ALLHAT:Are Newer Antihypertensives Better Than ”Old Hat” Agents?

Q:Should we be prescribing α-blockers to control hypertension in lightof the ALLHAT findings of adverse effects?A:This question regarding the findings of the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT)¹ raisesa crucial issue: do all antihypertensive drugs provide similar benefits if they reduceblood pressures to a similar degree? Various antihypertensives havediffering mechanisms of action and they exert different physiologic effects. Ifthese agents all lower blood pressure to the same level, will there still bedifferences in morbidity and mortality?Is newer better? Some of the newer classes of antihypertensive drugs--including calcium antagonists, ACE inhibitors, and α-blockers--appearto have mechanisms of action and beneficial effects, apart from blood pressurereduction, that would make them better choices than older agents. Forexample:

• Diuretics and β-blockers that lack intrinsic sympathomimetic activity increaseserum cholesterol levels. The newer antihypertensives do not have this effect, and α-blockers lower cholesterol levels.
• Diuretics decrease serum potassium and magnesium and raise blood glucoselevels. Newer antihypertensives do not have either of these effects, andα-blockers may actually improve insulin sensitivity.
• Some of the newer agents (ACE inhibitors, for example) may reduceleft ventricular hypertrophy (LVH) to a greater extent than diuretics andβ-blockers.

Are the newer agents really as good as the traditional ones? Clinical trialswere needed to find out--and to determine potential advantages of selectedagents. One such trial, ALLHAT, began enrolling patients in February 1994.That study will conclude early this year.

All about ALLHAT.

ALLHAT is a randomized, double-blind, activecontrolledcomparison of 4 antihypertensive agents:

• The diuretic, chlorthalidone (12.5 to 25 mg/d).
• The α-blocker, doxazosin (2 to 8 mg/d).
• The calcium antagonist, amlodipine (2.5 to 10 mg/d).
• The ACE inhibitor, lisinopril (10 to 40 mg/d).

One fourth of the ALLHAT patients are also participating in a randomized,open-label trial to determine whether lowering serum low-densitylipoprotein cholesterol levels with pravastatin, an HMG-CoA reductase inhibitor,reduces all-cause mortality. The control group is receiving usual care.Participants are 55 years or older with hypertension plus at leastone other risk factor for coronary heart disease (CHD). These include previousmyocardial infarction (MI) or stroke, LVH, type 2 diabetes mellitus,current cigarette smoking, or a low high-density lipoprotein cholesterollevel.A total of 42,448 men and women were recruited and randomized, 15,268to receive chlorthalidone, 9067 to receive doxazosin. The remainder were ran-domized to receive the other drugs.The predefined outcome measuresinclude:

• CHD, including both coronarydeath and nonfatal MI.
• All-cause mortality.
• Stroke.
• Combined CHD (coronary death,nonfatal MI).
• Revascularization procedure.
• Hospitalization for angina.

Finally, combined cardiovasculardisease (coronary death, nonfatal MI,stroke, revascularization, angina, congestiveheart failure [CHF], and peripheral arterial disease)would be assessed.

Early termination of a treatment arm.

Followingindependent data reviews in early 2000 by the data andsafety monitoring board, the doxazosin treatment arm inthe blood pressure component of the trial was terminatedafter a median follow-up of 3.3 years.

²

Total mortalitydid not differ between the doxazosin and chlorthalidonearms. However:

• There was a 25% increase in the incidence of major cardiovascularevents in the doxazosin group-a statisticallysignificant difference.
• The risk of CHF was double that of the chlorthalidonegroup.

Moreover, the investigators determined that thelikelihood that doxazosin would prove to be more beneficialthan chlorthalidone by the end of the trial was very low.

What are the implications?

It is not possible todetermine from ALLHAT whether the incidence of CHFin doxazosin recipients is the same as, less than, or morethan would be expected without any antihypertensivetreatment. There was a difference of 2 to 3 mm Hg inachieved blood pressure at the time of termination. (At 1year, mean blood pressure in doxazosin recipients was140/79 mm Hg versus 137/79 mm Hg in chlorthalidonerecipients.) Recent trials in older patients suggest that a3 mm Hg difference in pressure could explain a 10% to20% increase in CHF; however, results of those studies didnot support the doubled risk seen in ALLHAT.

^3-5

Previous studies have suggested that α-blockers mayaffect LVH less than diuretics. As yet, however, we haveno data from ALLHAT on the effects of the differentagents on LVH. α-Blockers may increase plasma volumeand possibly plasma norepinephrine levels, but the significanceof these effects is unknown.Most important: the above observations fromALLHAT refer only to patients receiving monotherapywith doxazosin. They do

not

apply to patients receivingallowable combination therapy with other agents toachieve the study treatment goal of a blood pressure levelbelow 140/90 mm Hg.

Recommendations for practice.

Based on theALLHAT observations, it is appropriate to recommendthat doxazosin not be used as monotherapy for patientswith stage 1 or 2 hypertension (ie, blood pressures between140 to 179/90 to 109 mm Hg). The ALLHAT resultsdo not address patients receiving doxazosin as part ofcombination therapy for hypertension, however. Thus, itmay be appropriate to continue doxazosin in patients alsoreceiving a diuretic and, possibly, other classes of antihypertensiveagents concurrently.This study did not address use of doxazosin (orother α-blockers) as an adjunct to the treatment of elevatedcholesterol levels or benign prostatic hypertrophy innormotensive patients. If you prescribe α-blockers forthese conditions, it would be appropriate to know whetherthe patient had a normal ejection fraction and was not inan early stage of heart failure. Given the other classes ofdrugs available to treat hypertension, elevated cholesterol,and benign prostatic hyperplasia, it may be reasonable toavoid α-blockers in these settings.Finally, ALLHAT suggests the need for some modificationof the notion that the lowering of blood pressurewill reduce morbidity and mortality--regardless of theagent used. Different antihypertensive agents may, in fact,have different physiologic effects that can contribute todifferences in morbidity and mortality. Ongoing large clinicaltrials may help to clarify these issues.