Reprinted from AMERICAN HEART JOURNAL. St. LOUIS
   Vol. 95, No. 5, pp. 611.fJlX. May. 1978
   ICopyrluht 5 1978 by The C. 1'. Mash? I:ompany)
    (Prmted in the 1K.4.1

Mechanism of antihypertensive effect of
thiazide diuretics

Shaukat Shah, M.D.
Ibrahim Khatri, M.D.
Edward D. Freis, M.D.
Washington, D. C.

The thiazide diurectics have become established
as the drugs of first choice in the treatment of
hypertension. Not only do they exert an indepen-
dent antihypertensive effect but they also
enhance the action of other antihypertensive
drugs and prevent the development of resistance
to various adrenergic blocking and vasodilator
agents used for treating hypertensive patients.

The mechanism of the antihypertensive effect
of the thiazides has not been clarified. Pharmaco-
logical textbooks'. Z ascribe the antihypertensive
action to two mechanisms as follows: (1) deple-
tion of extracellular fluid volume (ECF) includ-
ing plasma volume accounting for the initial fall
of blood pressure, followed by (2) decreased
peripheral vascular resistance due to a direct
vasodilator action of the drug. It is generally
believed that. the vasodilator action contributes
importantly to the long-term antihypertensive
effect of these drugs.`. " The presently accepted
theory of the antihypertensive action of the thia-
zides, therefore, defines two entirely separate
mechanisms, a diuretic and a vasodilator action.
Furthermore, they are said to act primarily in
sequence rather than together. Such a dual
theory, however, must be regarded with suspicion,
because (1) it is unusual for any drug to exert its
therapeutic effect via one mode of action initially
and via an entirely different mechanism later on,
and (2) a direct vasodilator effect of the thiazides
has never been convincingly demonstrated.

From the Department of Senior Medical Investigator, Veterans
Administration Hospital, and the Department of Medicine, George-
town University, Washington, D. C.
Received for publication April 5, 1977.
Accepted for publication July 8, 1977.
Reprint requests: Edward D. Freis, M.D., Senior Medical Investigator,
Veterans Administration Hospital, 50 Irving Street, N. W., Washing-
ton. D. C. 20422.

0002-8703/78/0595-0611$00.80/0~~~ 1978 The C. V. Mosby Co.

In the present investigation the mode of action
of the thiazides has again been investigated by
recording both the short and long-term effects of
hydrochlorothiazide. Among the hemodynamic
parameters investigated are cardiac output,
plasma volume, and ECF. Using these data as
well as previous reports in the literature a unitary
rather than the present dual theory of the mode
of action of the thiazides will be presented.

Methods

Thirteen patients with essential hypertension
of mild to moderate severity who were attending
the Hypertension Clinic of the Veterans Adminis-
tration Hospital consented to participate in the
study after being fully informed of its design and
purposes. The patients either were previously
untreated or were receiving thiazide diuretics
without other drug treatment. Prior to the
control determinations, the latter patients were
placed on placebos for a 2 to 3 week period. Pill
counts were carried out throughout the trial as a
test of compliance. One patient dropped out while
another was noncompliant as judged by the pill
counts. Both were dropped from the study.

The patients were studied on four occasions as
follows: before treatment, 48 hours after begin-
ning treatment with hydrochlorothiazide 50 mg.
twice daily to determine acute effects of the drug,
and 6 weeks as well as 8 weeks after beginning
treatment with hydrochlorothiazide for assessing
long-term effects.

During test days patients came to the labora-
tory in the fasting state. An 18 gauge Teflon
catheter was inserted into an antecubital vein to
facilitate blood sampling. Plasma volume and
extracellular fluid measurements were deter-
mined in eight patients by the method of
Gregersen and Stewart,.' adapted for use with the

American Heart Journal  611


Shah, Khatri, and Freis

21r

19c A\
171       

15' !     --.--
         I         /i----------L--i'
Control 48                 6       8
       hours               weeks     weeks

Fig. 1. Mean changes in cardiac output, blood pressure.
plasma volume, and extracellular fluid volume (thiocyanate
space) before as well as 48 hours, 6 weeks. and 8 weeks after
treatment with hydrochlorothiazide.

spectrophotometer. The procedures were carried
out after the patients had remained continuously
in the recumbent position for at least 30 minutes.
After drawing control samples of blood, 22.5 mg.
of Evans Blue dye in 5 ml. of distilled water and
18 ml. of 5 per cent sodium thiocyanate were
injected into an antecubital vein using calibrated
syringes. Blood samples were drawn from a vein
in the opposite arm for plasma volume determina-
tion at 10, 15, and 20 minutes after the injection
of Evans Blue dye. The concentrations at these
times were plotted on semilog paper and a best fit
line was extrapolated back to zero time. The
latter concentration was taken as the value
present after mixing and before any disappear-
ance of the dye.

Blood samples were drawn for thiocyanate
determinations at 2 hours following injection as
well as 10 minutes later. Thiocyanate concentra-
tion of t.he second sample was required to agree
within 5 per cent of the first. The average
concentration of these two samples corrected for
the plasma blank was used to determine the

thiocyanate space. While this :nethoi! rrt:!y
approximates the true extracellular fluid volume
since the thiocyanate ion enters red cells and
other compartments. prior experience indicated
that this method is reliable for measuring
changes in the ECF.'

Cardiac outputs were determined by the CO1
rebreathing method alone in six patients and by
the CO1 rebreathing method and the dye dilution
methods in five patients. Cardiac output was
determined by the CO, rebreathing method
before the administration of hydrochlorothiazide
and 48 hours, 6 weeks, and 8 weeks after initiating
treatment with the drug. After resting in the
supine position for at least ten minutes the blood
pressure was recorded by the same ohsrrver
throughout, using the standard arlscult,atc,t>
method. The fifth phase of the Korotkoff sounds
were used as the indication of' diwstoli(. blood
pressure. The blood pressure was ~co~~rird before
and after each determination of c,arc!iac cutput
and the rrs111~ were averaged. The initial dcter-
minations of blood pressure and cardiac output
were discarded in order to accustom the patient
to t.he procedure prior to the collection of data.
Cardiac output also was determined by the dye
method in five of the patients before beginning
treatment wit,h hydrochlorothiazide and X weeks
after beginning treatment. The significance of
differences after as compared to before treatment
was determined using Student's t test.

The technique of Franciosa and associates' was
used for estimat,ing cardiac output by the CO2
rebreathing method. Their t.echnique was modi-
fied slightly with respect to the point at which
CO, is measured in the expired air. Rather than
using forced expiration, which in our hands
frequently lead to an overestimate of alveolar
CO,, we recorded the CO, concentration in
expired air during normal or deep respiration at a
point where the expired volume approximated 400
ml. The CO, concentration and the respiratory
volume were recorded continuously and simulta-
neously. The 400 ml. point was chosen because it,
well exceeded the dead space volume of 150 to 250
ml. but was not so large as to reflect the CO,
concentration in underventilated alveoli such as
we had experienced using forced expiration.

For the dye dilut,ion method, a polyethylene
catheter was inserted into an ant,ecubital vein
and was threaded into a central vein. A Teflon
catheter was then inserted via a needle into the
brachial artery. Mean arterial blood pressure was

612

May, 1978, Vol. 95, No. 5


Antihypertensi~?e effect of thiazide diuretics

Table I. Hemodynamic changes from control at 48 hours, 6 weeks, and 8 weeks after beginning
treatment with hydrochlorothiazide

During treatment

Parameter

Control       4X Hours           6 Weeks           8 LVeerks


Mean     Mean      P*      Mean      P*      Mean      P*

Body weight (lbs.)             11
Hematocrit (`7 )               11
Plasma volume (I,.)            8
Extracellular fluid volume IL.!     8
Systolic blood pressure (mm.      1:
Ha
Diastolic blood pressure (mm.     11
Hg)
Heart rate (per min.)           11
Cardiac output (L./tnin.)        11
Total peripheral resistance       11
(Dynes cm. ,/SW.)

201
40.4
3.69
20.4
153

103

76       79
  6.4       t5.4
1479      171ti

197
43.1
:;.I?3
I a.0
1 35

95

<.005
<.005
<I .oos
r.005
CL.01

c .05

NS
< ,025
.us

78       NS
  6.4      NS
1:356       NS

<.05
<.0005
c.005
< .0005
c- .Ol

-: .025

196
42.7
3.47
15.4
131

91

77       NS
  6.0      NS
1393       NS

<.Ol
<.0005
NS
`: .005
< ,005

i .01

Table II. Simultaneous determination of cardiac output by the Cardiogreen dye and CO, rebreathing
methods

                Control period

       Cardiac   Cardiac    Mean        Total
      output    output    arterial     peripheral
Putient   CO,     Dye   pressurei      resistance
  1 0.
  v    (L.lmin.) (L.lmin.) (mm. Hg)  (Dynes-cm.~ -sec.*)


   6      7.85      6.83       115          1346
   8      7.39      6.37       151          1897
   9      8.06      6.71       113          1346
  10      5.21      4.90       133          2171
  11      7.02      7.19       92          1023
Mean     7.11      6.40       121          1557
SD.     0.11      0.89       22          466

*Dye outputs used to calculate total peripheral resistance.
?Direct ink-arterial recording of' blood pressure.

     After 8 weeks of thlazide treatment

Cardiac   Cardiac   Mean         Total
output   output   arterial     perLphera1
CO:    Dye    pressure      resistan cc
(L./minJ (L./min.) (mm. Hg)   (Dynes-cm. -sec.*)


6.80      8.37      110          1294
4.98      5.83       95          1303
6.02      6.15      105          1482
6.90      6.45      123          1525
6.22      6.01       a7          1158
6.50      6.25      104          1352
0.13      0.38       14           150

measured directly using a Statham strain gauge.
After injecting rapidly 5 mg. Cardiogreen in one
ml. distilled water into the central vein using a
flush of 5 ml. saline, blood was withdrawn from
the brachial artery through a Gilford densitom-
eter using a motor-driven withdrawal syringe.
The resulting dye-concentration curve was
recorded on a Hewlett-Packard oscillographic
recorder. Because of the more invasive nature of
the procedure three such dye curves were
recorded only during the control period and again
at 8 weeks following initiation of treatment. It
was not considered necessary to determine the
acute effects of thiazides on cardiac output with
the dye technique because such early effects as
were determined by the dye dilution method have

already been well defined.+' The dye dilution
method was used primarily to confirm the results
obtained with the noninvasive CO, method on the
effects of long-term treatment with the thia-
zides.

Results

Cardiac output, blood pressure, total periph-
eral resistance, and heart rate. Cardiac output as
determined by the CO2 rebreathing method fell
significantly (P < .05) from a mean value of 6.4 L.
per minute during the control period to 5.4 L. per
minute 48 hours after beginning hydrochlorothia-
zide (Table I, Fig. 1). However, 6 and 8 weeks
after beginning treatment with hydrochlorothia-
zide, cardiac output rose to or toward the

American Heart Journal

613


Shah, Khatri, and Freis

pretreatment values and was then insignificantly
different from the pretreatment control level.

Cardiac output determined by the dye method
in five patients agreed reasonably well with the
cardiac output determined by the CO, rebreath-
ing method (Table II). The correlation coefficient
comparing both methods was r = 0.84 (P < .05)
during the control period and r = 0.97 (P < .005)
after 8 weeks of treatment. After 8 weeks the
mean change in cardiac output from pretreat-
ment control was -0.2 L. and -0.6 L. for the dye
and CO, rebreathing methods, respectively.
Neither change was significant.

By contrast to the initial fall and later return of
cardiac output, the blood pressure fell early and
remained reduced (Table I, Fig. 1). The average
control blood pressure of 153/103 fell to 135/95
mm. Hg after 48 hours (P < .05). After 6 and 8
weeks of treatment. the blood pressure was even
lower, averaging 130/92 and 131/91 mm. Hg at 6
and 8 weeks, respectively (P < .025).
Total peripheral resistance increased initially
from a mean of 1479 dynes-cm. `-sec. in the
control period to 1716 dynes-cm. "-sec. 48 hours
after beginning hydrochlorothiazide. The change
was of borderline significance (P = .09). After 6
and 8 weeks, total peripheral resistance decreased
slightly but not significantly below the control
level, the mean values being 1356 and 1393 dynes
cm. i per second respectively at the 6 and 8 week
post-treatment periods.

Heart rate did not change significantly
throughout the treatment period. The mean
value for heart rate rose insignificantly from 76
per minube in the pretreatment period to 79 per
minute at 48 hours after beginning treatment.
The mean values for heart rate at 6 and 8 weeks
after beginning treatment remained within the
range of these values.

Extracellular fluid volume, plasma volume,
hematocrit, and body weight. ECF measured in
eight patients by the thiocyanate method fell
significantly and remained reduced throughout
the period of observation (Table I). From a mean
control value of 20.4 L. the ECF decreased 12 per
cent to 18.0 L. at 48 hours after beginning
hydrochlorothiazide. The reductions in ECF were
even greater at 6 and 8 weeks, averaging 20 and 24
per cent, respectively, below the control mean.

Plasma volume decreased from a mean of 3.69
L. in the control period to 3.23 L. 48 hours after
beginning treatment with hydrochlorothiazide.

The decrease averaged 13 per cent and was highly
significant (P < .005). Plasma volume remained
reduced at 6 and 8 weeks after treatment with
hydrochlorothiazides, although the change at-
tained statistical significance only at the 6 weeks
interval.

Reflecting the decrease in plasma volume the
hematocrit rose and remained significantly above
the control mean throughout the 8 week period of
treatment. The mean value in the control period
was 40.4 and this rose to 43.1 after 48 hours of
treatment, an increase of 7 per cent. The mean
values at 6 and 8 weeks of treatment were essen-
tially the same as at 48 hours (Table I).

Body weights decreased at 48 hours after t,reat-
ment and remained significantly below the
control for the remainder of the period of obser-
vation. The average loss of weight was 4 pounds
or 1.8 kilograms at 48 hours, increasing to 5 and 6
pounds (2.3 and 2.7 kilograms), respectively, at 6
and 8 weeks after beginning treatment.

Discussion

Three principal hypotheses have been ad-
vanced to explain the antihypertensive effects of
the thiazide diuretics. The first two hypotheses
are concerned with sodium depletion, of which
one ascribes the fall in blood pressure to a reduc-
tion in total ECF and the other to a loss of
sodium from the arterial walls. The latter hy-
pothesis probably can be discarded, because
studies from three independent sources found no
evidence that thiazides affect the sodium content,
of either large or small arteries in rats.`"' The
third and most popular hypothesis is that the
hypotensive effect, at least during chronic treat-
ment, is due to a direct vasodilat,or effect of these
drugs.`. "

Hollander, Chobanian, and Wilkins" found no
reduction in extracellular fluid volume after one
to six months of treatment with chlorothiazide.
On the basis of this and other data, they
concluded that the antihypertensive effect was
not due solely to sodium depletion. Their conclu-
sion gained support from the observation that the
chemically related compound, diazoxide,' i is a
vasodilator antihypertensive compound which
produces sodium retention rather than diure-
sis.' `. " According to Rowe and colleagues,`,' acute
administration of diazoxide is associated with a
considerable decrease in total peripheral
resistance in hypertensive patients and with a

614

May, 1978, Vol. 95, No. 5


Antihypertensive effect of thiazide diuretics

reflex increase in heart rate and cardiac output
similar to that seen with other vasodilator drugs,
such as hydralazine or amyl nitrite. Acute admin-
istration of the thiazides, however, is associated
with the opposite hemodynamic effects, that is,
with an increase in total peripheral resistance and
a fall in cardiac output.:. ' Also the initial obser-
vation" that hypotension preceded the diuresis
has not been confirmed, as three other investiga-
tions concluded that the fall in blood pressure
occurred only after an effective diuresis."`-`"
Furthermore, a direct vasodilator effect of the
thiazide diuretics has not been convincingly
demonstrated.

The vasodilator theory of the action of the
thiazides has been further questioned by the
observation that other diuretics, such as paren-
teral mercurials' ' and ethacrynic acid,"' also
exhibit an antihypertensive effect. These drugs
have in common volume depleting effects rather
than vasodilator properties. Also, the reduction of
blood pressure occurring in patients receiving the
low-sodium "rice diet" is associated with a signif-
icant reduction in ECF."`, "I All of the above
antihypertensive procedures decrease ECF, sug-
gesting that it is the latter which is causally
implicated in the reduction of blood pressure in
both the acute and chronic responses.

Volume depletion following acute administra-
tion of the thiazides has been documented by
many investigators. `. I'%, `"~ ii% Patients undergoing
continuous treatment with thiazides lose about 2
L. of ECF and about 300 to 400 ml. of plasma
volume during the first 72 hours of treatment.
After this time there is no additional volume
depletion. However, the reduction in ECF which
occurs during the first 72 hours is maintained.
Patients receiving diets severely restricted in
sodium show quantitatively similar reductions
averaging 10 per cent of the ECF and 15 per cent
of the plasma volume.`", A'

The importance of reduced volume is again
indicated by observations that the antihyperten-
sive effects of combined treatment with thiazide
diuretics and adrenergic blocking agents can be
reversed in many patients by re-expanding the
plasma volume with Sal&free dextran.";, `:. `j
However, in patients treated with thiazides
without adrenergic blocking agents, dextran will
only occasionally but not usually restore the
hypertension." Nevertheless, oral ingestion of 20
to 30 Gm. salt, which is sufficient to re-expand

total ECF despite the administration of thiazides,
does restore the hypertension even during long-
term treatmentZ"' This suggests that re-expan-
sion of total ECF rather than plasma volume
alone is required to overcome completely the
antihypertensive effects of the thiazides.

Three possible mechanisms can be postulated
to explain how thiazide-induced reduction of ECF
lowers blood pressure. Firstly, volume depletion
may blunt environmental pressor stimuli while at
the same time enhancing depressor influences.
For example, patients receiving ganglion-blocking
agents  often  exhibit  marked hypotensive
responses with only small reduction of blood
volume." It has further been observed that
following either thiazide or mercurial-induced
diuresis, the depressor responses to infused tri-
methaphan are increased.`, Also, expansion of
ECF in rats results in an enchanced pressor
response to angiotensin whereas ECF depletion
has the opposite effect."' Volume depletion, there-
fore, seems to dampen pressor reactions and
increase depressor responses, the combined effects
of which could result in a reduction of average
blood pressure. While this theory is attractive, it
does not explain the rise in cardiac output toward
control values and fall in total peripheral
resistance which occurs with long-term treat-
ment.

The second possible mechanism by which
volume depletion may lower blood pressure is via
the baroreceptor reflexes. It has been noted clini-
cally that elderly patients often exhibit an
enhanced hypotensive response to the thiazide
diuretics requiring reduction of dosage, often to
quite small amounts. On the other hand, young
normal adults do not show an antihypertensive
effect with the thiazide diuretics, although their
heart rates increase significantly consistent with
an active barorreceptor response to the lower
blood pressure associated with the volume deple-
tion." Elderly normal subjects, however, do show
a fall of blood pressure with the thiazide
diuretics. j"

Baroreceptor reflex responses depend upon the
compliance of the aorta and carotid artery. These
vessels become less distensible as a result of
structural changes induced by aging or by hyper-
tension. Homeostatic circulatory adjustment to
pressor stimuli are poorly moderated in long-
standing chronic hypertension.,"  While the
evidence is incomplete, it is consistent with the

American Heart Journal

615


Shah, Khaki, and Freis

concept that the antihypertensive effect of the
thiazide diurectics, at least in patients with
noncompliant arteries, could be due in part to an
inadequate baroreceptor compensation for the
volume depletion and reduced cardiac output.

The third, and we believe most likely mecha-
nism, for explaining the hemodynamic changes
seen with the thiazide diuretics invoices feedback
autoregulatory responses to a prolonged reduc-
tion in cardiac oubput as originally suggested by
Tobian:", .I' Such feedback autoregulatory sys-
tems have been postulated to expiain the rise of
blood pressure which occurs in terrain types of
experimental hypertension. In experimental reno-
vascular hypertension " or in the hypertension
caused by reduct,ion in renal mass plus salt
loading,,' ) or in the hypertension induced by
excess licorice ingestion in man,"' the following
sequence of events was observed during the devel-
opment of the hypertension: (1) salt and water
retention, leading to (2) expansion of the ECF
and plasma volume, resulting in (3) increased
venous pressure and venous return to the heart,
followed by (4) increased cardiac output, (5) rise
in blood pressure, leading to (6) increase in urine
volume preventing further rise in ECF, followed
by (7) gradual elevation of total peripheral
vascular resistance occurring over a period of
weeks. It was postulated that the rise in total
peripheral resistance was due to autoregulation in
response to the elevated peripheral blood flows
which were in excess of local metabolic needs
because of the increased cardiac out,put. At the
same time, there was (8) reduction of the elevated
cardiac output toward normal because of the
elevated afterload resulting from the increase in
total peripheral resistance. Thus, the hyperten-
sion which was initiated by an expansion of ECF
and resulting increase in cardiac output evolved
after approximately one month into a hypert,en-
sion characterized by an elevat,ed total peripheral
resistance and normal cardiac output.

Tobian (1. 1 L has postulated that, thiazides and
related diuretics induce "reverse autoregulation"
which leads to a decrease in total peripheral
resistance. Thus, reduction in ECF caused by
thiazides produce initially a fall in central venous
pressure, venous return, and cardiac output. At
first, this is associated with an increased total
peripheral resistance. However, with continued
treatment and maintained reduction in cardiac
output and total blood flow, autoregulation

occurs gradually over a period of weeks leading to
a decline in total peripheral resistance. Finally,
we suggest t,hat the resulting fall in afterload
permits a rise in cardiac output, toward normal.
This sequence of hemodynamic events is the same
as that actually observed during the early and
late phases of continuous treatment with the
thiazides, and they all follow from a single hemo-
dynamic effect of the thiazides which is a
sustained reduction in ECF. No direct vasodilator
action of the thiazides need to be postulated, as
the early fall and later return of cardiac output to
normal and the early rise and later fall in total
peripheral resistance can both be explained by the
reduction in ECF alone.

Most pharmacological textbooks differentiate
between the short-term and long-term effects of
thiazides.' According to these texts the short-
term antihypertensive effect is due primarily to
volume depletion while the long-term effect is
ascribed to a vasodilator action of the thiazides.
The principal evidence for such a dual mechanism
of action is the report of Conway and Lauwers,"
which indicates that plasma volume and ECF fall
initially but then return to essentially control
values after one month of continuous treatment
with chlorothiazide. These investigators con-
cluded that decreased ECF could not have an
influence on the long-term antihypertensive
effect of the thiazides.

Other investigators, however, have not corrobo-
rated the observations of Conway and Lauwers.
Wilson and Freis' found a significant reduction of
ECF after six months of continuous treatment
wit.h chlorothiazide. Plasma volume also was
reduced at six months, although not significantly.
However, at 12 mont,hs plasma volume was signif-
icantly lower than control. More recently, other
investigat,ors also found a continued reduction of
plasma volume and ECF during long-term treat-
ment with thiazides. `*m "' Furthermore, following
withdrawal of long-term treatment there was a
prompt and significant rise of plasma volume,
ECF, and blood pressure, providing further
evidence that the thiazides continue to maintain
their volume-reducing effect during chronic
administration. `. "

Conway and Lauwers:`: also observed that
while cardiac output falls initially after adminis-
tration of thiazide, it returns to normal after one
month. The latter observation has been con-
firmed in the present study and by Lund-Johan-

616

May, 1978, Vol. 95, No. 5


,-I t! iihypertensiue e,fecf of thinzitle diuretics

sen."' Ijlli'i:lg the fir71 fen days of ir~rmi~1~1.
cardiac output is reduced and t otai ~~eriphrrwi
resistance is increased?' However. al'ter a month
or more of trratmrnt, cardiac OII tput pi+- i:? :!:
toward c,ontrol values uhile total Iw:ipht~:.::!
resi-rance falls, These changes :`an !XS t .~pIaitl~d
by reverse autoregulation without Invoking a
separate  \2sodilator  effect  of the thiazide
diuretics.

AutoregulaGon may not be limited ta the thi::
zides alone. Other antihyperttbnsive agents frL(~,
induce differing hemodynamic effects during
acute and chronic administratior;. (.;il:lnc;lli:!ine.
an adrenergic blocking agent, (`auses an in it in! !`:I f 1
in cardiac output hut with ll,ng-term treatmrnt
cardiac output rises gradualI:; hil<`k ! r>icai.<.i
normal. " Hydralazine, a vasodilator agent,
initiall?, increases cardiac output but the iattet
rei IIMS toward normal during chronic trr:a t-
merit:- These observations suggest that aurorrg-
uhtmy  responses  which return the cardiac
olltpllt toward normal often occur during long-
term administration of antihypertensive agents.

Summary

Hemodynamic studies were carried out before
and during 8 weeks of treatment with hydrochio-
rothiazide 50 mg. twice daily in 11 hypertensive
patients. Forty-eight hours after beginning treat-
ment there was a significant reduction in blood
pressure, cardiac output, plasma volume, and
extracellular fluid volume (thiocyanate space)
while total peripheral resistance increased. After
6 and 8 weeks of treatment, the blood pressure
and the plasma  and extracellular volumes
remained reduced. However. total peripheral
resistance fell while cardiac output rose to control
levels. These results were consistent with the
"reverse autoregulation" theory of the action of
the thiazides as proposed by Tobian. The present
findings as well as other clinical and experimental
evidence discussed below makes it appear unlike-
ly that the thiazides have an important direct
vasodila tor effect.

REFERENCES

1. Aviado. D. M.: Pharmacologic principles of medical
practice, ed. 8, Baltimore, 1972. The Williams & Wilkins
Company. p. 54'.

2. Nickerson, M., and Ruedy, J.: Antihypertensive agents
and the drug thrrapy of hypertension, in The pharmaco-
logical hasis of therapeutics, edited by Goodman, L. S.,
and Gilman, A.. New York, 1975 McMillan Publishing
Co.. Inc.. p. 712.

H

1:).



I!.

I".



1X

14.

15.



j 6.

17.

18.



19.


20.



21.

1 ;:q+:,.~:+n. M. I., and Stewart, J. I).: Simultaneous
de:crminati*;r? of plasma volume with T-1824 and "avail-
chic, fi::i:i" volume with sodium thiocganate, Am. ,i.
/`hy.wd 175.' 2.2. l9X.
~\t-ilk::;. I ?I.. and Freis, E. D.: Relationship between
pia:+;i.,? :~;r: c r.!raceilular fluid volume depletion and the
.yu?ii; !it-l if.!lhi\e ef'ect of chlorothiazide, Circulation
20: l!L&. 133.

Franciosa. .J.. Kagan, D. O., and Rubenstone, S.: Valida-
:~OII of CO rrhreuthing method for measuring cardiac
output in patient!: with hypertension or heart failure. .i.
IJab. Clin. >lrd. 88:67?. 1976.
c:?os!r~: ;I. I'.. Jr., Castillo, C.. Freeman, D. ,J., White, 1).
H., .Jr.. and Rowe. (+. G.: The acute effects of carbonic
w~d~yr!~-ase in b:t:it io!! on systemic hemodynamics, J. Clin.
Irr\e.ic. 37:HH- I. !958.
I)ustan, H. I'., Gumming, G. H., and Corcoran, A. C.: A
mechanism of chlorothiazide-enhanced effectiveness of
antihypertension  ganglioplegic  drugs,   Circulation
19~:xi~~. :959.

Frohlic,h. E. I)., Schnaper, H. W., Wilson, I. M.. and Freis,
K D.: Hcmodynamic alterations in hypertensive patients
rlue trj chlurothiazide, N. Engl. .J. Med. 262:3X1, 1960.
Tobian. L., .Janecek, .J., Faker, d., and Ferreira, I).: Effect
ot chlorothiazide on renal juxtaglomerular cells and
tissue electrolytes, Am. J. Physiol. 202:905, 1962
I)nnirl, E. E.: On the mechanism of the antihypertensive
action, of h.~riroc,hlorrlthiazide in rats, Circ. Rrs. 11 :941.
lUtj2.

Weller., .J. M., and Haight, A. S.: Effect of chlorothiazidr
on blood pressure and electrolytes of normotenrive and
hypertensive rats. Prrx,. Sot,. Exp. Rio]. Med. 11 2:880,
1963.

Hollander. !l'.. Chobanian, A. V., and Wilkins. R. W.:
Stud& on the antihypertensive action of chlorothiazidr.
Clin. Rrs. 6:21. 1958.

Rubin. A. A., Roth, F. E., Taylor, R. M., and Rosenkilde.
H.: Pharmacology of tliazoridr. an antihypertensive
nondiuretl,' hrnzothintliazinr, .J. Pharmacol. Exp. Ther.
136:244, 1%".
Bartorc~lh, (`., Gaqano, N., Leonetti, G., and Zanchetti,
`2.: H~]~~~tcwwivr and renal effects of diazoxidr, a sodium-
retamlng  t)c~~l~c~thi:~diazine compound,  Circul._ltion
27:H95, l!JKj.
Rowe, CG. G.. I,eicht, T. K., Boake, M'. C.. Kyle, .J. C., and
Crumptw. (`. W.: The systemic and coronarv hemodv-
narnil: effects of diazoxide, A>t. HRAHT .J. 66:&s& 196;i.
Fws. E. Ii.: The effects of salt ant1 csxtracellular fluid
depletion on vascular responsiveness with particular
reference to chlorothiazide, in Hypertension. Proceedings
of the Council on High Blood Pressure Research, Novem-
ber, 1958, edited bg Skelton, F. R.. New York. 1959.
American Heart Association, p. 6.
Dollery, C. T., Harrington. M., and Kaufmann, G.: The
mode of action of chlorothiazide in hypertension with
special reference to potentiation of ganglion-blocking
agents, Lancet 1:1215, 1959.
Macleod. C., and Dustan, H. P.: Sequential changes
evoked by chlorothiazide in hypert,ensive patients, Arch
Intern. Med. 106:316, 1960.
Conway, J., and Leonetti, G.: Hypotensive effect of
ethacrynic acid, Circulation 3 1:661, 1965.
Murphy, R. J. F.: The effect of "rice diet" on plasma
volume and extracellular fluid space in hypertensive
subjects, ,J. Clin. Invest. 29:912, 1950.
Wat.kin, D. M., Fraeb, H. F., Hatch. F. T., and Gutman.
A. B.: Effects of diet in essential hypertension II. Results
with unmodified Kempner rice diet in fifty hospitalized
patients, Am. d. Med. 9:441, 1950.

American Heart Journal                                                         617


Shah, Khaki, and Freis

2".

"3.

24.


25.


26.

27.

28.

29.

30.

31.

32.

33.

Tapia, F. A., Dustan, H. P., Schneckloth, R. A., Corco-
ran, A. C., and Page, I. H.: Enhanced effectiveness of
ganglion-blocking agents in hypertensive patients during
administration of a saluretic agent (chlorothiazide),
Lancet 2:8X, 1957.
Freis, E. D., Wanko, A., Wilson, I. M., and Parrish, A. E.:
Chlorothiazide in hypertensive and normotensive pa-
tients, Ann. N. Y. Acad. Sci. 71:450, 1958.
Wine,, B. M.: The antihypertensive actions of benzothia-
diazines, Circulation 23:211, 1961.
McQueen, E. G., and Morrison, R. B. I.: The hypotensive
action of diuretic agents, Lancet 1: 1209, 1960.
Ronnov-Jessen, V.: Variations in blood volume during
treatment  of hypertension with ganglion-blocking
agents and thiazide perparation, Acta Med. Stand.
170:363, 1961.
Freis, E. D., Finnerty, F. A., Jr., Schnaper, H. U'.,
Stanton, J. R., Rath, C. S., Johnson, R. L., and Wilkins,
R. W.: The collapse produced by venous congestion of
the extremities or by venesection following certain hypo-
tensive agents, J. Clin. Invest. 30:453, 1951.
Wanko, A., and Freis, E. D.: Altered vascular responsives
following chlorothiazide or mercurial diuresis in normo-
tensive subjects, Circulation 18:792, 1958.
Reid, W. D., and Laragh, J. H.: Sodium and potassium
intake, blood pressure and pressor response to angioten-
sin, Proc. Sot. Exp. Biol. Med. 120:26, 1965.
Hoobler, S.: Antihypertensive effects of diuretics, in
Antihypertensive therapy principles and practice, edited
by Gross, F., Heidelberg, 1966, Springer-Verlag, p. 256.
Finnerty, F. A., .Jr., Tuckman, J., and Hajjar, G.: Change
in heart rate during levarterenol infusion, an index of
arterial elasticity, Circ. Res. 7:565, 1959.
Tobian, L.: How sodium and the kidney relate to the
hypertensive art.eriole, Fed. Proc. 33:138, 1974.
Tobian, L.: Hypertension and the kidney, Arch. Intern.
Med. 133:959, 1974.

34.

35.

36.

37.

38.

39.





40.



41.





42.

Ledingham, J. M.: Distribution of mater, sodium and
potassium in heart and skeletal muscle in experimental
renal hypertension in rats, Clin. Sci. 12337, 1953.
Guyton, A. C. Coleman, T. G., Cowley, A. W.. Manning,
R. D., Jr., Norman. R. A., and Ferguson, J. D.: A systems
analysis approach to understanding long-range arterial
blood pressure control and hypertension, Circ. Res.
35:159, 1974.
Borst. J. G. G., and De Borst, G. A.: Hypertension
explained by Starling's theory of circulatory homeosta-
sis, Lancet 1:677, 1963.
Conway, J., and Lauwers, P.: Hemodynamic and hypo-
tensive effects of long-term therapy with chlorothiazide,
Circulation 21:21, 1960.
Tarazi, R. C., Dustan, H. P., and Frohlich, E. D.: Long-
term thiazide therapy in essential hypertension, Circula-
tion 41:709, 1970.
Leth, A.: Changes in plasma and extracellular fluid
volumes in patients with essential hypertension during
long-term treatment with hydrochlorothiazide, Circula-
tion 42:479, 1970.
Lund-Johansen, P.: Hemodynamic changes in long-term
diuretic therapy of essential hypertension, Acta. Med.
Stand. 187:590, 1970.
Villarreal, H., Exaire, J. E., Rubio, V., and Davila, H.:
Effect of guanethidine and bretylium tosylate on
systemic and renal hemodynamics in essential hyperten-
sion. Am. J. Cardiol. 14:633, 1964.
Reubi, F. C., Gossweiler, N., Butikofer, E., and Blumberg,
A.: Quelques aspects du reajustement circulatoire au
tours du traitement antihypertensif de longue duree, in
International Club on Arterial Hypertension. Premiere
Reunion, July 5-7, 1965, edited by Millez, P., Tcherda-
koff, P., Paris, 1966, L`Expansion Scientifique Francaise,
p. 419.

618                                            May, 1978, Vol. 95, No. 5