ARTICLE IN PRESS

Different Clinical Phenotypes of Embolic Stroke of

Undetermined Source: A Subgroup Analysis of 86 Patients

D1X XSilvio Piffer, MD,D2X X* D3X XValeria Bignamini, MD,D4X X* D5X XUmberto Rozzanigo, MD,D6X X†
D7X XPiero Poletti, PhD,D8X X‡ D9X XStefano Merler,D10X X‡ D1X XElisabetta Gremes,D12X X* and
D13X XDomenico Marco Bonifati, MD, PhDD14X X*,§

Introduction and Study Aim: Embolic strokes of undetermined source (ESUS) represent
a rather recent diagnostic entity under clinical research for relapse prevention in crypto-
genic stroke patients. Despite strict diagnostic criteria, ESUS definition ignores major
clinical and radiological aspects, so including heterogeneous cases and probably influ-
encing trial results. This study researches clinically relevant phenotypes among ESUS
patients. Patients and Methods: We evaluated ESUS patients admitted at Trento Stroke
Unit over a 4-year period. Vascular risk factors (RFs), neurological deficit severity, pres-
ence of potential embolic sources, and ASCOD phenotype were recorded. Ischemic
lesions were categorized considering their extension in 4 groups. Subgroup compari-
sons by predefined differences in age, amount of RFs, history of previous stroke, deficit
severity, and stroke lesion extension were done. Results: ESUS cases were 86. Patients
younger than 50 years old (n = 17) had a lower prevalence of RFs, left atrial enlarge-
ment, left ventricle diastolic dysfunction, a higher proportion of ASCOD score A0
(P < .05). Patients without RFs (n = 18) differed from those with greater than or equal
to 3 RFs (n = 23) for a younger age and a lower prevalence of potential causes of embo-
lism (P < .05). Patients without a previous stroke (n = 70) were younger, had a lower
prevalence of RFs, left ventricle diastolic dysfunction, a higher prevalence of ASCOD
score A0 (P < .05). No differences were observed comparing minor and major clinical
and radiological strokes. Discussion and Conclusions: ESUS patients can be distinguished
in 2 opposite phenotypes defined by a lower and a higher load of atherosclerotic
pathology. They may suggest possible underlying pathogenic mechanisms and sup-
port interpretation of ongoing trials results.
Key Words: Embolic stroke of undetermined source (ESUS)—Cryptogenic stroke—
radiological features of ischemic lesion—TOAST stroke classification
© 2018 National Stroke Association. Published by Elsevier Inc. All rights reserved.

Introduction
Embolic strokes of undetermined source (ESUS) repre-
sent a rather recent diagnostic entity that was proposed in
2014 to foster clinical research in relapse prevention in that
large but still rather poorly studied subgroup of the crypto-
genic stroke patients.1 Cryptogenic strokes still account for
almost 25% of all stroke cases.1 ESUS entity has been built
on the evidence that most cryptogenic strokes are thrombo-
embolic and, regardless the possible arterial or cardiac
source, probably more prone to benefit from an anticoagu-
lation treatment in comparison with antiaggregation.1
ESUS definition is based on strict diagnostic criteria and
defined instrumental exams. Briefly the ischemic lesion
must not be a lacunar stroke and there must be no

From the *Department of Emergency, Section of Neurology, Santa Chiara Hospital, Trento, Italy; †Department of Radiology, Neuroradiology,
Santa Chiara Hospital, Trento, Italy; ‡Fondazione Bruno Kessler, Trento, Italy; and §Department of Medicine, Unit of Neurology, Azienda ULSS 2
Marca Trevigiana, Treviso, Italy.
Received July 11, 2018; revision received August 7, 2018; accepted August 12, 2018.
Grant Supports: no-ones.
Address correspondence to Silvio Piffer, Department of Neurological Disorders, Neurology, Santa Chiara Hospital, Largo Medaglie d'Oro 9,
38122 Trento, Italy. E-mail: silviopiffer@yahoo.it
1052-3057/$ - see front matter
© 2018 National Stroke Association. Published by Elsevier Inc. All rights reserved.
https://doi.org/10.1016/j.jstrokecerebrovasdis.2018.08.029

Journal of Stroke and Cerebrovascular Diseases, Vol. &&, No. && (&&), 2018: &&
&& 1
 ARTICLE IN PRESS
2 S. PIFFER ET AL.
evidence of atherosclerotic stenosis causing greater than
or equal to 50% luminal stenosis in arteries supplying the
infarcted brain area, major cardiac sources of embolism,
other specific causes of stroke.1 Although this strict diag-
nostic criteria, from an aetiopathogenetic point of view,
important clinical aspects such as age, presence of vascu-
lar risk factors, history of a previous stroke, severity of the
index stroke are not taken into account and, with the
unique exclusion of lacunar stroke, no relevance has been
given to radiological features, such as extension and
topography, of the ischemic lesions. As a consequence it is
likely that highly heterogeneous patients are grouped
under the umbrella diagnosis of ESUS. Till now only few
studies have described these patients in detail.2
From 2014 4 large international trials, comparing efficacy
and safety of the direct oral anticoagulants versus aspirin
in secondary prevention of stroke in patients with ESUS,
have started (NCT02239120, NCT02313909, NCT02427126,
and NCT03192215). One of them, NCT02313909—NAVI-
GATE ESUS trial, which evaluated rivaroxaban efficacy,
was stopped in advance in late 2017 for futility. The trial
showed comparable efficacy between rivaroxaban and
aspirin, with an increased, even if within the expected
range, bleeding rate in the rivaroxaban arm.3
It is possible that heterogeneity of ESUS patients may
hinder the success or even the interpretation of results of
clinical trials. Therefore, despite the still unknown thera-
peutic implications of ESUS construct for daily clinical
practice, additional clinical, and radiological features may
give important clues for advancement in research and
care of cryptogenic stroke patients.
Till now only few studies has analyzed possible ESUS
patient subgroup, prevalently according to the age.4,5 In
this study we analyze, from a clinical and radiological
point of view, all ESUS patients admitted in the stroke
unit (SU) of Santa Chiara Hospital (Trento) Registry from
2012 to 2015 looking for the presence of possible different
pathogenic phenotypes.
Methods
Study Population
The study population was derived from the Trento
Stroke Unit Registry which includes all consecutive
patients admitted to the SU of “Santa Chiara” Hospital of
Trento—Italy. The study period was from 1st January
2012 to 31th December 2015.
All patients with ESUS as defined by the Cryptogenic
Stroke/ESUS International Working Group criteria were
included.1
Items Recorded
The following clinical data were collected: gender, age,
vascular risk factors, comorbidities, CHA2DS2VASc,
ROPE score, severity of neurological deficit at onset
measured by NIH Stroke Scale (NIHSS), disability at dis-
charge measured by modified Rankin Scale (mRS), and
NIHSS.6,7,8-10 We defined a patient having dyslipidemia if
at least 1 of the following was present: (1) previous docu-
mented finding of hypercholesterolemia; (2) ongoing
treatment with statin; and (3) LDL level greater than
130 mg/dL (3.3 mmol/L). Hypertension and diabetes
mellitus were considered present if the patient was
known to be affected by these diseases or if a specific ther-
apy was prescribed. Diabetes mellitus was also consid-
ered present if it was diagnosed during hospital stay.
Minor stroke was defined as a National Institutes of
Health Stroke Scale (NIHSS
3).11
Diagnostic work-up was evaluated to look for possible
ESUS causes as defined by the Cryptogenic Stroke/ESUS
International Working Group.1 We also looked for the
presence of: nonsignificant atherosclerosis of the aortic
arch, supra-aortic and intracranial arteries; minor sources
of cardio-embolism and other possible causes of stroke in
order to classify ESUS patients according to the ASCOD
classification.1,12 Further detailed analysis of transthoracic
echocardiography was performed considering the pres-
ence of left atrial enlargement, reduced left ventricle ejec-
tion fraction, segmental left ventricle hypo/akinesia and
left ventricle diastolic dysfunction.13
The radiological analysis considered firstly the involve-
ment of the anterior or posterior cerebral circulation and
then the type of the ischemic lesion distinguishing the fol-
lowing subtypes:14
- Cortical-subcortical stroke (CSS): brain infarct
involving, completely or partially, both the corti-
cal and subcortical vascularization territory of a
leptomeningeal artery.
- Subcortical stroke: brain infarct involving,
completely or partially, the subcortical vasculari-
zation territory of a leptomeningeal artery.
- Cortical stroke: single or multiple brain infarcts
involving the cortical vascularization territory of
a leptomeningeal artery.
- Multiple ischemic lesions: evidence of more than
1 recent ischemic lesions in different vascular
territories regardless the subtype.
Subgroup Analysis
Firstly, subgroup analysis was performed comparing all
the collected items in patients younger and older than
50 years of age ( <50 and 50), with and without 3 or more
RFs, with and without a previous stroke and with and with-
out a minor stroke. We considered the total amount of RFs
among these 7: hypertension, dyslipidemia, diabetes melli-
tus, smoking, history of coronaropathy or peripheral arterio-
pathy, previous transient ischemic attack (TIA) or stroke,
and OSAS.
 ARTICLE IN PRESS
CLINICAL PHENOTYPES OF EMBOLIC STROKE OF UNDETERMINED SOURCE
Secondly, a radiological subgroup analysis was con-
ducted comparing patients grouped according to the dif-
ferent ischemic lesions types for all the collected items.
Statistical Analysis
Statistical analysis was performed using the Wilcoxon-
Mann-Whitney test for quantitative variables, the chi-
square test for categorical variables and the proportional
test for dichotomic variables. Difference was considered
significant with a P value <.05.
Results
From 1st January 2012 to 31st December 2015, 439
patients with an ischemic stroke were admitted to the SU,
86 (20%) met criteria for ESUS diagnosis, male:female
ratio 1,7. Clinical and radiological data are summarized in
Tables 1
Table 3.
ESUS Subgroup Results
Young ESUS
The median age of ESUS patients was 63 years (inter-
quartile ratio IQR 52.7-69.7). Among patient with ESUS,
17 (20%) were older than 50 years of age, 19 (22%) were
between 51 and 60 years, 29 (34%) between 61 and
70 years, and 21 (24%) had more than 71 years of age.
In comparison with older ESUS patients (69-80.2%),
those younger than 50 years of age had a lower preva-
lence of dyslipidemia (29% versus 65%), lower median
value of CHA2DS2VASc (0 versus 2), lower prevalence of
left atrial enlargement (12% versus 29%) and left ventricle
diastolic dysfunction (0% versus 35%), but a higher pro-
portion of patent foramen ovale (PFO) (53% versus 17%)
and of ASCOD score A0 corresponding to absence of ath-
erosclerosis (82% versus 45%) (P < .05).
ESUS Patients Without Common Vascular Risk
Factors
The most frequent risk factors in ESUS patients were:
hypertension (59.3%), dyslipidemia (58.1%) and smoking
(24.4%).
Considering the total number of risk factors among
hypertension, dyslipidemia, diabetes mellitus, smoking,
coronaropathy, or peripheral arteriopathy, previous TIA/
stroke and OSAS, 18.6% of ESUS patients (n = 16) had no
risk factors, 1 risk factor was present in the 24.4% (n = 21),
2 in the 30.2 % (n = 26), 3 or more in the 26.8% (n = 23).
Patients without risk factors differed significantly from
all the others for a lower presence of potential causes of
ESUS: 25% versus 67%, 65%, and 74% of patients with 1, 2
and 3 risk factors respectively (P < .05).
Patients without risk factors in comparison with those
with 3 risk factors were younger (mean 54 versus 64
3
years) and had a lower prevalence of arteriogenic-emboli
sources (0% versus 30%). Moreover considering ASCOD
classification, the combined phenotypes A0C0 (absence of
atherosclerotic and cardiac pathology) and A0C2/3
(absence of atherosclerotic pathology and minor cardio-
embolic sources corresponding in our cases to PFO and
atrial septal aneurysm) were both significantly higher in
ESUS patients without risk factors (44% versus 18% and
44% versus 13% respectively, P <.05). On the opposite, in
ESUS patients without risk factors, the combined pheno-
types A2/3C0 (mild nonsignificant atherosclerotic pathol-
ogy and absence of cardiac pathology), and A2/3C2/3
(mild nonsignificant atherosclerotic pathology and minor
cardioembolic sources such as mitral annulus calcifica-
tion, calcification aortic valve, segmental akinesia of the
left ventricle) were lower (6% versus 27% and 6% versus
44% respectively, P < .05).
ESUS in Patients with Previous TIA/Stroke
A previous stroke was present in 9 ESUS patients
(10.4%) while a TIA in 7 cases (8.1%).
Patients without a previous TIA/stroke were youn-
ger (mean age 59 versus 67 years), had a lower mean
number of risk factors (2 versus 3), a lower mean
CHA2DS2VASc (1.6 versus 4.6), a higher ROPE score (5
versus 3), a lower prevalence of arteriogenic emboli
sources (16% versus 38%), a higher prevalence of A0
phenotype with ASCOD classification (60% versus
19%), and a lower prevalence of diastolic dysfunction at
the TTE (20% versus 63%), (P < .05).
Minor ESUS
ESUS patients with minor stroke were 46 (53.5%).
Minor and severe strokes did not differ for demo-
graphic features, vascular risk factors, and possible causes
of stroke. Patients with minor stroke had a lower mean
NIHSS score at discharge (.8 versus 5.4), a higher preva-
lence of territorial cortical lesions (37% versus 10%), and a
lower prevalence of multiple ischemic lesions (11% versus
20%), (P < .05).
Radiology in ESUS Patients
Stroke lesions involved brain anterior circulation in 59
cases (69%), posterior circulation in 21 cases (24%), and
both circulations in 5 cases (6%). One patient (1.2%), with
a history of acute and persistent left hemianopsia and
hemiparesis suggestive for a territorial stroke, had no
clear ischemic lesion on head CT scan but a diffuse leu-
koencephalopathy, no MRi was performed.
ESUS patients had several types of ischemic lesions
without a clear prevalence of one of them. Patients with a
CSS were 33 (38.4%): the vascular territory of the index
cerebral artery was involved completely in 9 cases
 4
Table 1. Clinical, instrumental, and radiological data of all ESUS cohort

CLINICAL DATA. N = 86. n/Med %/IQR POTENTIAL CAUSES OF ESUS. # N n %

ANAGRAPHICS No evidence of potential causes of ESUS 86 26 30.2
Sex [f] 31 36.0 Arteriogenic emboli 86 17 19.8
Age at ictus onset [years] 63 52.7-69.7 Aortic arch atherosclerotic plaques 86 16 18.6
VASCULAR RISK FACTORS Cerebral artery nonstenotic plaques with ulceration 86 1 1.2
hypertension 51 59.3 Minor-Risk Potential Cardioembolic Sources 86 24 27.9
Diabetes mellitus 6 7.0 Mitral valve 86 2 2.3
Dyslipidaemia 50 58.1 Myxomatous valvulopathy with prolapse 86 1 1.2
Smoke 21 24.4 Mitral annular calcificatio 86 1 1.2
Previous smoke 20 23.3 Aortic valve 86 2 2.3
Estroprogestinic therapy 3 3.5 Calcific aortic valve 86 2 2.3
Previous TIA/Stroke 16 18.6 Non-atrial fibrillation atrial dysrhythmias and stasis 86 2 2.3

ARTICLE IN PRESS
Cardiopathy 16 18.6 Atrial asystole and sick-sinus syndrome 86 1 1.2
Sclero-hypertensive 8 9.3 Atrial high-rate episodes 86 1 1.2
Ischemic 7 8.1 Atrial structural abnormalities 86 14 16.3
Valvulary 1 1.2 Atrial septal aneurysm (ASA) 86 12 14.0
Coronaropathy or peripheral arteriopathy 8 9.3 Chiari network 86 2 2.3
Obstructive sleep apnea syndrome (OSAS) 0 0 Left ventricle 86 8 9.3
Vascular risk factors sum* Moderate systolic/diastolic dysfunction (global/regional) 86 8 9.3
0 16 18.6 Paradoxical embolism 31 21 67.7
1 21 24.4 Patent foramen ovale 31 21 67.7
2 26 30.2 RADIOLOGICAL ANALYSIS. N = 86. n %
3 23 26.7 BRAIN CIRCULATION INVOLVEMENT
CHA2DS2VASc 2 1-3 Anterior circulation 59 69
STROKE ONSET Posterior circulation 21 24
NIHSS in emergency room 3 1-6 Anterior and Posterior circulations 5 6
Minor Stroke [NIHSS
3] 46 53.5 Vascular leucoencephalopathy 1 1
DISABILITY A DISCHARGE ISCHEMIC LESION EXTENSION
NIHSS at discharge 1 0-3 Territorial complete cortical-subcortical stroke 9 10.5
mRS before index stroke 0 0-0 Territorial partial cortical-subcortical stroke 24 27.9
mRS at discharge 1 1-4 Territorial cortical stroke 21 24.4
Territorial subcortical stroke 18 21.0
Multiple stroke lesions 13 15.1
Vascular leucoencephalopathy 1 1.2
Abbreviations: Med, median; IQR, interquartile range 25%-75%. mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; SU, stroke unit; TIA, transient ischemic attack.

S. PIFFER ET AL.
*Sum of the following vascular risk factors: hypertension, dyslipidaemia, diabetes mellitus, smoke, coronaropathy or peripheral arteriopathy, previous TIA/stroke, and OSAS.
†
Valvular biological prosthesis. # from Hart et al. Lancet Neurology 2014; 13:429-38.
 CLINICAL PHENOTYPES OF EMBOLIC STROKE OF UNDETERMINED SOURCE
Table 2. Subgroup analysis. N = 86

N/Med (%/IQR)
50 yy n = 17 >50 yy n = 69 S No RFs n = 16 RFs  3 n = 23 S No PreS n = 70 PreS n = 16 S

Age [yy] 43.6 (36.3-46.9) 65.5 (58.5-71.4) 52.0 (44.1-65.5) 63.4 (58.5-69.2) * 61.6 (50.7-69.2) 65.6 (62.6-72.6) *

ARTICLE IN PRESS
CHA2DS2VASc 0 (0-1) 2 (1-3) * .5 (0-1) 3 (2.5-4) 2 (1-2) 4 (3.8-6) *
RFs sum= 0 7 (41.2) 9 (13.0) * 16 (100) 0 (0) 16 (22.9) 0 (0) *
RFs sum = 1 or 2 9 (52.9) 38 (55.1) 0 (0) 0 (0) 41 (58.6) 6 (37.5)
RFs sum  3 1 (5.9) 22 (31.9) 0 (0) 23 (100) 13 (18.6) 10 (62.5)
No evident potential sources 5 (29.4) 21 (30.4) 9 (56.3) 5 (21.7) 23 (32.9) 3 (18.8)
Arteriogenic sources 0 (0) 17 (24.6) 0 (0) 7 (30.4) * 11 (15.7) 6 (37.5) *
Minor-Risk Cardioembolic Sources 8 (47.1) 34 (49.3) 4 (25) 14 (60.9) * 30 (42.9) 12 (75) *
PFO 9 (52.9) 12 (17.4) 5 (31.3) 5 (21.7) 19 (27.1) 2 (12.5)
ASCOD A0 14 (82.4) 31 (44.9) * 14 (87.5) 7 (30.4) * 42 (60) 3 (18.8) *
ASCOD C0 4 (23.5) 39 (56.5) * 8 (50) 10 (43.5) 33 (47.1) 10 (62.5)
ASCOD A0C0 3 (17.6) 17 (24.6) * 7 (43.8) 4 (17.4) * 20 (28.6) 0 (0) *
ASCOD A2/3C0 1 (5.9) 22 (31.9) 1 (6.3) 6 (26.1) 13 (18.6) 10 (62.5)
ASCOD A0C2/3 11 (64.7) 14 (20.3) 7 (43.8) 3 (13.0) 22 (31.4) 3 (18.8)
ASCOD A2/3C2/3 2 (11.8) 16 (23.2) 1 (6.3) 10 (43.5) 15 (21.4) 3 (18.8)
LEA 2 (0) 29 (42.0) * 3 (18.8) 10 (43.5) 23 (32.9) 8 (50.0)
LVDD 0 (11.8) 24 (34.8) * 1 (6.3) 7 (30.4) 14 (20.0) 10 (62.5) *
Abbreviations: IQR, interquartile range 25%-75%; LEA, left atrial enlargement; LVDD, left ventricular diastolic dysfunction; Med, median; Min, minimum; Max, maximum; PreS, previous
stroke; PFO, patent foramen ovale; RFs, (vascular) risk factors; S, statistically significant difference in comparison with all other radiological type of ischemic lesions; YY, age at stroke onset in
years.
*P value <.05.

5
 ARTICLE IN PRESS
6 S. PIFFER ET AL.

Table 3. Radiological analysis. N = 85

N / Med (% / IQR) Total n=86 CSS n=33 CS n=21 SS n=18 MS n=13

Age [yy] 63 (52.7-69.7) 63.4 64.5 59.9 63.3 (56.3-69.2)
(52.5-70.4) (54.1-70.8) (48.9-67.8)
CHA2DS2VASc 2 (1-3) 2 (1-4) 2 (1-3) 2 (1-3) 1 (0-2)
Hypertension 51 (59) 19 (56) 13 (62) 11 (61) 7 (54)
Diabetes mellitus 6 (7) 6 (18) 0 (0) 0 (0) 0 (0)
Active smoke 21 (24) 10 (30) 3 (14) 4 (22) 3 (23)
Dyslipidaemia 50 (58) 19 (56) 13 (62) 9 (50) 8 (62)
Previous TIA/Stroke 16 (19) 6 (18) 4 (19) 4 (22) 1 (8)
Coronary/peripheral artheriopathy 8 (9) 4 (12) 0 (0) 3 (17) 1 (8)
NIHSS at onset 3 (1-6) 3 (1-7) 2 (1-3)* 4 (2-6.75) 6 (2-7)
Minor Stroke (NIHSS
3) 46 (54) 17 (52) 17 (81) 6 (33) 5 (38)
mRS at discharge 1 (1-4) 2 (1-4) 1 (0-1)* 3 (1-4) 1 (1-4)
No evident potential sources 26 (30) 10 (30) 6 (29) 4 (22) 5 (38)
Arteriogenic sources 17 (20) 7 (21) 2 (10) 6 (33) 2 (15)
Minor-Risk Cardioembolic Sources 24 (28) 17 (52) 11 (52) 8 (44) 6 (46)
PFO (evaluated in 31 Pt) 21 out of 31(68) 6 out of 11 7 out of 9 4 out of 6 4 out of 5
(54) (78) (67) (80)
ASCOD A0C0 20 (23) 10 (30) 6 (29) 3 (17) 1 (8)
ASCOD A2/3C0 23 (27) 11 (33) 3 (14) 7 (39) 1 (8)
ASCOD A0C2/3 25 (29) 7 (21) 9 (43) 4 (22) 5 (38)
ASCOD A2/3C2/3 18 (21) 5 (15) 3 (14) 4 (22) 6 (46)
Abbreviations: CSS, cortical-subcortical stroke; CS, cortical stroke; Med, median; IQR, interquartile range 25%-75%; Min, minimum;
Max, maximum; MS, multiple ischemic lesions; PFO, patent foramen ovale; SS, subcortical stroke; YY, age at stroke onset in years.
*BOLD: value significantly different in comparison with all other radiological subgroups P < .05.

(10.5%), partially in 24 (27.9%); with a subcortical stroke
(SS) were 18 (21%); with a cortical stroke (CS) 21 (24.4%)
and with multiple ischemic lesions (MS) 13 (15.1%).
CS patients differed from all other subgroups for a
lower value of NIHSS score at admission and of mRS
score at discharge (P < .05), Table 3.
Discussion
This cohort of ESUS patients shows similar clinical fea-
tures to other previously reported case-series presenting:
a slightly male prevalence, a comparable median age of
onset, hypertension as the most frequent vascular risk fac-
tor, a low median score at NIHSS with an absolute preva-
lence of minor strokes, a low median score at mRS at
discharge, a comparable percentage of patients without a
potential causes of ESUS as defined by the Cryptogenic
Stroke/ESUS International Working Group, a comparable
number of patients with evidence of PFO.1,2,15-18 More-
over the majority of cases had small ischemic lesion lim-
ited to cortical or cortical-subcortical extension. At the
same time, as expected, it appears heterogeneous regard-
ing age, amount of vascular risk factors, amount of ath-
erosclerotic pathology as stratified by the ASCOD
phenotype and radiological ischemic lesion extension.
In the study this heterogeneity was analyzed compar-
ing subgroups of patients with predefined differences in
major clinical and radiological features as age, the amount
of common vascular risk factors, history of previous
stroke, clinical severity, and stroke lesions extension.
Whereas the comparisons between patients with minor
and major stroke (NIHSS
3 versus NIHSS>3) and with
different extensions of the stroke lesions did not show
apparent significant differences, the comparisons between
ESUS patients younger and older than 50 years of age,
without or with greater than or equal to 3 common vascu-
lar risk factors, without or with a previous TIA/stroke
have showed a possible dichotomy between 2 opposite
ESUS pathogenic phenotypes: the “no-atherosclerotic”
one and the “atherosclerotic” one.
The usefulness of stroke clinical severity and of the
radiological pattern of ischemic lesions in the diagnostic
process of cryptogenic strokes are argument of discussion.
Because atrial fibrillation related strokes are more severe
than strokes from other etiologies, it has been hypnotized
that the severity of ESUS may be a clue for an occult par-
oxysmal atrial fibrillation.19 Moreover, it has been hypno-
tized that some radiological imaging-patterns such as
multiple infarcts, simultaneous involvement of different
circulations, infarcts of different ages, and isolated cortical
infarcts may suggest cardioembolic pathogenesis.12,20-22
This are still open questions but some emerging evidences
from ESUS cohorts studies do not support these ideas.19,22
In this study minor and major stroke comparison as well
as the comparison between patients with different exten-
sions of stroke lesions at imaging, including focal and
multifocal lesions, did not show significant differences
regarding the amount of vascular risk factors, possible
 ARTICLE IN PRESS
CLINICAL PHENOTYPES OF EMBOLIC STROKE OF UNDETERMINED SOURCE
causes of stroke or pathological finding at echocardiogra-
phy. This is probably due to the small sample size and the
large prevalence of stroke causing minor or moderate def-
icit, 75% of cases with NIHSS
6. These subgroup analysis
confirmed the direct linear correlation between clinical
neurological deficit as measured by NIHSS or mRS and
the size of the infarcted area.8
On the other side, subgroup comparisons considering
aspects knowing to be involved in vascular atherosclerotic
disease like age, vascular risk factors, and history of previ-
ous clinical vascular events have brought out diagnosti-
cally relevant phenotypic dichotomy among ESUS
patients.
The first phenotype, “the no-atherosclerotic one”, is
smaller in amount accounting in this study for less than
20% of cases. It is characterized by younger age, less
amount of vascular risk factor and less evidence of non-
significant atherosclerosis and of cardiac degenerative
involution such as left atrial enlargement and left ventricle
diastolic dysfunction. PFO associated or not with atrial
septal aneurism, has a higher prevalence but is not always
present. This phenotype includes the so defined true cryp-
togenic stroke patients that, despite a comprehensive
assessment, does not present either common vascular risk
factors or evidence of common stroke etiologies. In these
cases an occult arterial dissection may be the cause but
also paradoxical embolism due to PFO is an important
pathogenesis to take in account as suggested by the ROPE
score.6,23 However other specific conditions, such as a his-
tory of migraine, illicit drugs consumption or oral con-
traceptive, must be investigated.24,25
The second phenotype, “the atherosclerotic one”, is
more frequent representing more than 80% of cases in
this study. It is characterized by older age, presence of
vascular risk factors, major evidence of nonsignificant
atherosclerosis, and/or cardiac degenerative involution.
In this phenotype, it is compelling to look carefully for
occult paroxysmal atrial fibrillation, substenotic athero-
sclerotic plaques with evidence of intraplaque hamor-
rhage, aortic arch atherosclerotic plaques and evidence
of atrial cardiopathy. It has been demonstrated that reli-
able predictors of occult paroxysmal atrial fibrillation are
age over 60 years, radiographic evidence of prior cortical
or cerebellar stroke, left atrial enlargement on echocar-
diogram and premature atrial complex on ECG or inpa-
tient cardiac telemetry.26-29 Substenotic plaques with
intraplaque hemorrhage, visualized by high resolution
magnetic resonance angiography, have been demon-
strated to be a probable cause of cryptogenic stroke by a
mechanism of plaque rupture and artery-to-artery embo-
lization.30,31 In addition it has been demonstrated that
large but nonstenotic carotid artery plaque is consider-
ably more common ipsilateral than contralateral to cryp-
togenic stroke, suggesting that nonstenotic plaque is an
7
underestimated cause of stroke.32 Aortic arch atheroscle-
rosis has been showed to be linked to embolism and
cryptogenic stroke when plaques are 4 mm in size,
mobile and/or complex. It is more frequent in patients
older than 55 years of age and it is associated with tradi-
tional vascular risk factors such as cigarette smoking,
diabetes, hypertension, and hypercholesterolemia.33,34
Atrial cardiopathy, defined as severe left atrial enlarge-
ment, serum N-terminal probrain natriuretic peptide
(NT-proBNP) level greater than 250 pg/mL or PTFV1 on
ECG greater than 5000 mV¢ms, has recently been associ-
ated with cryptogenic stroke.35 Patients with atrial cardi-
opathy have an older age, a higher prevalence of
dyslipidemia, hypertension and coronaropathy and less
prevalence of PFO.34-38
This pragmatic dichotomic distinction of ESUS patients
in atherosclerotic and nonatherosclerotic phenotypes is
not simply amenable to a distinction among young and
not-young stroke. Traditional vascular risk factors are in-
fact more common in young patients with stroke, espe-
cially those over 35 years of age, than was previously
thought, and that old patients without traditional risk fac-
tors may suffer from other pathogenetical condition such
as malignancy.24,34
However, dealing with a patient with ESUS, this prag-
matic approach cannot be exhaustive if according to
patient's age, history, and laboratory findings some spe-
cific but uncommon stroke etiologies are not ruled out as:
hypercoagulable states including inherited coagulopa-
thies, antiphospholipid syndrome, myeloproliferative dis-
orders, sickle cell disease, hyperhomocisteinemia; genetic
diseases as Fabry disease, mitochondrial diseases, CADA-
SIL; other autoimmune diseases including systemic lupus
erythematosus, Susac Syndrome, Giant cell arteritis, Poly-
arteritis nodosa, Kawasaki's arteritis, microscopic polyan-
giitis, Eosinophilic granulomatosis, Behcets's, Cogan's
syndrome, granulomatosis with polyangiits, Sarcoidosis,
etc; malignancies.24,25,34 In this study no patient presented
features suggesting these conditions.
Despite some limits as the sample size, the single center
registry and the observational transverse descriptive
research design, this study, based on an accurate single
patient evaluation, tries to investigate the presence of clin-
ically relevant different pathogenic phenotypes among
ESUS patients.
Age seems an important factor. Interestingly subgroup
analysis in the NAVIGATE ESUS trial shows a clear
advantage of aspirin in patients less than 60 years that is
lost in older patients. Vascular risk factors have been less
studied. Considering all this and the negative results of
NAVIGATE ESUS trial, it may be worth to give more
importance to clinical heterogeneity among ESUS patients
and take it in account in trial result evaluation and in
planning future trials (Fig 1).
 ARTICLE IN PRESS
8 S. PIFFER ET AL.

Figure 1. Legend. ESUS patients phenotypes—DIPOLE SCHEME. This schematic representation tries to simplify ESUS patients heterogeneity according to
their most probable potential etio-pathogenesis emerged from clinical and laboratory findings. It may help further diagnostic and therapeutic management in
such patients. Abbreviations: Athero, atherosclerotic phenotype; CNS, central nervous system; ESUS, embolic strokes undetermined source; LAE, left atrial
enlargement; LV diastolic dysf., left ventricular diastolic dysfunction; No-Athero., no-atherosclerotic phenotype; RFs, risk factors; pAF, paroxysmal atrial fibrilla-
tion.

Conclusions 4. Ladeira F, Barbosa R, Caetano A, et al. Embolic stroke of

ESUS represents a specific diagnosis for patients
affected by cryptogenic stroke that can be made after a
specific work-up and is currently under investigation in
clinical trials. ESUS patients are characterized prevalently
by mild neurological deficit and small cortical or CSS
lesions but besides their strict diagnostic criteria they
show a clinical and radiological heterogeneity. Consider-
ing age, amount of common vascular risk factors and his-
tory of previous stroke 2 opposite phenotypes of ESUS
patients seem to come out distinguished by a lower and a
higher load of atherosclerotic pathology. These 2 pheno-
types may suggest different underlying pathogenic mech-
anisms in ESUS patients. Eventually stratifying ESUS
patients in more homogeneous subgroups may be useful
to interpret the ongoing clinical trials results.
References
1. Hart RG, Diener HC, Coutts SB, et al. Embolic strokes of
undetermined source: the case for a new clinical con-
struct. Lancet Neurol 2014;13:429-438. https://doi.org/
10.1016/S1474-4422(13)70310-7.
2. Hart RG, Catanese L, Perera KS, et al. Embolic stroke of
undetermined source. Stroke 2017;116:016414. https://doi.
org/10.1161/STROKEAHA.116.016414. STROKEAHA.
3. Hart RG, Sharma M, Mundl H, et al. Rivaroxaban for
stroke prevention after embolic stroke of undetermined
source. 2018. doi:10.1056/NEJMoa1802686.
unknown source (ESUS) in young patients. Int J Stroke
2015;10:165. https://doi.org/10.1111/ijs.12596.
5. Ntaios G, Lip GYH, Vemmou A, et al. Age- and sex-spe-
cific analysis of patients with embolic stroke of undeter-
mined source. 2017.
6. Kent DM, Ruthazer R, Weimar C, et al. An index to iden-
tify stroke-related vs incidental patent foramen ovale in
cryptogenic stroke. Neurology 2013;81:619-625.
7. Lip GYH, Nieuwlaat R, Pisters R, et al. Refining clinical risk
stratification for predicting stroke and thromboembolism in
atrial fibrillation using a novel risk factor-based approach:
the Euro Heart Survey on atrial fibrillation. Chest
2010;137:263-272. https://doi.org/10.1378/chest.09-1584.
8. Kasner SE. Clinical interpretation and use of stroke
scales. Lancet Neurol 2006;5:603-612. https://doi.org/
10.1016/S1474-4422(06)70495-1.
9. Sulter G, Steen C, Keyser JDe. Use of the Barthel
index and modified rankin scale in acute stroke trials.
Stroke 1999;30:1538-1541. https://doi.org/10.1161/
01.STR.0000018645.04200.7D.
10. van Swieten JC, Koudstaal PJ, Visser MC, et al. Interob-
server agreement for the assessment of handicap in
stroke patients. Stroke 1988;19:604-607. https://doi.
org/10.1161/01.STR.20.6.828.
11. Fischer U, Baumgartner A, Arnold M, et al. What is a
minor stroke . Stroke 2010;41:661-666. https://doi.org/
10.1161/STROKEAHA.109.572883.
12. Amarenco P, Bogousslavsky J, Caplan LR, et al. The
ASCOD phenotyping of ischemic stroke (updated ASCO
phenotyping). Cerebrovasc Dis 2013;36:1-5. https://doi.
org/10.1159/000352050.
13. Lang RM, Badano LP, Mor-avi V, et al. Recommenda-
tions for cardiac chamber quantification by
 ARTICLE IN PRESS
CLINICAL PHENOTYPES OF EMBOLIC STROKE OF UNDETERMINED SOURCE
echocardiography in adults: an update from the Ameri-
can Society of Echocardiography and the European
Association of Cardiovascular Imaging. J Am Soc
Echocardiogr 2015;28:1-39. https://doi.org/10.1016/j.
echo.2014.10.003.
14. Rovira Ar, Grive E, Rovira A, et al. Distribution territories
and causative mechanisms of ischemic stroke. Eur Radiol
2005;15:416-426. https://doi.org/10.1007/s00330-004-
2633-5.
15. Li L, Yiin GS, Geraghty OC, et al. Incidence, outcome,
risk factors, and long-term prognosis of cryptogenic tran-
sient ischaemic attack and ischaemic stroke: a popula-
tion-based study. Lancet Neurol 2015;14:903-913.
https://doi.org/10.1016/S1474-4422(15)00132-5.
16. Li L, Yiin GS, Geraghty OC, et al. Supplementary webap-
pendix, incidence, outcome, risk factors, and long-term
prognosis of cryptogenic transient ischaemic attack and
ischaemic stroke: a population-based study. Lancet
Neurol 2015 http://dx.doi.org/10.1016/S1474-4422(15)
00132-5.
17. Ntaios G, Papavasileiou V, Milionis H, et al. Embolic
strokes of undetermined source in the athens stroke regis-
try: a descriptive analysis. Stroke 2015;46:176-181.
https://doi.org/10.1161/STROKEAHA.115.009334.
18. Perera KS, Vanassche T, Bosch J, et al. Embolic strokes of
undetermined source: prevalence and patient features in
the ESUS Global Registry. Int J Stroke 2016;11:526-533.
https://doi.org/10.1177/1747493016641967.
19. Ntaios G, Papavasileiou V, Lip GYH, et al. Embolic stroke
of undetermined source and detection of atrial fibrillation
on follow-up: how much causality is there. J Stroke Cere-
brovasc Dis 2016;25:2975-2980. https://doi.org/10.1016/
j.jstrokecerebrovasdis.2016.08.015.
20. Kang D-W, Chalela JA, Ezzeddine MA, Warach S. Associ-
ation of ischemic lesion patterns on early diffusion-
weighted imaging with TOAST stroke subtypes. Arch
Neurol 2003;60:1730-1734.
21. Barber PA. Is diffusion-weighted imaging helpful in
determining the source of stroke . Nat Clin Pract Neurol
2006;2:424-425. https://doi.org/10.1038/ncpneuro0258.
22. Maier IL, Schregel K, Karch A, et al. Association between
embolic stroke patterns, ESUS etiology, and new diagno-
sis of atrial fibrillation: a secondary data analysis of the
find-AF trial. Stroke Res Treat 2017;2017. https://doi.
org/10.1155/2017/1391843.
23. Bang OY, Lee J, Ryoo S, et al. Patent Foramen ovale and
stroke
current status. J Stroke 2015;17:229-237.
24. Maaijwee NAMM, Rutten-jacobs LCA, Schaapsmeerders
P, et al. Ischaemic stroke in young adults: risk factors and
long-term consequences. Nat Rev Neurol 2014;10:315-
325. https://doi.org/10.1038/nrneurol.2014.72.
25. Ferro JM, Massaro AR, Mas J. Aetiological diagnosis of
ischaemic stroke in young adults. Lancet Neurol 2010;9:1085-
1096. https://doi.org/10.1016/S1474-4422(10)70251-9.
9
26. Favilla CG, Ingala E, Jara J, et al. Predictors of finding
occult atrial fibrillation after cryptogenic stroke.
Stroke 2015;46:1210-1215. https://doi.org/10.1161/
STROKEAHA.114.007763.
27. Gladstone DJ, Dorian P, Spring M, et al. Atrial premature
beats predict atrial fibrillation in cryptogenic stroke:
results from the EMBRACE trial. Stroke 2015;46:936-941.
https://doi.org/10.1161/STROKEAHA.115.008714.
28. Miller DJ, Khan MA, Schultz LR, et al. Outpatient cardiac
telemetry detects a high rate of atrial fibrillation in cryp-
togenic stroke. J Neurol Sci 2013;324:57-61. https://doi.
org/10.1016/j.jns.2012.10.001.
29. Perera KS, Vanassche T, Bosch J, et al. Global survey of
the frequency of atrial fibrillation—associated stroke
embolic stroke of undetermined source global registry.
Stroke 2016;47:2197-2202. https://doi.org/10.1161/
STROKEAHA.116.013378.
30. Gupta A, Gialdini G, Lerario MP, et al. Magnetic reso-
nance angiography detection of abnormal carotid
artery plaque in patients with cryptogenic stroke. J Am
Heart Assoc 2015;4:e002012. https://doi.org/10.1161/
JAHA.115.002012.
31. Freilinger TM, Schindler A, Schmidt C, et al. Preva-
lence of nonstenosing, complicated atherosclerotic
plaques in cryptogenic stroke. JACC Cardiovasc
Imaging 2012;5:397-405. https://doi.org/10.1016/j.
jcmg.2012.01.012.
32. Coutinho JM, Derkatch S, Potvin ARJ, et al. Nonstenotic
carotid plaque on CT angiography in patients with cryp-
togenic stroke. Neurology 2016;87:665-672. https://doi.
org/10.1212/WNL.0000000000002978.
33. Amarenco P, Cohen A, Tzourio C, et al. Atherosclerotic
disease of the aortic arch and the risk of ischemic stroke.
N Engl J Med 1994;331:1474-1479.
34. Nouh A, Hussain M, Mehta T, et al. Embolic strokes of
unknown source and cryptogenic stroke: implications in
clinical practice. Front Neurol 2016;7:1-16. https://doi.
org/10.3389/fneur.2016.00037.
35. Yaghi S, Boehme AK, Hazan R, et al. Atrial cardiopathy
and cryptogenic stroke : a cross-sectional pilot study. J
Stroke Cerebrovasc Dis 2016;25:110-114. https://doi.
org/10.1016/j.jstrokecerebrovasdis.2015.09.001.
36. Yaghi S, Moon YP, Mora-Mclaughlin C, et al. Left atrial
enlargement and stroke recurrence: the Northern Man-
hattan Stroke Study. Stroke 2015;46:1488-1493. https://
doi.org/10.1161/STROKEAHA.115.008711.
37. Kamel H, O'Neal WT, Okin PM, et al. Electrocar-
diographic left atrial abnormality and stroke subtype in
the atherosclerosis risk in communities study. Ann Neu-
rol 2015;78:670-678. https://doi.org/10.1002/ana.24482.
38. Kamel H, Okin PM, Longstreth WT, et al. Atrial cardiopa-
thy: a broadened concept of left atrial thromboembolism
beyond atrial fibrillation. Future Cardiol 2015;11:323-331.
https://doi.org/10.2217/fca.15.22.