https://api.elsevier.com/content/article/pii/S0167527317311075doi:10.1016/j.ijcard.2017.02.0781-s2.0-S016752731731107510.1016/j.ijcard.2017.02.078S0167-5273(17)31107-5Serially measured circulating miR-22-3p is a biomarker for adverse clinical outcome in patients with chronic heart failure: The Bio-SHiFT study International Journal of CardiologyJournalfla01675273235124132124-132text/xml2017-05-1515 May 2017© 2017 The Authors. Published by Elsevier Ireland Ltd.The Authors. Published by Elsevier Ireland Ltd.van Boven, NickAkkerhuis, K. MartijnAnroedh, Sharda S.Rizopoulos, DimitrisPinto, YigalBattes, Linda C.Hillege, Hans L.Caliskan, Kadir C.Germans, TjeerdManintveld, Olivier C.Cornel, Jan-HeinConstantinescu, Alina A.Boersma, EricUmans, Victor A.Kardys, Isabella
Background
Several studies have suggested circulating microRNAs (miRs) are associated with heart failure, but these studies were small, and limited to single miR measurements.
We examined 7 miRs which were previously linked to heart failure, and tested whether their temporal expression level predicts prognosis in a prospective cohort of chronic heart failure (CHF) patients.
Methods and results
In 2011–2013, 263 CHF patients were included. At inclusion and subsequently every 3months, we measured 7miRs. The primary endpoint (PE) comprised heart failure hospitalization, cardiovascular mortality, cardiac transplantation and LVAD implantation. Associations between temporal miR patterns and the PE were investigated by joint modelling, which combines mixed models with Cox regression.
Mean age was 67±13years, 72% were men and 27% NYHA classes III–IV. We obtained 873 blood samples (median 3 [IQR 2–5] per patient). The PE was reached in 41 patients (16%) during a median follow-up of 0.9 [0.6–1.4] years. The temporal pattern of miR-22-3p was independently associated with the PE (HR [95% CI] per doubling of level: 0.64 [0.47–0.77]). The instantaneous change in level (slope of the temporal miR pattern) of miR-22-3p was also independently associated with the PE (HR [95% CI] per doubling of slope: 0.33 [0.20–0.51]). These associations remained statistically significant after adjustment for temporal patterns of NT-proBNP, Troponin T and CRP.
Conclusions
The temporal pattern of circulating miR-22-3p contains important prognostic and independent information in CHF patients. This concept warrants further investigation in larger series with extended follow-up.
1trueFullfalseVSNUFundingBodyhttp://creativecommons.org/licenses/by/4.0/MicroRNAsBiomarkersChronic heart failurePrognosticationMortality850142154582-s2.0-8501421545828274577
serial
JL
271057
291210
291896
291900
31
90
International Journal of Cardiology
INTERNATIONALJOURNALCARDIOLOGY
2017-02-22
2017-02-22
2017-03-22
2017-03-22
2017-03-22T19:50:36
1-s2.0-S0167527317311075
S0167-5273(17)31107-5
S0167527317311075
10.1016/j.ijcard.2017.02.078
S300
S300.1
FULL-TEXT
1-s2.0-S0167527317X00065
2017-03-22T21:11:06.209699-04:00
0
0
20170515
2017
2017-02-22T17:53:21.709486Z
articleinfo articletitlenorm authfirstinitialnorm authfirstsurnamenorm cid cids contenttype copyright crossmark dateloaded dateloadedtxt datesearch datesort dateupdated dco docsubtype doctype doi eid ewtransactionid fundingbodyid hubeid indexeddate issn issnnorm itemstage itemtransactionid itemweight oauserlicense openaccess openarchive pg pgfirst pglast pii piinorm pubdatestart pubdatetxt pubyr sectiontitle sortorder sponsoredaccesstype srctitle srctitlenorm srctype ssids alllist content oa subj subheadings suppl tomb volfirst volissue volumelist webpdf webpdfpagecount yearnav figure table e-component body acknowledge affil articletitle auth authfirstini authfull authkeywords authlast grantsponsor orcid primabst ref
0167-5273
01675273
UNLIMITED
VSNUGOLD
true
235
235
C
Volume 235
21
124
132
124
132
20170515
15 May 2017
2017-05-15
2017
Orginal Articles
article
fla
© 2017 The Authors. Published by Elsevier Ireland Ltd.
SERIALLYMEASUREDCIRCULATINGMIR223PABIOMARKERFORADVERSECLINICALOUTCOMEINPATIENTSCHRONICHEARTFAILUREBIOSHIFTSTUDY
VANBOVEN
N
1
Introduction
2
Methods
2.1
Patients
2.2
Baseline assessment
2.3
Follow-up visits
2.4
Blood sampling and miR measurement
2.5
Clinical study endpoints
2.6
Statistical analysis
3
Result
3.1
Baseline characteristics
3.2
Clinical endpoints
3.3
Temporal miR patterns and the primary endpoint
3.4
Slope and area under the curve of the temporal miR patterns and the primary endpoint
3.5
Temporal patterns of miRs, NT-proBNP, Hs-TNT and CRP combined in a multiple-marker model
3.6
Repeated miR measurements and other biomarkers
4
Discussion
Funding sources
Conflicts of interest
Acknowledgements
References
PONIKOWSKI
2016
891
975
P
YANCY
2013
e147
e239
C
JANUZZI
2013
500
507
J
DECOUTO
2010
334
344
G
SAVARESE
2013
e58287
G
PORAPAKKHAM
2010
507
514
P
MILLER
2007
249
257
W
YAN
2005
698
704
R
TROUGHTON
2014
16
24
R
BARTEL
2004
281
297
D
GILAD
2008
e3148
S
CHEN
2008
2111
2116
J
COSTAMARTINS
2008
1567
1576
P
CREEMERS
2012
483
495
E
VOGEL
2013
410
418
B
ZAMPETAKI
2012
290
299
A
CORSTEN
2010
499
506
M
OLIVIERI
2013
531
536
F
GOLDRAICH
2014
135
141
L
BAUTERS
2013
1837
1840
C
TIJSEN
2010
1035
1039
A
DICKINSON
2013
650
659
B
NAGAPRASAD
2009
27487
27499
S
GOREN
2012
147
154
Y
MCMURRAY
2012
1787
1847
J
DICKSTEIN
2008
2388
2442
K
PAULUS
2007
2539
2550
W
RAO
2009
585
594
P
RIZOPOULOS
2010
1
33
D
RIZOPOULOS
2012
D
JOINTMODELSFORLONGITUDINALTIMETOEVENTDATAAPPLICATIONSINR
LI
2010
543
548
J
HUANG
2014
267
272
Z
HUANG
2013
1234
1243
Z
GURHA
2012
2751
2761
P
GUPTA
2016
1557
1571
S
KNEZEVIC
2012
12913
12926
I
NAGALINGAM
2013
11216
11232
R
NAGALINGAM
2014
27199
27214
R
GANESAN
2013
2097
2106
J
FANG
2012
410
423
J
LI
2013
5267
5275
X
SMALL
2010
4218
4223
E
BESLER
2016
C
VEGTER
2016
457
468
E
VANBOVENX2017X124
VANBOVENX2017X124X132
VANBOVENX2017X124XN
VANBOVENX2017X124X132XN
Full
2017-02-23T16:49:18Z
FundingBody
VSNU
http://creativecommons.org/licenses/by/4.0/
2018-03-22T00:00:00.000Z
UnderEmbargo
http://creativecommons.org/licenses/by-nc-nd/4.0/
This is an open access article under the CC BY license.
© 2017 The Authors. Published by Elsevier Ireland Ltd.
item
S0167-5273(17)31107-5
S0167527317311075
1-s2.0-S0167527317311075
10.1016/j.ijcard.2017.02.078
271057
2017-03-22T15:56:08.934608-04:00
2017-05-15
UNLIMITED
VSNUGOLD
1-s2.0-S0167527317311075-main.pdf
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/MAIN/application/pdf/853882e716e989980df4b0c990696662/main.pdf
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/MAIN/application/pdf/853882e716e989980df4b0c990696662/main.pdf
main.pdf
pdf
true
661455
MAIN
9
1-s2.0-S0167527317311075-main_1.png
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/PREVIEW/image/png/f2d8e62a7ab53d8269d54420bb6c2bc9/main_1.png
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/PREVIEW/image/png/f2d8e62a7ab53d8269d54420bb6c2bc9/main_1.png
main_1.png
png
59085
849
656
IMAGE-WEB-PDF
1
1-s2.0-S0167527317311075-gr1.sml
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/gr1/THUMBNAIL/image/gif/d5d6c1e3ced2b63c5e53a33bc735f43e/gr1.sml
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/gr1/THUMBNAIL/image/gif/d5d6c1e3ced2b63c5e53a33bc735f43e/gr1.sml
gr1
gr1.sml
sml
15297
164
181
IMAGE-THUMBNAIL
1-s2.0-S0167527317311075-gr2.sml
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/gr2/THUMBNAIL/image/gif/5da9128a2e0a822f56160ae8632b173e/gr2.sml
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/gr2/THUMBNAIL/image/gif/5da9128a2e0a822f56160ae8632b173e/gr2.sml
gr2
gr2.sml
sml
16484
135
219
IMAGE-THUMBNAIL
1-s2.0-S0167527317311075-gr3.sml
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/gr3/THUMBNAIL/image/gif/04ccf74f922f2d47740b866fde933ab2/gr3.sml
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/gr3/THUMBNAIL/image/gif/04ccf74f922f2d47740b866fde933ab2/gr3.sml
gr3
gr3.sml
sml
6403
76
219
IMAGE-THUMBNAIL
1-s2.0-S0167527317311075-gr1.jpg
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/gr1/DOWNSAMPLED/image/jpeg/bd1181f951e83c0858f7dd875683038e/gr1.jpg
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/gr1/DOWNSAMPLED/image/jpeg/bd1181f951e83c0858f7dd875683038e/gr1.jpg
gr1
gr1.jpg
jpg
88866
566
625
IMAGE-DOWNSAMPLED
1-s2.0-S0167527317311075-gr2.jpg
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/gr2/DOWNSAMPLED/image/jpeg/1f9419c12bba6c3dbe9cd17c8c70a91e/gr2.jpg
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/gr2/DOWNSAMPLED/image/jpeg/1f9419c12bba6c3dbe9cd17c8c70a91e/gr2.jpg
gr2
gr2.jpg
jpg
30718
207
335
IMAGE-DOWNSAMPLED
1-s2.0-S0167527317311075-gr3.jpg
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/gr3/DOWNSAMPLED/image/jpeg/e6dbe97db71b3177c8d078f1f7eeab1d/gr3.jpg
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/gr3/DOWNSAMPLED/image/jpeg/e6dbe97db71b3177c8d078f1f7eeab1d/gr3.jpg
gr3
gr3.jpg
jpg
52619
373
1069
IMAGE-DOWNSAMPLED
1-s2.0-S0167527317311075-gr1_lrg.jpg
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/gr1/HIGHRES/image/jpeg/7525836472ea11c5ea0a1f3936a17dd5/gr1_lrg.jpg
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/gr1/HIGHRES/image/jpeg/7525836472ea11c5ea0a1f3936a17dd5/gr1_lrg.jpg
gr1
gr1_lrg.jpg
jpg
309448
1504
1660
IMAGE-HIGH-RES
1-s2.0-S0167527317311075-gr2_lrg.jpg
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/gr2/HIGHRES/image/jpeg/46fb594de7c89f897a7a1e8d1ac60dc4/gr2_lrg.jpg
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/gr2/HIGHRES/image/jpeg/46fb594de7c89f897a7a1e8d1ac60dc4/gr2_lrg.jpg
gr2
gr2_lrg.jpg
jpg
174302
919
1486
IMAGE-HIGH-RES
1-s2.0-S0167527317311075-gr3_lrg.jpg
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/gr3/HIGHRES/image/jpeg/d4747ab479e85f198d012df86488ecf8/gr3_lrg.jpg
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/gr3/HIGHRES/image/jpeg/d4747ab479e85f198d012df86488ecf8/gr3_lrg.jpg
gr3
gr3_lrg.jpg
jpg
497681
1651
4734
IMAGE-HIGH-RES
1-s2.0-S0167527317311075-mmc1.doc
https://s3-eu-west-1.amazonaws.com/prod-ucs-content-store-eu-west/content/pii:S0167527317311075/mmc1/MAIN/application/msword/fa1f37a4a01e8afe7c0ffa862f916858/mmc1.doc
https://s3.amazonaws.com/prod-ucs-content-store-us-east/content/pii:S0167527317311075/mmc1/MAIN/application/msword/fa1f37a4a01e8afe7c0ffa862f916858/mmc1.doc
mmc1
mmc1.doc
doc
72192
APPLICATION
IJCA
24604
S0167-5273(17)31107-5
10.1016/j.ijcard.2017.02.078
The Authors
Fig. 1
Inclusion and exclusion criteria.
Fig. 1
Fig. 2
Associations between baseline MicroRNA levels and presence of ischemic cardiomyopathy.
Fig. 2
Fig. 3
Temporal evolution of microRNA levels. Legend: Temporal evolution was estimated by using mixed models. Y-axis: arbitrary units [a.u.] representing relative miRNA expression levels. X-axis: time in years as the endpoint approaches (time=0 denotes the moment that the endpoint occurs for the patients reaching the endpoint, and the moment of right-censoring for the patients not reaching the endpoint). Red solid line: miR expression level in patients reaching the endpoint. Red dotted lines: 95% confidence interval. Black solid line: miR expression level in patients not reaching the endpoint. Black dotted line: 95% confidence interval.
Fig. 3
Table 1
Baseline characteristics.
Table 1
Total (n=263)
Demographics
Age
67 (±13)
Male gender
189 (72)
Caucasian ethnicity
244 (94)
Clinical characteristics
Body mass index kg/m2
28 (±5)
Heart rate, bpm
67 (±12)
Systolic blood pressure, mmHg
122 (±20)
Diastolic blood pressure, mmHg
73 (±11)
Biomarker level
NT-proBNP (pmol/L)
139.6 (51.9 – 272.9)
HsTNT (ng/l)
18.0 (9.6 – 33.2)
CRP (mg/l)
2.2 (0.9 – 4.8)
Features of heart failure
Duration of heart failure, years
4.6 (1.7 – 9.9)
NYHA class I
75 (29)
NYHA class II
115 (44)
NYHA class III or IV
69 (27)
Left ventricular function
Systolic dysfunction
250 (95)
HFPEF
13 (5)
LVEF
a
32 (±10)
Etiology of heart failure
Ischemic heart disease
117 (44)
Hypertension
34 (13)
Secondary to valvular heart disease
12 (5)
Cardiomyopathy
68 (26)
Dilated
49 (19)
Hypertrophic
12 (5)
Non compaction
4 (1)
Unclassified
3 (1)
Unknown
19 (7)
Other
13 (5)
Medical history
Myocardial infarction
94 (36)
PCI
82 (31)
CABG
43 (16)
Valvular heart disease
136 (53)
Atrial fibrillation
105 (40)
Other arrhythmia
82 (32)
ICD
151 (59)
CRT
78 (30)
Pacemaker
38 (15)
CVA
41 (16)
Chronic renal failure
136 (53)
Diabetes Mellitus
81 (31)
Known hypercholesterolemia
93 (35)
Hypertension
120 (46)
Sleep apnea
26 (10)
Intoxications
Alcohol consumption (>1unit/day)
108 (42)
Smoking
185 (71)
Ever
186 (72)
Current
26 (10)
Medication use
ACE-i
173 (67)
ARB
75 (29)
Aldosteron antagonist
178 (68)
Diuretic
237 (90)
Beta-blocker
232(88)
Aspirin
45 (17)
Vitamin K antagonist
200 (77)
Nitrates
44 (17)
Digoxin
59 (23)
Antiarrhythmics
39 (15)
Normally distributed continuous variables are presented as mean (±standard deviation). Non-normally distributed continuous variables are expressed as median (25th–75th percentile). Categorical variables are expressed as count (percentage). Valid percentages may vary for some counts, because of missing values. ACE-I=ace inhibitor; ARB=angiotensin II receptor blocker; CABG=coronary artery bypass grafting; COPD=chronic obstructive pulmonary disease; CRP=C-reactive protein; CRT=cardiac resynchronization therapy; CVA=cerebrovascular accident; HFPEF=heart failure with preserved ejection fraction; HsTNT=high sensitive cardiac troponin T; ICD=implantable cardioverter/defibrillator; LVEF=left ventricular ejection fraction; NT-proBNP=N-terminal pro-B-type natriuretic peptide NYHA=New York heart association; PCI=percutaneous coronary intervention.
a
Baseline echocardiograms were available in 72% of all patients because of logistic reasons.
Table 2
Associations between baseline miRs and the primary endpoint and secondary endpoint.
Table 2
Primary endpoint
Secondary endpoint
HR
⁎
(95% CI)
P
HR (95% CI)
P
miR-1254
1.01 (0.92–1.09)
0.91
1.00 (0.91–1.09)
0.99
miR-22-3p
1.02 (0.93–1.12)
0.61
1.00 (0.90–1.09)
0.85
miR-345-5p
0.99 (0.89–1.09)
0.81
0.95 (0.86–1.06)
0.38
miR-378a-3p
1.00 (0.91–1.11)
0.93
0.99 (0.89–1.09)
0.77
miR-423-5p
1.01 (0.91–1.11)
0.92
0.97 (0.87–1.07)
0.53
miR-486-5p
1.04 (0.94–1.15)
0.47
1.00 (0.89–1.11)
0.93
miR-320a
1.04 (0.95–1.13)
0.44
1.02 (0.93–1.12)
0.71
miR=microRNA.
Primary endpoint: cardiac death, cardiac transplantation, left ventricular assist device implantation, and hospitalization for the management of acute or worsened heart failure; secondary endpoint: hospitalization for the management of acute or worsened heart failure.
⁎
Hazard ratios (HRs) and corresponding confidence intervals (CIs) were calculated using Cox regression analyses and are given per doubling of level of miR.
Table 3
Associations between temporal miR patterns and the primary endpoint.
Table 3
miR-1254
miR-22-3p
miR-345-5p
miR-378a-3p
miR-423-5p
miR-486-5p
miR-320a
HR
a
(95% CI)
p
HR
a
(95% CI)
p
HR
a
(95% CI)
p
HR
a
(95% CI)
p
HR
a
(95% CI)
p
HR
a
(95% CI)
p
HR
a
(95% CI)
p
Temporal pattern of miR value
Adjusted for age
and gender
1.02
(0.71–1.26)
0.782
0.64
(0.47–0.77)
<0.001
1.01
(0.97–1.04)
0.624
0.84
(0.69–1.03)
0.085
1.01
(0.97–1.05)
0.800
1.00
(1.00–1.01)
0.069
0.92
(0.74–1.19)
0.516
Multivariable
adjusted
b
0.97
(0.78–1.22)
0.757
0.71
(0.44–0.94)
0.005
1.00
(0.97–1.03)
0.921
0.88
(0.71–1.06)
0.148
1.00
(0.95–1.08)
0.845
1.00
0.99–1.01)
0.380
1.01
(0.78–1.22)
0.896
Slope of temporal pattern
Adjusted for age
and gender
1.29
(0.75–2.49)
0.456
0.20
(0.09–0.50)
<0.001
NA
NA
NA
NA
NA
NA
NA
NA
0.51
(0.29–0.90)
0.012
Multivariable
adjusted
b
1.30
(0.79-2.35)
0.288
0.33
(0.20–0.51)
<0.001
NA
NA
NA
NA
NA
NA
NA
NA
0.59
(0.33–1.12)
0.102
Area under the curve of temporal pattern
Adjusted for age
and gender
1.09
(0.97–1.27)
0.206
1.00
(0.80–1.14)
0.898
NA
NA
NA
NA
NA
NA
NA
NA
1.16
(0.94–1.42)
0.257
Multivariable
adjusted
b
1.26
(0.95–1.53)
0.100
0.99
(0.88–0.17)
0.846
NA
NA
NA
NA
NA
NA
NA
NA
1.19
(0.88–1.44)
0.330
miR=microRNA.
a
Hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) are given per doubling of value, slope or area under the curve at any point in time.
b
Adjusted for age, gender, New York Heart Association class and ischemic cardiomyopathy.
Table 4
Associations between repeatedly measured miRs and B-type natriuretic peptide, high sensitive cardiac troponin T and C- reactive protein.
Table 4
NT-proBNP
HsTNT
CRP
Beta coefficient (95% CI)
P
Beta coefficient (95% CI)
P
Beta coefficient (95% CI)
P
miR-1254
0.07 (−0.05–0.02)
0.27
0.00 (−0.01–0.01)
0.59
0.00 (−0.02–0.02)
0.92
miR-22-3p
−0.01 (−0.022 to −0.00)
0.14
−0.01 (−0.02 to −0.00)
0.014
−0.03 (−0.06–0.00)
0.051
miR-345-5p
−0.04 (−0.02–0.01)
0.57
−0.01 (−0.02–0.01)
0.35
−0.04 (−0.07 to −0.01)
0.012
miR-378a-3p
−0.01 (−0.02–0.00)
0.12
0.00 (−0.01–0.01)
0.91
−0.02 (−0.04–0.01)
0.27
miR-423-5p
0.00 (−0.02–0.01)
0.81
−0.01 (−0.02–0.00)
0.17
−0.01 (−0.04–0.03)
0.63
miR-486-5p
−0.01 (−0.02–0.01)
0.43
−0.01 (−0.02–0.01)
0.37
−0.03 (−0.06–0.00)
0.070
miR-320a
0.00 (−0.01–0.02)
0.69
0.00 (−0.01–0.01)
0.74
0.01 (−0.02–0.03)
0.71
Beta coefficients and corresponding 95% confidence intervals (CIs) were calculated using generalised estimating equations with 10log transformed microRNAs (miRs) as independent variables and 10log transformed N-terminal pro-B-type natriuretic peptide (NT-proBNP), high sensitive cardiac troponin T (HsTNT) and C-reactive protein (CRP) as dependent variables.
Serially measured circulating miR-22-3p is a biomarker for adverse clinical outcome in patients with chronic heart failure: The Bio-SHiFT study
Nick
van Boven
a
K. Martijn
Akkerhuis
b
Sharda S.
Anroedh
b
Dimitris
Rizopoulos
c
Yigal
Pinto
d
Linda C.
Battes
b
Hans L.
Hillege
e
Kadir C.
Caliskan
b
Tjeerd
Germans
a
Olivier C.
Manintveld
b
Jan-Hein
Cornel
a
Alina A.
Constantinescu
b
Eric
Boersma
b
Victor A.
Umans
a
Isabella
Kardys
b
⁎
a
Cardiology, Medical Centre Alkmaar, Alkmaar, The Netherlands
Cardiology
Medical Centre Alkmaar
Alkmaar
The Netherlands
b
Cardiology, Erasmus Medical Centre, Rotterdam, The Netherlands
Cardiology
Erasmus Medical Centre
Rotterdam
The Netherlands
c
Biostatistics, Erasmus Medical Centre, Rotterdam, The Netherlands
Biostatistics
Erasmus Medical Centre
Rotterdam
The Netherlands
d
Cardiology, Academic Medical Centre, Amsterdam, The Netherlands
Cardiology
Academic Medical Centre
Amsterdam
The Netherlands
e
Cardiology, University Medical Centre Groningen, Groningen, The Netherlands
Cardiology
University Medical Centre Groningen
Groningen
The Netherlands
⁎
Corresponding author at: Department of Cardiology, Erasmus Medical Centre, room Ba-561, 's Gravendijkwal 230, 3015 CE Rotterdam, The Netherlands.
Department of Cardiology
Erasmus Medical Centre
room Ba-561, 's Gravendijkwal 230
CE Rotterdam
3015
The Netherlands
Background
Several studies have suggested circulating microRNAs (miRs) are associated with heart failure, but these studies were small, and limited to single miR measurements.
We examined 7 miRs which were previously linked to heart failure, and tested whether their temporal expression level predicts prognosis in a prospective cohort of chronic heart failure (CHF) patients.
Methods and results
In 2011–2013, 263 CHF patients were included. At inclusion and subsequently every 3months, we measured 7miRs. The primary endpoint (PE) comprised heart failure hospitalization, cardiovascular mortality, cardiac transplantation and LVAD implantation. Associations between temporal miR patterns and the PE were investigated by joint modelling, which combines mixed models with Cox regression.
Mean age was 67±13years, 72% were men and 27% NYHA classes III–IV. We obtained 873 blood samples (median 3 [IQR 2–5] per patient). The PE was reached in 41 patients (16%) during a median follow-up of 0.9 [0.6–1.4] years. The temporal pattern of miR-22-3p was independently associated with the PE (HR [95% CI] per doubling of level: 0.64 [0.47–0.77]). The instantaneous change in level (slope of the temporal miR pattern) of miR-22-3p was also independently associated with the PE (HR [95% CI] per doubling of slope: 0.33 [0.20–0.51]). These associations remained statistically significant after adjustment for temporal patterns of NT-proBNP, Troponin T and CRP.
Conclusions
The temporal pattern of circulating miR-22-3p contains important prognostic and independent information in CHF patients. This concept warrants further investigation in larger series with extended follow-up.
Keywords
MicroRNAs
Biomarkers
Chronic heart failure
Prognostication
Mortality
1
Introduction
Contemporary treatment of chronic heart failure ((C)HF) generally aims to stabilize or at least decelerate disease progression. Adjustment of pharmacotherapy is largely based on clinical judgement and thus mostly relies on the worsening of symptoms or signs [1,2]. However, in the context of the complexity of HF therapy, considerable clinical skills, as well as cooperation of the patients, are required to recognize opportunities to titrate therapies and to implement such changes intervene early and timely [3]. Consequently, higher-risk patients may be undertreated [3]. Blood biomarkers have the potential to monitor subtle changes in the heart that reflect and possibly predict adverse changes before they become clinically apparent or reported [4]. The value of biomarkers, such as B-type natriuretic peptides (BNP), cardiac troponins and C-reactive protein (CRP), for risk stratification of CHF patients has already been demonstrated [5–8]. Moreover, natriuretic peptide-guided HF therapy has recently been given a class IIa recommendation in US HF guidelines to achieve guideline-directed medical therapy [2,9].
Nevertheless, the predictive capability of the aforementioned biomarkers for worsening of CHF still leaves room for improvement. MicroRNAs (miRs) are upcoming novel biomarkers that seem promising for early diagnosis and treatment of HF. MiRs are non-coding, ~22 nucleotide long RNA sequences, which target messenger RNAs for cleavage or translational repression and thereby influence a great variety of biological processes [10]. The stability of miRs in plasma, and consequently their reliable assessment in easily accessible samples, potentially makes them attractive biomarkers for a wide range of diseases [11]. Studies revealing that deletion of Dicer, a gene encoding an RNase III endonuclease essential for miR processing, leads to cardiac remodelling and dilation, were the first to show involvement of miRs in HF [12,13] and to suggest that miRs might be used as biomarkers for cardiovascular diseases [14]. Several other studies have subsequently shown associations between miRs and myocardial infarction [15–18] and HF [19–24]. However, most studies pertaining to HF were performed in case-control settings and had a limited sample size. Furthermore, these studies usually assessed miRs only once. Repeated, longitudinal miR assessment in CHF patients may, however, provide insight into individual, temporal patterns and the patient's ensuing risk of disease progression. The temporal patterns of miRs in patients with known CHF have not yet been investigated.
In the current prospective, observational study, we have performed frequent (up to 8), repeated measurements of multiple miRs that were previously linked to HF (miR-1254, miR-22-3p, miR-423-5p, miR-486-5p and miR-320a) or have been shown to be cardiac-enriched (miR-345-5p, miR-378a-3p) in a cohort of 263 outpatients with CHF, and have subsequently investigated the associations of the obtained temporal patterns with adverse clinical outcome during follow-up.
2
Methods
2.1
Patients
The Serial biomarker measurements and new echocardiographic techniques in chronic heart failure patients result in tailored prediction of prognosis (Bio-SHiFT) study was designed to investigate the relationship between temporal patterns of biomarkers involved in CHF and prognosis. Bio-SHiFT is an ongoing prospective, observational study of stable outpatients with CHF, conducted in Erasmus MC, Rotterdam, The Netherlands and Medical Centre Alkmaar, The Netherlands. Patients were recruited during their regular outpatient visits and were in clinically stable condition. Patients were eligible for inclusion if aged 18years or older, capable of understanding and signing informed consent, and if CHF (including HF with preserved ejection fraction (HFPEF)) was diagnosed ≥3months ago according to the guidelines of the European Society of Cardiology (ESC) [25–27]. Detailed inclusion and exclusion criteria are shown in Fig. 1
. The study was approved by the medical ethics committees of the participating hospitals and was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all patients.The study is registered in ClinicalTrials.gov, number NCT01851538. In the current paper, we have performed an analysis on the 263 patients who were enrolled during the first inclusion period between October 2011 and June 2013. Follow-up for this analysis lasted from October 2011 until November 2013.
2.2
Baseline assessment
At baseline patients were evaluated by trained research physicians, who collected information on HF related symptoms including NYHA class [25,26] and performed a physical examination, including blood pressure, heart rate and body mass index. Information on aetiology of heart failure, presence of systolic dysfunction, cardiovascular risk factors, medical history and medical treatment was retrieved primarily from hospital records and was completed by anamnesis in case of ambiguities. History of myocardial infarction, percutaneous coronary intervention (PCI), coronary artery bypass grafting, valvular heart disease, atrial fibrillation or other arrhythmias, device implantation, cerebrovascular accident, diabetes mellitus, hypercholesterolemia, hypertension, and sleep apnea were defined as a clinical diagnosis of these conditions, as reported by the treating physician in the medical chart. History of chronic renal failure was defined as glomerular filtration rate less than 60mL/min/1.73m2. Alcohol consumption was defined as drinking ≥1 alcoholic consumption per day. Electrocardiography and echocardiography were performed. Data were entered into electronic case report forms. Non-fasting blood and urine samples were collected.
2.3
Follow-up visits
All patients continued routine outpatient visits during the study. Study follow-up visits were scheduled every 3months (a window of ±1month was allowed), for a maximum of 30months. For patients' convenience, study visits and routine outpatient visits were combined when possible. At each tri-monthly study visit, a short medical evaluation was performed and blood and urine samples were collected. Adverse cardiovascular events and changes in medication were recorded in electronic case report forms.
2.4
Blood sampling and miR measurement
Blood samples were collected at baseline and at each follow-up visit, and were processed and stored locally at a temperature of −80°C within 2h after blood collection. When applicable, samples were transported to the central laboratory (Erasmus MC, Rotterdam, The Netherlands) under controlled conditions (at a temperature of −80°C), until batch analysis took place. Accordingly, results of the biomarker assays were not available to treating physicians at the time of the outpatient visits. Thus, the biomarker measurements did not lead to treatment adjustments, and all patients received usual care based on European guidelines [25,26].
For the purpose of the current investigation, stored plasma samples were transported under controlled conditions to ACS Biomarker, Amsterdam, The Netherlands, and seven miRs were measured in one batch: miR-1254, miR-22-3p, miR-345-5p, miR-378a-3p, miR-423-5p, miR-486-5p and miR320a. MiR-1254, miR-22-3p, miR-423-5p, miR-486-5p and miR-320a were selected because they were associated with HF in previous studies [11,21,24]. MiR-378a-3p and miR-345-5p were selected because of their presence in cardiomyocytes [28]. Plasma was thawed on ice and RNA isolation was performed using the TRIZOL LS reagent (Life Technology) according to the manufacturer's protocol, with a starting volume of 200μl plasma. Subsequently, 8μl of the eluate from the RNA isolation was used for the reverse transcription reaction. The transcription followed the manufacturer's protocol of the miScript Reverse Transcription Kit (Qiagen). For real-time PCR, 2μl of 10× diluted cDNA was used in a total volume of 10μll. The real-time PCR reaction was performed on a LightCycler 480 using the following program: 5min of pre-incubation at 95°C; 10s of denaturation at 95°C, 20s of annealing at 58°C and 30s of elongation at 72°C, in a total of 45 cycles. Data were analysed with LinRegPCR quantitative PCR data analysis software. MiR values were normalized using exogenous C. elegans miR-39 as a spike-in control, which was applied prior to the RNA isolation step. The forward primers used were: miR-423-5p: TGAGGGGCAGAGAGCGAGACTTT; miR-22-3p: AAGCTGCCAGTTGAAGAACTGT; miR-378a-3p: ACTGGACTTGGAGTCAGAAGG; miR-1254: CTGGAAGCTGGAGCCTGC; miR-345-5p: GCTGACTCCTAGTCC; miR-486-5p: TCCTGTACTGAGCTG; miR-320a: AAAAGCTGGGTTGAGAGGGCGA.
Batch analysis of N-terminal pro B-type natriuretic peptide (NT-proBNP), high-sensitive cardiac troponin T (HsTNT) and C-reactive protein was performed in the Clinical Chemistry Laboratory of the Erasmus MC, Plasma NT–proBNP and cardiac Troponin T were analysed using electrochemiluminesence immunoassays (Roche Diagnostics, Elecsys 2010, Indianapolis, Indiana, USA), measuring concentrations ranging from 5 to 35,000ng/l and 3 to 10,000ng/l, respectively. CRP was measured using an immunoturbidimetric assay (Roche Hitachi 912 chemistry analyser, Basel, Switzerland), which measures concentrations ranging from 0.3 to 350mg/l.
2.5
Clinical study endpoints
During follow-up, endpoints were recorded in the electronic case report forms by trained research physicians, and associated hospital records and discharge letters were collected. A clinical event committee blinded to the biomarker results subsequently reviewed all collected information and adjudicated primary and secondary endpoints.
The primary endpoint comprised the composite of cardiac death, cardiac transplantation, left ventricular assist device implantation, and hospitalization for the management of acute or worsened HF. Secondary endpoints included individual components of the primary endpoint, as well as myocardial infarction, percutaneous coronary intervention, coronary artery bypass grafting, cerebrovascular accident, and all-cause mortality.
Cardiac death was defined as death from myocardial infarction or other ischemic heart disease (ICD-10: I20–I25), death from other heart disease including HF (I30–I45 and I47–I52), sudden cardiac death (I46), sudden death undefined (R96) or unwitnessed or ill-described death (R98, R99). Hospitalisation for acute or worsened HF was defined as exacerbation of symptoms typical of HF, in combination with 2 of the following: BNP or NT-proBNP >3× ULN, signs of worsening HF, such as pulmonary rales, raised jugular venous pressure or peripheral oedema, increased dose or intravenous administration of diuretics, or administration of positive inotropic agents.
2.6
Statistical analysis
Distributions of continuous variables, including biomarker and miR concentrations, were tested for normality using the Kolmogorov-Smirnov test. Variables with normal distributions are presented as mean±SD. Variables with non-normal distributions are presented as median and interquartile range (IQR). Categorical data are displayed as count and percentage.
In case of skewed distributions, continuous variables were logarithmically transformed for further analyses. Associations between temporal patterns of each separate miR and the primary endpoint were assessed using a joint modelling approach, which combines a linear mixed-effects model for the serial miR measurements with a Cox proportional hazards model for the occurrence of the primary endpoint. Results are presented as hazard ratios (HRs) per doubling of the miR concentration at any point in time, along with the corresponding 95% CIs. All analyses were adjusted for age and gender. Subsequently, additional multivariable adjustment for potential confounding variables was performed. For this purpose, associations between patient characteristics and baseline miR levels, as well as associations between patient characteristics and the primary endpoint, were identified using univariable linear- or Cox regression analyses, when appropriate. All characteristics associated with the miRs or the primary endpoint were subsequently used as covariates in the mixed-effects and Cox models used to construct the joint models. Adjustment for repeatedly assessed NYHA class was performed by entering NYHA class as a time-dependent variable into the joint model.
The above-described analysis assesses the predictive value of the temporal pattern of the actual miR measurements. However, additional features of the miR's trajectory may better predict the primary endpoint [29,30]. Therefore, we investigated whether the instantaneous rate of change in miR (slope of the miR trajectory) is associated with the risk of the primary endpoint. Furthermore, to assess the predictive value of long-term elevation of miR levels, we investigated the associations between the area under the miRs' trajectory and the primary endpoint. This area can be considered as a weighted average of the total longitudinal profile of each patient. The results are presented as HRs per doubling of the area at any point in time, with 95% CIs.
Subsequently, we assessed whether the predictive value of the temporal miR patterns was independent of temporal patterns of NT-proBNP, HsTNT and CRP. For this purpose, all individual temporal miR patterns (of miRs significantly associated with the primary endpoint), as well as NT-proBNP, HsTNT and CRP patterns, derived from the adjusted joint models were saved and subsequently entered simultaneously as time-varying covariates into an extended Cox analysis. The same approach was used to investigate the independent predictive value of the slope and the area under the curve of the temporal miR patterns associated with the primary endpoint.
To investigate the associations between repeated miR measurements and repeated NT-proBNP, HsTNT and CRP measurements, we used generalised estimating equations. Herewith, we examined whether the miR level at a certain time-point is associated with the level of the other biomarkers at that same time-point.
All analyses were performed with R Statistical Software using package JM [29,30]. All tests were two-tailed and p-values<0.05 were considered statistically significant.
3
Result
3.1
Baseline characteristics
From October 2011 to June 2013, 263 patients were included. Mean age was 67years (SD±13), 72% were men, and 73% were in New York Heart Association (NYHA) class I or II (Table 1
). Median baseline NT-proBNP was 139.6pmol/L (IQR 51.9–272.9), HsTNT 18.0ng/l (IQR 9.6–33.2) and CRP 2.2mg/l (IQR 0.9–4.8). Median duration of HF at inclusion was 4.6years (IQR 1.7–9.9). Associations between patient characteristics and baseline miR measurements are presented in the Supplementary table. Of note is the presence of a significant association between ischemic cardiomyopathy (ICM) and several miRs. This is further illustrated in Fig. 2
. No significant associations were present with other baseline characteristics.
From October 2011 to June 2013, 263 patients were included. Mean age was 67years (SD±13), 72% were men, and 73% were in New York Heart Association (NYHA) class I or II (Table 1). Median baseline NT-proBNP was 139.6pmol/L (IQR 51.9–272.9), HsTNT 18.0ng/l (IQR 9.6–33.2) and CRP 2.2mg/l (IQR 0.9–4.8). Median duration of HF at inclusion was 4.6years (IQR 1.7–9.9). Associations between patient characteristics and baseline miR measurements are presented in the Supplementary table. Of note is the presence of a significant association between ischemic cardiomyopathy (ICM) and several miRs. This is further illustrated in Fig. 2. No significant associations were present with other baseline characteristics.
3.2
Clinical endpoints
The primary, composite endpoint was reached by 41 patients (16%), during a median follow-up of 0.9 [0.6–1.4] years: 5 patients died from a cardiovascular cause, 35 patients were re-hospitalized for worsened HF and 1 patient underwent heart transplantation. Of the 35 patients reaching the primary endpoint because of re-hospitalisation for HF, 16 patients died eventually during further follow-up, of whom 12 died from a cardiovascular cause. Overall all-cause mortality was thus 21 (8.0%).
Of all baseline variables, NYHA class (HR 1.98; CI 1.24–3.16 for III/IV vs I/II), baseline NT-proBNP (HR per two-fold difference of baseline NT-proBNP 1.42; CI 1.09–1.86), and baseline CRP (HR per two-fold difference of baseline CRP 1.29; CI 1.05–1.60) were independently associated with the primary endpoint. As shown in Table 2
, none of the baseline miR values were associated with the primary endpoint, or the secondary endpoint comprising HF hospitalisations only.
3.3
Temporal miR patterns and the primary endpoint
During follow-up, we collected 923 blood samples, of which 885 were drawn before the occurrence of the primary endpoint (median of 3 (IQR 2–5) samples per patient). MiRs were successfully determined in 873 of these samples. The temporal pattern of miR-22-3p was inversely associated with the primary endpoint (age and gender adjusted HR per doubling of miR-22-3p level, 0.64; CI 0.47–0.77; p<0.001), i.e. higher miR-22-3p was associated with lower risk of events (Table 3
). As described above, baseline measurements of several miRs were associated with ICM. On the other hand, NYHA class was independently associated with the primary endpoint. Therefore, the baseline variables ICM and NYHA class were added to the models. After adjustment for age, gender, NYHA class and ICM, the association between the temporal pattern of miR-22-3p and the primary endpoint remained present (HR per doubling of miR-22-3p level 0.71; CI 0.44–0.94; p=0.005). When serial NYHA class assessments were added to the model instead of baseline NYHA assessments only, the association between miR miR-22-3p and the primary endpoint also remained present (HR per doubling of miR-22-3p level, 0.52; CI 0.44–0.61; p<0.001). The temporal patterns of the other 6 miRs were not significantly associated with the primary endpoint.
Baseline echocardiograms were available in 72% of all patients due to logistic reasons. After correction for left ventricular ejection fraction in this subgroup, miR-22-3p remained significantly associated with the primary endpoint (HR per doubling of miR-22-3p level, 0.91; CI 0.85–0.95; p<0.001).
Fig. 3
shows that in the patients who reached the endpoint, miR-22-3p values declined as this endpoint approached, while in the patients who did not reach the endpoint, miR-22-3p values remained stable.
3.4
Slope and area under the curve of the temporal miR patterns and the primary endpoint
The hazard ratios for the doubling of the slopes and areas under the curve of the temporal miR patterns are displayed in Table 3. For 4 of the miRs (miR-345-5p, miR-378a-3p, miR-423-5p, and miR-486-5p), slopes remained virtually constant over time, and hazard ratios were not calculable. The slope of the temporal pattern of miR-22-3p was inversely associated with the primary endpoint after multivariable adjustment (HR per doubling of miR-22-3p slope at any given time point, 0.33; CI 0.20–0.51; p<0.001), while the slopes of the patterns of miR-1254 and miR-320a were not. None of the areas under the curves of the temporal miR patterns were associated with the primary endpoint.
3.5
Temporal patterns of miRs, NT-proBNP, Hs-TNT and CRP combined in a multiple-marker model
After constructing the multivariable models that corrected for age, sex, ICM and NYHA class, we proceeded to additionally adjust the association between the temporal miR-22-3p pattern and the primary endpoint for temporal patterns of the other biomarkers which were measured in this study (NT-proBNP, HsTNT and CRP). We did this by entering the multivariably adjusted temporal pattern of miR-22-3p, as well as the multivariably adjusted temporal patterns of NT-proBNP, HsTNT and CRP simultaneously into one model. We found an independent association of miR-22-3p (HR per doubling of miR-22-3p level at any given time point, 0.61; CI 0.51–0.73; p<0.001) with the primary endpoint. Temporal patterns of NT-proBNP and CRP level were also independently associated with the primary endpoint in the multiple-marker model (HRs: 1.34 (1.04–1.71) and 2.14 (1.59–2.88), respectively), while the temporal pattern of HsTNT was not (HR: 1.30 (0.91–1.84)). The slope of the temporal pattern of miR-22-3p also remained significantly associated with the primary endpoint after adjustment for the slopes of the temporal patterns of NT-proBNP, Troponin T and CRP (HR per doubling of miR-22-3p slope at any given time point, 0.55 (0.46–0.65); p<0.001).
3.6
Repeated miR measurements and other biomarkers
Associations between repeatedly measured miR and repeatedly measured NT-proBNP, HsTNT and CRP, calculated using generalised estimating equations, are displayed in Table 4
. Of note are the significant inverse associations between miR-345-5p and CRP (β-coefficient [95% CI]: −0.04 [−0.07 to −0.01] 10log(mg/l) per 10log increase in miR-345-5p level; p=0.012) and miR-22-3p and HsTNT (β-coefficient [95% CI]: −0.01 [−0.02 to −0.00] 10log(mg/l) per 10log increase in miR-22-3p level; p=0.014; as well as the borderline significant inverse association between miR-22-3p and CRP (β-coefficient [95% CI]: −0.02 [−0.04–0.00] 10log (mg/l) per 10log increase in miR-22-3p level; p=0.051). No further associations between repeatedly measured miR and repeatedly measured NT-proBNP, HsTNT and CRP were found.
4
Discussion
In this prospective, observational cohort of CHF patients, temporal miR-22-3p patterns were inversely and independently associated with adverse outcome. The association was independent of temporal NT-proBNP, HsTNT and CRP patterns. The instant rate of change in miR-22-3p level (in terms of the slope of the temporal pattern) was also inversely associated with adverse outcome. Moreover, we found inverse associations of temporal patterns of miR-22-3p with temporal patterns of HsTNT and CRP. An additional finding of this study was the significant, inverse association between ICM and miR-1254, miR486-5p and miR-320a.
Our study has several strengths. Although the number of incident endpoints was limited, this is the largest study to date that has examined miRs in a prospective cohort of patients with chronic HF. NT-proBNP, HsTNT and CRP were measured concomitantly, and, in addition, frequent repeated measurements were performed during follow-up, both of the miRs and of the established blood biomarkers. Statistical models containing all these, repeatedly measured, biomarkers were constructed and modern statistical methods were applied (‘joint modelling’), which take into account the continuous, dynamic changes in biomarker patterns and thus result in less bias than simpler analyses such as delta's between two measurements. Herewith, the study expands existing evidence with regard to miRs and cardiovascular disease in several respects.
MiR-22-3p has previously received attention in the cancer field; it has been implicated in the regulation of various cellular processes, including cell growth, apoptosis, motility, and the cell cycle [31]. With regard to cardiovascular disease, fundamental experiments have demonstrated that expression of miR-22-3p is enriched in striated muscle tissues (i.e. cardiac and skeletal muscles); and miR-22-3p is thought to be required for normal cardiac remodelling in response to stresses [32]. A recent study in miR-22 deficient mice demonstrated that miR-22 depleted hearts quickly progress to adverse cardiac remodelling and develop cardiac dilation, with increased cardiomyocyte loss and deposition of fibrotic tissue, and signs of heart failure [33]. Another study also found that mice without miR-22 develop dilatation of myocardium more quickly [34]. Interestingly, several animal models have shown that miR-22 may potentially serve as a therapeutic target in cardiac hypertrophy [33]. Recently, Gupta et al found that pharmacological inhibition of miR-22 post-infarction activated cardiac autophagy, led to improved cardiac function, and inhibited cardiac remodelling in older mice [35]. They also found that circulating miR-22 provides prognostic information in HF patients. Specifically, they demonstrated an association between higher baseline levels of circulating miR-22 and mortality in154 HF patients. Repeated measurements were not performed. In the present study, although we saw a temporal decline of miR-22-3p levels in patients with adverse clinical outcome, we could not demonstrate a significant association between higher baseline miR-22-3p and clinical outcome. The shorter follow-up duration of our study (median of approx. 1year, versus 3years follow-up in the study of Gupta et al) may have contributed to this. Furthermore, in an earlier study, Goren et al, demonstrated that elevated serum levels of miR-22-3p identify patients with systolic heart failure using 30 stable chronic systolic heart failure patients and 30 controls [24]. This seeming discrepancy with our study could possibly be explained by differences in study design. While Goren et al focused on diagnostic properties of miR-22-3p by comparing cases and controls, we focused on prognostic value of miR-22-3p in patients with established CHF. To the best of our knowledge, other cardiovascular studies on miR-22-3p in human subjects are lacking. Altogether, our study is in line with previous fundamental experiments which demonstrated that loss of miR-22-3p leads to dilatation of myocardium and progressive HF. Our study confirms that lower levels of miR-22-3p are associated with worsening of CHF, and that miR-22-3p may complement established HF biomarkers in prognostication.
Previously, several studies using animal models have shown decreased levels of miR-378a-3p during cardiac remodelling [36] and in HF due to cardiac hypertrophy [37–39]. Furthermore, a significant down-regulation of miR-378a-3p in hearts with dilated cardiomyopathy compared to non-failing control hearts has been observed [23]. Since the presence of miR-378a-3p is required to resist ventricular remodelling during cardiac stress, it may offer therapeutic potential for the management of HF [38]. It has been suggested, that up-regulation of miR-378a-3p could serve as a potential novel treatment for apoptosis and ischemic heart disease, since up-regulation of this miR enhanced cell viability, and inhibited apoptosis and necrosis in myocardium of rats [40]. In the current study, we also observed a tendency towards an inverse association between the temporal pattern of miR378a-3p and the primary endpoint in univariable models. These results concur with the aforementioned studies.
Another miR that has previously been associated with HF, is miR-423-5p [19,21,22,24]. However, in our study, there was no association between the temporal pattern of miR-423-5p and the primary endpoint. Existing studies on miR-423-5p in humans all compared single measurements in (C)HF patients with measurements in healthy controls, and imply that this miR has diagnostic value for HF [19,21,24]. We did not evaluate the diagnostic value of miRs, but the prognostic value of serial miR measurements in patients with known CHF. Although miR-423-5p may discriminate between HF patients and healthy controls, it may lack prognostic value in patients with longstanding CHF. Earlier findings support this hypothesis [20]. Bauters et al. serially measured serum levels of circulating miR-423-5p in 246 patients with a first anterior Q-wave myocardial infarction, at discharge and subsequently at 3months and 1year. They found no association between miR-423-5p levels and indices of left ventricular function and left ventricular remodelling, which were serially assessed during a 1year period after the index myocardial infarction, nor with B-type natriuretic peptide. They concluded that miR-423-5p is not a useful biomarker of left ventricular remodelling after myocardial infarction [20].
The selection of the miRs for the current investigation was based on previous studies, which used animal models and relatively small patient groups. Some of these studies have shown associations between miR-1254, miR-345-5p, miR-486-5p and miR-320a and heart failure [19,21,24,41,42]. Again, these were mostly studies that compared HF patients to healthy controls. In our study, we did not find any associations between these miRs and poor outcome in CHF. Multiple other miRs, such as miR-133a [43], have previously been associated with HF [44]. Serial measurements of these other miRs, which were not investigated in this study, warrant further investigation. This could lead to further improvements in CHF prognostication, as well as tailored adjustment of treatment.
Some aspects of this study warrant consideration. For the current investigation we only used the first round of inclusions of the ongoing Bio-SHiFT study (n=263). It should be noted that the full Bio-SHiFT cohort is designed to have sufficient statistical power to enable correction for multiple testing in case of application of modern, large-scale biomarker measurement methods. However, the current investigation examines only 7 miRs, which were chosen based on pathophysiological considerations. Thus, the current investigation was hypothesis-driven and not data-driven, rendering correction for multiple testing unnecessary in this particular case. Furthermore, with an average follow-up ~1year, the number of primary endpoints in the current investigation (n=41) was limited, and consequently so was the number of covariables that could be entered into the models. The lack of a replication population is another limitation of the study. Finally, the majority of the patient population was in NYHA functional class I or II, and thus at relatively low risk. Accordingly, the primary outcome was predominantly composed of heart failure hospitalizations. This may have hampered the prognostic information of the other miRs tested in this study. Future studies using high-risk patient populations should therefore re-investigate the prognostic information carried by repeated assessment of these miRs.
Remarkably, in the present study, the rate of HFPEF was very low compared to other epidemiological studies. This can most likely be attributed to the fact that in the Netherlands, most HFPEF patients are treated by the general practitioner or in secondary referral centres, while the current study was performed in two centres which were both tertiary referral centres. Potential inclusion bias is not a likely reason for the low HPEF rate, because consecutive patients were screened in both participating centres. Moreover, the number of patients on vitamin K antagonists was quite high in this study, most likely due to the high number of patients with left ventricular aneurysms and thrombi, as well as the high number of patients with atrial fibrillation.
In conclusion, in patients with stable CHF, the temporal pattern of circulating miR-22-3p is a strong and independent predictor of prognosis. Thus, the use of individual patterns of change of circulating miR-22-3p for CHF prognostication, as well as for tailored adjustment of treatment, warrants further investigation.
The following are the supplementary data related to this article.
Supplemental table
Associations between baseline characteristics and miRNAs measured at baseline.
Supplemental table
Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.ijcard.2017.02.078.
Funding sources
This work was supported by the Jaap Schouten Foundation and the Foreest Medical School.
Conflicts of interest
MiR measurements were performed by ACS Biomarker BV. ACS Biomarker had no role in the study design, analysis and interpretation of data, writing of the manuscript, or in the decision to submit the manuscript for publication. Dr. Y. Pinto has a commercial interest in ACS Biomarker (<5%).
Acknowledgements
We would like to thank the following staff for their contribution to this study: Lara Emilsdottir, Ymkje Hendriksma, Chris Jansen, Ellen Klaassen, Agnes Muskens, and Wout de Ruiter.
References
[1]
P.
Ponikowski
A.A.
Voors
S.D.
Anker
H.
Bueno
J.G.
Cleland
A.J.
Coats
V.
Falk
J.R.
Gonzalez-Juanatey
V.P.
Harjola
E.A.
Jankowska
M.
Jessup
C.
Linde
P.
Nihoyannopoulos
J.T.
Parissis
B.
Pieske
J.P.
Riley
G.M.
Rosano
L.M.
Ruilope
F.
Ruschitzka
F.H.
Rutten
M.P.
van der
2016 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure: the Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC). Developed with the special contribution of the Heart Failure Association (HFA) of the ESC
Eur. J. Heart Fail.
18
8
2016
891
975
[2]
C.W.
Yancy
M.
Jessup
B.
Bozkurt
J.
Butler
D.E.
Casey
Jr.
M.H.
Drazner
G.C.
Fonarow
S.A.
Geraci
T.
Horwich
J.L.
Januzzi
M.R.
Johnson
E.K.
Kasper
W.C.
Levy
F.A.
Masoudi
P.E.
McBride
J.J.
McMurray
J.E.
Mitchell
P.N.
Peterson
B.
Riegel
F.
Sam
L.W.
Stevenson
W.H.
Tang
E.J.
Tsai
B.L.
Wilkoff
2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines
J. Am. Coll. Cardiol.
62
16
2013
e147
e239
[3]
J.L.
Januzzi
R.
Troughton
Are serial BNP measurements useful in heart failure management? Serial natriuretic peptide measurements are useful in heart failure management
Circulation
127
4
2013
500
507
[4]
G.
de Couto
M.
Ouzounian
P.P.
Liu
Early detection of myocardial dysfunction and heart failure
Nat. Rev. Cardiol.
7
6
2010
334
344
[5]
G.
Savarese
B.
Trimarco
S.
Dellegrottaglie
M.
Prastaro
F.
Gambardella
G.
Rengo
D.
Leosco
P.
Perrone-Filardi
Natriuretic peptide-guided therapy in chronic heart failure: a meta-analysis of 2,686 patients in 12 randomized trials
PLoS One
8
3
2013
e58287
[6]
P.
Porapakkham
P.
Porapakkham
H.
Zimmet
B.
Billah
H.
Krum
B-type natriuretic peptide-guided heart failure therapy: a meta-analysis
Arch. Intern. Med.
170
6
2010
507
514
[7]
W.L.
Miller
K.A.
Hartman
M.F.
Burritt
D.E.
Grill
R.J.
Rodeheffer
J.C.
Burnett
Jr.
A.S.
Jaffe
Serial biomarker measurements in ambulatory patients with chronic heart failure: the importance of change over time
Circulation
116
3
2007
249
257
[8]
R.T.
Yan
M.
White
A.T.
Yan
S.
Yusuf
J.L.
Rouleau
A.P.
Maggioni
C.
Hall
R.
Latini
R.
Afzal
J.
Floras
S.
Masson
R.S.
McKelvie
Usefulness of temporal changes in neurohormones as markers of ventricular remodeling and prognosis in patients with left ventricular systolic dysfunction and heart failure receiving either candesartan or enalapril or both
Am. J. Cardiol.
96
5
2005
698
704
[9]
R.
Troughton
F.G.
Michael
J.L.
Januzzi
Jr.
Natriuretic peptide-guided heart failure management
Eur. Heart J.
35
1
2014
16
24
[10]
D.P.
Bartel
MicroRNAs: genomics, biogenesis, mechanism, and function
Cell
116
2
2004
281
297
[11]
S.
Gilad
E.
Meiri
Y.
Yogev
S.
Benjamin
D.
Lebanony
N.
Yerushalmi
H.
Benjamin
M.
Kushnir
H.
Cholakh
N.
Melamed
Z.
Bentwich
M.
Hod
Y.
Goren
A.
Chajut
Serum microRNAs are promising novel biomarkers
PLoS One
3
9
2008
e3148
[12]
J.F.
Chen
E.P.
Murchison
R.
Tang
T.E.
Callis
M.
Tatsuguchi
Z.
Deng
M.
Rojas
S.M.
Hammond
M.D.
Schneider
C.H.
Selzman
G.
Meissner
C.
Patterson
G.J.
Hannon
D.Z.
Wang
Targeted deletion of Dicer in the heart leads to dilated cardiomyopathy and heart failure
Proc. Natl. Acad. Sci. U. S. A.
105
6
2008
2111
2116
[13]
P.A.
Costa Martins
M.
Bourajjaj
M.
Gladka
M.
Kortland
R.J.
van Oort
Y.M.
Pinto
J.D.
Molkentin
L.J.
De Windt
Conditional dicer gene deletion in the postnatal myocardium provokes spontaneous cardiac remodeling
Circulation
118
15
2008
1567
1576
[14]
E.E.
Creemers
A.J.
Tijsen
Y.M.
Pinto
Circulating microRNAs: novel biomarkers and extracellular communicators in cardiovascular disease?
Circ. Res.
110
3
2012
483
495
[15]
B.
Vogel
A.
Keller
K.S.
Frese
W.
Kloos
E.
Kayvanpour
F.
Sedaghat-Hamedani
S.
Hassel
S.
Marquart
M.
Beier
E.
Giannitis
S.
Hardt
H.A.
Katus
B.
Meder
Refining diagnostic microRNA signatures by whole-miRNome kinetic analysis in acute myocardial infarction
Clin. Chem.
59
2
2013
410
418
[16]
A.
Zampetaki
P.
Willeit
L.
Tilling
I.
Drozdov
M.
Prokopi
J.M.
Renard
A.
Mayr
S.
Weger
G.
Schett
A.
Shah
C.M.
Boulanger
J.
Willeit
P.J.
Chowienczyk
S.
Kiechl
M.
Mayr
Prospective study on circulating MicroRNAs and risk of myocardial infarction
J. Am. Coll. Cardiol.
60
4
2012
290
299
[17]
M.F.
Corsten
R.
Dennert
S.
Jochems
T.
Kuznetsova
Y.
Devaux
L.
Hofstra
D.R.
Wagner
J.A.
Staessen
S.
Heymans
B.
Schroen
Circulating MicroRNA-208b and MicroRNA-499 reflect myocardial damage in cardiovascular disease
Circ. Cardiovasc. Genet.
3
6
2010
499
506
[18]
F.
Olivieri
R.
Antonicelli
M.
Lorenzi
Y.
D'Alessandra
R.
Lazzarini
G.
Santini
L.
Spazzafumo
R.
Lisa
L.
La Sala
R.
Galeazzi
R.
Recchioni
R.
Testa
G.
Pompilio
M.C.
Capogrossi
A.D.
Procopio
Diagnostic potential of circulating miR-499-5p in elderly patients with acute non ST-elevation myocardial infarction
Int. J. Cardiol.
167
2
2013
531
536
[19]
L.A.
Goldraich
N.C.
Martinelli
U.
Matte
C.
Cohen
M.
Andrades
M.
Pimentel
A.
Biolo
N.
Clausell
L.E.
Rohde
Transcoronary gradient of plasma microRNA 423-5p in heart failure: evidence of altered myocardial expression
Biomarkers
19
2
2014
135
141
[20]
C.
Bauters
R.
Kumarswamy
A.
Holzmann
J.
Bretthauer
S.D.
Anker
F.
Pinet
T.
Thum
Circulating miR-133a and miR-423-5p fail as biomarkers for left ventricular remodeling after myocardial infarction
Int. J. Cardiol.
168
3
2013
1837
1840
[21]
A.J.
Tijsen
E.E.
Creemers
P.D.
Moerland
L.J.
De Windt
A.C.
van der Wal
W.E.
Kok
Y.M.
Pinto
MiR423-5p as a circulating biomarker for heart failure
Circ. Res.
106
6
2010
1035
1039
[22]
B.A.
Dickinson
H.M.
Semus
R.L.
Montgomery
C.
Stack
P.A.
Latimer
S.M.
Lewton
J.M.
Lynch
T.G.
Hullinger
A.G.
Seto
E.
van Rooij
Plasma microRNAs serve as biomarkers of therapeutic efficacy and disease progression in hypertension-induced heart failure
Eur. J. Heart Fail.
15
6
2013
650
659
[23]
S.V.
Naga Prasad
Z.H.
Duan
M.K.
Gupta
V.S.
Surampudi
S.
Volinia
G.A.
Calin
C.G.
Liu
A.
Kotwal
C.S.
Moravec
R.C.
Starling
D.M.
Perez
S.
Sen
Q.
Wu
E.F.
Plow
C.M.
Croce
S.
Karnik
Unique microRNA profile in end-stage heart failure indicates alterations in specific cardiovascular signaling networks
J. Biol. Chem.
284
40
2009
27487
27499
[24]
Y.
Goren
M.
Kushnir
B.
Zafrir
S.
Tabak
B.S.
Lewis
O.
Amir
Serum levels of microRNAs in patients with heart failure
Eur. J. Heart Fail.
14
2
2012
147
154
[25]
J.J.
McMurray
S.
Adamopoulos
S.D.
Anker
A.
Auricchio
M.
Bohm
K.
Dickstein
V.
Falk
G.
Filippatos
C.
Fonseca
M.A.
Gomez-Sanchez
T.
Jaarsma
L.
Kober
G.Y.
Lip
A.P.
Maggioni
A.
Parkhomenko
B.M.
Pieske
B.A.
Popescu
P.K.
Ronnevik
F.H.
Rutten
J.
Schwitter
P.
Seferovic
J.
Stepinska
P.T.
Trindade
A.A.
Voors
F.
Zannad
A.
Zeiher
ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure 2012: The Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association (HFA) of the ESC
Eur. Heart J.
33
14
2012
1787
1847
[26]
K.
Dickstein
A.
Cohen-Solal
G.
Filippatos
J.J.
McMurray
P.
Ponikowski
P.A.
Poole-Wilson
A.
Stromberg
D.J.
van Veldhuisen
D.
Atar
A.W.
Hoes
A.
Keren
A.
Mebazaa
M.
Nieminen
S.G.
Priori
K.
Swedberg
ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure 2008: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2008 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association of the ESC (HFA) and endorsed by the European Society of Intensive Care Medicine (ESICM)
Eur. Heart J.
29
19
2008
2388
2442
[27]
W.J.
Paulus
C.
Tschope
J.E.
Sanderson
C.
Rusconi
F.A.
Flachskampf
F.E.
Rademakers
P.
Marino
O.A.
Smiseth
G.
De Keulenaer
A.F.
Leite-Moreira
A.
Borbely
I.
Edes
M.L.
Handoko
S.
Heymans
N.
Pezzali
B.
Pieske
K.
Dickstein
A.G.
Fraser
D.L.
Brutsaert
How to diagnose diastolic heart failure: a consensus statement on the diagnosis of heart failure with normal left ventricular ejection fraction by the Heart Failure and Echocardiography Associations of the European Society of Cardiology
Eur. Heart J.
28
20
2007
2539
2550
[28]
P.K.
Rao
Y.
Toyama
H.R.
Chiang
S.
Gupta
M.
Bauer
R.
Medvid
F.
Reinhardt
R.
Liao
M.
Krieger
R.
Jaenisch
H.F.
Lodish
R.
Blelloch
Loss of cardiac microRNA-mediated regulation leads to dilated cardiomyopathy and heart failure
Circ. Res.
105
6
2009
585
594
[29]
D.
Rizopoulos
JM: an R package for the joint modelling of longitudinal and time-to-event data
J. Stat. Softw.
35
9
2010
1
33
[30]
D.
Rizopoulos
Joint Models for Longitudinal and Time-to-Event Data: With Applications in R
2012
Chapman & Hall/CRC
Boca Raton
[31]
J.
Li
S.
Liang
H.
Yu
J.
Zhang
D.
Ma
X.
Lu
An inhibitory effect of miR-22 on cell migration and invasion in ovarian cancer
Gynecol. Oncol.
119
3
2010
543
548
[32]
Z.P.
Huang
D.Z.
Wang
miR-22 in cardiac remodeling and disease
Trends Cardiovasc. Med.
24
7
2014
267
272
[33]
Z.P.
Huang
J.
Chen
H.Y.
Seok
Z.
Zhang
M.
Kataoka
X.
Hu
D.Z.
Wang
MicroRNA-22 regulates cardiac hypertrophy and remodeling in response to stress
Circ. Res.
112
9
2013
1234
1243
[34]
P.
Gurha
C.
Abreu-Goodger
T.
Wang
M.O.
Ramirez
A.L.
Drumond
D.S.
van
Y.
Chen
N.
Bartonicek
A.J.
Enright
B.
Lee
R.J.
Kelm
Jr.
A.K.
Reddy
G.E.
Taffet
A.
Bradley
X.H.
Wehrens
M.L.
Entman
A.
Rodriguez
Targeted deletion of microRNA-22 promotes stress-induced cardiac dilation and contractile dysfunction
Circulation
125
22
2012
2751
2761
[35]
S.K.
Gupta
A.
Foinquinos
S.
Thum
J.
Remke
K.
Zimmer
C.
Bauters
P.
de Groote
R.A.
Boon
L.J.
De Windt
S.
Preissl
L.
Hein
S.
Batkai
F.
Pinet
T.
Thum
Preclinical development of a microRNA-based therapy for elderly patients with myocardial infarction
J. Am. Coll. Cardiol.
68
14
2016
1557
1571
[36]
I.
Knezevic
A.
Patel
N.R.
Sundaresan
M.P.
Gupta
R.J.
Solaro
R.S.
Nagalingam
M.
Gupta
A novel cardiomyocyte-enriched microRNA, miR-378, targets insulin-like growth factor 1 receptor: implications in postnatal cardiac remodeling and cell survival
J. Biol. Chem.
287
16
2012
12913
12926
[37]
R.S.
Nagalingam
N.R.
Sundaresan
M.P.
Gupta
D.L.
Geenen
R.J.
Solaro
M.
Gupta
A cardiac-enriched microRNA, miR-378, blocks cardiac hypertrophy by targeting Ras signaling
J. Biol. Chem.
288
16
2013
11216
11232
[38]
R.S.
Nagalingam
N.R.
Sundaresan
M.
Noor
M.P.
Gupta
R.J.
Solaro
M.
Gupta
Deficiency of cardiomyocyte-specific MicroRNA-378 contributes to the development of cardiac fibrosis involving a transforming growth factor beta (TGFbeta1)-dependent paracrine mechanism
J. Biol. Chem.
289
39
2014
27199
27214
[39]
J.
Ganesan
D.
Ramanujam
Y.
Sassi
A.
Ahles
C.
Jentzsch
S.
Werfel
S.
Leierseder
X.
Loyer
M.
Giacca
L.
Zentilin
T.
Thum
B.
Laggerbauer
S.
Engelhardt
MiR-378 controls cardiac hypertrophy by combined repression of mitogen-activated protein kinase pathway factors
Circulation
127
21
2013
2097
2106
[40]
J.
Fang
X.W.
Song
J.
Tian
H.Y.
Chen
D.F.
Li
J.F.
Wang
A.J.
Ren
W.J.
Yuan
L.
Lin
Overexpression of microRNA-378 attenuates ischemia-induced apoptosis by inhibiting caspase-3 expression in cardiac myocytes
Apoptosis
17
4
2012
410
423
[41]
X.
Li
X.
Zhang
T.
Wang
C.
Sun
T.
Jin
H.
Yan
J.
Zhang
X.
Li
T.
Geng
C.
Chen
A.
Ma
S.
Li
Regulation by bisoprolol for cardiac microRNA expression in a rat volume-overload heart failure model
J. Nanosci. Nanotechnol.
13
8
2013
5267
5275
[42]
E.M.
Small
J.R.
O'Rourke
V.
Moresi
L.B.
Sutherland
J.
McAnally
R.D.
Gerard
J.A.
Richardson
E.N.
Olson
Regulation of PI3-kinase/Akt signaling by muscle-enriched microRNA-486
Proc. Natl. Acad. Sci. U. S. A.
107
9
2010
4218
4223
[43]
C.
Besler
D.
Urban
S.
Watzka
D.
Lang
K.P.
Rommel
R.
Kandolf
K.
Klingel
H.
Thiele
A.
Linke
G.
Schuler
V.
Adams
P.
Lurz
Endomyocardial miR-133a levels correlate with myocardial inflammation, improved left ventricular function, and clinical outcome in patients with inflammatory cardiomyopathy
Eur. J. Heart Fail.
2016
[44]
E.L.
Vegter
M.P.
van der
L.J.
De Windt
Y.M.
Pinto
A.A.
Voors
MicroRNAs in heart failure: from biomarker to target for therapy
Eur. J. Heart Fail.
18
5
2016
457
468
SimpleXMLElement Object
(
[coredata] => SimpleXMLElement Object
(
[eid] => 1-s2.0-S0167527317311075
[pii] => S0167-5273(17)31107-5
[pubType] => fla
[openaccess] => 1
[openaccessArticle] => true
[openaccessType] => Full
[openArchiveArticle] => false
[openaccessSponsorName] => VSNU
[openaccessSponsorType] => FundingBody
[openaccessUserLicense] => http://creativecommons.org/licenses/by/4.0/
[link] => Array
(
[0] => SimpleXMLElement Object
(
[@attributes] => Array
(
[href] => https://api.elsevier.com/content/article/pii/S0167527317311075
[rel] => self
)
)
[1] => SimpleXMLElement Object
(
[@attributes] => Array
(
[href] => https://www.sciencedirect.com/science/article/pii/S0167527317311075
[rel] => scidir
)
)
)
)
[objects] => SimpleXMLElement Object
(
[object] => Array
(
[0] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-gr1.sml?httpAccept=%2A%2F%2A
[1] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-gr2.sml?httpAccept=%2A%2F%2A
[2] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-gr3.sml?httpAccept=%2A%2F%2A
[3] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-gr1.jpg?httpAccept=%2A%2F%2A
[4] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-gr2.jpg?httpAccept=%2A%2F%2A
[5] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-gr3.jpg?httpAccept=%2A%2F%2A
[6] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-gr1_lrg.jpg?httpAccept=%2A%2F%2A
[7] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-gr2_lrg.jpg?httpAccept=%2A%2F%2A
[8] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-gr3_lrg.jpg?httpAccept=%2A%2F%2A
[9] => https://api.elsevier.com/content/object/eid/1-s2.0-S0167527317311075-mmc1.doc?httpAccept=%2A%2F%2A
)
)
[scopus-id] => 85014215458
[scopus-eid] => 2-s2.0-85014215458
[pubmed-id] => 28274577
[link] => SimpleXMLElement Object
(
[@attributes] => Array
(
[href] => https://api.elsevier.com/content/abstract/scopus_id/85014215458
[rel] => abstract
)
)
[originalText] => SimpleXMLElement Object
(
)
)
Visual Composer #161 - Kentucky Medical Directors Association
Visual Composer #161 - Kentucky Medical Directors Association