Journal of the American Society of Hypertension 6(4) (2012) 291–295

Research Article
Relationship between blood pressure and persistent epistaxis
at the emergency department: a retrospective study
Moriyuki Terakura, MD*, Ryuichi Fujisaki, MD, Takaoki Suda, MD, Toshio Sagawa, MD,
and Tetsuya Sakamoto, MD
Department of Emergency Medicine, Teikyo University School of Medicine, Tokyo, Japan
Manuscript received March 1, 2012 and accepted May 8, 2012

Abstract

Background: Persistent nosebleed episodes have occurred in patients with idiopathic epistaxis from Kiesselbach’s area
despite confirmed location of the bleeding site, but the cause remains unclear. We tried to determine whether persistent
epistaxis was associated with blood pressure.
Methods and Results: Between May 2009 and May 2010, the records for 133 adult patients with idiopathic epistaxis from
Kiesselbach’s area were obtained from the emergency department of our hospital. The bleeding site was pressed with a cotton
strip for about 30 minutes, followed by checking for nosebleed. Comparison of background factors by the presence or
absence of persistent epistaxis revealed a significantly higher systolic blood pressure in patients with persistent nosebleed
than in those without (181.3
26.9 vs. 156.6
26.1 mm Hg; P < .0001). Persistent epistaxis was significantly more frequent
in patients with hypertension than in those without (26% vs. 8%; P ¼ .002). Multivariate logistic analysis revealed systolic
blood pressure to be an independent factor associated with epistaxis persistence (odds ratio, 1.03; 95% confidence interval,
1.01–1.06; P ¼ .002).
Conclusion: Proper blood pressure management is necessary for the prevention of persistent epistaxis from Kiesselbach’s area
in the clinical setting of emergency care practice. J Am Soc Hypertens 2012;6(4):291–295. Ó 2012 American Society of
Hypertension. All rights reserved.
Keywords: Persistent epistaxis; Kiesselbach’s area; systolic blood pressure.

Introduction emergency care activity increase is anticipated for this ED,

which serves as a core facility for epistaxis treatment in
In May 2009, an all-around medical aid emergency northwestern Tokyo.
department (ED) was established at Teikyo University As for causes of nosebleed, there have been reports from
Hospital to provide primary and secondary emergency settings of emergency care practice suggesting associations
care, mainly accepting patients requiring ambulance service with hypertension,1,2 yet the etiology remains largely
in the ED during general outpatient service hours and both unknown.3,4 In particular, persistent nosebleed episodes
patients brought to the ED and ambulatory patients seen have occurred among cases of idiopathic epistaxis from
outside of regular service hours. In total, there were 313 Kiesselbach’s area despite confirmed location of the
patients who were initially examined/treated for idiopathic bleeding site, but the cause remains unclear in many
epistaxis by emergency physicians at our ED during the respects. Furthermore, reported studies on patients with
1-year period between May 2009 and May 2010, and a future idiopathic epistaxis initially examined/treated by emer-
gency physicians are as yet few, and nearly all such reports
There has been no significant support for this work that could
were by otorhinolaryngologists.5,6
have influenced the outcome. None of the authors have declared This study aimed to determine associations of back-
any financial conflicts of interest regarding this study. ground characteristics with the presence or absence of
*Corresponding author: Moriyuki Terakura, MD, 2-11-1 Kaga, persistent nosebleed episodes in adult patients with idio-
Itabashi, Tokyo, Japan 173-8606. pathic epistaxis from Kiesselbach’s area (except for sponta-
E-mail: [email protected] neous hemostasis).
1933-1711/$ - see front matter Ó 2012 American Society of Hypertension. All rights reserved.
doi:10.1016/j.jash.2012.05.001
292 M. Terakura et al. / Journal of the American Society of Hypertension 6(4) (2012) 291–295

Methods the area under the curve (AUC) and optimal cutoff value
were calculated. In all interpretations for statistical signifi-
The study population comprised 133 consecutive adult cance, differences were considered significant at P < .05.
patients with idiopathic epistaxis from Kiesselbach’s area The SPSS statistical package for windows, version 11.0.1J
(excluding cases in which bleeding stopped spontaneously) (SPSS Inc., Chicago, Illinois) was used in all of the previ-
for whom documented blood pressure (BP) data between ously described statistical analyses.
May 2009 and May 2010 were available at the ED of out
hospital.
Patients with any of the following criteria were excluded: 1) Results
bleeding other than idiopathic epistaxis (ie, traumatic epistaxis, Table 1 shows comparisons of background factors by
hemorrhage from tumor), 2) age younger than 20 years, 3) whether or not persistent epistaxis occurred. Systolic BP
bleeding site other than Kiesselbach’s area or unknown, 4) was significantly higher in patients who had persistent nose-
initially examined by an otorhinolaryngologist, 5) BP data bleed episodes than in those without persistent bleeding
undetermined, 6) bleeding that subsided spontaneously, and (181.3
26.9 vs 156.6
26.1 mm Hg; P < .0001). Persistent
7) use of a drug or drugs that stimulate epinephrine release. epistaxis incidence was significantly higher in patients with
For the control of nasal bleeding, an emergency physician than in those without hypertension (26% vs. 8%; P ¼ .002).
first located the site of bleeding through a rhinoscope, the As shown in Table 2, pertinent data were subjected first to
bleeding site was then pressed with a cotton strip moistened univariate logistic analysis taking whether or not persistent
with a 1:10,000 dilution of epinephrine (Bosmin) plus 4% epistaxis occurred as a dependent variable and also taking
lidocaine hydrochloride (Xylocaine) for about 30 minutes, the following parameters as independent variables: gender
followed by checking for persistent nosebleed episodes. (female, male), age, the presence or absence of a past history
The purpose of this study was to determine associations of epistaxis, systolic BP, diastolic BP, whether or not the
of background characteristics with the presence or absence patient was on anticoagulant medication, was a known
of persistent episodes of nosebleed from Kiesselbach’s area hypertensive, was under treatment for hypertension, was
in adult patients and thereby to contribute to prevention of receiving dialysis, had any history of cerebral infarction,
persistent epistaxis. any cardiac disorder, diabetes mellitus, any malignant tumor,
The study protocol was in accordance with the tenets of any hepatic disorder, bronchial asthma, thyroid disorder, or
the Declaration of Helsinki. This study was approved by the hyperlipidemia, and whether a bleeding point was locatable
Ethics Committee of Teikyo University Hospital (Teikyo or not. Subsequently, independent factors found to show P <
Ethics Committee Approval No. 11-031). .2 and the presence or absence of persistent epistaxis (depen-
dent variable) was subjected to multivariate logistic analysis,
Statistical Analysis and the results are presented herein.
The univariate logistic analysis revealed systolic BP and
Continuous variables were expressed as mean
standard whether the patient had hypertension to be significant vari-
deviation. Two groups were compared using the t-test or ables. The independent factors showing P < .2 included:
Mann-Whitney test. Intergroup comparisons were carried systolic BP, diastolic BP, whether the patient had hyperten-
out using Fisher’s exact test or the chi-squared (c2) test. The sion, was being treated for hypertension, had any malignant
correlation coefficient was Spearman’s rank correlation coef- tumor or had bronchial asthma. Two variables, whether the
ficient. Relationships of persistent epistaxis with individual patient was receiving anticoagulant medication and whether
background factors (gender [female, male], age, any history the patient had heart disease, showed multicollinearity (R ¼
of nosebleed episodes, systolic and diastolic BPs, whether 0.917, P ¼ .0001) and therefore were excluded from the
currently on anticoagulants, whether hypertensive, whether independent variables for multivariate logistic analysis.
under treatment for hypertension, whether on dialysis therapy, In the multivariate logistic analysis, the c2 test showed
any history of cerebral infarction, any cardiac disorder, dia- P < .0001, the Hosmer-Lemeshow test was as good as
betes mellitus, malignant tumor(s), any hepatic disorder, 0.774, and the discriminatory predictive value was 77.8%,
bronchial asthma, any thyroid disorder or hyperlipidemia, indicating that systolic BP constituted an independent
and whether a bleeding point was confirmed) were evaluated factor associated with whether or not persistent epistaxis
using univariate logistic analysis. Subsequently, background occurred (P ¼ .002; OR, 1.03 [95%CI, 1.01–1.06]).
factors found to show P < .2 and the presence or absence of For evaluation of the predictability and diagnosability of
persistent nosebleed episodes were subjected to multivariate nasal persistent bleeding based on systolic BP, found to be
logistic analysis, and results of the analysis were expressed in a significant independent associative variable by multivariate
terms of odds ratios (OR) with 95% confidence intervals (CI). logistic analysis, and for calculation of the optimal cutoff
Furthermore, for each significant independent variable value, we constructed a ROC curve (Fig. 1). The AUC of
derived from the multivariate logistic analysis, a receiver ROC for evaluating predictability and diagnosability from
operating characteristic (ROC) curve was constructed and the ROC curve was calculated to be 0.746 (95%CI,
 M. Terakura et al. / Journal of the American Society of Hypertension 6(4) (2012) 291–295 293

Table 1
Comparison of background factors by the presence or absence of persistent epistaxis
Number Persistent Epistaxis (þ) Persistent Epistaxis (-) P
Total 133 34 99 —
Gender (female:male) 47:86 11:23 36:63 NS
Age (mean
SD) 63.9
12.8 64.2
11.5 63.8
13.2 NS
History of epistaxis (absence:presence) 117:16 30:4 87:12 NS
Systolic blood pressure (mm Hg) 162.9
28.3 181.3
26.9 156.6
26.1 <.0001
Diastolic blood pressure (mm Hg) 91.0
16.4 95.6
15.1 89.4
16.6 NS
Anticoagulant medication (absence:presence) 101:32 24:10 77:22 NS
History of hypertension (absence:presence) 61:72 8:26 53:46 .002
Current hypertensive treatment (absence:presence) 84:49 17:17 67:32 NS
Receiving dialysis (absence:presence) 130:3 33:1 97:2 NS
History of
Cerebral infarction (absence:presence) 123:10 32:2 91:8 NS
Cardiac disorder (absence:presence) 105:28 24:10 81:18 NS
Diabetes mellitus (absence:presence) 121:12 30:4 91:8 NS
Malignant tumor (absence:presence) 122:11 33:1 89:10 NS
Hepatic disorder (absence:presence) 130:3 33:1 97:2 NS
Bronchial asthma (absence:presence) 129:4 32:2 97:2 NS
Thyroid disorder (absence:presence) 131:2 34:0 97:2 NS
Hyperlipidemia (absence:presence) 122:11 31:3 91:8 NS
Bleeding point (not locatable:locatable) 2:131 1:33 1:98 NS

0.653–0.840). The optimal cutoff value calculated in terms patients with idiopathic epistaxis from Kiesselbach’s area.
of the Youden index (maximum of [sensitivity þ speci- Furthermore, systolic BP was significantly higher in patients
ficity-1])7 was a systolic BP of 160.5 mm Hg. with than in those without persistent epistaxis.
It is recognized that about 60% of individuals experience
Discussion epistaxis at least once in their lives, 6% of whom require
some form of treatment.8 Although there are sporadic reports
The present study showed systolic BP to be an indepen- suggesting a causative association with hypertension,1,2 the
dent factor associated with persistent epistaxis in adult etiology of epistaxis still remains a controversial.3,4 The

Table 2
Results of logistic analysis
Univariate Logistic Analysis P Multivariate Logistic Analysis P
Odds Ratio (95% CI) Odds Ratio (95% CI)
Gender (female:male) 1.20 (0.52–2.73) .909
Age (mean
SD) 1.00 (0.97–1.03) .713
History of epistaxis (absence:presence) 0.97 (0.29–3.23) .922
Systolic blood pressure (mm Hg) 1.04 (1.02–1.05) .0001 1.03 (1.01–1.06) .002
Diastolic blood pressure (mm Hg) 1.22 (0.10–1.05) .071 1.0 (0.97–1.03) .799
Anticoagulant medication (absence:presence) 1.49 (0.61–3.51) .162
History of hypertension (absence:presence) 3.75 (1.54–9.10) .003 1.55 (0.04–6.00) .522
Current hypertensive treatment (absence:presence) 2.10 (0.95–4.63) .068 1.63 (0.48–5.54) .436
Receiving dialysis (absence:presence) 1.47 (0.13–16.7) .426
History of
Cerebral infarction (absence:presence) 0.77 (0.14–3.53) .378
Cardiac disorder (absence:presence) 1.88 (0.764–4.60) .064
Diabetes mellitus (absence:presence) 1.52 (0.43–5.40) .277
Malignant tumor (absence:presence) 0.27 (0.03–2.19) .187 0.34 (0.038–3.00) .327
Hepatic disorder (absence:presence) 1.47 (0.13–16.7) .426
Bronchial asthma (absence:presence) 3.03 (0.41–22.4) .099 8.39 (0.64–110.4) .106
Thyroid disorder (absence:presence) 0.006 (0.00–1.44) .402
Hyperlipidemia (absence:presence) 1.01 (0.28–4.41) .741
Bleeding point (not locatable:locatable) 2.30 (0.18–48.8) .426
294 M. Terakura et al. / Journal of the American Society of Hypertension 6(4) (2012) 291–295

assume that the same pathophysiological events take place

in the nasal cavity, thereby giving rise to epistaxis.
To evaluate the predictability and diagnosability of persis-
tent epistaxis based on systolic BP, which was noted to be
a significant independent variable by multivariate logistic
analysis as well as to calculate an optimal cutoff value, we
constructed a ROC curve (Fig. 1). An AUC of 0.746 thus
determined was moderately accurate for predictability and
diagnosability. The optimal cutoff value of systolic BP for
predicting persistent epistaxis was 160.5 mm Hg. Inasmuch
as the systolic BP for the group of patients with nasal persis-
tent bleeding was 181.3
26.9 mm Hg (Table 1), it would be
of value to undertake treatment of epistaxis by keeping in
mind that patients showing a systolic BP higher than
155 mm Hg have a risk of persistent epistaxis.
A report on persistent epistaxis from Kiesselbach’s area
by Mladina13 dealt with associations of this syndrome
Figure 1. Receiver operating characteristic (ROC) curve for with (a) deformity/deviation of the nasal septum, (b) vaso-
evaluation of predictability and diagnosability of persistent dilatation in Kiesselbach’s plexus, (c) infection of the nasal
epistaxis based on systolic blood pressure and for calculation mucosa, and (d) inheritance. Endoscopic examination of
of optimal cutoff value. The area under the ROC curve was
the nose is reportedly required with respect to (a) and (b);
calculated to be 0.746 (95% confidence interval [CI], 0.653–
however, no endoscopic examination was performed in
0.840), indicating moderately accurate predictability and diag-
nosability. The optimal cutoff value calculated in terms of the the present series because details about such examination
Youden index (maximum of [sensitivity þ specificity-1]) was are lacking. Future assessments including nasal endoscopy
a systolic blood pressure of 160.5 mm Hg. are thus needed.
The frequency with which epistaxis recurred was 12%
nose is endowed with moisturizing functions to prevent to 18.5% for patients treated by otorhinolaryngological
foreign body intrusion from the outside, for which purpose specialists as reported in Japan,5,6 and was 25.6% for the
the nasal mucosa has a structure allowing plasma constitu- present series (treated by emergency physicians). Hence,
ents to permeate capillary walls below the mucosal epithe- the latter appears to be somewhat higher. Although this
lium of the nose so that they can readily reach the difference might be attributable at least in part to relatively
epithelial surface of the nasal mucosa. These plasma constit- insufficient technical skill in this specialty on the part of
uents are liable to be effused not only onto the epithelial emergency physicians, improved future prevention of
surface of the nasal mucosa but into the intercellular matrix persistent epistaxis would be expected from appropriate
of the nasal mucosal epithelium as well, thereby increasing BP management based partly on the present study results.
inflammatory substances with a consequent state that may
give rise to inflammation of the nasal mucosa.9 In the nasal Limitations
mucosa, moreover, there is an abundance of porous arterioles
that are structurally fragile as regards hemodynamic changes As for limitations of this investigation, inter-individual
because the nasal mucosa is devoid of elastica interna.10 differences in the treatment ability of surgeons/physicians
These anatomical structural features peculiar to the nose were not reflected and adequate assessments for inflamma-
seem to etiologically account for the bleeding liability. tion of the nasal mucosa and vasculature by nasal endos-
Systolic BP was noted to be an independent factor associ- copy were not performed. It is considered preferable that
ated with persistent epistaxis in the present series. It is endoscopic examination of the nose be conducted wherever
generally recognized that systolic BP increases progres- feasible.
sively with advancing age whereas diastolic BP rises Causes of nosebleed are currently in dispute, as noted
progressively up to an age of about 50 years and remains previously. The BP measurements in this investigation
practically unchanged thereafter. A majority of subjects represent pretreatment data obtained just after arrival at
with hypertension who were older than 60 years of age are our hospital. As such, these BP data do not include changes
hypertensive in terms of systolic BP alone.11 There is a report that might have occurred during the treatment process.
demonstrating that sheer stress on blood vessels is increased However, anxiety arising in connection with sudden
by elevated pulse pressure resulting from systolic hyperten- bleeding from the nose can undeniably influence BP. Reas-
sion,12 whereas another report describes the spread of sessments employing a prospective study design encom-
unstable blood microcirculation within the vascular system passing objective indices that reflect aspects of anxiety
as leading to a risk of stroke.11 It might be reasonable to are needed in the future.
 M. Terakura et al. / Journal of the American Society of Hypertension 6(4) (2012) 291–295 295

Conclusion end-organ damage in patients with hypertension?

Laryngoscope 1999;109:1111–5.
The present data demonstrate systolic BP to be an inde- 5. Nario K, Miyahara H, Sasai H, Kamakura A,
pendent factor associated with persistent epistaxis in adult Kajikawa H, Matushiro N. A clinical study of hospital-
patients with idiopathic epistaxis from Kiesselbach’s area. ized epistaxis patients [in Japanese]. Nihon Bika Gak-
Furthermore, systolic BP was significantly higher in kai Kaishi 2008;47:1–7.
patients with than in those without persistent epistaxis. 6. Ishii Y, Ohki M, Yamaguchi S, Ohkubo H, Sakurai S,
The present study suggests that proper management of Ogoshi T. Temporary aspirin discontinuation to stop
BP is necessary for the prevention of persistent epistaxis epistaxis precipitating transient ischemic attack-a case
from Kiesselbach’s area in the clinical setting of emergency report [in Japanese]. Nihon Bika Gakkai Kaishi 2010;
care practice. 49:108–11.
7. Akobeng AK. Understanding diagnostic tests 3:
Acknowledgments Receiver operating characteristic curves. Acta Paediatr
2007;96:644–7.
The authors are heartily thankful to the otolaryngologist 8. Petruson B, Rudin R. The frequency of epistaxis in
in Teikyo University School of Medicine for helping with a male population sample. Rhinology 1975;13:129–33.
the hemostatic treatment. 9. Cauna N. The fine structure of the arteriovenous anas-
tomosis and its nerve supply in the human nasal respi-
References ratory mucosa. Anat Rec 1970;168:9–21.
10. Mladina R, Skitarelic NB, Skitarelic NP. Is recurrent
1. Herkner H, Laggner AN, M€ ullner M, Formanek M, epistaxis from Kiesselbach’s area (REKAS) in any rela-
Bur A, Gamper G, et al. Hypertension in patients present- tionship to the hemorrhoidal disease? Med Hypotheses
ing with epistaxis. Ann Emerg Med 2000;35:126–30. 2009;73:955–7.
2. Herkner H, Havel C, M€ ullner M, Gamper G, Bur A, 11. Lee HY, Oh BH. Aging and arterial stiffness. Circ J
Temmel AF, et al. Active epistaxis at ED presentation 2010;74:2257–62.
is associated with arterial hypertension. Am J Emerg 12. Giannakoulas G, Giannoglou G, Soulis J, Farmakis T,
Med 2002;20:92–5. Papadopoulou S, Parcharidis G, et al. A computational
3. Celik T, Iyisoy A, Yuksel UC, Karahatay S, Tan Y, model to predict aortic wall stresses in patients with
Isik E. A new evidence of end-organ damage in the systolic arterial hypertension. Med Hypotheses 2005;
patients with arterial hypertension: epistaxis? Int J Car- 65:1191–5.
diol 2010;141:105–7. 13. Mladina R. REKAS (recurrent epistaxis from Kiessel-
4. Lubianca Neto JF, Fuchs FD, Facco SR, Gus M, bach’s area syndrome. Chir Maxillofac Plast 1985;15:
Fasolo M, Mafessoni R, et al. Is epistaxis evidence of 91–5.