Sulfonylurea Treatment of Type 2 Diabetic Patients Does Not Reduce the Vasodilator Response to Ischemia

Home , Glibenclamide, Glimepiride, Sulfonylurea

Pathophysiology/Complications ORIGINAL ARTICLE

1 1 PAOLO SPALLAROSSA, MD LUCA OLIVOTTI, MD vided by the discovery that sulfonylurea 2 2 MARA SCHIAVO, MD RENZO CORDERA, MD drugs stimulate insulin secretion by 1 1 ϩ PIERFRANCO ROSSETTIN, MD CLAUDIO BRUNELLI, MD ␤ 1 blocking -cell ATP-sensitive K (KATP) STEFANO CORDONE, MD channels (5). KATP channels are also abundantly present in the cardiac muscle and in vascular smooth muscle cells in coronary and peripheral arteries, where they play a key role in the protective OBJECTIVE — Sulfonylureas block the activation of vascular potassium-dependent ATP channels and impair the vasodilating response to ischemia in nondiabetic individuals, but it is mechanisms that are triggered during not known whether this occurs in type 2 diabetic patients under chronic treatment with these ischemia (6,7). Soon after the onset of drugs. Glimepiride, a new sulfonylurea, apparently has no cardiovascular interactions. The aim ischemia, the activation of KATP channels of our study was to compare the effect of the widely used compound glibenclamide, the pan- decreases vascular resistance and in- creas-specific glimepiride, and diet treatment alone on brachial artery response to acute forearm creases blood flow (ischemic vasodila- ischemia. tion) and renders myocardial cells more resistant to a subsequent bout of ischemia RESEARCH DESIGN AND METHODS — Brachial artery examination was performed Ϯ Ϯ (the so-called ischemic preconditioning) by a high-resolution ultrasound technique on 20 type 2 diabetic patients aged (mean SD) 67 (8,9). A great deal of experimental evi- 2 years and on 18 nondiabetic patients matched for age, hypertension, and dislipidemia. Diabetic subjects underwent three separate evaluations at the end of each 8-week treatment period, dence shows that glibenclamide and the during which they received glibenclamide, glimepiride, or diet alone according to crossover other sulfonylurea compounds, which are design. Scans were obtained before and after 4.5 min of forearm ischemia. Postischemic vaso- not ␤-cell–specific, block cardiovascular dilation and hyperemia were expressed as percent variations in vessel diameter and blood flow. KATP channels and prevent both protective phenomena (8–11). Glimepiride is a RESULTS — Postischemic vasodilation and hyperemia were, respectively, 5.42 Ϯ 0.90 and new sulfonylurea compound that inter- 331 Ϯ 38% during glibenclamide, 5.46 Ϯ 0.69 and 326 Ϯ 28% during glimepiride, and 5.17 Ϯ ␤ acts more selectively with the -cell KATP 0.64 and 357 Ϯ 35% during diet treatment (NS). These results were similar to those found in the Ϯ Ϯ channels (12). In agreement with experi- nondiabetic patients (6.44 0.68 and 406 42%, NS). mental evidence, studies in humans have CONCLUSIONS — In type 2 diabetic patients, the vasodilating response to forearm isch- shown that acute administration of glib- emia was the same whether patients were treated with diet treatment alone or with glibenclamide enclamide, but not of the pancreas- or glimepiride at blood glucose–lowering equipotent doses. specific glimepiride, induces potentially harmful cardiovascular effects in both Diabetes Care 24:738–742, 2001 nondiabetic patients with coronary artery disease and healthy volunteers (13–17). Less is known about the clinical situ- ulfonylureas represent the most might increase cardiovascular mortality, ation of type 2 diabetic patients who are commonly prescribed pharmaco- but the more recent U.K. Prospective Di- chronically treated with sulfonylureas. S logical treatment for type 2 diabetic abetes Study (UKPDS) showed no evi- The evidence of poor outcome in diabetic patients, despite early observations sug- dence that intensive treatment with patients undergoing percutaneous coro- gesting a possible deleterious effect of these sulfonylureas had any specific adverse ef- nary angioplasty for acute myocardial in- drugs (1–3). However, the question is still fect on cardiovascular disease in type 2 farction while taking sulfonylureas is open. In fact, in the early 1970s, the Uni- diabetic patients (3,4). A physiological consistent with the hypothesis that long- versity Group Diabetes Program (UGPD) basis explaining the potentially detrimen- term inhibition of KATP channels may im- suggested that the use of sulfonylureas tal effects of sulfonylureas has been pro- pair myocardial tolerance to severe ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● ischemia (18). This hypothesis contrasts with the early observations by Paasikivi From the 1Division of Cardiology, Department of Internal Medicine and the 2Department of Endocrinology and Metabolism (MS, RC) University of Genova, Genova, Italy. and with a recent retrospective analysis Address correspondence and reprint requests to Claudio Brunelli, MD, Division of Cardiology, Internal showing that treatment with gliben- Medicine, University of Genova, Viale Benedetto XV, 6, 16132 Genova, Italy. E-mail: [email protected]. clamide does not affect infarct size and unige.it. in-hospital mortality in type 2 diabetic Received for publication 12 June 2000 and accepted in revised form 3 January 2001. ϩ patients with acute myocardial infarction Abbreviations: KATP, ATP-sensitive K . A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion (19,20). However, to our knowledge, no factors for many substances. studies have ever examined whether ther-

738 DIABETES CARE, VOLUME 24, NUMBER 4, APRIL 2001 Spallarossa and Associates

Table 1—General characteristics of the study population Protocol All subjects were examined in the after- Characteristics Diabetic patients Nondiabetic patients noon, 120–180 min after the main meal of the day and intake of the study drug. n 20 18 High-resolution ultrasound (Acuson Se- Age (years) 67.4 Ϯ 1.9 67.5 Ϯ 1.8 quoia C256) with an 8-MHz linear array Sex (M/F) 14/6 13/5 transducer was used to measure the bra- BMI (kg/m2) 28.8 Ϯ 0.7* 26.9 Ϯ 0.5 chial artery diameter. Machine operating Hypertension (yes/no) 14/6 13/5 parameters were kept constant during Cholesterol (mmol/l) 5.17 Ϯ 0.17 5.24 Ϯ 1.17 each study. The subject was required to HDL cholesterol (mmol/l) 1.0 Ϯ 0.3 1.02 Ϯ 0.3 recline on an examination bed for 10 min LDL cholesterol (mmol/l) 3.27 Ϯ 0.16 3.03 Ϯ 0.12 in a quiet setting. The brachial artery was Triglycerides (mg/dl) 130 Ϯ 14 119 Ϯ 13 longitudinally imaged ϳ5 cm above the Glucose (mmol/l) 7.55 Ϯ 0.2† 5.22 Ϯ 0.14 antecubital crease. Once an optimal posi- Ace inhibitors (yes/no) 7/13 5/13 tion was obtained, the transducer was no ␤-blockers (yes/no) 5/15 5/13 longer moved. A baseline scan was ob- Calcium antagonists (yes/no) 2/18 2/16 tained and pulsed Doppler blood flow ve- Lipid-lowering drugs (yes/no) 7/13 5/13 locity was determined with the signal at a History of cardiovascular disease (yes/no) 5/15 4/14 70° angle to the vessel and with the range Data are means Ϯ SE or n.*P Ͻ 0.05 compared to nondiabetic patients; †P Ͻ 0.001 compared to gate adjusted to 1.5 mm and positioned in nondiabetic patients. the center of the artery. A cuff placed on the forearm was then inflated to 300 mmHg for 4.5 min. A second scan was apeutic doses of sulfonylureas in diabetic with acute coronary syndromes were also carried out continuously for 30 s before patients may compromise vascular blood excluded. and 90 s after cuff deflation. Measure- flow regulation in response to ischemia. ments were calculated on the basis of S- The aim of this work was to investigate VHS video recordings by two observers the effects of chronic administration of Design who were blinded to clinical data and the the widely used sulfonylurea gliben- Regardless of previous treatment, diabetic type of therapy being administered. Di- clamide and the more ␤-cell–specific patients were randomized to receive glib- ameter measurements were made at end- glimepiride on the reactive vasodilating enclamide (10 patients) or glimepiride diastole, incident with the R wave on a response to a brief episode of arterial oc- (10 patients) as monotherapy for 8 weeks, continuously recorded electrocardio- clusion in type 2 diabetic patients. after which they were crossed over to the gram, 45–60 s after cuff deflation. Dopp- alternative treatment for a further 8-week ler flow signals were recorded during the RESEARCH DESIGN AND period. During both phases, dosage titra- first 15 s after cuff release. Volume flow METHODS tion took place at 2-week intervals on the was calculated by multiplying the velocity Twenty type 2 diabetic subjects aged basis of blood glucose. Hypoglycemic time integral of the Doppler signal (cor- 57–80 years were recruited from the dia- agents were then discontinued, and ef- rected for angle) by the heart rate and the betes clinic at the University of Genova, forts were made to control hyperglycemia artery cross-sectional area. Changes in Genova, Italy. The known duration of di- vessel diameter and blood flow after cuff for at least 4 weeks by adjusting their diet. abetes (means Ϯ SD) was 5.4 Ϯ 0.9 years, deflation are expressed as the percentage Hypoglycemic drugs could not be discon- and the mean level of HbA was 7.59 Ϯ relative to the baseline scan. 1c tinued in four patients, and three patients 0.28%. Subjects received glibenclamide dropped out because they found the re- (n ϭ 2), glibenclamide plus metformin Statistical analysis (n ϭ 13), or glicazide (n ϭ 3) or were on strictive diet too demanding. At the end of Comparisons between subject groups diet treatment alone (n ϭ 2). One sub- each treatment period with glibenclamide were performed by using paired or un- ject had diabetic retinopathy, three had and glimepiride, all diabetic patients un- paired t tests for continuous variables and microalbuminuria, and two had both derwent noninvasive assessment of bra- a ␹2 test for categorical variables. Post- conditions. A total of 18 nondiabetic pa- chial artery response to acute ischemia. Of ischemic vasodilation and hyperemia tients matched with respect to age, sex, the diabetic patients, 13 on diet therapy were examined by nonparametric statis- and hypertension were enrolled from the alone were also evaluated a third time. tics (Wilcoxon’s test for paired data and local community. Hypertension was de- Nondiabetic patients were only evaluated Mann-Whitney U test, as appropriate). fined as blood pressure Ն150/90 mmHg once. A total of 11 diabetic subjects and 9 Data are given as means Ϯ SEM. A value or current antihypertensive treatment. nondiabetic control subjects were on car- of P Ͻ 0.05 was considered statistically Exclusion criteria for both groups were diovascular drug therapy, and this treat- significant. current tobacco use, renal failure, heart ment was not discontinued. However, failure, history of cerebrovascular disease, every effort was made to maintain a con- RESULTS — Baseline characteristics of peripheral vascular disease, valvular heart stant drug regimen throughout the study the study patients are provided in Table 1. disease, uncontrolled hypertension, or in all diabetic patients undergoing re- The characteristics of diabetic and nondi- any relevant chronic disease. Patients peated examinations. abetic patients were comparable. The two

DIABETES CARE, VOLUME 24, NUMBER 4, APRIL 2001 739 Sulfonylurea treatment of type 2 diabetes

Table 2—Baseline brachial artery diameter and flow, blood pressure, and metabolic determi- neutralize the vascular protective mecha- nations during brachial artery examination nism that is triggered during ischemia (8, 13,14). By contrast, the new-generation sulfonylurea, glimepiride, is more pancreas- Diabetic subjects (n ϭ 20) Nondiabetic patients specific and does not interfere with vascu- Glibenclamide Glimepiride Diet (n ϭ 18) lar KATP channel activation (12,16). How- ever, it must be stressed that this data was Baseline diameter (mm) 4.12 Ϯ 0.19 4.12 Ϯ 0.14 4.05 Ϯ 0.20 4.08 Ϯ 0.25 derived from studies in which sulfonyl- Baseline flow (ml/min) 164 Ϯ 27 131 Ϯ 12 115 Ϯ 9 116 Ϯ 11 ureas were acutely administered in ani- Insulin (pmol/l) 233 Ϯ 24* 229 Ϯ 21* 213 Ϯ 19* 152 Ϯ 14 mals and in nondiabetic humans. It is not Glucose (mmol/l) 11.60 Ϯ 0.47† 11.42 Ϯ 0.40† 11.14 Ϯ 0.42† 6.74 Ϯ 0.26 known what would happen in the clinical Ϯ Ϯ Ϯ Ϯ HbA1c(%)‡ 7.23 0.25† 7.29 0.28† 7.17 0.28† 5.1 0.1 setting if ischemia occurred in diabetic Systolic blood pressure 142 Ϯ 4 141 Ϯ 5 144 Ϯ 5 145 Ϯ 7 patients chronically treated with sulfonyl- (mmHg) ureas (2). The salient information derived Diastolic blood pressure 84 Ϯ 283Ϯ 383Ϯ 285Ϯ 2 from the present study is that the vascular (mmHg) response to ischemia in the forearm skel- Data are means Ϯ SEM. Note that only 13 diabetic patients were treated with diet therapy alone. *P Ͻ 0.05 etal muscle vascular bed was the same compared with nondiabetic patients; † P Ͻ 0.01 compared with nondiabetic patients; ‡nondiabetic range ϭ whether diabetic patients were being 4–5.7%. treated with glibenclamide or glimepiride at blood glucose–lowering equipotent groups that were matched for age, sex, and hyperemia were, respectively, 5.42 Ϯ doses or with diet treatment alone. Sev- and hypertension did not differ with re- 0.90% (95%CI 3.53–7.30) and 331 Ϯ eral reasons may explain the discrepan- gard to total cholesterol, LDL cholesterol, 38% (252–410) during glibenclamide, cies among the findings of previous HDL cholesterol, triglyceride levels, and 5.46 Ϯ 0.69% (4.01–6.91) and 326 Ϯ studies and our results. First, the doses used cardiovascular therapy. They did, how- 28% (266–384) during glimepiride, and in animal models produced blood con- ever, differ significantly with respect to 5.17 Ϯ 0.64% (3.79–6.56) and 357 Ϯ centrations that were, in general, much BMI and fasting blood glucose. Three di- 35% (280–434) during diet treatment higher than the clinical range. Second, in abetic patients and two nondiabetic pa- (NS). No differences were observed be- humans, sulfonylureas are largely (90– tients had prior myocardial infarctions, tween diabetic patients with and without 99%) bound to plasma proteins (2); thus, and two patients in each of the groups had microvascular complications. These re- the actual free drug concentrations in the stable effort angina. sults were similar to the ones observed in steady state are probably lower than those All brachial artery examinations were nondiabetic patients (6.44 Ϯ 0.68% [5– reached during acute administration. A performed in the afternoon, 136 Ϯ 14 min 7.88], NS, and 406 Ϯ 42% [280–434], third possible reason is the presence of after the main meal of the day. In diabetic NS). confounding variables in a clinical study. subjects, the mean dose of glibenclamide There is evidence that hypertension, obe- Ϯ was 7.2 0.9 mg/day, whereas the dose CONCLUSIONS —KATP channels sity, and hypercholesterolemia (21) affect of glimepiride was 2.35 Ϯ 0.19 mg/day. play a key role in the regulation of insulin endothelium-mediated vasodilation and Both sulfonylureas were equally effective secretion and in the modulation of vascu- impair vascular response to ischemia in achieving glycemic control (Table 2). lar tone (5,8). In the resting pancreatic through a mechanism that differs from ␤ Baseline brachial artery diameter and -cells, KATP channels are normally open, KATP channel blockade. We decided a pri- blood flow values in diabetic patients but when plasma glucose rises, the en- ori not to exclude subjects with coronary were similar to those in nondiabetic pa- hanced glucose metabolism brings about risk factors, because these variables often tients (Table 2), and in the diabetic sub- their closure. The decrease of potassium cluster with hyperglycemia and may be jects, there as no difference between those permeability leads to an increase of cyto- considered components of the syndrome. values recorded during treatment with solic calcium that triggers the release of Excluding diabetic patients with coronary glibenclamide or glimepiride and those insulin. On the other hand, KATP channels risk factors in an attempt to eliminate con- recorded during diet treatment alone. in the vascular smooth muscle cells are founders might have made it easier to There was no difference in mean arterial normally closed or inactive. When isch- document differences between the two blood pressure between diabetic and non- emia occurs, the decline in the intracellu- sulfonylurea drugs; however, it would diabetic patients. Assays performed on lar concentration of ATP activates these have encompassed only a subset of the blood drawn immediately before brachial channels. The subsequent hyperpolariza- type 2 diabetic patients who were not rep- artery examination showed that glucose, tion of the cells decreases the inward flow resentative of the type 2 diabetic popula- insulin, and HbA1c levels were higher in of calcium, resulting in vasodilation that, tion. This reason, together with evidence diabetic patients than in nondiabetic sub- in turn, combats the effects of hypoxia (8). that strongly supports the chronic use of jects. However, diabetic subjects showed Sulfonylureas, as a class, interact spe- ACE inhibitors or antihypertensive or lip- no changes in these values during any of cifically with KATP channels of cell mem- id-lowering drugs in most patients with the three study treatments (Table 2). branes. Evidence shows that though type 2 diabetes, is why we did not stop the Brachial artery responses after cuff glibenclamide and other sulfonylureas administration of these drugs for the deflation are shown in Fig. 1. Among di- promote insulin secretion by the block- study (22–24). By the second week of the abetic subjects, postischemic vasodilation ade of pancreatic KATP channels, they also study, small increments in the doses of a

740 DIABETES CARE, VOLUME 24, NUMBER 4, APRIL 2001 Spallarossa and Associates

ischemia is clinically relevant. However, an invasive technique is necessary to explore coronary circulation, because a noninvasive method is not currently available. Even though noninvasive assessment of vascular response in the forearm arterial bed can be considered a convenient indicator of the responsive- ness of coronary arteries to myocardial ischemia (13,14,21,25), caution is needed before concluding that the lack of detect- able effects of hypoglycemic therapy with glibenclamide can be extrapolated to the cardiac level. Because the coronary vascu- lature has a higher density of sulfonylurea- sensitive receptors than the peripheral vascular bed, we cannot rule out a difference between glibenclamide and glimepiride with regard to their influence on coronary reactivity (26). It must also be taken into account that besides vascular smooth muscle cells, sulfonylureas target the myocardium and prevent preconditioning, the endogenous protective mechanism by which a brief episode of ischemia can protect the myocardium from subsequent ischemic insults. Indeed, the acute administration of glibenclamide in nondiabetic patients during coronary angioplasty attenuates preconditioning, whereas glimepiride has no effect (17). However, despite these results, it is not known whether chronic administration of sulfonylureas in diabetic individuals has similar effects on preconditioning. To our knowledge, the only study that investi- gated the effects of prolonged sulfonylurea treatment demonstrated that chronic glicazide treatment improves, rather than worsens, basal and postischemic cardiac function in diabetic rats (27). In conclusion, we found that 8 weeks of treatment with glibenclamide does not Figure 1—Plots show postischemic vasodilation (A) and postischemic hyperemia (B) assessed in determine effects on the vascular response 20 type 2 diabetic patients (F) during treatment with glibenclamide (GBC), glimepiride (GMP), to ischemia in diabetic patients. Although or diet treatment alone (Diet) and in 18 nondiabetic patients (Ⅺ). Note that only 13 diabetic this result may be considered quite reassur- patients were treated with diet therapy alone. ing as far as the use of sulfonylureas is con- cerned, we would like to emphasize that the study refers to the phenomenon of postisch- calcium antagonist were necessary in one tistically significant (14). The present emic vasodilation. Further clinical studies patient receiving glibenclamide and in one study involves an elderly population with are needed to clarify whether treatment patient receiving glimepiride. No other impaired vascular response to acute isch- with the new compound glimepiride is su- changes occurred throughout the study. emia. It may be that, in this context, the perior to less pancreas-specific sulfonyl- The fourth possible reason for the dis- contribution of a vascular KATP channel– ureas in terms of enhanced myocardial crepancies among previous findings and blocking agent that further impairs post- tolerance to ischemia. our results is that previous studies were ischemic dilation is too small to be de- performed on healthy volunteers with tected in a similar sample-size of patients. high vascular response to ischemia. Inter- Diabetic patients are often exposed to Acknowledgments—This study was sup- estingly, the effects of glibenclamide on myocardial ischemia, and investigation of ported in part by Ministero della Universita e healthy subjects was slight, although sta- the coronary vasodilating response to Ricerca Scientifica e Technologica (MURST)

DIABETES CARE, VOLUME 24, NUMBER 4, APRIL 2001 741 Sulfonylurea treatment of type 2 diabetes grants no. 9806467004_006 and no. the canine coronary circulation. Circ Res ment after myocardial infarction: a clini- 9906277284_007. 69:618–622, 1991 cal and biochemical study of 178 patients 11. Gross GJ, Auchampach JA: Blockade of without overt diabetes. Acta Med Scand ATP-sensitive potassium channels pre- Suppl 507:1–82, 1970 References vents myocardial preconditioning in 20. Klamann A, Sarfert P, Launhardt V, 1. Melander A, Lebovitz HE, Faber OK. Sul- dogs. Circ Res 70:223–233, 1992 Schulte G, Schmiegel WH, Nauck MA: fonylureas: why, which, and how? Diabe- 12. Geisen K, Vegh A, Krause E, Papp JG: Car- Myocardial infarction in diabetic vs. non- tes Care 13 (Suppl. 3):18–25, 1990 diovascular effects of conventional sulfo- diabetic subjects: survival and infarct size 2. Leibowitz G, Cerasi E: Sulphonylurea nylureas and glimepiride. Horm Metab Res following therapy with sulfonylureas (gli- treatment of NIDDM patients with cardio- 28:496–507, 1996 benclamide). Eur Heart J 21:220–229, vascular disease: a mixed blessing? Diabe- 13. Kosmas EN, Levy RD, Hussain SN: Acute 2000 tologia 39:503–514, 1996 effects of glyburide on the regulation of 21. Celermajer DS, Sorensen KE, Gooch VM, 3. Meinert CL, Knatterud GL, Prout TE, Klimt peripheral blood ﬂow in normal humans. Spiegelhalter DJ, Miller OI, Sullivan CR: A study of the effects of hypoglyce- Eur J Pharmacol 274:193–199, 1995 mic agents on vascular complications in 14. Bijlstra PJ, den Arend JA, Lutterman JA, ID, Lloyd JK, Deanﬁeld JE: Non-inva- patients with adult-onset diabetes. II. Mor- Russel FG, Thien T, Smits P: Blockade of sive detection of endothelial dysfunc- tality results. Diabetes 19:789–830, 1970 vascular ATP-sensitive potassium chan- tion in children and adults at risk of 4. U.K. Prospective Diabetes Study (UKPDS) nels reduces the vasodilator response to atherosclerosis. Lancet 340:1111–1115, Group: Intensive blood-glucose control ischaemia in humans. Diabetologia 39: 1992 with sulphonylureas or insulin compared 1562–1568, 1996 22. MacLeod MJ, McLay J: Drug treatment of with conventional treatment and risk of 15. Tomai F, Crea F, Gaspardone A, Versaci hypertension complicating diabetes mel- complications in patients with type 2 di- F, De Paulis R, Penta de Peppo A, Chia- litus. Drugs 56:189–202, 1998 abetes (UKPDS 33). Lancet 352:837–853, riello L, Gioffre PA: Ischemic precondi- 23. Haffner SM: Management of dyslipidemia 1998 tioning during coronary angioplasty is in adults with diabetes (Review). Diabetes 5. Sturgess NC, Ashford ML, Cook DL, prevented by glibenclamide, a selective Care 21:160–178, 1998 ϩ Hales CN: The sulphonylurea receptor ATP-sensitive K channel blocker. Circu- 24. Heart Outcomes Prevention Evaluation may be an ATP-sensitive potassium chan- lation 90:700–705, 1994 Study Investigators: Effects of ramipril on nel. Lancet 2:474–475, 1985 16. Bijlstra PJ, Lutterman JA, Russel FG, ϩ cardiovascular and microvascular out- 6. Noma A: ATP-regulated K channels in Thien T, Smits P: Interaction of sulpho- comes in people with diabetes mellitus: cardiac muscle. Nature 305:147–148, 1983 nylurea derivatives with vascular ATP- results of the HOPE study and MICRO- 7. Standen NB, Quayle JM, Davies NW, sensitive potassium channels in humans. HOPE substudy. Lancet 355:253–259, Brayden JE, Huang Y, Nelson MT: Hyper- Diabetologia 39:1083–1090, 1994 2000 polarizing vasodilators activate ATP-sen- 17. Klepzig H, Kober G, Matter C, Luus H, ϩ 25. Anderson TJ, Uehata A, Gerhard MD, sitive K channels in arterial smooth Schneider H, Boedeker KH, Kiowski W, Meredith IT, Knab S, Delagrange D, muscle. Science 245:177–180, 1989 Amann FW, Gruber D, Harris S, Burger Lieberman EH, Ganz P, Creager MA, 8. Daut J, Maier-Rudolph W, von Becker- W: Sulfonylureas and ischaemic precon- Yeung AC, Selwyn AP: Close relation of ath N, Mehrke G, Gunther K, Goedel- ditioning; a double-blind, placebo-con- Meinen L: Hypoxic dilation of coronary trolled evaluation of glimepiride and endothelial function in the human cor- arteries is mediated by ATP-sensitive glibenclamide. Eur Heart J 20:439–446, onary and peripheral circulations. JAm potassium channels. Science 247:1341– 1999 Coll Cardiol 26:1235–41, 1995 1344, 1990 18. Garratt KN, Brady PA, Hassinger NL, 26. Ashcroft SJ, Ashcroft FM: The sulfonyl- 9. Cole WC, McPherson CD, Sontag D: Grill DE, Terzic A, Holmes DR Jr: urea receptor. Biochim Biophys Acta 1175: ATP-regulated Kϩ channels protect the Sulfonylurea drugs increase early mor- 45–59, 1992 myocardium against ischemia/reperfu- tality in patients with diabetes mellitus 27. Shimabukuro M, Nagamine F, Murakami sion damage. Circ Res 69:571–581, 1991 after direct angioplasty for acute myo- K, Oshiro K, Mimura G: Chronic glicla- 10. Aversano T, Ouyang P, Silverman H: cardial infarction. J Am Coll Cardiol 33: zide treatment affects basal and post-isch- Blockade of the ATP-sensitive potassium 119–124, 1999 emic cardiac function in diabetic rats. Gen channel modulates reactive hyperemia in 19. Paasikivi J: Long-term tolbutamide treat- Pharmacol 25:697–704, 1994

742 DIABETES CARE, VOLUME 24, NUMBER 4, APRIL 2001