Beta-blockers for the Treatment of Hypertensionin Patients with Diabetes: Exploring theContraindication Myth

Beta-blockers, Hypertension, and DiabetesMajumdar

Sumit R. Majumdar, MD, MPHDepartment of Ambulatory Care and Prevention, HarvardMedical School and Harvard Pilgrim Health Care, Boston MA

Abstract. Purpose: To review the evidence supporting the

“contraindications” (hypoglycemic unawareness, insulin

resistance, and dyslipidemia) usually given as the reasons

for not using beta blockers for treating hypertension in

patients with diabetes mellitus. Methods: A research syn-

thesis based on MEDLINE (January 1966 through January

1999), hand searches of pertinent references and text-

books, and consultation with experts. Results: There is

little evidence to support the assertion that beta blockers

should be routinely contraindicated in diabetes. Beta block-

ers have few clinically important effects on hypoglycemic

awareness and recovery, insulin resistance and hyperglyce-

mia, or lipid pro~les. Moreover, when diabetics have been

treated with beta blockers for hypertension or for the sec-

ondary prevention of myocardial infarction, they bene~t as

much, if not more, than nondiabetic patients. There may be

circumstances (e.g., hypertensive patients with coronary

disease) under which beta blockers are the drugs of ~rst

choice for diabetic patients. Recommendations to use

agents other than beta blockers for the treatment of hyper-

tension in diabetes are based on these agents’ effectiveness

against surrogate endpoints, and not their proven bene~t

in preventing important clinical endpoints. Conclusions:

Except for patients with brittle glycemic control, manifest

hypoglycemic unawareness, renal parenchymal disease, or

documented intolerance, beta blockers should no longer be

considered routinely contraindicated in the presence of

diabetes.

Key Words. hypertension, diabetes mellitus, treatment,

beta blockers, guidelines

Introduction

Diabetes mellitus (90% of which is non-insulin depend-ent diabetes mellitus [NIDDM]) and hypertension aretwo of the most common chronic conditions in theUnited States, and each predisposes to acceleratedatherosclerosis, cardiovascular disease, and death.They occur together more commonly than chancewould dictate: The prevalence of hypertension in dia-betics is twice that in non-diabetics, and at least threemillion Americans carry both diagnoses [1]. Moreover,the risk of cardiovascular death in a patient with dia-betes and hypertension is double that of a nondiabetichypertensive and four times that of a nondiabetic with-

out hypertension [2,3]. Hypertension also acceleratesthe microvascular complications of diabetes, such asretinopathy and nephropathy [1,3]. Distressingly,fewer than half of those with both diagnoses have theirblood pressure adequately treated [3].

The treatment of high blood pressure has beenshown in randomized controlled trials (RCTs) to de-crease total mortality, cardiovascular mortality,stroke, and myocardial infarction. Only low dose thiaz-ide diuretics (“diuretics”) and beta adrenergic antag-onists (“beta blockers”) have been proven both safeand effective, based on the results of 18 RCTs withalmost 50,000 patients [4]. However, diabetic patientshave been systematically excluded from all but twoof these trials, and constitute only 3% of all patientsenrolled [5].

Although a subgroup analysis of a recent RCT dem-onstrated that diuretics were both safe and effective inelderly patients with diabetes and isolated systolic hy-pertension [5], previous cohort studies [6,7] had shownsigni~cantly increased risks of excess mortality (be-tween two-and ~ve-fold) in diabetic patients treatedwith these agents. Considering the strength of theselatter ~ndings, and the known adverse metabolic se-quelae of diuretics (e.g., hypokalemia, hypomagne-semia, hyperlipidemia, and hyperglycemia), physicianshave avoided these drugs and national guidelines haverecommended against their use as ~rst line agents indiabetic patients [8].

One might think that beta blockers, by default,would become the drugs of ~rst choice, but this is notthe case. Why? Because pharmaceutical compendiums[9], textbooks [10], review articles [11], national guide-

Supported by: National Research Service Award (training grantPE 11001-10), Public Health Service, Bethesda MD.

Address for correspondence: Dr. S. R. Majumdar, Department ofAmbulatory Care and Prevention, 126 Brookline Avenue, Suite200, Boston MA 02215, Phone 617-421-3765, FAX 617-859-8112.email sumit majumdar@hphc.org

Received 9 February 1999; receipt/review time 2 mos; revisedmanuscript accepted 16 April 1999

435

Cardiovascular Drugs and Therapy 1999;13:435–439

© Kluwer Academic Publishers. Boston. Printed in U.S.A.

lines [12], and editorials [13] state that beta blockers“should not be used,” or “must only be used with cau-tion,” or “are relatively contraindicated,” or “used onlyunder special circumstances,” in the presence of diabe-tes. The purpose of this article is to describe the com-mon reasons (i.e., relative contraindications) given forwithholding beta blockers from diabetic patients, andevaluate the evidence available to support these rea-sons. Furthermore, it examines the evidence support-ing the possible bene~ts of using beta blockers in pa-tients with diabetes and hypertension, and concludeswith some treatment recommendations.

Relative ContraindicationsHypoglycemic unawareness and

delayed recovery

The ~rst reason usually given by physicians for with-holding beta blocker therapy is that these agents ob-scure the warning symptoms of hypoglycemia, a dan-gerous complication of diabetes treatment [9–14].Symptoms such as palpitations, tremor, and anxietyare mediated by sympathetic out_ows of adrenalinethat stimulate beta receptors and serve to warn pa-tients of hypoglycemia. While these particular symp-toms of hypoglycemia may be blunted, other symp-toms, such as sweating, may be enhanced [9,10,15].Even if some symptoms are blunted by beta blockade,hypoglycemic unawareness to a dangerous extent isonly a problem in a minority of IDDM patients, and isan uncommon problem in the vast majority of NIDDMpatients. Furthermore, blunting of hypoglycemicawareness is much less of a concern with the use ofmore cardioselective agents such as atenolol and meto-prolol [3,10,14].

Nonselective beta blockade during an acute episodeof hypoglycemia may also delay the physiological cor-rection of hypoglycemia. This is because glucagon andadrenaline mediate glycogenolysis and glucose produc-tion via beta2 receptors in liver and muscle [11,13].However, in controlled metabolic studies, cardioselec-tive beta blockers have been reported to neither pro-long hypoglycemia nor interfere with glucagon and ad-renaline mediated recovery [16,17].

Although hypoglycemic unawareness and delayedrecovery are potential adverse events attributable tobeta blockers, there is little evidence that this is animportant clinical problem. In a nested case controlstudy of diabetic patients (94 cases of hypoglycemiaand 654 controls), no association was found between anincreased risk of hospitalization for hypoglycemia andthe use of beta blockers [18]. Shorr and colleaguesreported a retrospective cohort study with 13,559 eld-erly diabetic patients, and found no increase in the riskof serious hypoglycemia with the use of beta blockers[19]. They concluded that antihypertensive therapyshould be chosen based on “other” considerations ofsafety and effectiveness [19]. Finally, the UK Prospec-

tive Diabetes Study Group recently completed a RCTof blood pressure therapy in NIDDM (UKPDS-39),comparing atenolol (n 5 358) with captopril (n 5 400)[20]. These investigators found no difference betweentreatment arms for either the number or severity ofhypoglycemic episodes, over nine years of follow-up[20].

Increased insulin resistance

Increased insulin resistance and the possibility of hy-perglycemia are also frequently cited, but insuf~cient,grounds for withholding beta blockers from diabeticpatients. Beta blockers appear to induce insulin resis-tance, by reducing beta2 receptor mediated secretion ofinsulin and by decreasing insulin mediated peripheraluptake of glucose [10,16]. This effect is attenuated, butnot abolished, by the use of cardioselective beta block-ers [3,10,14]. NIDDM itself is characterized by insulinresistance. It is assumed that the addition of a betablocker will further exacerbate this state and lead tocompromised blood sugar control. This hyperglycemictendency may be potentiated when beta blockers anddiuretics are used in combination [12,14,21]. Good gly-cemic control is nonetheless achievable in beta blockerusers through a regimen of weight loss, exercise, diet,oral hypoglycemic agents and insulin; small perturba-tions of blood sugar can, and should, be treated bymanipulating these factors [22].

Although insulin resistance and hyperglycemia arepotential adverse events attributable to beta blockers,there is little evidence that this is an important clini-cal problem. For example, in a prospective cohort of2,723 patients with diabetes and established coronaryartery disease, patients on beta blockers (n 5 911)had lower fasting blood glucose, the same use of hy-poglycemic therapies, and lower mortality than pa-tients not using beta blockers [23]. In the previouslydescribed UKPDS-39, there was no difference in he-moglobin A1C levels between the atenolol and capto-pril arms after nine years of follow-up [20]. However,patients in the atenolol arm did require an additionalhypoglycemic medication more often than captoprilpatients (81% of the time, compared with 71% [p ,0.05]) [20]. This lends support to the contention thatangiotensin converting enzyme inhibitors decreaseinsulin resistance and improve glycemic control[10,14,16], but this remains questionable clinical sig-ni~cance.

Dyslipidemia

Small treatment induced changes in serum lipids alsoseem to be insuf~cient grounds for withholding betablockers from diabetic patients. By interfering withhepatic lipoprotein synthesis and peripheral lipopro-tein lipase activity, beta blockers (older nonselectiveagents moreso than cardioselective agents) have con-sistently been shown to induce a state of dyslipidemia.This is characterized by an elevation of triglycerides, a

436 Majumdar

decrease in high density lipoprotein levels, and nochange in low density lipoprotein or total cholesterollevels [10,11,16]. A recent meta-analysis of 65,000 pa-tients in 474 trials (mostly short term studies) foundthat beta blockers, on average, increased triglyceridelevels about 30 mg per dL and decreased high densitylipoprotein levels about 4 mg per dL, without any otherchanges in serum lipids [24].

Although these small changes are statisticallysigni~cant, whether they persist over time or are clini-cally signi~cant are different questions [4,25].Speci~cally, whether or not treatment induced changeshave the same associated risks that “baseline” (i.e.,pre-treatment) lipid measurements have is not wellde~ned. For example, in a prospective cohort oftreated hypertensives (n 5 686, 76% using beta block-ers), investigators found that baseline elevation of totalcholesterol or triglycerides independently predictedsubsequent cardiovascular events [26]. However,treatment induced hypertriglyceridemia that devel-oped over 15 years of follow-up conferred no additionalrisk [26]. In addition, both total cholesterol and serumtriglycerides decreased over time for the entire cohort,even though patients were treated almost exclusivelywith beta blockers and diuretics. Beta blockers havealso consistently been shown to prevent reinfarctionand death after acute myocardial infarction [27,28], abene~t that are essentially undiminished in the pres-ence of adverse changes in serum lipids [29].

Beta Blockers for the Treatment ofHypertension in DiabetesPossible indications

In addition to lowering blood pressure, beta blockershave anti-anginal, anti-ischemic, and anti-arrhythmicproperties, and are thought to be especially effective inthe prevention of morbidity and mortality from coro-nary heart disease. This is borne out by secondaryprevention studies of myocardial infarction: The re-sults of a recent meta-analysis indicate a risk reductionof 20% for mortality and 25% for reinfarction [28].Moreover, a subgroup analysis of a RCT of timolol forthe secondary prevention of myocardial infarction sug-gested that diabetic patients derived even greaterbene~t than seen for the trial overall (67% decrease inmortality for diabetic patients, compared with a 40%decrease for all patients [p , 0.05]) [30].

The distinction between primary and secondaryprevention appears to be somewhat blurred in diabe-tes. These patients have more severe and diffuse coro-nary artery disease, a worse prognosis, and a propen-sity to myocardial infarctions without symptoms[1,2,10,11,23]. Half of all patients with diabetes will dieof coronary heart disease [3], and their baseline risk ismuch higher than the patients typically enrolled inRCTs. Therefore, combining direct evidence fromstudies of hypertension treatment and indirect evi-

dence from secondary prevention trials leads to a com-pelling argument for the use of beta blockers as thedrugs of ~rst choice for treating hypertension in somepatients with diabetes mellitus.

And yet, beta blockers are almost routinely with-held from these patients because of the assertion thatthey are contraindicated [3,9–13,16]. Furthermore,many experts argue that angiotensin converting en-zyme inhibitors or calcium channel blockers should be~rst line agents in diabetic patients [10,13,14,27]. Asstated by Kaplan and colleagues: “the availability ofvarious equally effective agents that do not have ad-verse effects seems evidence enough to validate theiruse in this large population of vulnerable people [13].”Such recommendations are based on the fact that nei-ther angiotensin converting enzyme inhibitors nor cal-cium channel blockers adversely affect a constellationof surrogate endpoints [9–14].

Surrogate endpoints

A surrogate endpoint is a laboratory measurement(e.g., serum cholesterol) or a physical sign (e.g., bloodpressure) used as a substitute for a clinically meaning-ful endpoint that measures directly how a patient feels,functions, or survives [31]. Changes induced by ther-apy on a surrogate endpoint (e.g., lowering serum cho-lesterol or blood pressure) are assumed to directlyre_ect changes in clinically meaningful endpoints suchas myocardial infarction or death. The assumption thattherapy induced improvement in a surrogate endpointwill result in improved outcomes has been repeatedlytested, and it has often failed [31,32].

A recent example is the Appropriate Blood Pres-sure Control in Diabetes (ABCD) trial, comparing anangiotensin converting enzyme inhibitor (enalapril, n5 235) to a calcium channel blocker (nislodipine, n 5235) [33]. Both treatment arms had equivalent bloodpressure control, and if this RCT had used blood pres-sure alone as its (surrogate) endpoint, the investiga-tors would have concluded that both drugs wereequally effective. However, consideration of clinicalendpoints led to the early termination of the nislodip-ine arm by the ABCD Data and Safety MonitoringCommittee because of a seven-fold increased risk ofmyocardial infarction. Although there was no placeboor standard treatment arm in this RCT, and enalaprilmay exert some protective effect against myocardialinfarction independent of lowering blood pressure, it ismore likely that nislodipine increased the risk of myo-cardial infarction [33]. This increased risk is consistentwith recent observational studies [34,35] and other re-cent RCTs [36,37] of calcium channel blockers.

Non-traditional beta blockers

If physicians remain concerned about the possibility ofadverse surrogate endpoints, there are newer betablockers (e.g., acebutolol, labetalol, or carvedilol) thatcould be considered. For example, acebutolol is a

Beta-blockers, Hypertension, and Diabetes 437

cardioselective beta blocker with intrinsic sym-pathomimetic activity. This latter property makes ace-butolol fairly unique. It is a beta blocker that does notaffect hypoglycemic awareness; does not cause insulinresistance; and does not affect serum lipid levels [38].Nevertheless, except for the secondary prevention ofmyocardial infarction [39], acebutolol has not yet beenshown to decrease cardiovascular morbidity and mor-tality. This drug is dissimilar enough, structurally andfunctionally, to conventional beta blockers that a “classeffect” should not necessarily be assumed. With thiscaveat in mind, using acebutolol to treat hypertensionin patients with diabetes may be a reasonable compro-mise between the available evidence and expert opin-ion.

Conclusions

This article is not an argument for placing every pa-tient with diabetes and hypertension on a beta blocker.There are many important and practical exceptions,especially in patients with severe manifestations ofIDDM. In the minority of IDDM patients with brittleglycemic control or manifest hypoglycemic unaware-ness, beta blockers may be contraindicated, and ther-apy with an angiotensin converting enzyme inhibitoror a long acting calcium channel blocker consideredappropriate [10,11,12,27]. In IDDM patients with mi-croalbuminuria, or any patient with diabetic neph-ropathy, an angiotensin converting enzyme inhibitorshould be considered to prevent or delay the onset ofend-stage renal disease and dialysis [27,40,41], al-though even this conventional recommendation has re-cently been questioned [42]. Beta blockers are usuallyavoided in patients with peripheral vascular diseaseand cold extremities [3,12,20,21,43], although a recentmeta-analysis concluded that these agents do not affectsymptoms or walking capacity [43]. Patients with auto-nomic dysfunction (e.g., orthostatic hypotension orsexual dysfunction) may not tolerate any antihyper-tensive agent, but beta blockers may be especiallytroublesome [3,10,13,14]. Patients known to be intoler-ant of beta blockers, or with other contraindications(e.g., reactive airways disease, or AV conductionblock), should also be offered alternate therapy.

This article is an argument for not routinely with-holding a proven effective therapy from patients justbecause they are diabetic. The reluctance of physiciansto prescribe beta blockers to diabetic patients withhypertension is based on the anticipation of putativeadverse effects. As reviewed, there is little evidence tosupport this reluctance, and even evidence to supportthe contrary position, that beta blockers might be thepreferred treatment for hypertension in some patientswith diabetes. Beta blockers should no longer be con-sidered routinely contraindicated in the presence ofdiabetes—to do so denies high risk patients the oppor-

tunity to bene~t from a therapy proven to decreasemortality, stroke, and myocardial infarction.

References

1. Epstein M, Sowers JR. Diabetes mellitus and hypertension.Hypertension 1992;19:403–418.

2. Kannel WB, McGee DL. Diabetes and cardiovascular riskfactors: The Framingham Study. Circulation 1979;59:8–13.

3. Statement on hypertension in diabetes mellitus: Final re-port. Arch Int Med 1987;147:830–842.

4. Psaty BM, Smith NL, Siscovick DS, et al. Health outcomesassociated with antihypertensive therapies used as ~rst-lineagents: A systematic review and meta-analysis. JAMA1997;277:739–745.

5. Curb JD, Pressel SL, Cutler JA, et al. Effect of diuretic-based antihypertensive treatment on cardiovascular diseaserisk in older diabetic patients with isolated systolic hyper-tension. JAMA 1996;276:1886–1892.

6. Klein R, Moss SE, Klein BE and Demets DL. Relation ofocular and systemic factors to survival in diabetes. Arch IntMed 1989;149:266–272.

7. Warram JH, Laffel LMB, Valsania P, Chrislieb AR,Krolewski AS. Excess mortality associated with diuretictherapy in diabetes mellitus. Arch Int Med 1991;151:1350–1356.

8. Myers MG, Carruthers SG, Leenen FHH, et al. Recommen-dations from the Canadian Hypertension Society ConsensusConference on the pharmacologic treatment of hyperten-sion. Can Med Assoc J 1989;140:1141–1146.

9. Physicians Desk Reference, 50th Ed. Medical EconomicsData Production Company, 1996.

10. Kaplan NM. Clinical Hypertension, 6th Ed. Williams andWilkins, 1994: chapter 7.

11. Leese GP, Savage MW, Chattington PD, Vora JP. The dia-betic patient with hypertension. Postgrad Med J 1996;72:263–268.

12. National High Blood Pressure Expert Panel WorkingGroup. Hypertension in diabetes. Hypertension 1994;23:147–162.

13. Kaplan NM, Rosenstock J, Raskin P. A differing view oftreatment of hypertension in patients with diabetes melli-tus. Arch Int Med 1987;147:1160–1162.

14. Sowers JR, Zemel MB. Clinical implications of hypertensionin the diabetic patient. Am J Hyper 1990;3:415–424.

15. Molnar GW, Read RC, Wright FE. Propranolol enhance-ment of hypoglycemic sweating. Clin Pharmacol Ther 1974;15:490–496.

16. Giugliano D, Acampora R, Marfella R, et al. Metabolic andcardiovascular effects of carvedilol and atenolol in non-insu-lin dependent diabetes mellitus and hypertension: A ran-domized controlled trial. Ann Intern Med 1997;126:955–959.

17. Lager I, Blohme G, Smith U. Effects of cardioselective andnon-selective beta blockade on the hypoglycemic response ininsulin dependent diabetes. Lancet 1979;1:458–462.

18. Herings RM, de Boer A, Striker BH, Leufkens HG, PorsiusA. Hypoglycemia associated with use of inhibitors of angio-tensin converting enzyme. Lancet 1995;345:1195–1198.

19. Shorr RI, Ray WA, Daugherty JR, Grif~n MR. Antihyper-tensives and the risk of serious hypoglycemia in older per-sons using insulin or sulfonylureas. JAMA 1997;278:40–43.

20. UK Prospective Diabetes Study Group. Ef~cacy of atenololand captopril in reducing risk of macrovascular and mi-

438 Majumdar

crovascular complications in type 2 diabetes: UKPDS 39.BMJ 1998;317:713–720.

21. Sowers JR, Epstein M. Diabetes mellitus and associatedhypertension, vascular disease, and nephropathy—an up-date. Hypertension 1995;26:869–879.

22. Manolio TA, Cutler JA, Furberg CD, et al. Trends in phar-macologic management of hypertension in the UnitedStates. Arch Int Med 1995;155:829–837.

23. Jonas M, Reicher-Reiss H, Boyko V, et al. Usefulness of betablocker therapy in patients with non-insulin dependent dia-betes mellitus and coronary artery disease. Am J Cardiol1996;77:1273–1277.

24. Kasiske BL, Ma JZ, Kalil RS, Louis TA. Effects of antihy-pertensive therapy on serum lipids. Ann Intern Med 1995;122:133–141.

25. Lasser NL, Grandits G, Caggiula AW, et al. Effects of anti-hypertensive therapy on plasma lipids and lipoproteins inthe Multiple Risk Factor Intervention Trial. Am J Med1984;76(2A):52–66.

26. Samuelsson O, Hedner T, Persson B, et al. The role of diabe-tes mellitus and hypertriglyceridemia as coronary risk fac-tors in treated hypertension: 15 years of follow- up of anti-hypertensive treatment in middle aged men in the PrimaryPrevention Trial in Goteborg, Sweden. J Int Med 1994;235:217–227.

27. The Sixth Report of the Joint National Committee on Pre-vention, Detection, Evaluation, and Treatment of HighBlood Pressure. Arch Intern Med 1997;157:2413–2444.

28. Yusuf S, Peto R, Lewis JA, Collins R, Sleight P. Beta block-ade during and after myocardial infarction: a review of therandomized controlled trials. Prog Cardiovasc Dis 1987;27:335–343.

29. Byington RP, Worthy J, Craven T, Furberg CD. Propra-nolol-induced changes and their prognostic signi~cance aftermyocardial infarction: The Beta-Blocker Heart Attack Trialexperience. Am J Cardiol 1990;65:1287–1291.

30. Gundersen T, Kuekshus J. Timolol treatment after myocar-dial infarction in diabetic patients. Diabetes Care 1983;6:285–290.

31. Fleming TR, Demets DL. Surrogate endpoints in clinical tri-als: Are we being misled? Ann Intern Med 1996;125:605–613.

32. Echt DS, Liebson PR, Mitchell LB, et al. Mortality andmorbidity in patients receiving encainide, _ecainide, or pla-cebo. The Cardiac Arrhythmia Suppression Trial. N Engl JMed 1991;324:781–788.

33. Estacio RO, Jeffers BW, Hiatt WR, et al. The effect of nis-lodipine as compared with enalapril on cardiovascular out-comes in patients with non-insulin dependent diabetes andhypertension. N Engl J Med 1998;338:645–652.

34. Psaty BM, Heckbert SR, Koepsell TD, et al. The risk ofmyocardial infarction associated with antihypertensive drugtherapies. JAMA 1995;274:620–625.

35. Furberg CD, Psaty BM, Meyer JV. Nifedipine: Dose-relatedincrease in mortality in patients with coronary heart dis-ease. Circulation 1995;92:1326–1331.

36. Borhani NO, Mercuri M, Borhani PA, et al. Final outcomeresults of the Multicentre Isradipine Diuretic Athero-sclerosis Study (MIDAS): A randomized controlled trial.JAMA 1996;276:785–791.

37. Tatti P, Pahor M, Byington RP, DiMauro P, Guarisco R,Strollo F. Results of the Fosinopril Amlodipine Cardiovascu-lar Events Trial (FACET) in hypertensive patients withnon-insulin dependent diabetes mellitus (NIDDM). Circula-tion 1997;96(Suppl I):I764.

38. De Bono G, Kaye CM, Roland E, Summers AJH. Acebutolol:Ten years of experience. Am Heart J 1985;109:1211–1223.

39. Boissel JP, Leizorovicz A, Piclolet H, et al. Secondary pre-vention after high-risk acute myocardial infarction with lowdose acebutolol. Am J Cardiol 1990;66:251–260.

40. Lewis EJ, Hunsicker LG, Bain RP, et al. The effect ofangiotensin converting enzyme inhibition on diabetic neph-ropathy. N Engl J Med 1993;329:1456–1462.

41. Maschio G, Alberti D, Janin G, et al. Effect of the angioten-sin converting enzyme inhibitor benazepril on the progres-sion of chronic renal insuf~ciency. N Engl J Med 1996;334:939–945.

42. Sawicki PT. Do ACE inhibitors offer speci~c bene~ts in theantihypertensive treatment of diabetic patients? 17 years ofunful~lled promises. Diabetologia 1998;41:598–602.

43. Radack K, Deck C. Beta-adrenergic blocker therapy doesnot worsen intermittent claudication in subjects with pe-ripheral arterial disease: A meta-analysis of randomizedcontrolled trials. Arch Intern Med 1991;151:1769–1776.

Beta-blockers, Hypertension, and Diabetes 439