Hypertensive Patients Using

Thiazide Diuretics as PrimaryStroke Prevention Make Better Functional Outcome after

Ischemic Stroke

Hong-Mo Shih, MD, Wei Chun Lin, MD, Cheng-Hsien Wang, MD,

and Leng-Chieh Lin, MD

From the Departmen

Memorial Hospital, Chia

Medicine, Puzi City, Chia

Received March 28, 20

May 20, 2014.

Supported by grants

690443, CMRPG 690431,

the Chang Gung Medica

Address corresponden

cian of Emergency Med

Chang Gung Memorial H

College ofMedicine, Chia

County 613, Taiwan (R.O

1052-3057/$ - see front

� 2014 by National Str

http://dx.doi.org/10.1

2414

Background: Thiazides have been used for the control of blood pressure and primary

prevention of ischemic stroke. No previous studies have assessed the influence of

thiazides on functional prognosis after ischemic stroke. Methods: Demographics,

prestroke conditions, poststroke National Institutes of Health Stroke Scale score,

and clinical and laboratory parameters were prospectively registered in 216 Taiwa-

nese patients. One hundred forty patients who completed follow-up 3 months after

experiencing ischemic stroke were assessed with the modified Rankin scale as

functional prognoses. Results: Twenty-one patients used thiazide to control hyper-

tension before experiencing ischemic stroke. No differences of stroke subtypes

and comorbidities before stroke were observed between the 2 groups. The emer-

gency department National Institutes of Health Stroke Scale was lesser among thia-

zide users (4 [2-7] versus 6 [4-16], P 5 .02). Among 140 patients who completed

follow-up in 90 days, thiazide users had more favorable functional status (modified

Rankin scale#2: 42.4% versus 26.9%, P5 .02, odds ratio 3.34, 95%, confidence inter-

val .130-.862). Conclusion: Hypertensive patients treated with thiazides long term

had a lesser severity of stroke and better functional outcomes after ischemic

stroke. Key Words: Thiazide-diuretics—ischemic stroke—hypertension—

functional prognosis—stroke in evolution—stroke primary prevention.

� 2014 by National Stroke Association

Introduction

Ischemic stroke is one of the major causes of death

worldwide. Epidemiologic studies have revealed that

arterial hypertension is the most likely risk factor for

t of Emergency Medicine, Chang Gung

yi and Chang Gung University College of

yi County, Taiwan (R.O.C.).

14; revision received May 15, 2014; accepted

CMRPG690441, CMRPG690442, CMRPG

CMRPG 690432, and CMRPG 690433 from

l Research Council.

ce to Leng-Chieh Lin, MD, Attending Physi-

icine, Department of Emergency Medicine,

ospital, Chiayi and Chang Gung University

yi No. 6,W. Sec., Jiapu Rd., Puzi City, Chiayi

.C.). E-mail: a3456711@ms65.hinet.net.

matter

oke Association

016/j.jstrokecerebrovasdis.2014.05.021

Journal of Stroke and Cerebrov

stroke.1,2 Approximately 54% of stroke and 47% of

ischemic heart disease cases worldwide are attributable

to high blood pressure.1 The risk of hypertension in

ischemic stroke patients, however, can be reversed with

antihypertensive drug treatment, especially in elderly pa-

tients who have isolated systolic hypertension.3,4

Thiazide diuretics have been amainstay of essential hy-

pertension therapy. A recent Cochrane review of 19 ran-

domized controlled trials demonstrated that thiazide

diuretics reduced overall mortality and risk of stroke.5

In the Systolic Hypertension in the Elderly Program,

chlorthalidone caused a 36% reduction in the incidence

of stroke.3 Without contraindication, thiazide diuretics

may serve as a first-line antihypertensive medication for

primary stroke prevention.6

Only a small amount of medical literature has been

published about the prognosis of long-term thiazide users

after acute ischemic stroke. We performed this prospec-

tive observational study to evaluate the severity and

ascular Diseases, Vol. 23, No. 9 (October), 2014: pp 2414-2418

STROKE IN THIAZIDE USERS 2415

functional outcome after ischemic stroke in patients who

used thiazides to control hypertension.

Methods and Materials

Study Design

This was a prospective observational study of consecu-

tive patients admitted to the Chia-Yi Chang Gung Memo-

rial Hospital, a stroke referral center, in the period from

between October 2007 and June 2010. All patients with a

measurable National Institutes of Health Stroke Scale

(NIHSS) score were recruited to participate. In addition,

patients transferred from other hospitals also were

eligible for enrollment. Patients were excluded if they

had a previous acute ischemic stroke, unknown drug

history, recorded time .12 hours between the onset of

neurologic symptoms and presentation to the emergency

department (ED), evidence of hemorrhagic stroke

assessed by brain computed tomography (CT), fibrino-

lytic therapy requirement, diagnosed transient ischemic

attack, or lack of ischemic stroke diagnosis on discharge.

The study protocol was approved by the Chang Gung

Memorial Hospital institutional review board. Written

informed consent was obtained from all approached

patients.

Diagnostic Studies

Data were collected from the ED participants via a

standardized data collection form, including age, sex,

arterial blood pressure on ED admission and every

8 hours for the first 3 days, admission blood glucose,

complete blood count, blood urea nitrogen, creatine,

blood urea nitrogen/creatine ratio, triglycerides, and total

cholesterol. All patients received brain CT scans within

6 hours of ED admission. A radiologist from the stroke

team interpreted the CT findings. Previous medication

histories also were reviewed.

Clinical Assessment

Stroke severity was assessed via the NIHSS and

performed by physicians trained in NIHSS assessment.

This assessment was carried out immediately after a

patient’s ED admission and every 24 hours within the first

3 days during hospitalization. NIHSS of ED and 24 hours

after admission were compared to see whether there was

early improvement clinically. Stroke in evolution (SIE)

was diagnosed in those patients who experienced

worsening neurologic condition as indicated by an

increase of 4 or more points on the NIHSS within 72 hours

after stroke onset. Stroke subtype using TOASTclassifica-

tion was made before discharge after surveyed possible

cause of stroke. Neurologists evaluated patient functional

independent status with modified Rankin scale (mRS)

3 months after stroke. A good functional outcome was

defined as mRS #2.7

Statistical Analysis

Patients were divided into 2 groups according to use or

nonuse of thiazides. The differences between the 2 groups

were analyzed using the chi-square (c2) test for

categorical variables; they were expressed as frequencies

and percentages. Continuous data were analyzed with

independent two sample t tests for continuous variables

expressed as a mean 6 SD. NIHSS was analyzed with

the Mann-Whitney U test and presented via median (in-

terquartile range). Variables that were associated with a

prognosis of stroke, including age, sex, stroke subtype

(lacunar or nonlacunar), risk of cardiac embolism (atrial

fibrillation or congestive heart failure), use or nonuse of

thiazide, SIE, and glucose level on admission, were used

in multivariate analysis.8 All statistical assessments

were 2-sided, and differences with P-values , .05 were

considered statistically significant. Statistical analyses

were performed using SPSS 17.0 statistical software

(SPSS Institute, Chicago, IL).

Results

Therewere 216patientswhomet the enrolled criteria and

agreed to join this study.Only 140patient completed follow-

up in90days.Patientsweredivided into2groupsaccording

to their thiazide use for controlling blood pressure. Twenty-

one patients used thiazides. The mean age of 140 patients

was 70.1 6 10.3 years of age with a range between 46 and

96 years of age. Table 1 summarizes both the demographic

and clinical characteristics of 21 thiazide users and 119 thia-

zide nonusers. There were no differences found between

the 2 groups inmultiple variables, including associated un-

derlying disease, age, and sex, as well as clinical features,

including blood pressure measurements, heart rate, blood

glucose levels, and triglyceride levels.

Table 2 compares initial NIHSS at ED, better NIHSS

when followed 24 hours after admission, SIE within

72 hours, favorable outcome as mRS #2 90 days after

stroke, and stroke subtypes using the TOAST (ie, Trial of

Org 10172 in Acute Stroke Treatment) classification. There

was nodifference in stroke subtypes between the 2 groups.

The ED NIHSS was lower among thiazide users (4 [2-7]

versus 6 [4-16], P 5 .02). There also was more neurologic

improvement at 24 hours after admission and less SIE

within 72 hours in thiazide users, although not statistically

significant.

There was a statistically significant increase in favor-

able function outcome among thiazide users as seen by

the mRS followed at 90 days later (mRS #2: 42.4% versus

26.9%, P5 .02, odds ratio [OR] 3.34, 95%, confidence inter-

val [95% CI] .130-.862).

Table 3 shows the result of multivariate logistic regres-

sion of variables that may influence functional outcome of

stroke patients, including age, sex, use or nonuse of thia-

zides, glucose levels on admission, SIE $4, the presence

Table 1. Demographic data, clinical feature and

comorbidities

Demographic

Thiazide

nonusers,

n 5 119

Thiazides

users,

n 5 21

P-value

Age* 70.6 6 10.5 67.4 6 9.6 .199

Maley 71 (59.7%) 15 (71.4%) .307

Diabetes

mellitusy54 (45.4%) 13 (61.9%) .162

Hyperlipidemiay 7 (5.9%) 2 (9.5%) .530

Coronary artery

diseasey8 (6.7%) 1 (4.8%) .736

Atrial

fibrillationy30 (25.2%) 6 (28.6%) .745

Systolic

blood pressure,

* mmHg

169.7 6 35.1 178.5 6 33.5 .289

Diastolic blood

pressure,

* mmHg

91.1 6 21.3 93.5 6 17.5 .623

Heart rate,

* beats/min

82.2 6 16.8 80.0 6 17.2 .589

Hemoglobin,

* g/dL

13.4 6 2.1 13.8 6 1.9 .274

Glucose,* mg/dL 161.7 6 95.1 191.2 6 91.1 .190

Triglycerides,*

mg/dL

118.7 6 66.8 117.6 6 69.8 .948

Cholesterol,

* mg/dL

178.3 6 44.4 168.2 6 5.36 .384

Uric acid,

* mg/dL

5.4 6 1.8 6.0 6 1.4 .207

BUN/Cr . 15y 51 (42.9%) 5 (23.8%) .100

Abbreviations: BUN, blood urea nitrogen; Cr, creatine.

*Continuous data expressed as mean 6 SD.

yCategorical data expressed as number (%).

Table 2. Comparison of stroke subtypes and stroke severity,

functional outcomes

Stroke subtypes,

severity and

outcomes

Thiazide

nonusers,

n 5 119

Thiazide

users,

n 5 21 OR

P-value

ED NIHSS* 6 (4-16) 4 (2-7) .02

NIHSS 24

hours , ED

25 (21%) 7 (33.3%) .26

SIE $4 17 (14.3%) 1 (4.8%) 3.33 .310

mRS #2* 32 (26.9%) 11 (42.4) .02

Large-artery

atherosclerosis

26 (21.8%) 4 (19.0%) 1.19 .773

Cardioembolism 16 (13.4%) 2 (9.5%) 1.47 .621

Lacunar stroke 33 (27.7%) 7 (33.3%) .77 .607

Stroke of

undetermined

etiology

44 (37%) 8 (38.1%) .95 .922

Abbreviations: ED, emergency department; NIHSS, National In-

stitutes of Health Stroke Scale; mRS, modified Rankin Scale; OR,

odds ratio; SIE, stroke in evolution.

Categorical data are expressed as number (%); NIHSS presented

with median (interquartile range).

*P , .05.

Table 3. Multivariate logistic regression analysis of

favorable functional outcome in 90 days (mRS # 2)

H.-M. SHIH ET AL.2416

of lacunar stroke, and the risk of cardioembolic stroke.8

After multivariate adjusted analyses, age (OR 1.084, 95%

CI 1.037-1.134, P , .001), cardioembolic stroke (OR .539,

95% CI .296-.984, P 5 .044), and use of thiazides (OR

.326, 95% CI .111-.956, P 5 .041) remained independent

predictors.

Variables

Multivariate (adjusted)

OR 95% CI P-value

Age* 1.084 1.037-1.134 ,.001

Sex 1.165 .487-2.785 .732

Thiazide user* .326 .111-.956 .041

Glucose 1.004 .999-1.009 .164

SIE $4 6.762 .815-56.119 .077

Lacunar stroke .872 .640-1.189 .386

Cardioembolic stroke* .539 .296-.984 .044

Abbreviations: CI, confidence interval; mRS, modified Rankin

Scale; OR, odds ratio; SIE, stroke in evolution.

*P , .05.

Discussion

We found that long-term thiazide uses who experi-

enced their first acute ischemic stroke experienced it

less severely, with lower initial NIHSS and better 3-month

mRS score results. No previous studies have assessed the

severity of stroke and functional outcomes after ischemic

stroke in long-term thiazide users. The exact mechanism

to achieve this benefit remains unclear; we have discussed

possible contributing mechanisms subsequently.

Several factors may contribute. First, long-term thiazide

users may have a better collateral cerebral blood supply

after ischemic stroke that can reduce infarct volumes

and lead to lesser stroke severity and favorable functional

outcomes. Long-term thiazide therapy lowers blood

pressure through a vasodilation effect,9,10 and this

vasodilation effect persists for weeks after stopping

thiazide therapy.11 As Fujii et al12 discussed in their study,

treating hypertension with hydrochlorothiazide can lead

to smaller infarct volumes. Thiazides may restore the dila-

tory capacity of cerebral arterioles, consequently

improving the cerebral blood flow reserve and the collat-

eral vascular capacity after acute ischemic stroke.12 This

collateral cerebral blood supply may attenuate arterial oc-

clusion effects associated with infarct growth.13-16 As a

result, patients with long-term thiazide use might have

better collateral circulation and relative smaller infarct

STROKE IN THIAZIDE USERS 2417

volume, which accounted for less stroke severity and bet-

ter functional outcome, as seen in our study. Although no

statistically significant, the greater ratio of decreased

24 hours NIHSS among thiazides users also may support

our hypothesis about better collateral circulation.

On the other hand, the effect of fluid retention from

thiazide withdrawal may help provide adequate postis-

chemic stroke cerebral circulation. Increased blood pres-

sure during the acute stage of ischemic stroke is

necessary to maintain brain perfusion in borderline

ischemic areas.17 Thus, we often discontinued antihyper-

tensive agents during the early stage of ischemic strokes.

As Tarazi et al11 have discussed, patients who withdrew

from long-term thiazide diuretics increased their body

weight and plasma volume during the first week. The

fluid retentive effect avoided dehydration status that in-

creases blood viscosity and elevates hematocrit, which

have been shown to be associated with a larger infarct

volume in patients with cerebral infarction.18 In addition,

the relative hemodilutory effect on such fluid retention

status also may play a neuroprotective role postischemic

stroke.19 Stroke patients who withdraw from thiazides

may achieve adequate plasma volume through this fluid

retention effect, which leads to better cerebral circulation

and better functional outcomes.

In our study, the finding that long-term thiazide users

experienced lesser amounts of SIE may support the

concept ofmore adequate fluid status from thiazideswith-

drawal. Therewere a total of 18 patients (12.86%)who had

early neurologic deterioration, similar to previous

studies.20-22 As a previous study discussed, dehydration

status increased an early stroke-in-evolution.23,24

Although long-term thiazide use caused depletion of

plasma volumes, volumes returned to nearly normal

levels after long-term use.9-11,25 In our observational

study, the slight degree of dehydration did not cause

greater amounts of SIE. On the contrary, there was

decreased risk of SIE among thiazide users, which may

support the hypothesis of fluid retention effect after

thiazide withdrawal. The expanded plasma volume

protected stroke patients from dehydration, resulting in

a decreased risk of SIE and better functional outcomes.

Limitations

Our results are limited by sample size and thus statisti-

cal power. The use of a single institution also may limit

the results, because there is no comparison with other

geographic locations. The hypothesis of collateral circula-

tion and fluid retention could not be proven directly. In

addition, we did not compare thiazides with other antihy-

pertensive drugs, such as angiotensin-converting enzyme

inhibitors, angiotensin receptor blockers, calcium channel

blockers, and even combination therapy for primary

stroke prevention. Further randomized control studies

may be needed to verify functional progress results of

different antihypertensive treatments for primary stroke

prevention.

Conclusion

Our findings suggest that hypertensive patients treated

long term with thiazides had lesser stroke severity, better

functional outcomes after ischemic stroke, and showed no

difference in SIE risk. Hence, it is safe to use thiazides as a

first-line antihypertensive treatment. There may be addi-

tional benefits from thiazides along with lowering blood

pressure and stroke prevention.

References

1. Lawes CM, Vander Hoorn S, Rodgers A, InternationalSociety of Hypertension. Global burden of blood-pressure-related disease, 2001. Lancet 2008;371:1513-1518.

2. D’Agostino RB, Wolf PA, Belanger AJ, Kannel WB. Strokerisk profile: adjustment for antihypertensive medication.The Framingham Study. Stroke 1994;25:40-43.

3. SHEP Cooperative Research Group. Prevention of strokeby antihypertensive drug treatment in older persons withisolated systolic hypertension. Final results of the SystolicHypertension in the Elderly Program (SHEP). JAMA1991;265:3255-3264.

4. Staessen JA, Gasowski J, Wang JG, et al. Risks ofuntreated and treated isolated systolic hypertension inthe elderly: meta-analysis of outcome trials. Lancet2000;355:865-872.

5. Wright JM, Musini VM. First-line drugs for hypertension.Cochrane Database Syst Rev 2009;(3):CD001841.

6. Ravenni R, Jabre JF, Casiglia E, Mazza A. Primary strokeprevention and hypertension treatment: which is thefirst-line strategy? Neurol Int 2011;3:e12.

7. Leifer D, Bravata DM, Connors JJ 3rd, et al. Metrics formeasuring quality of care in comprehensive stroke centers:detailed follow-up to Brain Attack Coalition comprehen-sive stroke center recommendations: a statement forhealth-care professionals from the American Heart Association/American Stroke Association. Stroke 2011;42:849-877.

8. Saposnik G, Raptis S, Kapral MK, et al. Investigators ofthe Registry of the Canadian Stroke Network and theStroke Outcome Research Canada Working Group TheiScore predicts poor functional outcomes early after hos-pitalization for an acute ischemic stroke. Stroke 2011;42:3421-3428.

9. Shah S, Khatri I, Freis ED. Mechanism of antihyperten-sive effect of thiazide diuretics. Am Heart J 1978;95:611-618.

10. van Brummelen P, Man in’t Veld AJ, Schalekamp MA.Hemodynamic changes during long-term thiazide treat-ment of essential hypertension in responders and nonre-sponders. Clin Pharmacol Ther 1980;27:328-336.

11. Tarazi RC, Dustan HP, Frohlich ED. Long-term thiazidetherapy in essential hypertension. Evidence for persistentalteration in plasma volume and renin activity. Circula-tion 1970;41:709-717.

12. Fujii K,Weno BL, Baumbach GL, et al. Effect of antihyper-tensive treatment on focal cerebral infarction. Hyperten-sion 1992;19:713-716.

13. Shuaib A, Butcher K, Mohammad AA, et al. Collateralblood vessels in acute ischaemic stroke: a potential ther-apeutic target. Lancet Neurol 2011;10:909-921.

H.-M. SHIH ET AL.2418

14. Campbell BC, Christensen S, Tress BM, et al, EPITHETInvestigators. Failure of collateral blood flow is associ-ated with infarct growth in ischemic stroke. J Cereb BloodFlow Metab 2013;33:1168-1172.

15. Kucinski T, Koch C, Eckert B, et al. Collateral circulationis an independent radiological predictor of outcome afterthrombolysis in acute ischaemic stroke. Neuroradiology2003;45:11-18.

16. Lima FO, Furie KL, Silva GS, et al. The pattern of lepto-meningeal collaterals on CT angiography is a strongpredictor of long-term functional outcome in strokepatients with large vessel intracranial occlusion. Stroke2010;41:2316-2322.

17. Aiyagari V, Gorelick PB. Management of blood pressurefor acute and recurrent stroke. Stroke 2009;40:2251-2256.

18. HarrisonMJ, Pollock S, Kendall BE, et al. Effect of haema-tocrit on carotid stenosis and cerebral infarction. Lancet1981;2:114-115.

19. Ginsberg MD. Neuroprotection for ischemic stroke: past,present and future. Neuropharmacology 2008;55:363-389.

20. Tei H, Uchiyama S, Ohara K, et al. Deteriorating ischemicstroke in 4 clinical categories classified by the OxfordshireCommunity Stroke Project. Stroke 2000;31:2049-2054.

21. Arenillas JF, Rovira A, Molina CA, et al. Prediction ofearly neurological deterioration using diffusion- andperfusion-weighted imaging in hyperacute middlecerebral artery ischemic stroke. Stroke 2002;33:2197-2203.

22. Yamamoto H, Bogousslavsky J, van Melle G. Differentpredictors of neurological worsening in different causesof stroke. Arch Neurol 1998;55:481-486.

23. Lin LC, Lee TH, Chang CH, et al. Predictors of clinicaldeterioration during hospitalization following acuteischemic stroke. Eur Neurol 2012;67:186-192.

24. Schrock JW, Glasenapp M, Drogell K. Elevated bloodurea nitrogen/creatinine ratio is associated with pooroutcome in patients with ischemic stroke. Clin NeurolNeurosurg 2012;114:881-884.

25. Freis ED, Reda DJ, Materson BJ. Volume (weight) loss andblood pressure response following thiazide diuretics.Hypertension 1988;12:244-250.