Aljadhey et al.

BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

RESEARCH ARTICLE

Open Access

Comparative effects of non-steroidal

anti-inflammatory drugs (NSAIDs) on blood
pressure in patients with hypertension
Hisham Aljadhey1*, Wanzhu Tu2,3, Richard A Hansen4, Susan J Blalock5, D Craig Brater2 and Michael D Murray3,6

Abstract
Background: Nonsteroidal anti-inflammatory drugs (NSAIDs) may disrupt control of blood pressure in hypertensive
patients and increase their risk of morbidity, mortality, and the costs of care. The objective of this study was to
examine the association between incident use of NSAIDs and blood pressure in patients with hypertension.
Methods: We conducted a retrospective cohort study of adult hypertensive patients to determine the effects of
their first prescription for NSAID on systolic blood pressure and antihypertensive drug intensification. Data were
collected from an electronic medical record serving an academic general medicine practice in Indianapolis, Indiana,
USA. Using propensity scores to minimize bias, we matched a cohort of 1,340 users of NSAIDs with 1,340 users of
acetaminophen. Propensity score models included covariates likely to affect blood pressure or the use of NSAIDs.
The study outcomes were the mean systolic blood pressure measurement after starting NSAIDs and changes in
antihypertensive therapy.
Results: Compared to patients using acetaminophen, NSAID users had a 2 mmHg increase in systolic blood
pressure (95% CI, 0.7 to 3.3). Ibuprofen was associated with a 3 mmHg increase in systolic blood pressure compared
to naproxen (95% CI, 0.5 to 4.6), and a 5 mmHg increase compared to celecoxib (95% CI, 0.4 to 10). The systolic
blood pressure increase was 3 mmHg in a subgroup of patients concomitantly prescribed angiotensin converting
enzyme inhibitors or calcium channel blockers and 6 mmHg among those prescribed a beta-adrenergic blocker.
Blood pressure changes in patients prescribed diuretics or multiple antihypertensives were not statistically
significant.
Conclusion: Compared to acetaminophen, incident use of NSAIDs, particularly ibuprofen, is associated with a small
increase in systolic blood pressure in hypertensive patients. Effects in patients prescribed diuretics or multiple
antihypertensives are negligible.
Keywords: NSAIDs, Hypertension, Blood pressure, Propensity score

Background
Cardiovascular diseases are the most common cause of
death in the world [1], and uncontrolled hypertension is
a harbinger of such poor outcomes. Seven million
deaths worldwide each year are attributed to hypertension [2], and in the United States alone, about 73 million people, or one in three adults, have high blood
pressure. Only 35% of these patients have adequate
* Correspondence: haljadhey@ksu.edu.sa
1
Medication Safety Research Chair, College of Pharmacy, King Saud
University, Riyadh, Saudi Arabia
Full list of author information is available at the end of the article

blood pressure control [3]. Chronic administration of

medications that increase blood pressure is presumed to
be among a variety of factors responsible for poorly
controlled blood pressure that can lead to deleterious
cardiovascular effects.
The effect of nonsteroidal anti-inflammatory drugs
(NSAIDs) on the incidence of hypertension has been
previously investigated [4-8]; however, little information
is available about the magnitude of changes in blood
pressure in populations of patients chronically taking
antihypertensive medications. The effect of NSAIDs on
blood pressure has been investigated in clinical trials

2012 Aljadhey et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.

Aljadhey et al. BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

[9-24] but not observational studies, which are more

representative to real-world clinical settings. Therefore,
we examined the association between NSAIDs and blood
pressure in patients with hypertension who were taking
antihypertensive medications. This observational study
included incident users of NSAIDs and used propensity
scores to minimize selection bias. Using these methods,
we aimed to determine the comparative effects on systolic blood pressure of NSAIDs that are commonly used
in primary care. Because acetaminophen is an oft-used
alternative to NSAIDs, we determined the effects of
NSAID relative to acetaminophen. Our objectives were
to: (1) examine the association between NSAIDs and
blood pressure compared to acetaminophen in patients
with hypertension; (2) compare the effects of various
NSAIDs on blood pressure in patients with hypertension; and, (3) examine changes in antihypertensive therapy after starting NSAIDs.

Methods
Design and subjects

This retrospective cohort study included adult patients

who had received their first prescription for NSAID
from the general medicine practice of Wishard Health
Services in Indianapolis, Indiana. Wishard is a citycounty health center affiliated with the Indiana University School of Medicine. The Regenstrief Medical
Record System (RMRS) was used to identify eligible
patients within Wishard and to collect data on relevant variables. The RMRS is an electronic medical
record system that captures prescriptions, laboratory
test results (including electrographic results), and
other clinical data [25].
We included incident users in order to prevent underascertainment of adverse effects that occur early in treatment [26]. In this design only patients nave to the study
medication are included after the date of their first prescription (index date). Patients were eligible for inclusion
if they had received a prescription for any NSAID between 1993 and 2006, were aged 18 years or older, and
had a clinical diagnosis of hypertension on the index
date. Patients were excluded if they had an active prescription for any NSAID during the year preceding the
index date. Included patients had at least one measurement of sitting systolic blood pressure the year after the
index date and no changes in their antihypertensive
therapy until the measurement of blood pressure.
Patients prescribed acetaminophen who met the same
criteria described above formed a non-NSAID comparison group. Users of acetaminophen are similar to users
of NSAIDs in most characteristics and comorbidities
since acetaminophen is usually prescribed as a first line
therapy for patients with osteoarthritis [27]. The study
was approved by the Institution Review Boards at the

Page 2 of 10

University of North Carolina at Chapel Hill and Indiana

University-Purdue University at Indianapolis.
Data analysis

Systolic blood pressure measurements were collected

one year before and one year after the date of the first
prescription for NSAID or acetaminophen. Information
was collected on relevant confounders at baseline. The
end date of follow-up was defined as one year after
the index date or 30 days after the last dispensed
prescription, whichever came first.
Propensity score matching was used to prevent selection bias by balancing covariates between comparison
groups. Propensity scores estimate the probability of
each subjects exposure to treatment A versus treatment
B, based on measured covariates. The score combines all
confounding covariates into a single composite factor.
Mahalanobis metric matching without replacement was
used since it produces a good balance in covariates
between comparison groups [28,29]. A caliper of onequarter of the standard deviation of the propensity score
was used in the match. To assess the covariate balance,
we used chi-square tests for categorical variables and
t-tests for continuous variables. Separate propensity
scores were estimated for each comparison. Propensity
scores were calculated using SAS PROC LOGISTIC
(SAS Institute, Inc., Cary, North Carolina).
After covariate balance, a multiple linear regression
model was used to study the effect of NSAIDs on systolic blood pressure and changes in antihypertensive
therapy. The primary independent variable was the use
of NSAIDs. Covariates unbalanced after propensity score
matching were included in the model. Also, the model
included the time from index date until blood pressure
measurement, since this time occurred after the index
date it could not be included in the propensity score
matching. Another advantage of using a regression
model is that it reduces the standard error and, hence,
improves the precision of the estimate [29].
We compared individual commonly used NSAIDs.
Non-selective NSAIDs included ibuprofen and naproxen. The effects of naproxen and ibuprofen on blood
pressure have not been compared in observational studies of patients using antihypertensive medications. Previous observational studies have not compared the effect
of selective COX-2 inhibitors on blood pressure to nonselective NSAIDs. Since celecoxib is the only selective
COX-2 inhibitor on the market, it was compared to the
non-selective NSAIDs included in this study. We converted the odds ratios and related confidence intervals
into relative risks using the conversion formula by
Zhang and Yu [30]. We verified the model assumptions
for linear regression analysis by examining the residuals
of the model for signs of deviation from normality and

Aljadhey et al. BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

unequal variability. As expected, the model assumptions

were satisfied.

Page 3 of 10

angiotensin II receptor antagonists, and other antihypertensive medications. The MPR was used to control for
adherence with antihypertensive medications.

Covariates

Based on previous literature, the models included covariates likely to affect blood pressure or the use of NSAIDs
including age, race, gender, and baseline systolic blood
pressure. Baseline systolic blood pressure was defined as
the last measurement before the index date. The models
controlled for the presence of the following diagnoses at
the index date: rheumatoid arthritis, osteoarthritis, coronary artery disease or myocardial infarction, stroke
(cerebrovascular accident or transient ischemic attack),
arrhythmia, asthma or chronic obstructive pulmonary
disease, renal insufficiency, cirrhosis with ascites, systemic lupus erythematosus, diabetes mellitus, and congestive heart failure. We controlled for the use of
medications known to increase blood pressure including
venlafaxine, a high dose of oral glucocorticoids, and the
use of oral contraceptives.
To minimize bias introduced by variations in the time
between baseline systolic blood pressure assessment and
the index date and the time between the index date to
the first systolic blood pressure, we included in the models indicator variables for these times and the index year.
Variations in exposure to NSAID or acetaminophen
were controlled by using the medication possession ratio
(MPR) and the number of refills per month. The MPR
assesses refill adherence and was calculated by dividing
the sum of the days between the last refill and the next
expected refill (i.e. days supply) by the number of days
between the last refill and the next actual refill and then
multiplied by 100. For each patient, an average MPR was
calculated for the index drug. As a proxy for as-needed
versus regular NSAID use, the number of refills per
month was included in the model.
Sensitivity analyses were conducted to assess potential
variations in exposure to the index drug. In one analysis,
the model included the extent of exposure as the doseMPR interaction. In the other, the analysis was restricted
to only those patients who had a blood pressure measurement within 30 days of the index date. These results
did not change our conclusions. Furthermore, we examined the dose effect of NSAIDs by stratifying patients
into low and high dose groups. Patients who were prescribed less than 75% of the maximum daily dose listed
in Facts and Comparisons [31] were included in the low
dose category and those prescribed 75% or more were
included in the high dose category.
Baseline use of antihypertensive medications was
included as covariates in the models. Five groups of antihypertensive medications were formed: beta-adrenergic
antagonists, calcium channel blockers (CCBs), diuretics,
angiotensin-converting enzyme inhibitors (ACE-I) or

Endpoints

The outcomes of the study were the first and average

systolic blood pressure measurements after starting
NSAIDs, included as a continuous variable, and changes
in antihypertensive therapy. Systolic blood pressure was
assessed because it is associated with morbidity and
mortality more so than diastolic blood pressure and is
targeted in the treatment of hypertension [2]. The first
blood pressure measurement following the drug index
date was selected because physicians change antihypertensive therapy based on this measurement. Although, it
was found in the same study population that one blood
pressure reading has significant prognostic value [32],
the results using an average of all blood pressure measurements were compared to the results when using
only a single measurement. To prevent any potential
effect of changing antihypertensive therapy on blood
pressure, measurements of blood pressure were included
only up to the date when the antihypertensive regimen
was changed. The outcome was measured within one
year of first NSAID prescription. Another analysis was
conducted to investigate whether the increase in systolic
blood pressure associated with NSAIDs is clinically important. A clinically important increase was defined as
systolic blood pressure increase from baseline by at least
20 mmHg since the risk of mortality doubles for such an
increase [2,33].
Changes in antihypertensive therapy post-index were
considered as intensified when: (1) the dose of any currently prescribed antihypertensive medication was
increased; or (2) the patient was started on a new antihypertensive medication from another class.

Results
A total of 3,928 patients prescribed NSAIDs (n=2,181)
or acetaminophen (n=1,747) met the inclusion criteria.
Before matching on propensity score, many relevant
baseline characteristics differed between the NSAID and
acetaminophen cohorts (Table 1). Patients in the acetaminophen group were older and had a higher baseline
mean systolic blood pressure compared to those in the
NSAID group. They were also more likely to have renal
insufficiency, congestive heart failure, diabetes, coronary
artery disease or myocardial infarction, and stroke than
those in the NSAIDs group Figure 1.
A total of 1,340 patients prescribed NSAIDs were
matched to the same number of patients who had been
prescribed acetaminophen. Matching on propensity
scores balanced all covariates between the two groups
(Table 1). Compared to acetaminophen, NSAIDs were

Aljadhey et al. BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

Page 4 of 10

Table 1 Comparison of Covariate Balance between NSAIDs and Acetaminophen before and after propensity score
matching
Variable

Age (yrs) mean

Sample

NSAIDs*

Acetaminophen*

P Value**

Standardized
Difference

Unmatched

55

60

<.001

42.0

Matched

56

57

0.119

6.0

Unmatched

70

70

0.687

1.3

Matched

72

70

0.157

5.5

Bias
Reduction
(%)
86%

Gender:
Female

323%

Race:
African American
Other
Baseline systolic blood pressure (mmHg) mean

Unmatched

58

63

0.003

9.4

Matched

63

61

0.353

3.6

Unmatched

0.007

7.5

Matched

0.837

0.8

Unmatched

139

141

0.006

8.8

Matched

140

140

0.949

0.2

Unmatched

53

65

<.001

23.8

Matched

62

61

0.596

2.0

Unmatched

14

10

<.001

11.6

Matched

10

11

0.380

3.4

Unmatched

33

25

<.001

17.4

Matched

28

28

0.975

0.1

62%
89%
97%

Time from baseline SBP to index:

7 days
> 7 days and 30 days
> 30 days

91%
71%
99%

Year of index date:

1993 - 1996
1997-2002
2003 - 2006

Unmatched

50

58

<.001

17.1

Matched

57

56

0.713

1.4

Unmatched

39

40

0.477

2.3

Matched

41

42

0.741

1.3

Unmatched

11

<.001

40.8

Matched

0.881

0.6

92%
44%
99%

Diagnosis of:
Unmatched

22

23

0.768

0.9

Matched

21

22

0.295

4.0

Unmatched

0.381

2.8

Matched

0.819

0.9

Renal Insufficiency

Unmatched

<.001

24.5

Matched

0.922

0.4

Cirrhosis with Ascites

Unmatched

0.4

0.343

3.0

Matched

0.807

0.9

Osteoarthritis
Rheumatoid Arthritis

Systemic Lupus Erythematosus

Unmatched

0.277

3.5

Matched

0.465

2.8

Unmatched

28

35

<.001

14

Matched

28

31

0.124

5.9

Unmatched

11

19

<.001

20.9

Matched

14

14

0.651

1.7

Coronary Artery Disease or History of

Myocardial Infarction

Unmatched

13

19

<.001

16.0

Matched

14

15

0.266

4.3

Stroke

Unmatched

12

<.001

15.4

Diabetes
Congestive Heart Failure

327%
69%
98%
69%
20%
58%
92%

73%

Aljadhey et al. BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

Page 5 of 10

Table 1 Comparison of Covariate Balance between NSAIDs and Acetaminophen before and after propensity score
matching (Continued)
Arrhythmia
Asthma or Chronic Obstructive Pulmonary Disease

Matched

0.784

1.1

Unmatched

0.055

6.1

Matched

0.489

2.7

Unmatched

17

19

0.121

5.0

Matched

18

19

0.420

3.1

93%
56%
37%

Medications:
ACE-I or Angiotensin II blocker
Beta- Blocker

Unmatched

36

43

<.001

15.7

Matched

37

38

0.594

2.1

Unmatched

18

17

0.270

3.5

Matched

15

17

0.338

3.7

Calcium Channel Blocker

Unmatched

34

38

0.003

9.7

Matched

36

36

0.829

0.8

Diuretic

Unmatched

44

48

0.005

9.0

Matched

46

45

0.921

0.4

Other BP medications

Unmatched

12

<.001

15.2

Matched

0.835

0.8

Oral high dose glucocorticoid***

Unmatched

0.068

5.8

Matched

0.850

0.7

Unmatched

0.528

2.0

Matched

0.998

0.0

Unmatched

0.3

0.075

5.8

Matched

0.4

0.782

1.1

Unmatched

69

74

<.001

10.6

Matched

72

70

0.320

3.8

Oral Contraceptives
Venlafaxine

87%
4%
91%
96%
95%
87%
100%
82%

Adherence to antihypertensive medications:

MPR 80%
MPR < 80%
Not using antihypertensives (reference group)

Unmatched

13

12

0.262

3.6

Matched

12

13

0.159

5.4

Unmatched

18

14

0.002

9.8

Matched

16

16

0.965

0.2

Unmatched

32

24

<.001

19.5

Matched

26

27

0.750

1.2

64%
51%
98%

Exposure to index drug:

MPR > 80%
MPR 2080 %
MPR < 20%

Unmatched

44

43

0.278

3.5

Matched

44

43

0.859

0.7

Unmatched

23

34

<.001

23.3

Matched

30

30

0.909

0.4

Unmatched

52

36

<.001

30.8

Matched

44

43

0.301

4.0

94%
80%
98%

Number of refills per month:

1 refills

87%

* % unless indicated as mean. Because of rounding values may not add to 100 %.
** P-value of t-tests for continuous variables and chi-square tests for categorical variables.
*** High dose was defined as 10mg for prednisonse, 50mg for cortisone, and 1.5 mg for dexamethasone.
Standardized difference: 100 ( treated  control)/ {(s2 treated + s2 control)/2}. A positive value means the treated group is higher in % (or mean) compared to the
control group and negative value means the control is higher than the treated.
Bias reduction (%) =1- {|Standardized difference matched|/| Standardized difference unmatched |} x 100. A positive value means bias is reduced by propensity score
matching and negative means bias increased.
ACE-I: Angiotensin converting enzyme inhibitor; MPR: Medication Possession Ratio; SBP: Systolic blood pressure. Unmatched: all patients before propensity score
matching, N=3,928 (2,181 NSAIDs and 1,747 acetaminophen). Matched: only matched patients, N=2,680 (1,340 in each group).

Aljadhey et al. BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

Page 6 of 10

associated with a moderate mean increase in systolic

blood pressure of 2 mmHg in patients with hypertension
(95% confidence interval, 0.7 to 3.3) (Table 2). A prescription for NSAIDs was associated with a 3 mmHg increase in average systolic blood pressure in patients who
were concurrently prescribed ACE-I or CCB, and a 6
mmHg increase in those prescribed beta-adrenergic
blockers. However, no effect of NSAIDs on blood pressure was found in patients who were concurrently taking
diuretics (Table 2). Changes in blood pressure were not
associated with NSAIDs in patients prescribed various
combinations of two or more antihypertensive medications (Table 2).
Based on their propensity scores, 472 patients from
the naproxen group were matched to the same number
of patients from the ibuprofen group. Matching on propensity scores resulted in balanced covariates between
the two treatment groups. Compared to naproxen, ibuprofen was associated with a 2.5 mmHg increase in average systolic blood pressure (95% confidence interval, 0.5
to 4.6), and ibuprofen was associated with a clinically
important increase in systolic blood pressure as defined
above (relative risk, 1.47; 95% confidence interval, 1.09
to 1.96). The absolute risk of clinically important blood
pressure increase in the ibuprofen group was 20.6% and
in the naproxen group it was 14.6% and the calculated
number needed to harm was twelve patients.
Ibuprofen was associated with a 5.9 mmHg increase in
average systolic blood pressure in patients who were
prescribed a beta-adrenergic blocker (95% confidence
interval, 0.0 to 11.7; N=130) compared to naproxen
(Table 3). A prescription for ibuprofen in patients who
were prescribed various combinations of two or more
antihypertensive medications was not associated with
significant changes in systolic blood pressure (Table 3).

Dose effects of ibuprofen or naproxen on blood pressure were not statistically significant. A high dose of ibuprofen increased systolic blood pressure, 2.3 mmHg
(95% confidence interval, 1.3 to 5.1) and a high dose
of naproxen slightly decreased in systolic blood pressure
by 3.3 mmHg (95% confidence interval, 9.6 to 3.1).
Based on their propensity scores, 113 patients prescribed celecoxib were matched to 113 patients from the
ibuprofen group. A prescription for ibuprofen was associated with a 5.2 mmHg increase in the mean systolic
blood pressure compared to celecoxib (95% confidence
interval, 0.4 to 10.0) (Table 4). Compared to ibuprofen
or naproxen, celecoxib was not associated with a clinically important increase in systolic blood pressure.
When the outcome was defined as a change in antihypertensive therapy, 2,494 patients in the NSAID group
were matched based on their propensity score to the
same number of patients in the acetaminophen group.
Compared to acetaminophen, receipt of NSAID was not
associated with a change in antihypertensive therapy
(odds ratio, 0.95; 95% confidence interval, 0.84 to 1.08;
p = 0.4).

Discussion
In the current study, patients receiving NSAIDs showed
a 2 mmHg increase in systolic blood pressure compared
to acetaminophen recipients. The systolic blood pressure
increase was 3 mmHg in a sub-sample of those who
were concomitantly prescribed ACE-I or CCB and 6
mmHg in those prescribed a beta-adrenergic blocker.
Ibuprofen was associated with a systolic blood pressure
increase, compared to both naproxen and celecoxib, of 3
and 5 mmHg, respectively. Despite these effects we did
not detect significant changes in antihypertensive therapy in the NSAIDs users suggesting clinicians were not

Eligible Subjects
NSAIDs: 2181
Acetaminophen: 1747

NSAIDs (1340) vs.

Acetaminophen (1340)

Celecoxib (113) vs.

Ibuprofen (113)

Figure 1 Study flow chart.

Celecoxib (102) vs.

Naproxen (102)

Ibuprofen (472) vs.

Naproxen (472)

Aljadhey et al. BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

Page 7 of 10

Table 2 Difference in systolic blood pressure between NSAIDs and acetaminophen after propensity score matching
Sample

Dependent Variable*

Estimate of SBP (mmHg)**

95% Confidence Interval

All Patients (n=2,680)

First SBP

1.8

0.3 to 3.3

ACE-I (n=768)

CCB (n=804)

BB (n=340)

Diuretics (n=1,022)

CCB & ACE-I (n=202)

CCB & BB (n=104)

CCB & diuretics (n=328)

ACE-I & BB (n=108)

ACE-I & diuretics (n=366)

BB & diuretics (n=156)

CCB & ACE-I & diuretics (n=100)

BB & ACE-I & diuretics (n=42)

Average SBP

2.0

0.7 to 3.3

First SBP

2.8

0.2 to 5.8

Average SBP

2.8

0.2 to 5.4

First SBP

2.5

0.4 to 5.4

Average SBP

3.2

0.6 to 5.7

First SBP

6.3

1.7 to 10.8

Average SBP

5.5

1.4 to 9.6

First SBP

0.2

2.3 to 2.8

Average SBP

1.3

0.8 to 3.4

First SBP

1.1

5.6 to 7.8

Average SBP

3.1

2.8 to 8.9

First SBP

4.0

5.0 to 13.0

Average SBP

6.0

2.0 to 14.1

First SBP

1.6

3.3 to 6.5

Average SBP

3.5

0.8 to 7.9

First SBP

7.5

1.0 to 16.0

Average SBP

6.7

1.1 to 14.5

First SBP

1.0

3.4 to 5.5

Average SBP

1.2

2.7 to 5.1

First SBP

3.8

3.9 to 11.5

Average SBP

4.2

2.8 to 11.3

First SBP

1.5

7.9 to 10.8

Average SBP

3.8

4.9 to 12.4

First SBP

6.8

9.4 to 22.9

Average SBP

5.4

10.3 to 21.1

ACE-I: Angiotensin converting enzyme inhibitor; BB: Beta-blocker; CCB: Calcium channel blocker; SBP: Systolic blood pressure.
* First SBP is the first systolic blood pressure measurement after the index date. Average SBP is the average of all systolic blood pressure measurements after the
index date and prior to any changes in the antihypertensive therapy.
** Estimate of SBP is the estimate difference between NSAIDs and acetaminophen after controlling for baseline SBP. A higher value means NSAIDs is associated
with higher increase in systolic blood pressure compared to acetaminophen.

noticing these blood pressure changes that were admittedly subtle though sufficient to increase risk.
Similar to the current results, previous studies
reported an up to 7 mmHg increase in blood pressure in
patients who were stable on beta-adrenergic antagonists
and had started NSAID therapy [34,35]. Interestingly, in
our study the blood pressure increase associated with
NSAIDs was greatest in patients prescribed a betaadrenergic antagonist compared to other antihypertensive medications. The reason for this variation in blood
pressure among antihypertensives could be related to
the degree of prostaglandin (PGs) inhibition and the differences among these medications in their antihypertensive mechanisms. A proposed mechanism to explain this
effect with beta-adrenergic antagonists is that inhibition
of PGs by NSAIDs could increase sensitivity to the

vasoconstrictor effects of sympathetic nervous system

stimulation. Blocking beta receptors increases this sensitivity to the alpha sympathetic nervous system, resulting
in abolishment of the blood pressure lowering effect of
beta-adrenergic antagonists [36]. Further, some betaadrenergic antagonists reduce the glomerular filtration
rate [37]. In the long-term, this could increase the sensitivity to blood pressure increases by NSAIDs. This effect
has important implications for those patients with heart
failure and hypertension who have been prescribed betaadrenergic antagonists.
The blood pressure increase we observed with
NSAIDs in ACE-I users agrees with previous studies that
reported a 5 to 10 mmHg increase in systolic blood
pressure [13,18,19,23]. The inhibition of PGs by NSAIDs
is proposed as the mechanism that explains the loss of

Aljadhey et al. BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

Page 8 of 10

Table 3 Difference in systolic blood pressure between naproxen and ibuprofen after propensity score matching
Sample

Dependent Variable*

Estimate of SBP (mmHg)**

95% Confidence Interval

All Patients (n=944)

First SBP

2.0

4.4 to 0.4

Average SBP

2.5

4.6 to 0.5

First SBP

0.7

4.1 to 5.4

Average SBP

1.1

5.3 to 3.0

First SBP

2.3

6.8 to 2.2

Average SBP

2.2

6.1 to 1.7

First SBP

4.3

10.6 to 2.0

Average SBP

5.9

11.7 to 0.01

First SBP

3.2

7.6 to 1.2

Average SBP

3.3

7.0 to 0.5

First SBP

0.4

11.9 to 12.7

Average SBP

0.8

11.7 to 10.1

First SBP

4.5

18.8 to 9.7

Average SBP

4.4

18.3 to 9.6

First SBP

3.2

11.3 to 5.0

Average SBP

3.2

9.5 to 3.2

First SBP

5.8

19.1 to 7.5

Average SBP

10.1

23.0 to 2.8

First SBP

3.3

10.5 to 4.0

ACE-I (n=276)

CCB (n=268)

BB (n=130)

Diuretics (n=340)

CCB & ACE-I (n=60)

CCB & BB (n=34)

CCB & diuretics (n=118)

ACE-I & BB (n=40)

ACE-I & diuretics (n=124)

BB & diuretics (n=70)

CCB & ACE-I & diuretics (n=28)

BB & ACE-I & diuretics (n=22)

Average SBP

4.3

11.0 to 2.5

First SBP

7.0

16.3 to 2.4

Average SBP

5.4

14.0 to 3.2

First SBP

4.6

22.9 to 13.8

Average SBP

0.5

16.5 to 15.6

First SBP

7.4

29.2 to 14.4

Average SBP

8.9

30.8 to 13.0

ACE-I: Angiotensin converting enzyme inhibitor; BB: Beta-blocker; CCB: Calcium channel blocker; SBP: Systolic blood pressure.
* First SBP is the first systolic blood pressure measurement after the index date. Average SBP is the average of all systolic blood pressure measurements after the
index date and prior to any changes in the antihypertensive therapy.
** Estimate of SBP is the estimate difference between naproxen and ibuprofen after controlling for baseline SBP. A positive value means naproxen is associated
with higher increase in systolic blood pressure compared to ibuprofen. A negative value means ibuprofen is associated with higher increase in systolic blood
pressure compared to naproxen.

Table 4 Difference in systolic blood pressure between celecoxib and ibuprofen or naproxen after propensity score
matching
Comparison

Dependent Variable*

Estimate of SBP(mmHg)**

95% Confidence Interval

Celecoxib vs. Ibuprofen (n=226)

First SBP

5.4

10.8 to 0.0

Average SBP

5.2

10.0 to 0.4

First SBP

0.3

5.5 to 4.9

Average SBP

0.3

5.1 to 4.5

Celecoxib vs. Naproxen (n=204)

SBP: Systolic blood pressure.

* First SBP is the first systolic blood pressure measurement after the index date. Average SBP is the average of all systolic blood pressure measurements after the
index date and prior to any changes in the antihypertensive therapy.
** Estimate of SBP is the estimate difference between celecoxib and the comparator drug (ibuprofen or naproxen) after controlling for baseline SBP. A positive
value means celecoxib is associated with higher increase in systolic blood pressure compared to the comparator drug. A negative value means the comparator
drug is associated with higher increase in systolic blood pressure compared to celecoxib.

Aljadhey et al. BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

the blood pressure lowering effect of ACE-I. Because

PGs mediate the antihypertensive effect of ACE-I at least
in part, inhibition of PGs by NSAIDs could disrupt the
blood pressure control achieved by ACE-I [19,22,38].
These observations may be particularly important in
patients with diabetes. Antihypertensive treatment is
often intensified in patients with diabetes mellitus [39];
in addition, patients with diabetes mellitus who are diagnosed with hypertension are more likely to receive an
ACE-I rather than other antihypertensive medications to
preserve renal function. Therefore, it is important to
monitor blood pressure closely in diabetic patients who
are prescribed NSAIDs to ensure adequate blood pressure control.
No statistically significant changes in systolic blood
pressure were associated with a prescription for NSAID
in patients who were prescribed multiple antihypertensive medications. This can be explained by small sample
size in some of these combinations. Because some of the
combinations with beta-adrenergic blockers involved
only small number of patients, it is possible that this
study was not powered to detect small effects.
Similar to previous studies [10,11], the current study
found no effect of NSAIDs on blood pressure in patients
who were using diuretics. Current hypertension guidelines recommend starting patients on thiazide diuretics
because they are associated with better clinical outcomes
and lower mortality rates than other antihypertensive
medications [2]. In addition, diuretics are often less expensive than other antihypertensive medications. The
absence of an effect of NSAIDs is further reinforcement
for use of diuretics to control blood pressure in patients
who were using NSAIDs.
The results of the current study may have some clinical implications. Blood pressure is often poorly controlled in hypertensive patients [3], Our results raise the
prospect that NSAID use contributes to that poor control found in numerous epidemiologic surveys. It is possible that more attention to the effects of NSAIDs on
maintaining or achieving blood pressure control could
lower morbidity and mortality and in so doing reduce
health care costs [40]. For example, it was estimated
that in the United States achieving or maintaining blood
pressure control in users of selective COX-2 inhibitors
would prevent more than 70,000 deaths from stroke
and 60,000 others from coronary heart disease; such
control would also result in direct health care cost savings of more than 3.8 billion dollars [40]. The small increase in systolic blood pressure associated with
NSAIDs seen in this study may not affect a physicians
decision to change antihypertensive therapy. However,
in the long-term, such an increase could be associated
with significant comorbidity consequences. For example,
decreasing systolic blood pressure by just 2 mmHg

Page 9 of 10

lowers stroke mortality by 10% and ischemic heart disease mortality by 7% [41]. Future studies are needed to
assess the long term effect of such small increase in
blood pressure.
This study has limitations that should be considered
when interpreting the results. Patients included in this
study came from a single health system and may not be
representative of other practices. Hence, this study
should be replicated in other clinical settings. Although
propensity score matching balances many covariates at
baseline, unobserved covariates could still differ between the groups. Bias is a threat to the validity of these
results especially when comparing NSAID and acetaminophen groups. Acetaminophen has mild pressor
effects that may have dampened the relative effects of
NSAID [42]. Nonetheless, acetaminophen is often used
as an alternative to NSAIDs and we therefore believed
it was a reasonable non-NSAID comparator. Finally,
several NSAIDs are available over the counter (OTC) as
well as by prescription and this database captures only
the use of prescription NSAIDs. However, because
patients included in this study were provided with
needed OTC NSAIDs through a prescription assistance
program, it is less likely that they would have purchased
additional OTC NSAIDs. Furthermore, sensitivity analysis research suggests that missing OTC drug exposure
is not a significant source of bias [43].
In conclusion, compared to acetaminophen, incident
use of NSAIDs (particularly ibuprofen) is associated
with a small increase in systolic blood pressure in
hypertensive patients. Effects in patients prescribed
diuretics or multiple antihypertensives are negligible.
Competing interest
The author(s) declare that they have no competing interests.

Authors contribution
All authors contributed to the study idea, design, and methods. HA carried
out the statistical analyses and drafted the manuscript. All authors
contributed to the manuscript writing. All authors reviewed and edited the
final version of the manuscript. All authors read and approved the final
manuscript.

Acknowledgement
This study was funded partially by the National Plan for Science and
Technology (09-BIO708-02).
Author details
1
Medication Safety Research Chair, College of Pharmacy, King Saud
University, Riyadh, Saudi Arabia. 2Indiana University School of Medicine,
Indianapolis, IN, USA. 3Regenstrief Institute, Inc, Indianapolis, IN, USA.
4
Department of Pharmacy Care Systems, Harrison School of Pharmacy,
Auburn University, Auburn, AL, USA. 5Division of Pharmaceutical Outcomes &
Policy, School of Pharmacy, University of North Carolina, Chapel Hill, NC, USA.
6
Purdue University College of Pharmacy, West Lafayette, IN, USA.
Received: 8 June 2012 Accepted: 15 October 2012
Published: 24 October 2012

Aljadhey et al. BMC Cardiovascular Disorders 2012, 12:93

http://www.biomedcentral.com/1471-2261/12/93

References
1. Yach D, Hawkes C, Gould CL, Hofman KJ: The global burden of chronic
diseases: overcoming impediments to prevention and control. JAMA
2004, 291(21):26162622.
2. Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, et al:
Seventh report of the Joint National Committee on Prevention,
Detection, Evaluation, and Treatment of High Blood Pressure.
Hypertension 2003, 42(6):12061252.
3. American Heart Association: High Blood Pressure Statistics. 2008. Available at: http://
www.americanheart.org/presenter.jhtml?identifier=2139. Accessed 01/15/2008.
4. Johnson AG, Simons LA, Simons J, Friedlander Y, McCallum J: Non-steroidal
anti-inflammatory drugs and hypertension in the elderly: a communitybased cross-sectional study. Br J Clin Pharmacol 1993, 35(5):455459.
5. Dedier J, Stampfer MJ, Hankinson SE, Willett WC, Speizer FE, Curhan GC:
Nonnarcotic analgesic use and the risk of hypertension in US women.
Hypertension 2002, 40(5):604608.
6. Kurth T, Hennekens CH, Sturmer T, Sesso HD, Glynn RJ, Buring JE, et al:
Analgesic use and risk of subsequent hypertension in apparently healthy
men. Arch Intern Med 2005, 165(16):19031909.
7. Solomon DH, Schneeweiss S, Levin R, Avorn J: Relationship between COX-2
specific inhibitors and hypertension. Hypertension 2004, 44(2):140145.
8. Wang J, Mullins CD, Mamdani M, Rublee DA, Shaya FT: New diagnosis of
hypertension among celecoxib and nonselective NSAID users: a
population-based cohort study. Ann Pharmacother 2007, 41(6):937943.
9. Whelton A, Fort JG, Puma JA, Normandin D, Bello AE, Verburg KM:
Cyclooxygenase-2specific inhibitors and cardiorenal function: a
randomized, controlled trial of celecoxib and rofecoxib in older
hypertensive osteoarthritis patients. Am J Ther 2001, 8(2):8595.
10. Davies JG, Rawlins DC, Busson M: Effect of ibuprofen on blood pressure
control by propranolol and bendrofluazide. J Int Med Res 1988, 16(3):173181.
11. Wright JT, McKenney JM, Lehany AM, Bryan DL, Cooper LW, Lambert CM:
The effect of high-dose short-term ibuprofen on antihypertensive
control with hydrochlorothiazide. Clin Pharmacol Ther 1989, 46(4):440444.
12. Klassen D, Goodfriend TL, Schuna AA, Young DY, Peterson CA: Assessment
of blood pressure during treatment with naproxen or ibuprofen in
hypertensive patients treated with hydrochlorothiazide. J Clin Pharmacol
1993, 33(10):971978.
13. Polonia J, Boaventura I, Gama G, Camoes I, Bernardo F, Andrade P, et al:
Influence of non-steroidal anti-inflammatory drugs on renal function and
24h ambulatory blood pressure-reducing effects of enalapril and
nifedipine gastrointestinal therapeutic system in hypertensive patients.
J Hypertens 1995, 13(8):925931.
14. Houston MC, Weir M, Gray J, Ginsberg D, Szeto C, Kaihlenen PM, et al: The
effects of nonsteroidal anti-inflammatory drugs on blood pressures of
patients with hypertension controlled by verapamil. Arch Intern Med 1995,
155(10):10491054.
15. Klassen DK, Jane LH, Young DY, Peterson CA: Assessment of blood
pressure during naproxen therapy in hypertensive patients treated with
nicardipine. Am J Hypertens 1995, 8(2):146153.
16. Gurwitz JH, Everitt DE, Monane M, Glynn RJ, Choodnovskiy I, Beaudet MP,
et al: The impact of ibuprofen on the efficacy of antihypertensive
treatment with hydrochlorothiazide in elderly persons. J Gerontol A Biol
Sci Med Sci 1996, 51(2):M74M79.
17. Olsen ME, Thomsen T, Hassager C, Ibsen H, Dige-Petersen H: Hemodynamic
and renal effects of indomethacin in losartan-treated hypertensive
individuals. Am J Hypertens 1999, 12(2):209216.
18. Morgan TO, Anderson A, Bertram D: Effect of indomethacin on blood
pressure in elderly people with essential hypertension well controlled on
amlodipine or enalapril. Am J Hypertens 2000, 13(11):11611167.
19. Fogari R, Zoppi A, Carretta R, Veglio F, Salvetti A: Effect of indomethacin on
the antihypertensive efficacy of valsartan and lisinopril: a multicentre
study. J Hypertens 2002, 20(5):10071014.
20. Whelton A, White WB, Bello AE, Puma JA, Fort JG: Effects of celecoxib and
rofecoxib on blood pressure and edema in patients > or =65 years of
age with systemic hypertension and osteoarthritis. Am J Cardiol 2002,
90(9):959963.
21. White WB, Kent J, Taylor A, Verburg KM, Lefkowith JB, Whelton A: Effects of
celecoxib on ambulatory blood pressure in hypertensive patients on
ACE inhibitors. Hypertension 2002, 39(4):929934.
22. Palmer R, Weiss R, Zusman RM, Haig A, Flavin S, MacDonald B: Effects of
nabumetone, celecoxib, and ibuprofen on blood pressure control in

Page 10 of 10

23.

24.

25.

26.
27.

28.

29.

30.
31.
32.

33.

34.

35.

36.

37.
38.
39.
40.

41.

42.

43.

hypertensive patients on angiotensin converting enzyme inhibitors.

Am J Hypertens 2003, 16(2):135139.
Izhar M, Alausa T, Folker A, Hung E, Bakris GL: Effects of COX inhibition on
blood pressure and kidney function in ACE inhibitor-treated blacks and
hispanics. Hypertension 2004, 43(3):573577.
Sowers JR, White WB, Pitt B, Whelton A, Simon LS, Winer N, et al: The
Effects of cyclooxygenase-2 inhibitors and nonsteroidal antiinflammatory therapy on 24-hour blood pressure in patients with
hypertension, osteoarthritis, and type 2 diabetes mellitus. Arch Intern Med
2005, 165(2):161168.
Murray MD, Smith FE, Fox J, Teal EY, Kesterson JG, Stiffler TA, et al:
Structure, functions, and activities of a research support informatics
section. J Am Med Inform Assoc 2003, 10(4):389398.
Ray WA: Evaluating medication effects outside of clinical trials: new-user
designs. Am J Epidemiol 2003, 158(9):915920.
American College of Rheumatology Subcommittee on Osteoarthritis
Guidelines: Recommendations for the medical management of
osteoarthritis of the hip and knee: 2000 update. Arthritis Rheum 2000,
43(9):19051915.
DAgostino RB Jr: Propensity score methods for bias reduction in the
comparison of a treatment to a non-randomized control group. Stat Med
1998, 17(19):22652281.
Baser O: Too much ado about propensity score models? Comparing
methods of propensity score matching. Value Health 2006,
9(6):377385.
Zhang J, Yu KF: Whats the Relative Risk. JAMA 1998, 280(19):16901691.
Drug Facts and Comparisons: Facts & Comparisons 4.0. St. Louis, Mo:
Wolters Kluwer Health; 2001.
Tierney WM, Brunt M, Kesterson J, Zhou XH, LItalien G, Lapuerta P:
Quantifying risk of adverse clinical events with one set of vital signs
among primary care patients with hypertension. Ann Fam Med 2004,
2(3):209217.
Nietert PJ, Ornstein SM, Dickerson LM, Rothenberg RJ: Comparison of
changes in blood pressure measurements and antihypertensive therapy
in older, hypertensive, ambulatory care patients prescribed celecoxib or
rofecoxib. Pharmacotherapy 2003, 23(11):14161423.
Lopez-Ovejero JA, Weber MA, Drayer JI, Sealey JE, Laragh JH: Effects of
indomethacin alone and during diuretic or beta-adrenoreceptorblockade therapy on blood pressure and the renin system in essential
hypertension. Clin Sci Mol Med Suppl 1978, 4:203s205s.
Johnson AG, Nguyen TV, Day RO: Do nonsteroidal anti-inflammatory
drugs affect blood pressure? A meta-analysis. Ann Intern Med 1994,
121(4):289300.
Davis A, Day RO, Begg EJ: Interactions between non-steroidal antiinflammatory drugs and antihypertensives and diuretics. Aust N Z J Med
1986, 16(4):537546.
Epstein M, Oster JR: Beta blockers and renal function: a reappraisal. J Clin
Hypertens 1985, 1(1):8599.
MacFarlane LL, Orak DJ, Simpson WM: NSAIDs, antihypertensive agents
and loss of blood pressure control. Am Fam Physician 1995, 51(4):849856.
American Diabetes Association: Standards of medical care in diabetes
2007. Diabetes Care 2007, 30(Suppl 1):S4S41.
Grover SA, Coupal L, Zowall H: Treating osteoarthritis with
cyclooxygenase-2-specific inhibitors: what are the benefits of avoiding
blood pressure destabilization? Hypertension 2005, 45(1):9297.
Lewington S, Clarke R, Qizilbash N, Peto R, Collins R: Age-specific relevance
of usual blood pressure to vascular mortality: a meta-analysis of
individual data for one million adults in 61 prospective studies. Lancet
2002, 360(9349):19031913.
Sudano I, Flammer AJ, Priat D, Enseleit F, Hermann M, Wolfrum M, et al:
Acetaminophen increases blood pressure in patients with coronary
artery disease. Circulation 2010, 122(18):17891796. Nov 2.
Yood MU, Campbell UB, Rothman KJ, Jick SS, Lang J, Wells KE, et al: Using
prescription claims data for drugs available over-the-counter (OTC).
Pharmacoepidemiol Drug Saf 2007, 16(9):961968.

doi:10.1186/1471-2261-12-93
Cite this article as: Aljadhey et al.: Comparative effects of non-steroidal
anti-inflammatory drugs (NSAIDs) on blood pressure in patients with
hypertension. BMC Cardiovascular Disorders 2012 12:93.