EDITORIALS

Borderline Mild Systemic ' Hypertension:
   Should It Be Treated?

EDWARD D.

S ince public$tion of the Veterans Adm~n~stratlo~
Cooperative Study,' it has been generally accepted
that antiity~rtenstve drug treatment is effective in
preventing cardiovascular complications in severe
and moderate hypertension. However, evidence for
the effectiveness of treating patients with a diastolic
blood pressure (BP) of 90 to 94 mm Hg is controversialS2
Some authorities claim that the available data fully
justify reducing BP in ail patients with diastolic levels
of 90 mm Hg or higher.3 Others, including this authorv2
do not find the evidence to be so convincing. The
results of the various therapeutic trials are confltcting
with respect to the protection afforded to patients with
90. to +&mm Hg diastolic hypertension.
Diastolic Blood Pressure Below 100 mm Hg

The most favorable results in treating patients with
initial diastolic BP of less tban 90 to 94 mm Hg were
reported by the Hype~ension Detection and Follow-
Up Program (HDFP).' Other studies, such as the Veter-
ans Administratton trial,' U.S. Public Heglth Service
hospitals triahs Oslo study" and Multiple Risk Factor
Intervention Trial (MRFIT).' yielded no significant
difference in morbidity or. mortality between control
and treated patients when the entry diastolic BP was
90 to 94 mm Hg. Significant protection, however, was
found when the diastolic BP was approximately 100
mm Hg or higher. Some of these studies were small,l*~s
and significant benefit might have been found in the
90- to 94.mm Hg group if the sample sizes had been
larger. The larger Australian trial Is not applicabfe'
because it did not include patients with borderline
levels of BP,"

While the HDFP was a large trial, involving over
10,000 pattents. the design failed to control irariebles

From ihe Volorans Adminislrntion and Ceorgelown I.,Jniversify
Mcdicai Centers, Washlnglon, D.C. Manus&pl recelv& Febru-
ary 28,1986. accepted April 25.1988.

Address for reprinls: Edward D. Frels, MD, Hypertension
Research 115lEJ, Veterans Adminisltation MedIcal Center,
Wushinglon, D.C. 20422.

FFIEIS, MD

other than drug treatment that could have jnfluenced
m0rtality.s The "control" patients were sent out to   '
whatever health care facility was available to them
(referred care). They were managed differently from
the special treatment btepped care] patientq in most
aspects of medical care:The physicians, nurses. clini-
cal facilities and hospital back-up were generally su-
perior in the step~d-care cornered with the re-
ferred-care group. Education of the pbtients as to diet,
smoking, and so on, also was different. All costs of
medical care were provided free to the stepped-care
patients, but generally were not so provided to the
referred-care or control group, Because of these im-
portant differences, HDFP cannot be regarded as a
definitive trial of drug treatment per se. The results of
such a trial require confirmati<in from better-con-
trolied studies.

The Australian trial8 was not handicapped by the
same design problems as HDFP, Control and treated
groups were managed similarly except for the admin-
istration of active drugs in 1 group and placebos in the
other. Although the results indicated that treatment
was effective, the range of pretreatment diastolic BPS
included in the trial was 95 to 109 mm Hg, excluding
patients with borderline levels of 90 to 94 mm Hg. This
borderline hypertensive group, however, represents a
large population that, according to the National Health
Survey, comprises about 25 rnjl~lon patients ln the U.S.
alone.!@ The decision coticernlng treatment of this
group, therefore, has major medical and economic
consequences. Also, because the risk of complications
Is considerably less than in patients wlth more severe
hypertension, the cost/benefit ratito,of Jleatment at 90-
to 94-mm Hg levels of diastolic BP inust be carefully
evaluated.
Roes Treatment Protect against
Coronary Heart Disease?

One of the more im~rtant questions tn patients
with mild hypertension is the effectiveness of treat-
ment in preventing heart attack, My~ard~al jnfarctton
and sudden death constitute the melor complications

612


Saptember 15. 1996 THE AMRICAN JCWtNAL C%= CARDtCLCX3Y Volume 59

543

in these patients, as opposed to stroke and other "hy-
pertensive" complications that occur more frequently
in patients with moderate and severe degrees of hy-
pertension. Whether antihypertensive drug treatment
is effective in preventing myocardial infarction, how-
ever, is highly controversial.

A unique observation of the HDFP was that treat-
ment was as effective in preventing myocardial in-
farction as it was in preventing stroke. Only when
patients with ischemic heart disease were added to
those with myocardial infarction did the degree of
protection become less than against stroke. Another
unusual finding in HDFP was that treatment of bor-
derline hypertension resulted in greater protection
against mortality (special-care compared with regular-
care patients) than did treatment of moderate hy-
pertension.' The 3 smaller-scale but better-controlled
trialsr*s-s did not yield a reduction in the incidence
of myocardial infarction with treatment; nor was
there a significant decrease in all-cause morbidity or
mortality when diastolic BP at entry was less than 100
mm Hg.

MRFIT, which was similar in size and design
to HDFP. arrived at opposite conclusions.' Unlike
HDFP, among the MRFIT patients with.an initial dia-
stallc BP of 90 to 94 mm Hg the clinic-treated group
experienced more, rather than fewer, cardiovascular
deaths than the control group including deaths due to
coronary heart disease. In fact, it was only in patients
with a diastolic BP of at least 100 mm Hg that the
special-intervention group achieved a lower mortality
rate than the "control" patients. These contrasting re-
sults from MRFIT and HDFP illustrate the problems in
interpreting results of unconventionally designed clin-
ical trials.

Risk Versus Benefit

  The results of the trials dealing with treatment of
borderline hypertension are often contradictory. The
evidence from the HDFP that treatment had a greater
protective effect in borderline-mild than in moderate-
severe hypertension has not been confirmed by other
trials. Furthermore, the reduction of 45% in fatal myo-
cardial infarction reported by HDFP is not supported
by results of other trials. Neither HDFP nor any other
trial published provides a mandate for indiscriminate
drug treatment of borderline hypertension. The ques-
tion cannot be answered on the basis of the available
evidence.
While we have reason to suspect the efficacy of
treating borderline hypertension, there is no doubt
that long-term drug treatment can cause side effects
and toxicity. Although serious reactions. may be un-
common, they can add up to a considerable number
when the drugs are administered, as they are now, to
many millions of patients for very long periods, For
example, b-blocking drugs can precipitate severe and
even fatal attacks of bronchial asthma. They also may
aggravate`cardiac failure and produce other side ef-
fects. The thiazides have been accused, probably un-
justly, of increased risk associated with hypokalemia

and:-:hypercholesterolemia. Nevertheless, they have
other side effects, including hyperglycemia, hyperuri-
cemia with or without gout, and skin rashes. Prazosin
causes orthostatic hypotension. particularly after the
first dose of the drug. Reserpine can cause depression:
captopril may induce taste disturbance and, in high
doses, neutropenia. Therefore, no drug for reducing
blood pressure is innocuous. All of them can produce
severe and even fatal reactions. Other disadvantages
of drug treatment include the inconvenience of taking
medications every day, and the expense of medica-
tions can also be considerable. Patients are often given
the newest and most expensive of the current wonder
drugs, even though a much less costly tried-and-true
regimen may do as well.
Results of Treatment of Diastolic
Blood Pressure of 95 to 99 mm Hg

The evidence favoring drug treatment of patients
with a diastolic BP of 95 to 99 mm Hg is clearer than it
is in the patients with a diastolic BP 90 to 94 mm Hg. In
the Australian study, treated patients with initial dia-
stolic levels of 95 to 99 mm Hg had 30% fewer compli-
cations than patients who received p1acebo.s HDFP
reported a 23% reduction in mortality in stepped-care
as compared with referred-care patients with a dia-
stolic BP 95 to 99 mm Hg.at entry. On the other hand,
the somewhat similarly designed MRFIT' yielded no
difference in mortality between the special-interven-
tion and referred-care or usual-care patients. The Oslo
trial6 also yielded no benefit from treatment in patients
with entry diastolic BPS of less than 100 mm Hg. Thus,
while the evidence for the effectiveness of treating
patients with diastolic BPS of 95 to 99 mm Hg is some-
what stronger than in patients with borderline hyper-
tension it is by no means unanimous.

HDFP indicates that the decreased mortality in the
stepped-care patients was associated with an average
diastolic BP reduction to 83.4 mm Hg, compared with
87.8 mm Hg in the referred-care group. However, nei-
ther the Australian Hypertension Trial" nor the Vet-
erans Cooperative StudyI yielded a correlation be-
tween BP level and the incidence of complications
when the diastolic BP was 95 mm Hg or less during
treatment.

Recent Therapeutic Trials

The Medical Research Council (MRC] of Great
Britain has conducted the largest trial, with 17,354 pa-
tients.i3 Baseline diastolic BP ranged from 90 to 109
mm Hg. The study was carried out in family practices.
Patients were randomiped either to bendroflume-
thiazide, propranolol (40 to 240 mg/day) or placebo.
The number of total cardiovascular events was 352 in
the placebo group and 286 in the treated patients (p
CO.05 on sequential analysis). There was no differ-
ence, however in coronary events, the greatest reduc-
tion being in the incidence of strokes, which was lo9 in
the control group and 60 in the treated group (p <o.oI).
Retrospective subgroup analysis, a technique of quesi
tionable validity, indicated that nonsmokers who took


044   EDITORIALS

propranolol had a significant reduction in cardiovas-
cular complications (p <O.Ol), whereas the reduction c
in stroke rate was significantly greater in those receiv-
ing the diuretic (p - 9.902). Subgroup analysis by entry
diastolic BP has not been published for this study, and
it is therefore not possible to determine the effective-
ness of treatment in patients with diastolic BPS of less
than 95 or 1ess than 109 mm Hg.

The principal conclusion of the MRC trial is that
the major effect of treatment was prevention of stroke.
This result was similar to that of most other trials, in-
cluding the original'veterans Administration trial' and
the Austrelian tria1.s However, the degree of protec-
tion afforded against stroke in the MRC trial was low,
probably because the risk was low. Their results indi-
cated that it would be necessary to treat 850 mildly
hypertensive patients to prevent a single stroke over 1
year, To accomplish this a substantial percentage of
patients were subjected to chronic side effects, some of
which were more than minor. It is also of interest that
all-cause mortality wes the same in the treated end
control groups.

The European Working Party of High Blood Pres-
sure in the Elderly (EWPHE) was a smaller but well
controlled trial that included 840 patients older than
60 years with entry BP of 160 to 239/90 to 119 mm Hg."
Therefore, the study included patients with moder-
ately severe as well as mild hypertension. Total car=
diovascular mortality rate was lower in the treated
than in the placebo group. The results differed. how-
ever, from most other trials in that the effectiveness of
treatment was greater in prevention of cardiac mortal-
ity than of cerebrovascular mortality, although there
was an impressive reduction in both groups. However,
with respect to nonterminatirqqmorbid events the situ-
ation was reversed, with a greater reduction in cere-
brovasculer events than of cardiac events. While
EWPHE provides evidence for the effectiveness of
treatment, it must be recognized that this trial is not
limited to patients with mild hypertension, but in-
cludes patients with baseline diastolic levels as high as
119 mm Hg. Until more information is published by
the European Working Party on the refation between
entry BP strata end morbid events, the results of the
trial cannot be applied to patients with borderline-
mild hypertension.

A third, recently published trial, the International
Prospective Primary Prevention Study in Hyperten-
sion's did not include patients with diastolic BPS of
less than 100 mm Hg. Therefore, the results are not
germaine to the present discussion because patients
with diastolic BPS of 99 to 99 mm Hg were excluded.
Choosing Patients for Treatment

In view of these conflicting reports, what is the most
reasonable course to take in the management of pa-
tients with a diastolic BP of less than 100 mm Hg? The
1984 report of the Joint National Committeela states
that the benefit of treatment outweighs the risk in pa-
tients with a diastolic BP consistently elevated at or
above 95 mm Hg and also for those with lesser eleva-

tions who have risk factors such as target organ dam-
age, diabetes mellitus or other major risk factors for
coronary heart disease. The report states that "Opinion
varies as to the need for treatment in the 99 to 94 mm
H
fo lowed to detect any signs of progression." These
f range but if not treated they should be closely

guidelines probably represent the most reasonable
recommendations that can be made on the basis of the
current evidence. The World Health Organization also
recommends drug treatment only when the diastolic
BP is at least 95 mm Hg."

The Joint National Committee also stresses the im-
portance of recording BP during at least 3, and prefera-
sbly more, visits before making a therapeutic decision.
The exceptions, of course, are in patients with a dia-
stolic BP of 110 mm Hg or more, in whom early treat-
ment is mandatory, although if at all possible, it is
desirable even to have the patient return within a few
days of the initial visit for a second reading before
treatment. Even some of these patients may exhibit
sharp drops of BP sometimes to the normal range,
thereby calling for a longer period ,of pretreatment
observation..

Conclusions

Recommendations regarding selection of patients
for treatment have been liberalized greatly in recent
years, This has been the result of favorable evidence
accumulated by various intervention trials. However,
with respect to treatment of borderline hypertension-
a diastolic BP 90 to 94 mm Hg-the evidence for bene-
fit resulting from treatment is neither consistent nor
convincing, Patients with borderline hypertension
should be observed periodically for evidence of pro-
gression of hypertension. However, until more defini-
tive evidence of benefit becomes available, there is no
need to treat everyone in this population, especially
those without other risk factors. By using discretion,
millions of patients who are not likely to achieve much
if any benefit can be spared the potential adverse ef-
fects, inconvenience and expense of life-long drug
treatment.

References

1. Veterans Admfnistratlon Cooperative Study Croup on Antlhypertenslve
Agents. Effects rtf treetment on morbidity in hypertension. Il. ffesihs in
patients with dfostolfc bkmd pressure averaging 90 through 114 mm ffg.
IAMA 1970:213:1243-1152.
i. Frets ED. Soundfng board. Shot& mffd hypertension be treated? N Engf 1
Med 19tl2:307:3oa-309.

3. Moser U. "Lea severe" hypertension. Should it be treated? urn Ifeart \
1982;10!:465-472.           I
4. Hypertenalon DetectIon and Follow-up Program Cooperative Group. The
e@actof treolmenfan mortohty in"mtld" hypertension Rasuhsofths ffypcr-
tension Detection and Fotfow-Up Progrom. N Engl j Med 1982;207?976-980.
5. Smith WM. United States Public Health Service Ifospitols Cooperative
Study Group. Treatment of mifd hypertension. Rest&sofa ten-yeor fnterven-
lion trial. Circ flus 1977:4O:wppl 1:1-90-f-105.
6. ffelgeland A. Treetmenf of mild hypertension: o five yeor controlled drug
trial. The Osfo Study. Am 1 Med 12Otk69:72~-73~.
7. Multfple Risk Factor intervention Trfai Research Group. Multiple Risk
Factor fnterventfon Trfai; rbk /&or chenges end ~rtality rest&s. lAMA
t982:248:t48s-t477.
8. Management Committeeoftfte Australian Hypertenfson Trial. The Aurtro-
hart therapeutic trtal In mild hypertenston. honcat t~t10;~291-1~96.
0. Hypertension Detectfon and FoilowJfp Program Cooparatlvs Croup. Five


Septttrnber lb, 1988 1% hh%fVCAN JOUDNAL OF CADDKKOQY VOlVne 58     e4s

ycor fhrdingr of the hypertension detection and folfow-up program. 1. Reduc-
tion in mortality ofpereonm with hfgh blood preeeure tncfudfng mild hyperten-
sion. fAhfA ?97?4:242:2582-2572.
10. U.S. Dcpnilmcnt al Heeith end Humsn Senrkee. Public Heelth Service.
JJaolJh. Un#ted Sfotes, 1960 with Preventlort Profife. Washington. DC.: U.S.
Govt. Prinl. Off.. 190135. (DfllfS P&f No. (PIIS) ttJ-1232).
11. Menagoment CammllJeeof the Auetreillen Hypertenetan Trial. Untreated
mild hypertension. Lancer 1992:J:ltRt-191,
12. Tnguchi 1. Freta ED. Part fat redactton of blood pressure and prevention of
complicottoru in hypertension. N Engl f Med 1974:291!329-331.
13. Mctlicsf Rceeerch Councfl Working Perty. lLfRC trlof of treatment of mfld
hypertension: principal me&e. 5r Med 1 J965:2$1:97-194.
14. European Working Perry an High 5locd Pressure In the Elderly. Mortality

ond morbidity reeuhs from Europeon Worktng Potty on HfJth Slood Preeeure
In Elderly trial. Lancet J985:J~J349-1354.
19 IPPPSJI Collnborettve Group. Cardiovascular risk ond risk factors tn
o randomized trial of trsatmont baeed on the beta-blocker oxprenofol: the
tnternottonol Prospective Primary Freventfon Study in ffyportenston
(IPPPSffI. f ffyperteneion J905:3:379-392.
18. The folnt Nntional Commlttee an Detectton, Eveh~allon and Treetment ol
Htgh Bfoad Premure: the 1984 report of the foinf Nationof Committee on
Detection. Evafaafkm. and Treatment of high bland pressure. Arch Intern
h&d 1984:J44:194S-1057.
17. World Heelth Orgnnizetton/lnternatlonel Society of flyportension.
Guidelinea for the treatment of mild hypertension: memorandum from a
WHO/ISH Meeting. Lance1 J91W:J:4S7-4SS.