angina 3. Sick sinus syndrome or second or third degree heart block (unless treated with a permanent, functioning pacemaker) 4. Currently receiving treatment for heart failure 5. Myocardial infarction or stroke within three months of study screening 6. Bradycardia (heart rate b50 bpm) 7. Chronic obstructive pulmonary (COPDa) disease requiring continuous use of inhaled or oral bronchodilator therapy, as well as COPD treated with intermittent or continuous corticosteroids (oral or inhaled) 8. Bronchial asthma or related bronchospastic conditions 9. New onset/newly diagnosed Type II diabetes mellitus (onset within the three months prior to screening) 10. Clinically significant renal or liver disease: serum creatinine N2.5 mg/dl; ALT, AST, total bilirubin, or alkaline phosphatase more than 2.5 times the upper limit of the laboratory reference range [Note: microalbuminuria, if the only clinical sign of renal disease does not fulfill this exclusion criterion; fully developed macroalbuminuria (albumin–creatinine N300 mg/g), as an indicator of more serious renal disease constitutes a condition for which the patient should be excluded from participation in the study] 11. Endocrine disorders (e.g., pheochromocytoma, active and untreated hypo- or hyperthyroidism) 12. Use of anorectic or any other diet drugs, or any dietary measures inconsistent with recommendations for patients with Type II diabetes 13. Taking a h-blocker for any reason within three months of screening (eyedrops containing h-blockers are permissible throughout the trial) 14. Chronic disorder that requires chronic or intermittent use of corticosteroids 15. Systemic disease, including cancer, with reduced (b12 months) life expectancy
Table 1 (continued) Exclusion criteria 16. Psychological illness/condition that would interfere with the patient’s ability to understand or complete the requirements of the study 17. Use of an investigational drug within 30 days of entry to the study or within a time period equivalent to five half-lives of the investigational drug a
Added via amendment to the original protocol.
for the initial 4 weeks of the study. Two days prior to the completion of washout, patients return (bprerandomization visitQ) for a fasting HbA1c determination to fulfill the stability requirement and to serve as the baseline value. 2.3.2. Baseline/randomization At washout conclusion, and upon determination that all eligibility criteria are met, the patient is randomized (baseline/randomization visit) to Dose Level 1 of blinded study medication (carvedilol 6.25 mg b.i.d. or metoprolol tartrate 50 mg b.i.d.). Blood and urine samples are collected for the baseline determination of all laboratory-related secondary endpoints and routine safety assessments. Patients also undergo an interim medical history, ECG, cardiopulmonary exam, and evaluation of vital signs and body weight. At this and each visit throughout the study, patients are questioned regarding concomitant medications and adverse events. Randomization is stratified by whether patients are taking ARBs and by whether thiazolidinediones (TZDs) are part of their antidiabetic regimen. 2.3.3. Uptitration Patients are titrated at 1- to 2-week intervals with blinded study medication until they achieve target blood pressure for at least 1 week (i.e., at two consecutive weekly visits, see Study medication and Fig. 1). Patients return weekly for assessment of vital signs, tolerability to study medication, and medication compliance. 2.3.4. Maintenance The Maintenance phase is of 5-month duration. Patients enter maintenance after achieving the target blood pressure for at least 1 week and may enter at study medication Dose Levels 1, 2, or 3. Patients remaining above the goal blood pressure following 1 week at study medication Dose Level 3 are allowed to enter Maintenance but are required to return at weekly intervals for additional therapy. Hydrochlorothiazide (HCTZ) 12.5 mg and a dihydropyridine CCB may be added sequentially as necessary to achieve target blood pressure. This dose of HCTZ is chosen to minimize the potential for any impact on metabolic balance, and this sequence (HCTZ first, followed, if needed, by CCB), to emulate current prescribing guidelines. During the visit at which it is determined that a patient will enter Maintenance, patients undergo an interim medical history, cardiopulmonary exam, assessment of vital signs, body weight, tolerability to study medication and medication
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
77
Fig. 1. Schematic of the study design and timeline of the GEMINI trial. Following a 2-week screening period, there was a 2-week washout period, during which all patients remained on background ACE inhibitor or ARB, but discontinued all other antihypertensive medications. Fulfillment of HbA1c and blood pressure criteria following the washout allowed randomization to either the carvedilol or metoprolol treatment arm, as indicated. Patients were titrated to target blood pressure determined according to their baseline blood pressure. Each titration step was 1–2 weeks. Following one full week at goal blood pressure or at the highest dose level, patients were entered into a 5-month maintenance period. Patients not attaining target pressure on Dose Level 3 were further treated with hydrochlorothiazide and then calcium channel blocker during maintenance until the goal was reached. Following completion of Maintenance Month 5, but not reflected above, patients entered into a downtitration phase, dependent on the maintenance dose, and returned to the clinic 2 weeks later for a final follow-up visit. See the Study timelines and Study assessments sections for a detailed description.
compliance, and determination of serum potassium level. Once entering Maintenance, patients return monthly for 5 months for the assessment of vital signs, body weight, and tolerability, to study medication and medication compliance and to provide blood and urine samples for safety and efficacy assessments. 2.3.5. Downtitration and follow-up Patients taking study medication Dose Level 2 or 3 are required to gradually decrease their dosage prior to study completion. Downtitration is done at weekly intervals. Only blinded study medication is tapered; HCTZ and CCBs, if taken during Maintenance, are not discontinued. An end-ofdowntitration visit is done 1–3 days following the last study medication dose to assess patient safety and includes an interim medical history, vital signs, and cardiopulmonary exam. Blood pressure management thereafter is at the investigator’s discretion. Two weeks after this visit, patients return for final safety assessments including vital signs and evaluation of any laboratory values outside the normal range at the Maintenance Month 5 visit. 2.4. Study assessments Throughout the study, blood pressure and heart rate (sitting) are taken in triplicate, at least 2 min apart, after the participant sits quietly for at least 10 min. Clinical
laboratory efficacy assessments include HbA1c, glucose, insulin, total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides and aldosterone (blood), and ACR (urine). Safety assessments include routine hematology, blood chemistry, and urinalysis. All laboratory samples are drawn after a minimum of 9-h fast. Laboratory testing is performed by a central laboratory (Quest Diagnostics, Van Nuys, CA). Quality of Life (QoL) is assessed by patient completion of the Diabetes Symptoms Checklist (Grootenhuis et al., 1994; Van der Does et al., 1996) and the Sexual Symptoms Distress Index (Croog et al., 1988; males only) at baseline and Maintenance Months 3 and 5 visits. Questionnaires are marked with patient and site identification numbers prior to administration to the patient. Patients are instructed to complete the questionnaires, sealed in the preaddressed envelope provided, for mailing directly to a data coordinator at GlaxoSmithKline, without the assistance of site personnel and to return them to the study coordinator. Patients are required to maintain a home glucose diary in which fasting (a.m.) and preprandial (dinner) glucose values are recorded on alternating days. Study coordinators review the diaries at each on-therapy visit for incidences of hypoglycemia unaccompanied by patient self-reported symptoms. Patients are instructed also to obtain a blood glucose reading and complete the diary whenever symptoms of hypoglycemia (specifically, sweating, shaking, faintness, nervousness/anxiety, nausea, and headache) occur. Incidences
78
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
of symptomatic and/or asymptomatic hypoglycemia are recorded as adverse events. 2.5. Study completion A patient must undergo Maintenance Month 5 assessments to have completed the study. Patients may be permanently discontinued from study medication prior to completion due to the following: (1) an adverse experience (including failure to tolerate study medication); (2) insufficient therapeutic effect, defined as (a) mean sitting diastolic blood pressure exceeds 120 mm Hg at any visit, (b) mean sitting systolic blood pressure exceeds 200 mm Hg and remains so upon mandatory return within 3 days, (c) mean sitting blood pressure exceeds 160/95 mm Hg on two consecutive visits, and (d) diastolic or systolic blood pressure increases to a level unacceptable to the investigator; (3) loss of glycemic control, defined as (a) fasting plasma glucose N270 mg/dl at any time during the double-blind treatment period, confirmed by retest by the site with the result remaining N270 mg/dl, (b) hyperglycemia resulting in symptoms that, in the investigator’s opinion, are of clinical concern, including, but not limited to, polyuria, visual changes, and altered cognitive function, (c) the requirement of insulin (z7 days) for patients not previously on insulin or any additional alteration to the patient’s antidiabetic medication, (d) the requirement of a permanent (N7 days) dose increase for patients previously on insulin, and (e) recurrent episodes of symptomatic hypoglycemia, or recurrent episodes of moderate-to-severe hypoglycemia or asymptomatic hypoglycemia when associated with blood glucose values b50 mg/dl while on the lowest possible level of study medication to achieve the target blood pressure; (4) protocol deviation (including noncompliance); (5) lost to follow-up; and (6) other reasons (including withdrawal of consent). Ideally, patients withdrawing prior to study completion return for an early withdrawal visit for completion of interim history, cardiopulmonary exam, vital signs, body weight, blood and urine for all efficacy and safety evaluations, QoL assessments, study medication tolerability, compliance, adverse event history, and initiation of the downtitration phase. 2.6. Study medication Study medication dosages and administration are summarized in Fig. 1. Patients randomized to carvedilol begin uptitration at 6.25 mg b.i.d. (Dose Level 1). If blood pressure remains above target, the dose is increased to 12.5 mg b.i.d. (Dose Level 2). If the blood pressure still remains above target, the dose is increased to 25.0 mg b.i.d. (Dose Level 3). Patients randomized to metoprolol begin with 50 mg b.i.d. (Dose Level 1), followed, as necessary, by 100 (Dose Level 2) and 200 mg b.i.d. (Dose Level 3). Weekly titration intervals are encouraged, but 2-week intervals are allowed if medically necessary or preferred by the investigator, provided that the patients maintain
weekly visits during the uptitration phase, whether study medication is being adjusted. Patients on Dose Level 3 for 1 week, but still above target blood pressure, enter Maintenance as described above. Sequential addition of HCTZ up to 12.5 mg daily and a dihydropyridine CCB according to specific prescribing guidelines is allowed, as necessary, to achieve target blood pressure. The addition of these medications following entry into Maintenance represents an amendment to the original protocol, which called for their addition and attaining target blood pressure prior to Maintenance entry. Patients are instructed to withhold study medication on the morning of their clinic visit to allow the determination of trough blood pressure. Target blood pressure was initially =135/85 mm Hg; the blood pressure goals in Fig. 1 reflect amended targets consistent with newer recommendations of the American Diabetes Association (Arauz-Pancheco et al., 2002) and the 2003 JNC-7 report (Bakris, 1999). In both the original and amended protocols, a patient on study medication Dose Level 3, HCTZ 12.5 mg daily, and the maximum tolerated dose of a dihydropyridine CCB, who remains above target blood pressure, will have further treatment options discussed with the medical monitor or study-specific hypertension consultants. The dose of ACEI or ARB and antidiabetic medications, including insulin, must remain stable throughout the study. If medically appropriate, the dose levels of all other concomitant medications also should remain unchanged during the study. The protocol also specifies provisions for the management of patients whose blood pressure goes out of control and for patients who have not achieved target blood pressure but are experiencing tolerability issues with study medication Dose Level 2 or 3. 2.7. Target sample size and statistical analysis The target sample size of 1210 patients (484 carvedilol, 726 metoprolol; a 2:3 randomization) is required to satisfy the three principal study hypotheses: (1) that carvedilol is superior to metoprolol (primary study objective) with respect to change in HbA1c from baseline; (2) that carvedilol maintains or improves glycemic control; and (3) that metoprolol worsens glycemic control. For the primary hypothesis, sample size estimation was based on detecting a difference of 0.30% in the HbA1c change from baseline between carvedilol and metoprolol, using a two-sided test with 5% significance level and 90% power. The number of patients needed is 338 patients per arm, assuming that the standard deviation of the change from baseline HbA1c is 1.2%. For the secondary hypothesis, that carvedilol does not worsen glycemic control, a limit was set of +0.10% for the change from baseline HbA1c, beyond which glycemic control is said to have worsened. Using a one-sided bas good as or betterQ test with 2.5% significance level and 80% power (equivalent to using a two-sided 95% confidence
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
interval), 183 patients are needed, assuming that the expected change from baseline HbA1c for carvedilol is 0.15%, with standard deviation of 1.2%. For the other secondary hypothesis, to detect a mean change from baseline HbA1c of +0.15% for metoprolol, using a two-sided test with 5% significance level and 80% power, 505 patients are needed, assuming the same standard deviation. Thus, a minimum sample size of 338 patients for carvedilol and 505 for metoprolol (2:3 randomization, total 843) provides 94% power to test the primary hypothesis, and 96% and 80% power, respectively, for the secondary hypotheses. Target enrollment is adjusted to 1210 (i.e., 484 for carvedilol, 726 for metoprolol) to accommodate for dropouts. The patients were randomized via study site communication with an automated randomization and study medication dispensing system (RAMOS, GSK) using a block of 5 and according to the 2:3 carvedilol–metoprolol distribution described above. Final enrollment into the trial is 1235 patients.
3. Discussion The GEMINI trial will assess whether differences in glycemic control are seen when the second-generation hblocker metoprolol, or the third-generation h-blocker carvedilol, is utilized in patients with hypertension and Type 2 diabetes. Metoprolol is expected to increase insulin resistance and worsen glycemic control, whereas carvedilol is expected to be neutral or slightly improved glycemic control. Glycemic control, assessed by the change from baseline in HbA1c after 5 months of maintenance treatment, should be independent of differences in blood pressure lowering. Carvedilol is also expected to improve other parameters of insulin resistance, with the reverse occurring with metoprolol. Fasting plasma glucose, serum insulin, and insulin resistance assessed by the HOMA method will be compared, as will changes in triglycerides, HDL, LDL, and urinary ACR. The GEMINI trial will also compare the side-effect profile of these second- and third-generation h-blockers. Fatigue, cold extremities, and sexual dysfunction have been reported to be less with third-generation h-blockers. Because peripheral vascular disease and erectile impotence due to autonomic neuropathy are more common in the diabetic patient, the potential for their reduced frequency with carvedilol is of great importance. The common occurrence of hypoglycemia in the diabetic patient, although less of a problem in Type 2 than in Type 1 diabetic patients (Bell et al., 1997), could also potentially be lessened with a combined ah-blocker. In conclusion, the GEMINI trial should document whether the risks of h-blocker use in the Type 2 diabetic patient can be lessened with a third-generation (combined) ah-blocker while achieving antihypertensive control when added to angiotensin-converting enzyme (ACE) inhibition.
79
Acknowledgments This work was supported by GlaxoSmithKline, Philadelphia, PA.
References Arauz-Pancheco, C., Parot, M. A., & Raskin P., for the Professional Practice Committee and the Executive Committee of the American Diabetes Association. (2002). The treatment of hypertension in adults with diabetes (position paper). Diabetes Care, 25 (Suppl. 1), S71 – S73. Bakris, G. (1999). Maximizing cardiorenal benefit in the management of hypertension: achieve blood pressure goals. Journal of Clinical Hypertension, 1, 141 – 147. Bell, D. S. H. (1999). h-adrenergic blocking agents in patients with diabetes—friend and foe. Endocrine Practice, 5, 51 – 53. Bell, D. S. H. (2003). Use of beta blockers in the patient with diabetes. The Endocrinologist, 13, 116 – 123. Bell, D. S. H., & Yamuk, V. (1997). Frequency of severe hypoglycemia in patients with non-insulin-dependent diabetes mellitus treated with sulfonylureas or insulin. Endocrine Practice, 3, 281 – 283. Croog, S. H., Levine, S., Sudilovsky, A., Baume, R. M., & Clive, J. (1988). Sexual symptoms in hypertensive patients. A clinical trial of antihypertensive medications. Archives of Internal Medicine, 148, 788 – 794. Executive summary of the third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III), 2001. JAMA, 285, 2486 – 2497. Giugliano, D., Acampora, R., Marfella, R., De Rosa, N., Ziccardi, P., Ragone, R., De Angelis, L., & D’Onofrio, F. (1997). Metabolic and cardiovascular effects of carvedilol and atenolol in type II diabetes mellitus and hypertension. Annals of Internal Medicine, 126, 955 – 959. Gress, T. W., Nieto, F. J., Shahar, E., Wofford, M. R., & Brancati, F. L. (2000). Hypertension and antihypertensive therapy as risk factors for type 2 diabetes mellitus. New England Journal of Medicine, 342, 905 – 912. Grootenhuis, P. A., Snoek, F. J., Heine, R. J., & Bouter, L. M. (1994). Development of a type 2 diabetes symptom checklist: a measure of symptom severity. Diabetic Medicine, 11, 253 – 261. Jacob, S., Rett, K., Wicklmayr, M., Agrawal, B., Augustin, H. J., & Dietze, G. J. (1996). Differential effect of chronic treatment with two betablocking agents on insulin sensitivity: the carvedilol–metoprolol study. Journal of Hypertension, 14, 489 – 494. Janka, H. U., Holzgreve, H., & Nakov, R. (2000). Comparison of the effects of an ACE inhibitor/calcium antagonist and a betablocker/diuretic combination on a glycemic control in type 2 diabetic patients with hypertension: the AC-VT Study. Diabetes, 49 (Suppl. 1), A18. (74-OR). Jona, M., Reicher-Reiss, H., Boyko, V., Shotan, A., Mandelzweig, L, Glodbourt, U., Behar, S., for the Benzafibrate Infarction Prevention (BIP) Study Group. (1996). Usefulness of h-blocker therapy in patients with non-insulin dependent diabetes mellitus and coronary artery disease. American Journal of Cardiology, 77, 1273 – 1277. Kjekshus, J., Gilpin, E., Cali, G., Blackey, A. R., Henning, H., & Ross, J., Jr. (1990). Diabetic patients and beta-blockers after acute myocardial infarction. European Heart Journal, 11, 43 – 50. The seventh report of the Joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure. The JNC 7 report, 2003. JAMA, 289, 2560 – 2572. Tomlinson, B., Bompart, F., Graham, B. R., Liu, J. B., & Prichard, B. N. (1988). Vasodilating mechanism and response to physiological pressor stimuli of acute doses of carvedilol compared with labetalol, propanolol and hydralazine drugs. Drugs, 36 (Suppl. 6), 37 – 47. Van der Does, F. E., De Neeling, J. N., Snoek, F. J., Kostense, P. J., Grootenhuis, P. A., Bouter, L. M., & Heine, R. J. (1996). Symptoms and well-being in relation to glycemic control in type II diabetes. Diabetes Care, 19, 204 – 210.
The rationale and design of the Glycemic Effects in Diabetes Mellitus Carvedilol–Metoprolol Comparison in Hypertensives (GEMINI) trial George L. Bakrisa,*, David S.H. Bellb, Vivian Fonsecac, Richard Katholid, Janet McGille, Robert Phillipsf, Philip Rasking, Jackson T. Wright Jr.h, Malini Iyengari, Terry Holeslawi, Karen M. Andersoni, for the GEMINI Investigators a
Rush Hypertension Center, Rush Medical Center, Suite 470, 1700 W. Van Buren, Chicago, IL 60612, USA b University of Alabama at Birmingham, Birmingham, AL, USA c Tulane University, New Orleans, LA, USA d St. John’s Hospital, Springfield, IL, USA e Washington University School of Medicine, St. Louis, MO, USA f New York University, New York, NY, USA g University of Texas at Southwestern Medical Center, Dallas, TX, USA h University Hospitals of Cleveland, Cleveland, OH, USA i GlaxoSmithKline, Philadelphia, PA, USA Received 5 March 2004; received in revised form 6 July 2004; accepted 9 July 2004
Abstract h-blockers utilized in the Type 2 diabetic patient result in an even greater decrease in cardiac events than in the nondiabetic patient. Unfortunately, first-and second-generation h-blockers are associated with the worsening of insulin resistance, deterioration of glycemic control, peripheral vasoconstriction, potentially worsening peripheral vascular disease, and more frequent and severe hypoglycemia. The third-generation h-blockers have unique properties, including a1-blockade, and have been shown to lower insulin resistance, improve glycemic control, and vasodilate resistance arterioles. The Glycemic Effects in Diabetes Mellitus: Carvedilol–Metoprolol Comparison in Hypertensives (GEMINI) trial has been designed to compare a third-generation (cardevilol) with a second-generation h-blocker (metoprolol) in a cohort of participants with hypertension and Type 2 diabetes. The primary outcome measure of the study is change in the HbA1c. The study is powered to detect a difference in HbA1c of 0.3 units (%) between the groups. Secondary endpoints include changes in insulin resistance, fasting glucose, and the lipid profile. Differences in the side-effect profile (cold extremities, fatigue, impotence, and hypoglycemia) will also be assessed. The GEMINI trial, therefore, is the first large randomized trial to assess whether utilizing a thirdgeneration h-blocker yields a favorable metabolic profile in the patient with Type 2 diabetes and hypertension. D 2005 Elsevier Inc. All rights reserved. Keywords: Hypertension; Type 2 diabetes; h-blockers; Clinical trial; Carvedilol; Metoprolol
1. Introduction Type 2 diabetes is a cardiac disease equivalent—diabetic patients have a 20% chance of a cardiac event within 10 years
* Corresponding author. Tel.: +1 312 942 3140; fax: +1 312 942 8119. E-mail address: [email protected] (G.L. Bakris). 1056-8727/05/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jdiacomp.2004.07.003
(Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel, 2001). In patients with diabetes, h-blockade decreases mortality in post-MI patients and in patients with established coronary disease, even superimposed on current care standards of angiotensin converting enzyme inhibitors (ACEI) and fibrinolytics (Jona et al., 1996; Kjekshus et al., 1990). This may result from decreased myocardial workload, prevention and reversal of myocardial remodelling, prevention of
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
congestive heart failure, and antiatherogenic and antiinflammatory properties (Bell, 2003). The treatment of hypertension in diabetes, with the American Diabetes Association goal blood pressure of 130/80 mm Hg or less, often requires the use of several antihypertensive agents (Arauz-Pancheco, Parot, & Raskin, 2002; Bakris, 1999). After an ACEI or angiotensin receptor blocker (ARB), with their effects amplified by a thiazide diuretic, h-blockers should be the next choice rather than a calcium channel blocker (CCB) or vasodilator. Unfortunately, traditional h-blockers increase insulin resistance, lower HDL, and raise triglyceride and glucose levels (Bell, 1999). In nondiabetic patients, there is over a 20% increase in the development of diabetes with h-blockers but not with thiazide diuretics (Gress et al., 2000). In diabetic patients, increases in HbA1c of as much as 1% have been reported with a second-generation h-blocker (Janka, Holzgreve, & Nakor, 2000). The third-generation h-blocker, carvedilol, through its combined h- and a1adrenoceptor blockade, has been shown to lower insulin resistance and glucose and triglyceride levels and to increase HDL in both diabetic and nondiabetic subjects (Giugliano et al., 1997; Jacob et al., 1996). Peripheral vasoconstriction, also a potentially unacceptable side effect of first- and second-generation h-blockers in diabetic patients, does not occur because of carvedilol’s a1-blocking, vasodilating effect (Tomlinson, 1988). Because these encouraging preliminary data suggested potential advantages to using a third-generation h-blocker in diabetic hypertensive patients, the Glycemic Effects in Diabetes Mellitus: Carvedilol–Metoprolol Comparison in Hypertensives (GEMINI) trial is being performed.
75
in patients with controlled mild-to-moderate hypertension and Type 2 diabetes mellitus. The major secondary objectives are to determine whether carvedilol improves or maintains glycemic control and whether metoprolol worsens glycemic control, as measured by the change from baseline HbA1c. Additional secondary objectives are to assess effects on insulin sensitivity (HOMA analysis), fasting plasma glucose, serum insulin, triglycerides, total cholesterol, LDL cholesterol, HDL cholesterol, aldosterone, and body weight. The relationship between blood pressure and glycemic control will be examined, and the effects on urinary albumin excretion [albumin/creatinine ratio (ACR)] and the incidence of side effects (e.g., fatigue, cold extremities, and sexual dysfunction) are described. Safety parameters include the incidence of clinically defined symptomatic and asymptomatic hypoglycemia and the percentage of patients withdrawn from the trial. 2.2. Study population Participants recruited are males or females, 30 to 80 years of age, with Type 2 diabetes and a documented history of mild-to-moderate hypertension (systolic BP z140 to V179 mm Hg; diastolic BP z90 to V109 mm Hg). Patients must be taking an ACEI or ARB. Target enrollment is 1210 patients at approximately 250 sites. Specific inclusion and exclusion criteria are summarized in Table 1. 2.3. Study timelines and participant assessments The GEMINI trial includes six phases (Fig. 1): screening/ washout, baseline/randomization, uptitration, maintenance, downtitration, and follow-up. Maximum study length per participant is 35 weeks.
2. GEMINI trial design GEMINI is a randomized, double-blind, parallel-group, multicenter (United States) trial comparing the effects of carvedilol and metoprolol on glycemic control in hypertensive diabetic patients. Patients are titrated with blinded study medication to aggressive target blood pressures consistent with recently published recommendations (ArauzPancheco et al., 2002; The seventh report of the Joint National Committee on Prevention, Detection, and Treatment of High Blood Pressure, 2003) Following Institutional Review Board approval, all participants were consented according to the Declaration of Helsinki Recommendations Guiding Physicians in Biomedical Research Involving Human Subjects (1996). 2.1. Study objectives The primary objective of the GEMINI trial is to determine if the effect of carvedilol on glycemic control, assessed by differences in the change from baseline hemoglobin (Hb) A1c, is superior to the effect of metoprolol
2.3.1. Screening/washout A mandatory 4-week interval exists between screening and a prerandomization visit to verify the stability of glycemic control; HbA1c could not increase N0.5% prior to randomization. This interval comprises a 2-week screening interval and a 2-week washout period. During screening, after providing informed written consent, medical history, physical exam (including vital signs), electrocardiogram (ECG), and chest X-ray are done. Patients provide blood and urine samples for clinical chemistries to rule out any exclusionary disease and to assess whether C-peptide, HbA1c, and urinary ACR meet entry criteria. Once fulfilling the screening requirements, patients begin the 2-week washout phase. Patients on ACEI or ARB monotherapy are eligible for enrollment if they fulfill all other inclusion/ exclusion criteria. All patients are to remain on their ACEI or ARB, but additional antihypertensives are stopped. Patients are allowed to stay on other medications with antihypertensive properties only if prescribed for another indication (e.g., a diuretic for edema, an a-blocker for benign prostatic hypertrophy). Diabetes control must remain stable
76
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
Table 1 Inclusion and exclusion criteria to determine eligibility for enrollment into the GEMINI trial Inclusion criteria 1. z30 and V80 years of age. 2. A documented past history of mild to moderate hypertension (systolic BP z140 to V179 mm Hg; diastolic BP z90 to V109 mm Hg). 3. Taking an ACE inhibitor or an ARB either alone or as part of polypharmacy for control of blood pressure. 4. Stable antihypertensive regimen (no changes to medications and/or dosage) during the month prior to screening. 5. A documented history of Type II diabetes mellitus. Patients must be treating their diabetes with diet alone or with antidiabetic medications (may include stable insulina). 6. Stable antidiabetic regimen (no changes to diet, medication, dosage, or formulation) during the 3 months prior to screening. 7. Patients receiving antidiabetic medications must have an HbA1c of z6.5% and V8.5%; patients managing their diabetes by diet alone must have an HbA1c of z6.5% and V7.5%. 8. Body mass index of 22– 45 kg/m2. 9. Endogenous insulin production as manifested by a fasting C-peptide of z0.6 ng/ml at screening. 10. Evidence that diabetes is well controlled: HbA1c may not be increased by greater than 0.5 units (%) over the 4-week interval between screening and randomization. 11. Following washout of all antihypertensive medications, except ACE inhibitor or ARBs, the patient must have elevated blood pressure (systolic BP N130 to V179 mm Hg and/or diastolic BP N80 to V109 mm Hg). 12. If female of child-bearing potential, using adequate contraceptive measures. Exclusion criteria 1. Uncontrollable or symptomatic arrhythmias 2. Unstable angina 3. Sick sinus syndrome or second or third degree heart block (unless treated with a permanent, functioning pacemaker) 4. Currently receiving treatment for heart failure 5. Myocardial infarction or stroke within three months of study screening 6. Bradycardia (heart rate b50 bpm) 7. Chronic obstructive pulmonary (COPDa) disease requiring continuous use of inhaled or oral bronchodilator therapy, as well as COPD treated with intermittent or continuous corticosteroids (oral or inhaled) 8. Bronchial asthma or related bronchospastic conditions 9. New onset/newly diagnosed Type II diabetes mellitus (onset within the three months prior to screening) 10. Clinically significant renal or liver disease: serum creatinine N2.5 mg/dl; ALT, AST, total bilirubin, or alkaline phosphatase more than 2.5 times the upper limit of the laboratory reference range [Note: microalbuminuria, if the only clinical sign of renal disease does not fulfill this exclusion criterion; fully developed macroalbuminuria (albumin–creatinine N300 mg/g), as an indicator of more serious renal disease constitutes a condition for which the patient should be excluded from participation in the study] 11. Endocrine disorders (e.g., pheochromocytoma, active and untreated hypo- or hyperthyroidism) 12. Use of anorectic or any other diet drugs, or any dietary measures inconsistent with recommendations for patients with Type II diabetes 13. Taking a h-blocker for any reason within three months of screening (eyedrops containing h-blockers are permissible throughout the trial) 14. Chronic disorder that requires chronic or intermittent use of corticosteroids 15. Systemic disease, including cancer, with reduced (b12 months) life expectancy
Table 1 (continued) Exclusion criteria 16. Psychological illness/condition that would interfere with the patient’s ability to understand or complete the requirements of the study 17. Use of an investigational drug within 30 days of entry to the study or within a time period equivalent to five half-lives of the investigational drug a
Added via amendment to the original protocol.
for the initial 4 weeks of the study. Two days prior to the completion of washout, patients return (bprerandomization visitQ) for a fasting HbA1c determination to fulfill the stability requirement and to serve as the baseline value. 2.3.2. Baseline/randomization At washout conclusion, and upon determination that all eligibility criteria are met, the patient is randomized (baseline/randomization visit) to Dose Level 1 of blinded study medication (carvedilol 6.25 mg b.i.d. or metoprolol tartrate 50 mg b.i.d.). Blood and urine samples are collected for the baseline determination of all laboratory-related secondary endpoints and routine safety assessments. Patients also undergo an interim medical history, ECG, cardiopulmonary exam, and evaluation of vital signs and body weight. At this and each visit throughout the study, patients are questioned regarding concomitant medications and adverse events. Randomization is stratified by whether patients are taking ARBs and by whether thiazolidinediones (TZDs) are part of their antidiabetic regimen. 2.3.3. Uptitration Patients are titrated at 1- to 2-week intervals with blinded study medication until they achieve target blood pressure for at least 1 week (i.e., at two consecutive weekly visits, see Study medication and Fig. 1). Patients return weekly for assessment of vital signs, tolerability to study medication, and medication compliance. 2.3.4. Maintenance The Maintenance phase is of 5-month duration. Patients enter maintenance after achieving the target blood pressure for at least 1 week and may enter at study medication Dose Levels 1, 2, or 3. Patients remaining above the goal blood pressure following 1 week at study medication Dose Level 3 are allowed to enter Maintenance but are required to return at weekly intervals for additional therapy. Hydrochlorothiazide (HCTZ) 12.5 mg and a dihydropyridine CCB may be added sequentially as necessary to achieve target blood pressure. This dose of HCTZ is chosen to minimize the potential for any impact on metabolic balance, and this sequence (HCTZ first, followed, if needed, by CCB), to emulate current prescribing guidelines. During the visit at which it is determined that a patient will enter Maintenance, patients undergo an interim medical history, cardiopulmonary exam, assessment of vital signs, body weight, tolerability to study medication and medication
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
77
Fig. 1. Schematic of the study design and timeline of the GEMINI trial. Following a 2-week screening period, there was a 2-week washout period, during which all patients remained on background ACE inhibitor or ARB, but discontinued all other antihypertensive medications. Fulfillment of HbA1c and blood pressure criteria following the washout allowed randomization to either the carvedilol or metoprolol treatment arm, as indicated. Patients were titrated to target blood pressure determined according to their baseline blood pressure. Each titration step was 1–2 weeks. Following one full week at goal blood pressure or at the highest dose level, patients were entered into a 5-month maintenance period. Patients not attaining target pressure on Dose Level 3 were further treated with hydrochlorothiazide and then calcium channel blocker during maintenance until the goal was reached. Following completion of Maintenance Month 5, but not reflected above, patients entered into a downtitration phase, dependent on the maintenance dose, and returned to the clinic 2 weeks later for a final follow-up visit. See the Study timelines and Study assessments sections for a detailed description.
compliance, and determination of serum potassium level. Once entering Maintenance, patients return monthly for 5 months for the assessment of vital signs, body weight, and tolerability, to study medication and medication compliance and to provide blood and urine samples for safety and efficacy assessments. 2.3.5. Downtitration and follow-up Patients taking study medication Dose Level 2 or 3 are required to gradually decrease their dosage prior to study completion. Downtitration is done at weekly intervals. Only blinded study medication is tapered; HCTZ and CCBs, if taken during Maintenance, are not discontinued. An end-ofdowntitration visit is done 1–3 days following the last study medication dose to assess patient safety and includes an interim medical history, vital signs, and cardiopulmonary exam. Blood pressure management thereafter is at the investigator’s discretion. Two weeks after this visit, patients return for final safety assessments including vital signs and evaluation of any laboratory values outside the normal range at the Maintenance Month 5 visit. 2.4. Study assessments Throughout the study, blood pressure and heart rate (sitting) are taken in triplicate, at least 2 min apart, after the participant sits quietly for at least 10 min. Clinical
laboratory efficacy assessments include HbA1c, glucose, insulin, total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides and aldosterone (blood), and ACR (urine). Safety assessments include routine hematology, blood chemistry, and urinalysis. All laboratory samples are drawn after a minimum of 9-h fast. Laboratory testing is performed by a central laboratory (Quest Diagnostics, Van Nuys, CA). Quality of Life (QoL) is assessed by patient completion of the Diabetes Symptoms Checklist (Grootenhuis et al., 1994; Van der Does et al., 1996) and the Sexual Symptoms Distress Index (Croog et al., 1988; males only) at baseline and Maintenance Months 3 and 5 visits. Questionnaires are marked with patient and site identification numbers prior to administration to the patient. Patients are instructed to complete the questionnaires, sealed in the preaddressed envelope provided, for mailing directly to a data coordinator at GlaxoSmithKline, without the assistance of site personnel and to return them to the study coordinator. Patients are required to maintain a home glucose diary in which fasting (a.m.) and preprandial (dinner) glucose values are recorded on alternating days. Study coordinators review the diaries at each on-therapy visit for incidences of hypoglycemia unaccompanied by patient self-reported symptoms. Patients are instructed also to obtain a blood glucose reading and complete the diary whenever symptoms of hypoglycemia (specifically, sweating, shaking, faintness, nervousness/anxiety, nausea, and headache) occur. Incidences
78
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
of symptomatic and/or asymptomatic hypoglycemia are recorded as adverse events. 2.5. Study completion A patient must undergo Maintenance Month 5 assessments to have completed the study. Patients may be permanently discontinued from study medication prior to completion due to the following: (1) an adverse experience (including failure to tolerate study medication); (2) insufficient therapeutic effect, defined as (a) mean sitting diastolic blood pressure exceeds 120 mm Hg at any visit, (b) mean sitting systolic blood pressure exceeds 200 mm Hg and remains so upon mandatory return within 3 days, (c) mean sitting blood pressure exceeds 160/95 mm Hg on two consecutive visits, and (d) diastolic or systolic blood pressure increases to a level unacceptable to the investigator; (3) loss of glycemic control, defined as (a) fasting plasma glucose N270 mg/dl at any time during the double-blind treatment period, confirmed by retest by the site with the result remaining N270 mg/dl, (b) hyperglycemia resulting in symptoms that, in the investigator’s opinion, are of clinical concern, including, but not limited to, polyuria, visual changes, and altered cognitive function, (c) the requirement of insulin (z7 days) for patients not previously on insulin or any additional alteration to the patient’s antidiabetic medication, (d) the requirement of a permanent (N7 days) dose increase for patients previously on insulin, and (e) recurrent episodes of symptomatic hypoglycemia, or recurrent episodes of moderate-to-severe hypoglycemia or asymptomatic hypoglycemia when associated with blood glucose values b50 mg/dl while on the lowest possible level of study medication to achieve the target blood pressure; (4) protocol deviation (including noncompliance); (5) lost to follow-up; and (6) other reasons (including withdrawal of consent). Ideally, patients withdrawing prior to study completion return for an early withdrawal visit for completion of interim history, cardiopulmonary exam, vital signs, body weight, blood and urine for all efficacy and safety evaluations, QoL assessments, study medication tolerability, compliance, adverse event history, and initiation of the downtitration phase. 2.6. Study medication Study medication dosages and administration are summarized in Fig. 1. Patients randomized to carvedilol begin uptitration at 6.25 mg b.i.d. (Dose Level 1). If blood pressure remains above target, the dose is increased to 12.5 mg b.i.d. (Dose Level 2). If the blood pressure still remains above target, the dose is increased to 25.0 mg b.i.d. (Dose Level 3). Patients randomized to metoprolol begin with 50 mg b.i.d. (Dose Level 1), followed, as necessary, by 100 (Dose Level 2) and 200 mg b.i.d. (Dose Level 3). Weekly titration intervals are encouraged, but 2-week intervals are allowed if medically necessary or preferred by the investigator, provided that the patients maintain
weekly visits during the uptitration phase, whether study medication is being adjusted. Patients on Dose Level 3 for 1 week, but still above target blood pressure, enter Maintenance as described above. Sequential addition of HCTZ up to 12.5 mg daily and a dihydropyridine CCB according to specific prescribing guidelines is allowed, as necessary, to achieve target blood pressure. The addition of these medications following entry into Maintenance represents an amendment to the original protocol, which called for their addition and attaining target blood pressure prior to Maintenance entry. Patients are instructed to withhold study medication on the morning of their clinic visit to allow the determination of trough blood pressure. Target blood pressure was initially =135/85 mm Hg; the blood pressure goals in Fig. 1 reflect amended targets consistent with newer recommendations of the American Diabetes Association (Arauz-Pancheco et al., 2002) and the 2003 JNC-7 report (Bakris, 1999). In both the original and amended protocols, a patient on study medication Dose Level 3, HCTZ 12.5 mg daily, and the maximum tolerated dose of a dihydropyridine CCB, who remains above target blood pressure, will have further treatment options discussed with the medical monitor or study-specific hypertension consultants. The dose of ACEI or ARB and antidiabetic medications, including insulin, must remain stable throughout the study. If medically appropriate, the dose levels of all other concomitant medications also should remain unchanged during the study. The protocol also specifies provisions for the management of patients whose blood pressure goes out of control and for patients who have not achieved target blood pressure but are experiencing tolerability issues with study medication Dose Level 2 or 3. 2.7. Target sample size and statistical analysis The target sample size of 1210 patients (484 carvedilol, 726 metoprolol; a 2:3 randomization) is required to satisfy the three principal study hypotheses: (1) that carvedilol is superior to metoprolol (primary study objective) with respect to change in HbA1c from baseline; (2) that carvedilol maintains or improves glycemic control; and (3) that metoprolol worsens glycemic control. For the primary hypothesis, sample size estimation was based on detecting a difference of 0.30% in the HbA1c change from baseline between carvedilol and metoprolol, using a two-sided test with 5% significance level and 90% power. The number of patients needed is 338 patients per arm, assuming that the standard deviation of the change from baseline HbA1c is 1.2%. For the secondary hypothesis, that carvedilol does not worsen glycemic control, a limit was set of +0.10% for the change from baseline HbA1c, beyond which glycemic control is said to have worsened. Using a one-sided bas good as or betterQ test with 2.5% significance level and 80% power (equivalent to using a two-sided 95% confidence
G.L. Bakris et al. / Journal of Diabetes and Its Complications 19 (2005) 74–79
interval), 183 patients are needed, assuming that the expected change from baseline HbA1c for carvedilol is 0.15%, with standard deviation of 1.2%. For the other secondary hypothesis, to detect a mean change from baseline HbA1c of +0.15% for metoprolol, using a two-sided test with 5% significance level and 80% power, 505 patients are needed, assuming the same standard deviation. Thus, a minimum sample size of 338 patients for carvedilol and 505 for metoprolol (2:3 randomization, total 843) provides 94% power to test the primary hypothesis, and 96% and 80% power, respectively, for the secondary hypotheses. Target enrollment is adjusted to 1210 (i.e., 484 for carvedilol, 726 for metoprolol) to accommodate for dropouts. The patients were randomized via study site communication with an automated randomization and study medication dispensing system (RAMOS, GSK) using a block of 5 and according to the 2:3 carvedilol–metoprolol distribution described above. Final enrollment into the trial is 1235 patients.
3. Discussion The GEMINI trial will assess whether differences in glycemic control are seen when the second-generation hblocker metoprolol, or the third-generation h-blocker carvedilol, is utilized in patients with hypertension and Type 2 diabetes. Metoprolol is expected to increase insulin resistance and worsen glycemic control, whereas carvedilol is expected to be neutral or slightly improved glycemic control. Glycemic control, assessed by the change from baseline in HbA1c after 5 months of maintenance treatment, should be independent of differences in blood pressure lowering. Carvedilol is also expected to improve other parameters of insulin resistance, with the reverse occurring with metoprolol. Fasting plasma glucose, serum insulin, and insulin resistance assessed by the HOMA method will be compared, as will changes in triglycerides, HDL, LDL, and urinary ACR. The GEMINI trial will also compare the side-effect profile of these second- and third-generation h-blockers. Fatigue, cold extremities, and sexual dysfunction have been reported to be less with third-generation h-blockers. Because peripheral vascular disease and erectile impotence due to autonomic neuropathy are more common in the diabetic patient, the potential for their reduced frequency with carvedilol is of great importance. The common occurrence of hypoglycemia in the diabetic patient, although less of a problem in Type 2 than in Type 1 diabetic patients (Bell et al., 1997), could also potentially be lessened with a combined ah-blocker. In conclusion, the GEMINI trial should document whether the risks of h-blocker use in the Type 2 diabetic patient can be lessened with a third-generation (combined) ah-blocker while achieving antihypertensive control when added to angiotensin-converting enzyme (ACE) inhibition.
79
Acknowledgments This work was supported by GlaxoSmithKline, Philadelphia, PA.
References Arauz-Pancheco, C., Parot, M. A., & Raskin P., for the Professional Practice Committee and the Executive Committee of the American Diabetes Association. (2002). The treatment of hypertension in adults with diabetes (position paper). Diabetes Care, 25 (Suppl. 1), S71 – S73. Bakris, G. (1999). Maximizing cardiorenal benefit in the management of hypertension: achieve blood pressure goals. Journal of Clinical Hypertension, 1, 141 – 147. Bell, D. S. H. (1999). h-adrenergic blocking agents in patients with diabetes—friend and foe. Endocrine Practice, 5, 51 – 53. Bell, D. S. H. (2003). Use of beta blockers in the patient with diabetes. The Endocrinologist, 13, 116 – 123. Bell, D. S. H., & Yamuk, V. (1997). Frequency of severe hypoglycemia in patients with non-insulin-dependent diabetes mellitus treated with sulfonylureas or insulin. Endocrine Practice, 3, 281 – 283. Croog, S. H., Levine, S., Sudilovsky, A., Baume, R. M., & Clive, J. (1988). Sexual symptoms in hypertensive patients. A clinical trial of antihypertensive medications. Archives of Internal Medicine, 148, 788 – 794. Executive summary of the third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III), 2001. JAMA, 285, 2486 – 2497. Giugliano, D., Acampora, R., Marfella, R., De Rosa, N., Ziccardi, P., Ragone, R., De Angelis, L., & D’Onofrio, F. (1997). Metabolic and cardiovascular effects of carvedilol and atenolol in type II diabetes mellitus and hypertension. Annals of Internal Medicine, 126, 955 – 959. Gress, T. W., Nieto, F. J., Shahar, E., Wofford, M. R., & Brancati, F. L. (2000). Hypertension and antihypertensive therapy as risk factors for type 2 diabetes mellitus. New England Journal of Medicine, 342, 905 – 912. Grootenhuis, P. A., Snoek, F. J., Heine, R. J., & Bouter, L. M. (1994). Development of a type 2 diabetes symptom checklist: a measure of symptom severity. Diabetic Medicine, 11, 253 – 261. Jacob, S., Rett, K., Wicklmayr, M., Agrawal, B., Augustin, H. J., & Dietze, G. J. (1996). Differential effect of chronic treatment with two betablocking agents on insulin sensitivity: the carvedilol–metoprolol study. Journal of Hypertension, 14, 489 – 494. Janka, H. U., Holzgreve, H., & Nakov, R. (2000). Comparison of the effects of an ACE inhibitor/calcium antagonist and a betablocker/diuretic combination on a glycemic control in type 2 diabetic patients with hypertension: the AC-VT Study. Diabetes, 49 (Suppl. 1), A18. (74-OR). Jona, M., Reicher-Reiss, H., Boyko, V., Shotan, A., Mandelzweig, L, Glodbourt, U., Behar, S., for the Benzafibrate Infarction Prevention (BIP) Study Group. (1996). Usefulness of h-blocker therapy in patients with non-insulin dependent diabetes mellitus and coronary artery disease. American Journal of Cardiology, 77, 1273 – 1277. Kjekshus, J., Gilpin, E., Cali, G., Blackey, A. R., Henning, H., & Ross, J., Jr. (1990). Diabetic patients and beta-blockers after acute myocardial infarction. European Heart Journal, 11, 43 – 50. The seventh report of the Joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure. The JNC 7 report, 2003. JAMA, 289, 2560 – 2572. Tomlinson, B., Bompart, F., Graham, B. R., Liu, J. B., & Prichard, B. N. (1988). Vasodilating mechanism and response to physiological pressor stimuli of acute doses of carvedilol compared with labetalol, propanolol and hydralazine drugs. Drugs, 36 (Suppl. 6), 37 – 47. Van der Does, F. E., De Neeling, J. N., Snoek, F. J., Kostense, P. J., Grootenhuis, P. A., Bouter, L. M., & Heine, R. J. (1996). Symptoms and well-being in relation to glycemic control in type II diabetes. Diabetes Care, 19, 204 – 210.