The association between headache and elevated blood pressure among patients

presenting to an Emergency Department

Benjamin W. Friedman MD, MS, Binoy Mistry MD, Jason West MD,
Andrew Wollowitz MD

PII: S0735-6757(14)00336-2
DOI: doi: 10.1016/j.ajem.2014.05.017
Reference: YAJEM 54300

To appear in: American Journal of Emergency Medicine

Received date: 16 March 2014

Revised date: 13 April 2014
Accepted date: 11 May 2014

Please cite this article as: Friedman Benjamin W., Mistry Binoy, West Jason, Wollowitz
Andrew, The association between headache and elevated blood pressure among patients
presenting to an Emergency Department, American Journal of Emergency Medicine (2014),
doi: 10.1016/j.ajem.2014.05.017

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could aect the content, and all legal disclaimers that
apply to the journal pertain.
 ACCEPTED MANUSCRIPT

The association between headache and elevated blood pressure among patients presenting
to an Emergency Department

Benjamin W Friedman, MD, MS

Binoy Mistry, MD
Jason West, MD

PT
Andrew Wollowitz, MD

Department of Emergency Medicine, Albert Einstein College of Medicine, Montefiore

RI
Medical Center, Bronx, NY, USA
Department of Emergency Medicine, Johns Hopkins University School of Medicine,

SC
Baltimore, MD, USA

NU
Corresponding author:
Benjamin W. Friedman, MD, MS
Department of Emergency Medicine
MA
Montefiore Medical Center
Albert Einstein College of Medicine
111 East 210th Street
Bronx, NY, 10467
ED

bwfriedmanmd@gmail.com
(718) 920-6626
PT

Meetings at which this work has been presented: Society for Academic Emergency
Medicine, Atlanta, Georgia, May, 2013
Grant support: None
CE

Conflicts of interest: None

Word count for text excluding abstract: 3321
AC

Author contributions:
BWF conceived and designed the study. BM, JW, and AW abstracted data and reviewed
the literature. BWF analyzed the data. BWF drafted the manuscript, and all authors
contributed substantially to its revision. BWF takes responsibility for the paper as a
whole.
 ACCEPTED MANUSCRIPT

Abstract
Background. Elevated blood pressure (BP) and headache have long been linked in the
medical literature though data on association is conflicting. We used previously collected
data to address these related aims: 1) Using the National Hospital Ambulatory Medical
Care Survey (NHAMCS), we determined whether elevated BP is more likely in patients
who present to an ED with headache than in patients who present with other complaints;

PT
2) Using data collected in three ED-based migraine clinical trials, we determined the
association between improvement in headache pain and improvement in BP among
patients who present to an ED with migraine and elevated BP; 3)Using the data from the

RI
migraine clinical trials, we also determined if an elevated baseline BP identifies a group
of patients less likely to respond to standard migraine treatment.

SC
Methods. We analyzed two distinct datasets. The first, NHAMCS, is a national
probability sample of all U.S. ED visits. The second is a compilation of data gathered
during three ED-based migraine RCTs. We defined elevated BP as follows: moderate

NU
elevation= systolic BP > 150mmHg or diastolic BP > 95mmHg; marked elevation =
systolic BP >165mmHg or diastolic BP >100mmHg; severe elevation =systolic BP
>180mmHg or diastolic BP >110mmHg. We report the association between headache
MA
and elevated BP in NHAMCS using odds ratios (OR) with 95%CI. We report the
correlation coefficient and r2 for the association between improvement in BP and
improvement in migraine in our clinical trials dataset. Finally, using our clinical trials
database, we determined the influence of elevated BP at baseline on response to migraine
ED

medication by constructing a linear regression model in which the dependent variable

was improvement in zero to ten pain score between baseline and one hour and the
primary predictor variable was presence or absence of elevated BP at baseline.
PT

Results. Headache was the primary complaint in 3.7% (95%CI: 3.4, 4.0%) of all U.S. ED
visits, corresponding to 4.8 million (95%CI: 4.2, 5.4 million) patient visits. Among US
ED patients, those with headache were more likely than patients with other chief
CE

complaints to have markedly (OR 1.37 [95%CI 1.16, 1.61)] or severely elevated blood
pressure (OR1.49 [95%CI: 1.17, 1.90]). In our clinical trials dataset of migraine patients
with moderately elevated BP, there was no correlation between improvement in pain
AC

score and improvement in SBP (r=-0.07, r2=0, p=0.465) or DBP (r=-0.03, r2=0, p=0.75).
Similarly, there was no correlation between improvement in headache and improvement
in BP among migraine patients with markedly elevated blood pressure (for systolic BP
r=-0.19, r2=0.04, p=0.89; for diastolic BP r=-0.02, r2=0, p=0.87) nor among patients with
severely elevated BP (for systolic BP r=0.06, r2=0, p=0.81; for diastolic BP r=0.03, r2=0,
p=0.90). Patients with moderately elevated BP had slightly less improvement in their
zero to ten pain score than patients with BPs below this cutoff (-0.6, 95%CI: -1.2, -0.1,
p=0.03). This was more pronounced among patients with markedly elevated BP (-0.9,
95%CI: -1.7, -0.2).
Conclusions. While there is an association between elevated blood pressure and headache
among patients presenting to an emergency department, improvement in headache is not
associated with improvement in blood pressure.
 ACCEPTED MANUSCRIPT

Background
Elevated blood pressure (BP) and headache have long been linked in the medical
literature1 though data on association is conflicting. International guidelines stipulate that
headache should be attributed to elevated BP if the systolic BP rises rapidly to
>180mmHg or if the diastolic rises to >120mmHg and if the headache resolves with
normalization of BP.2 This guideline statement is supported by ambulatory blood

PT
pressure monitoring studies in hypertensive patients, which demonstrate that in general,
routine headache is not preceded by atypical fluctuations in blood pressure.3,4 Left
unclear by these data is the role of elevated BP in headaches of sufficient intensity to

RI
warrant a visit to a medical provider, a not uncommon scenario, and the question of how
to manage patients with both an acute headache, such as migraine, and elevated BP.

SC
To help guide care of these patients, we used previously collected data to address these
related aims: 1) Using the National Hospital Ambulatory Medical Care Survey

NU
(NHAMCS), a national database, we determined how frequently headache patients
presents to an emergency department (ED) with elevated BP and if elevated BP is more
likely in patients who present to an ED with headache than in patients who present with
MA
other complaints; 2) Using data collected in the course of three ED-based migraine
clinical trials, we determined if there is an association between improvement in headache
pain and improvement in BP among patients who present to an ED with migraine and
elevated BP; 3) Using the data from these three migraine clinical trials, we determined if
ED

an elevated baseline BP identifies a group of patients less likely to respond to standard

migraine treatment.
PT
CE
AC
 ACCEPTED MANUSCRIPT

Methods
Overview. Using two distinct databases, we assessed the relationship between headache
and elevated BP among patients presenting to an ED. In the first database, a national
probability sample designed to be representative of all U.S. ED visits, we determined 1)
the frequency of elevated BP among patients presenting to EDs across the U.S. with
headache, 2) the association between headache and elevated BP among all ED patients,

PT
and 3) ED treatment patterns with regard to elevated BP among headache patients. In the
second analysis, we use data gathered during three ED-based migraine RCTs to
determine 1) whether improvement of headache in migraine research subjects with

RI
elevated BP is associated with improvement in BP and 2) whether elevated BP at baseline
is associated with more refractory headaches. The Albert Einstein College of Medicine

SC
IRB reviewed this study administratively and exempted it from further review.

Databases.

NU
1) The National Hospital Ambulatory Medical Care Survey (NHAMCS) is a publically
available dataset, collected and distributed by the National Center for Health Statistics. It
is available at http://www.cdc.gov/nchs/ahcd.htm. Using a four-stage probability
MA
sampling method, trained data extractors identify randomly selected ED visits at
randomly selected emergency service areas from randomly selected hospitals from
randomly selected regions of the US. Trained analysts extract from the medical record the
patients socio-demographic characteristics, characteristics of the patients chief
ED

complaint, presenting vitals signs, work-up, and treatment. We used the 2010 dataset,
which was the most current dataset at the time of our analysis.
2) During the years 2005-2013, we completed three ED-based migraine randomized
PT

comparative efficacy trials in which patients presenting to one of four EDs with an acute
migraine headache were treated with intravenous metoclopramide.5-7 These migraine
clinical trials used nearly identical inclusion and exclusion criteria and enrolled patients
CE

with nearly identical socio-demographic characteristics. Therefore, we felt it reasonable

to combine the data from these similar studies into one dataset for the purpose of this
analysis. In each of these RCTs, research subjects had blood pressure and pain scores
AC

assessed at baseline and one hour later. In the first study, an examination of the efficacy
of dexamethasone, all subjects received metoclopramide 20mg IV + diphenhydramine
25mg IV. Half of the subjects also received dexamethasone 10mg IV while the other half
received intravenous placebo. This study was a negative studythere was no difference
between study arms in primary or secondary outcomes. The second study was a
metoclopramide dose finding study. ED migraine patients were randomized to receive
10mg, 20mg, or 40mg of IV metoclopramide. In this study, all research subjects also
received diphenhydramine 25mg IV. This study too was a negative study. The third study
was a randomized comparison of metoclopramide 10mg IV versus ketorolac 30mg IV
versus valproate 100mg IV. To minimize heterogeneity in our blood pressure analysis,
we only used the data of research subjects treated with metoclopramide. We felt
including the data from those subjects randomized to ketorolac or valproate would
introduce too much heterogeneity into our blood pressure analysis.

Measures.
NHAMCS. We utilized the following measures reported in NHAMCS:
 ACCEPTED MANUSCRIPT

1) Reason for visit. NHAMCS reports up to three reasons for visit for each patient visit.
We considered patients to have presented to the ED for headache if their primary stated
reason for visit was one of the following: i) Headache, pain in the head; ii) migraine; or
iii) sinus problems (pain and pressure). We included sinus problems (pain and pressure)
within our headache categorization because it is clear that many patients with migraine
headache consider their headaches to be attributable to sinus disease.8 As we did not wish

PT
to include patients presenting primarily with rhinorrhea and congestion, we did not
include patients whose primary chief complaint was one of the following: i) sinus
problems (unmodified); ii) sinus problems (sinus inflammation, infection); or iii) sinus

RI
problems (sinus congestion). Similarly, we did not include patients whose secondary or
tertiary reason for visit was headache because we did not wish to include patients who

SC
presented primarily for a febrile illness that included headache as one of a constellation of
symptoms. In one of our analyses, we compare headache patients to those patients who
presented to the ED for management of abdominal pain. For the abdominal pain analysis,

NU
we used the following reason for visit codes: i) stomach and abdominal pain, cramps and
spasms, including gastric pain; ii) abdominal pain, cramps, spasms NOS including
abdominal discomfort, gas pain, and intestinal colic; iii) lower abdominal pain, cramps,
MA
spasms, including right lower quadrant pain, left lower quadrant pain, and inguinal pain;
iv) upper abdominal pain, cramps, spasms, including epigastric pain, left upper quadrant
pain, pain in umbilical region, and right upper quadrant pain.
2) Systolic blood pressure. This was the patients presenting systolic blood pressure.
ED

3) Diastolic blood pressure. This was the patients presenting diastolic blood pressure.
4) Anti-hypertensive medication was defined to include any of the following: agents for
hypertensive emergencies; angiotensin converting enzyme inhibitors; antiadrenergic
PT

agents, peripherally acting; antiadrenergic agents, centrally acting; anti-anginal agents;

beta-adrenergic blocking agents; calcium channel blocking agents; diuretics; peripheral
vasodilators; antihypertensive combinations; angiotensin II inhibitors; and vasopressin
CE

antagonists
5) We included the following co-variates in the regression model described below:
-Age in years. This was a continuous value
AC

-Sex. Reported as male or female

-Race/ ethnicity. Reported as Hispanic, non-Hispanic White, non-Hispanic Black,
or non-Hispanic other
-Geographic region. Reported as Northeast, Midwest, Southeast, West
-Triage Acuity. Reported as non-urgent, semi-urgent, urgent, emergent,
immediate.

Clinical trials dataset. We extracted the following information from each of our datasets.
1) Pain score. Each research subject was asked to describe their headache intensity at
baseline and again 60 minutes later using a 0 to 10 verbal integer scale on which zero
signified no pain and ten signified the worst pain imaginable.
2). Systolic and diastolic blood pressure. Using an automated sphygmomanometer,
research associates specifically trained to measure blood pressure assessed it at baseline
and again 60 minutes later.

Analysis.
 ACCEPTED MANUSCRIPT

Definition of hypertension. We reviewed the medical literature to determine standardized

or evidence-based definitions for moderate and severe acute elevations in blood pressure.
We were unable to identify definitions that met these criteria. We therefore created the
following definitions for this analysis:
--Moderately elevated acute blood pressure: systolic BP > 150mmHg or diastolic BP >
95mmHg

PT
--Markedly elevated acute blood pressure: systolic BP >165mmHg or diastolic BP
>100mmHg
--Severely elevated acute blood pressure: systolic BP >180mmHg or diastolic BP

RI
>110mmHg

SC
NHAMCS.
In NHAMCS, each patient visit is assigned a weight, which incorporates the four stage
sampling strategy and missing data. Using this weight, representative national data can be

NU
calculated. Using the complex samples module from SPSS v. 21, we calculated the
frequency, with 95%CI, of elevated blood pressure among patients presenting with a
chief complaint of headache; the odds ratio (OR), with 95%CI, of elevated BP among
MA
patients who presented for headache versus those who presented for anything else; and
the frequency, with 95%CI of receipt of anti-hypertensive medications for each level of
elevated BP. We performed a second analysis, in which we compared the frequency of
elevated BP among those patients who presented with headache versus those patients
ED

who presented with a chief complaint of abdominal pain. We identified abdominal pain
as an appropriate comparator group because abdominal pain, like headache, can be
caused by many different benign or malignant process, and like headache, it is often
PT

associated with nausea, vomiting, and anorexia, which may result in dehydration and
lower blood pressure. The frequency of elevated blood pressure among headache patients
versus abdominal pain patients is also reported as ORs with 95%CI. Finally, to account
CE

for the influence of socio-demographic variables and severity of illness on the

relationship between chief complaint and blood pressure, we built a logistic regression
model in which we included chief complaint (headache or not headache) as the primary
AC

predictor variable and elevated BP (yes or no) as the dependent variable as well as the
following co-variates: age as a continuous value in years, sex, geographic region
(Northeast, Midwest, Southeast, West), race/ethnicity (Hispanic, non-Hispanic White,
non-Hispanic Black, non-Hispanic other) and the patients triage urgency, using a five
tiered system (Non-urgent,semi-urgent, urgent, emergent, immediate) .

Clinical trials dataset.

For analysis of our clinical trial data, we calculated one-hour improvement in headache as
baseline headache score minus one hour headache score. We calculated one hour
improvement in systolic BP as baseline systolic BP minus one hour systolic BP, and
similarly for diastolic BP. We calculated the correlation between these continuous data
using Pearsons r, reported the r2 and p value, and plotted the individual values using a
scatterplot. To determine the influence of elevated BP on improvement in headache, we
constructed a linear regression model in which presence of elevated BP was considered
the primary predictor variable, the improvement in headache score was the dependent
variable, and four dummy variables were created to account for the various doses of
 ACCEPTED MANUSCRIPT

intravenous metoclopramide and the adjuvant therapy used in each of the trials. The
influence of elevated BP on the improvement in pain score is reported with 95% CI.

PT
RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Results

In 2010, there were 129.8 million (95%CI: 116.0 million, 143.7 million) patient visits to
U.S. EDs. Headache was the primary complaint in 3.7% (95%CI: 3.4, 4.0%) of these
visits, corresponding to 4.8 million (95%CI: 4.2, 5.4 million) patient visits.

PT
Of patients who presented to the ED with a primary complaint of headache, 23.9%
(95%CI: 21.4%, 26.7%) had moderately elevated BP (SBP >150mmHg or a DBP
>95mmHg) versus 22.3% (95%CI: 21.3, 23.3%) of the non-headache population. Among

RI
the headache patients, 14.9% (95%CI: 13.0, 17.1%) had markedly elevated BP (SBP
>165mmHG or DBP >100mmHg) versus 11.4% (95%CI: 10.7, 12.0%) of the non-

SC
headache population. Similarly, 7.0% (5.6, 8.7%) had severely elevated BP (SBP >
180mmHg or DBP >110mmHg) versus 4.8% (4.4, 5.2%) of the non-headache
population. Odds ratios are presented in Table 1. The association between headache and

NU
elevated BP did not change substantially when we changed the comparison group from
all non-headache visits to abdominal pain visits nor when we performed a logistic
regression model controlling for triage acuity and socio-demographic characteristics, in
MA
which those with a chief complaint of headache were compared to all other chief
complaints (Table 1).

Among patients who presented to the ED primarily for headache with moderately
ED

elevated BP, 21.7% (16.1, 28.6%) were treated with an anti-hypertensive agent, versus
31.0% (22.4, 41.0%) with markedly elevated BP, and 41.1% (27.9, 55.8%) with severely
elevated BP.
PT

In the RCT dataset, 659 of 664 enrolled research subjects had baseline blood pressures
recorded. Of these, 17% (112) met our criteria for moderately elevated BP, 60 (9.1%) met
CE

our criteria for markedly elevated BP, and 21 (3%) met our criteria for severely elevated
BP.
AC

Among those with moderately elevated BP, there was no correlation between
improvement in pain score and improvement in systolic BP (r=-0.07, r2=0, p=0.465) or
diastolic BP (r=-0.03, r2=0, p=0.75) (Figure 1a, 1b). Similarly, among patients with
markedly elevated blood pressure, there was no correlation between improvement in
headache and improvement in systolic BP (r=-0.19, r2=0.04, p=0.89) or diastolic BP (r=-
0.02, r2=0, p=0.87) (Figure 2a, 2b), nor among patients with severely elevated BP, for
systolic BP (r=0.06, r2=0, p=0.81) or diastolic BP (r=0.03, r2=0, p=0.90) (Figures 3a, 3b).

As demonstrated in Figures 1,2, and 3, there was a reduction in BP between baseline and
one hour of 10-30mmHg regardless of improvement in pain score. Since pain
improvement was not driving this reduction, we sought alternative explanations.
Specifically, we determined whether baseline BP was associated with improvement in BP
among patients with elevated BP. Indeed, baseline systolic BP did explain some of the
variability in improvement in systolic BP (r=0.36, r2=0.13, p<0.01) and similarly with
baseline diastolic BP (r=0.57, r2=0.33, p<0.01) (Figures 4a &4b). The most plausible
explanation for these observations is regression of abnormal values to values that more
 ACCEPTED MANUSCRIPT

closely approximate population means.

After adjusting for investigational medication, the presence of elevated BP at baseline

had a small, but statistically significant downward trend on pain relief. Patients with
moderately elevated BP had slightly less improvement in their zero to ten pain score than
patients with BPs below this cutoff (-0.6, 95%CI: -1.2, -0.1, p=0.03). This effect was

PT
more pronounced among patients with markedly elevated BP (-0.9, 95%CI: -1.7, -0.2).
Because only 21 patients had severely elevated BP, we did not repeat this analysis for
those data.

RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Limitations

Limitations of working with NHAMCS have been well described previously.9 Primary
among these is the use of data that was collected not for research purposes but for clinical
care. Stated reasons for visit may not have captured the patients true concern. Reliability
and accuracy of the listed blood pressure are unknown. Nevertheless, we do not believe

PT
there are systematic errors in these variables that would affect our results. Some data
suggest that elevated blood pressures at triage tend to regress to the mean during
subsequent ED course.10,11 Thus, the triage BP may not accurately reflect the BP twenty

RI
minutes later. This is a limitation that cannot be addressed using NHAMCS data. Finally,
NHAMCS does not record data on a patients past history of hypertension. Thus, we

SC
cannot know how many of these patients with elevated BP had hypertension or were
being treated with anti-hypertensives.

NU
During the migraine clinical trials, we did not collect data regarding diagnosed history of
hypertension, previous blood pressures, or use of anti-hypertensive medications. It is
likely that patients with elevated blood pressure included some with diagnosed
MA
hypertension taking multiple anti-hypertensives and others who had never been
diagnosed with hypertension. Similarly, the normal blood pressure group likely contained
patients with a history of hypertension who were well controlled on anti-hypertensives. It
may be that we have lumped distinct populations. However, especially in an ED
ED

population, relying on markers of healthcare access to split groups of patients may result
in a socio-economic categorization more than a categorization based on biological
phenomena.
PT
CE
AC
 ACCEPTED MANUSCRIPT

Discussion

In this manuscript, we used two complimentary databases to explore the relationship

between elevated blood pressure and headache among patients presenting to an ED.
NHAMCS, a national probability sample database, provided a wealth of data regarding
ED practice across the country. Our migraine clinical trial database provided

PT
prospectively gathered data using a standardized methodology and validated instruments.
Using these databases, we found that 1) elevated BP is common among ED patients who
present with a chief complaint of headache; 2) ED patients with headache are more likely

RI
to have elevated BP than are ED patients with other chief complaints; 3) only a minority
of patients who present to an ED with headache and elevated BP are treated with anti-

SC
hypertensive agents; 4) Among patients who present to an ED with migraine and an
elevated BP, there is no correlation between improvement in headache and improvement
in systolic or diastolic BP; and 5) Among patients who present to an ED with migraine,

NU
elevated BP at baseline is associated with less headache relief.

The NHAMCS data demonstrate that markedly and severely elevated blood pressure is
MA
more common among patients who present to an ED with headache than among patients
who present to an ED for any chief complaint other than headache or among those who
present specifically for abdominal pain. Some have attributed elevated triage blood
pressures among ED headache patients to pain or anxiety, confounding factors that may
ED

drive both the elevated blood pressure and the need for an ED visit.12 However, our data
demonstrate that blood pressure is higher among headache patients even when compared
to other ED patients and that this relationship is directly related to degree of blood
PT

pressure elevation. Therefore, one can have more confidence that the association is in fact
genuine and not confounded by psychological distress, a finding reflected in the work of
Tanabe et.al., who compared ED BPs to home BPs in a cohort of patients, and found that
CE

the elevation of ED BP above home BP could not be explained by pain or psychological

distress.13 However, the directionality of this association between BP and headache, and
the causal pathway are still unknownone may hypothesize that elevated BP is causing
AC

headache though it is just as likely that headache is causing elevated blood pressure.
Alternatively, the relationship may be confounded by a third variable we have not
considered.

The NHAMCS data also demonstrate that even marked elevations of blood pressure in
patients with headache usually are not treated with anti-hypertensive medications.
However, the higher the blood pressure, the more likely the patient is to receive an anti-
hypertensive. Forty percent of headache patients who presented with systolic BP
>180mmHg or diastolic >110mmHg were treated with an anti-hypertensive, though 60%
were not. Heterogeneity in clinical practice reflects uncertainty. Until high quality
clinical trial data are available, it will remain unclear how best to manage these patients.
Some clinicians treat patients with acutely elevated BP with anti-hypertensives with the
goal of preventing morbidity or mortality. In our migraine clinical trials, we did not
observe any blood pressure related adverse outcomes,5,6,14 lending some support to the
practice of not treating elevated BP with anti-hypertensive in patients who present to an
ED with migraine.
 ACCEPTED MANUSCRIPT

Much to our surprise, among patients with migraine and elevated blood pressure, we did
not identify an association between improvement in blood pressure and improvement in
headache. An often-quoted aphorism in emergency medicine is that if one treats the pain,
the blood pressure will improve, an observation that may rely on regression of abnormal
values to the mean. As depicted in Figures 1, 2, and 3, there was a general improvement

PT
in BP of 10-30mmHg. This reflects a regression of abnormal values towards more normal
values, a phenomenon that explains normalization of blood pressure much better than an
association with relief of pain. Therefore, a therapeutic strategy aimed solely at lowering

RI
BP is unlikely to result in migraine improvement.

SC
Although improvement in blood pressure was not associated with improvement in
headache among patients with elevated BP, a higher baseline blood pressure was
associated with worse pain outcomes. Migraineurs with elevated blood pressure reported

NU
less improvement on a zero to ten pain scale than migraineurs without elevated blood
pressure. This phenomenon was more striking among patients with higher baseline blood
pressure elevations. However, even among patients with higher blood pressures, the
MA
difference in outcomes falls below standard thresholds for minimum clinically significant
difference.15 Therefore, the clinical impact of this finding is unlikely to be substantial. It
is clear that anti-hypertensives prevent headache in outpatient populations.16 Based on
our data, it seems less likely that anti-hypertensives will benefit migraine patients acutely.
ED

The relationship between blood pressure and pain is complex. Since initial reports of an
association between chronic severe hypertension and headache,1 there have been
PT

countless attempts to understand this association more completely. There are robust
outpatient data demonstrating that anti-hypertensive agents, regardless of class, prevent
headache, thereby suggesting that lowering blood pressure itself can prevent headache.16
CE

On the other hand, some population data and laboratory work suggest that chronic
hypertension as well as acutely elevated blood pressure moderate nociception, a
phenomenon named hypertension-associated hypalgesia.17 We do not have a ready
AC

explanation for the association between high blood pressure and headache we discovered
in the NHAMCS data. For all those patients who ask us, Is my high blood pressure
causing my headache? we can only answer maybe.

In conclusion, while there is an association between elevated blood pressure and

headache among patients presenting to an emergency department, it is not clear how best
to address patients who present to an acute care setting with both an elevated blood
pressure and headache.
 ACCEPTED MANUSCRIPT

References

1. Janeway TC. A Clinical Study of Hypertensive Cardiovascular Disease. Arch

Intern Med 1913;12:755-98.
2. The International Classification of Headache Disorders, 3rd edition (beta
version). Cephalalgia 2013;33:629-808.

PT
3. Gus M, Fuchs FD, Pimentel M, Rosa D, Melo AG, Moreira LB. Behavior of
ambulatory blood pressure surrounding episodes of headache in mildly

RI
hypertensive patients. Arch Intern Med 2001;161:252-5.
4. Kruszewski P, Bieniaszewski L, Neubauer J, Krupa-Wojciechowska B.
Headache in patients with mild to moderate hypertension is generally not

SC
associated with simultaneous blood pressure elevation. J Hypertens 2000;18:437-
44.
5. Friedman BW, Greenwald P, Bania TC, et al. Randomized trial of IV

NU
dexamethasone for acute migraine in the emergency department. Neurology
2007;69:2038-44.
6. Friedman BW, Mulvey L, Esses D, et al. Metoclopramide for acute migraine: a
MA
dose-finding randomized clinical trial. Ann Emerg Med 2011;57:475-82 e1.
7. Friedman BW, Garber L, Yoon A, et al. Randomized trial of IV valproate vs
metoclopramide vs ketorolac for acute migraine. Neurology 2014;82:976-83.
8. Eross E, Dodick D, Eross M. The Sinus, Allergy and Migraine Study (SAMS).
ED

Headache 2007;47:213-24.
9. McNaughton CD, Self WH, Pines JM. Annals of Emergency Medicine Journal
Club. Observational health services studies using nationwide administrative data
PT

sets: understanding strengths and limitations of the National Hospital Ambulatory

Medical Care Survey: answers to the May 2013 Journal Club questions. Ann Emerg
Med 2013;62:425-30.
CE

10. Pitts SR, Adams RP. Emergency department hypertension and regression to
the mean. Ann Emerg Med 1998;31:214-8.
11. Cienki JJ, Deluca LA, Feaster DJ. Course of untreated high blood pressure in
AC

the emergency department. West J Emerg Med 2011;12:421-5.

12. Denny CJ, Schull MJ. Headache and Facial Pain. In: Tintinalli JE, Strapcyznski
JS, Cline DM, Ma OJ, Cydulka RK, Meckler GD, eds. Tintinalli's Emergency Medicine A
Comprehensive Study Guide 7th Edition. New York: McGraw Hill 2011.
13. Tanabe P, Persell SD, Adams JG, McCormick JC, Martinovich Z, Baker DW.
Increased blood pressure in the emergency department: pain, anxiety, or
undiagnosed hypertension? Ann Emerg Med 2008;51:221-9.
14. Friedman BW, Garber L, Yoon A, et al. Randomized trial of IV valproate vs
metoclopramide vs ketorolac for acute migraine. Neurology 2014.
15. Todd KH, Funk JP. The minimum clinically important difference in physician-
assigned visual analog pain scores. Acad Emerg Med 1996;3:142-6.
16. Law M, Morris JK, Jordan R, Wald N. Headaches and the treatment of blood
pressure: results from a meta-analysis of 94 randomized placebo-controlled trials
with 24,000 participants. Circulation 2005;112:2301-6.
 ACCEPTED MANUSCRIPT

17. Ghione S. Hypertension-associated hypalgesia. Evidence in experimental

animals and humans, pathophysiological mechanisms, and potential clinical
consequences. Hypertension 1996;28:494-504.

PT
RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Table 1. Odds (OR, 95%CI) of headache patients having elevated blood pressure at triage
versus patients with other chief complaints

Blood pressure Headache versus any Headache versus Headache versus any
other chief complaint abdominal pain other chief complaint,

PT
controlling for triage
acuity and socio-

RI
demographic
characteristics
Moderately elevated 1.10 (0.95, 1.27) 1.06 (0.87, 1.28) 1.27 (1.09, 1.48)

SC
(SBP >150mmHg or
DBP >95mmHg)
Markedly elevated 1.37 (1.16, 1.61) 1.25 (1.01, 1.55) 1.63 (1.38, 1.93)

NU
(SBP >165mmHg or
DBP >100mmHg)
Severely elevated 1.49 (1.17, 1.90) 1.57 (1.12, 2.20) 1.87 (1.46, 2.41)
MA
(SBP > 180mmHg
or DBP
>110mmHg)
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Figure 1a. Improvement in systolic blood pressure versus improvement in pain score over
one hour among migraine patients with SBP >150mmHg or DBP >95mmHg. The
relationship between these variables is plotted as a line with 95%CI.

PT
RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Figure 1b. Improvement in diastolic blood pressure versus improvement in pain score
over one hour among migraine patients with SBP >150mmHg or DBP >95mmHg. The
relationship between these variables is plotted as a line with 95%CI.

PT
RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Figure 2a. Improvement in systolic blood pressure versus improvement in pain score over
one hour among migraine patients with SBP >165mmHg or DBP >100mmHg. The
relationship between these variables is plotted as a line with 95%CI.

PT
RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Figure 2b. Improvement in diastolic blood pressure versus improvement in pain score
over one hour among migraine patients with SBP >165mmHg or DBP >100mmHg. The
relationship between these variables is plotted as a line with 95%CI.

PT
RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Figure 3a. Improvement in systolic blood pressure versus improvement in pain score over
one hour among migraine patients with SBP >180mmHg or DBP >110mmHg. The
relationship between these variables is plotted as a line with 95%CI.

PT
RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Figure 3b. Improvement in diastolic blood pressure versus improvement in pain score
over one hour among migraine patients with SBP >180mmHg or DBP >110mmHg. The

PT
relationship between these variables is plotted as a line with 95%CI.

RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Figure 4a. Baseline systolic BP versus improvement in systolic BP between baseline and
one hour among migraine patients with SBP >150mmHg or DBP >95mmHg. The

PT
relationship between these variables is plotted as a line with 95%CI. The most likely
explanation for the association depicted here is regression of abnormal systolic blood
pressure values towards the mean systolic blood pressure.

RI
SC
NU
MA
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Figure 4b. Baseline diastolic BP versus improvement in diastolic BP between baseline

and one hour among migraine patients with SBP >150mmHg or DBP >95mmHg. The
relationship between these variables is plotted as a line with 95%CI. The most likely
explanation for the association depicted here is regression of abnormal diastolic blood
pressure values towards the mean diastolic blood pressure.

PT
RI
SC
NU
MA
ED
PT
CE
AC