Listen to this manuscript’s
audio summary by
JACC Editor-in-Chief
Dr. Valentin Fuster.
J O U R N A L O F T H E AM E R I C A N C O L L E G E O F C A R D I O L O G Y V O L . 7 1 , N O . 2 0 , 2 0 1 8
ª 2 0 1 8 T H E A U T H O R S . P U B L I S H E D B Y E L S E V I E R O N B E H A L F O F T H E A M E R I C A N
C O L L E G E O F C A R D I O L O G Y F OU N D A T I O N . T H I S I S A N O P E N A C C E S S A R T I C L E U N D E R
T H E C C B Y L I C E N S E ( h t t p : / / c r e a t i v e c o mm o n s . o r g / l i c e n s e s / b y / 4 . 0 / ) .Genetic Etiology for Alcohol-Induced
Cardiac Toxicity
James S. Ware, MRCP, PHD,a,b,c,* Almudena Amor-Salamanca, MD,d,* Upasana Tayal, MRCP, PHD,a,b,*
Risha Govind, MSC,a,b,e,* Isabel Serrano, MD,f Joel Salazar-Mendiguchía, MD,g,h
Jose Manuel García-Pinilla, MD, PHD,i,j Domingo A. Pascual-Figal, MD, PHD,i,k Julio Nuñez, MD, PHD,i,l
Gonzalo Guzzo-Merello, MD, PHD,d Emiliano Gonzalez-Vioque, PHD,m Alfredo Bardaji, MD, PHD,f
Nicolas Manito, MD, PHD,g Miguel A. López-Garrido, MD,i,j Laura Padron-Barthe, PHD,d,i Elizabeth Edwards, PHD,a,b
Nicola Whiffin, PHD,a,b,c Roddy Walsh, MSC, PHD,a,b Rachel J. Buchan, MSC,a,b William Midwinter, BSC,a,b
Alicja Wilk, BSC,a,b Sanjay Prasad, MD,a,b Antonis Pantazis, MD,b John Baski, MRCP, PHD,b
Declan P. O’Regan, MRCP, PHD,c Luis Alonso-Pulpon, MD, PHD,d,i Stuart A. Cook, MRCP, PHD,a,c,n,o
Enrique Lara-Pezzi, PHD,i,p Paul J. Barton, PHD,a,b,* Pablo Garcia-Pavia, MD, PHDd,i,q,*ABSTRACTISS
Fro
Ce
Me
De
Ne
me
Ca
Au
Dis
Ca
De
Ho
Sin
pM
qUBACKGROUND Alcoholic cardiomyopathy (ACM) is defined by a dilated and impaired left ventricle due to chronic
excess alcohol consumption. It is largely unknown which factors determine cardiac toxicity on exposure to alcohol.
OBJECTIVES This study sought to evaluate the role of variation in cardiomyopathy-associated genes in the patho-
physiology of ACM, and to examine the effects of alcohol intake and genotype on dilated cardiomyopathy (DCM) severity.
METHODS The authors characterized 141 ACM cases, 716 DCM cases, and 445 healthy volunteers. The authors
compared the prevalence of rare, protein-altering variants in 9 genes associated with inherited DCM. They evaluated the
effect of genotype and alcohol consumption on phenotype in DCM.
RESULTS Variants in well-characterized DCM-causing genes were more prevalent in patients with ACM than control
subjects (13.5% vs. 2.9%; p ¼ 1.2 105), but similar between patients with ACM and DCM (19.4%; p ¼ 0.12) and with a
predominant burden of titin truncating variants (TTNtv) (9.9%). Separately, we identified an interaction between TTN
genotype and excess alcohol consumption in a cohort of DCM patients not meeting ACM criteria. On multivariate analysis,
DCM patients with a TTNtv who consumed excess alcohol had an 8.7% absolute reduction in ejection fraction (95%
confidence interval: 2.3% to 15.1%; p < 0.007) compared with those without TTNtv and excess alcohol consumption.
The presence of TTNtv did not predict phenotype, outcome, or functional recovery on treatment in ACM patients.
CONCLUSIONS TTNtv represent a prevalent genetic predisposition for ACM, and are also associated with a worse left
ventricular ejection fraction in DCM patients who consume alcohol above recommended levels. Familial evaluation
and genetic testing should be considered in patients presenting with ACM. (J Am Coll Cardiol 2018;71:2293–302)
© 2018 The Authors. Published by Elsevier on behalf of the American College of Cardiology Foundation. This is an
open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).N 0735-1097 https://doi.org/10.1016/j.jacc.2018.03.462
m the aNational Heart and Lung Institute, Imperial College London, London, United Kingdom; bCardiovascular Research
ntre, Royal Brompton and Harefield NHS Foundation Trust London, London, United Kingdom; cMRC London Institute of
dical Sciences, Imperial College London, London, United Kingdom; dHeart Failure and Inherited Cardiac Diseases Unit,
partment of Cardiology, Hospital Universitario Puerta de Hierro, Madrid, Spain; eInstitute of Psychiatry, Psychology and
uroscience, Social Genetic and Developmental Psychiatry Centre, King’s College London, London, United Kingdom; fDepart-
nt of Cardiology, Hospital Universitario de Tarragona Joan XXIII, IISPV, Rovira Virgili University, Tarragona, Spain; gInherited
rdiac Diseases Unit, Department of Cardiology, Hospital Bellvitge, Barcelona, Spain; hGenetics Department, Universidad
tónoma de Barcelona, Barcelona, Spain; iCIBER in Cardiovascular Diseases, Madrid, Spain; jHeart Failure and Inherited Cardiac
eases Unit, Department of Cardiology, Hospital Universitario Virgen de la Victoria, IBIMA, Málaga, Spain; kDepartment of
rdiology, Hospital Universitario Virgen de la Arrixaca, IMIB-Arrixaca, University of Murcia, Murcia, Spain; lCardiology
partment, Hospital Clínico Universitario, INCLIVA Universitat de Valencia, Valencia, Spain; mDepartment of Biochemistry,
spital Universitario Puerta de Hierro, Madrid, Spain; nNational Heart Research Institute Singapore, National Heart Centre
gapore, Singapore; oDivision of Cardiovascular & Metabolic Disorders, Duke-National University of Singapore, Singapore;
yocardial Biology Programme, Centro Nacional de Investigaciones Cardiovasculares (CNIC), Madrid, Spain; and the
niversity Francisco de Vitoria (UFV), Pozuelo de Alarcón, Madrid, Spain. *Drs. Ware, Amor-Salamanca, Tayal, Govind, Barton, and
ABBR EV I A T I ON S
AND ACRONYMS
ACM = alcoholic
cardiomyopathy
DCM = dilated cardiomyopathy
ExAC = Exome Aggregation
Consortium
LVEF = left ventricular ejection
fraction
TTNtv = titin truncating variant
Garcia-Pavi
Spanish Min
Foundation
Biomedical
Biomedical
Leducq (11
Kingdom (H
Foundation
and Compe
of making E
European R
no role in th
Prof. Cook i
to the conte
Manuscript
Ware et al. J A C C V O L . 7 1 , N O . 2 0 , 2 0 1 8
Genetics and Alcohol Cardiotoxicity M A Y 2 2 , 2 0 1 8 : 2 2 9 3 – 3 0 2
2294A lcoholic cardiomyopathy (ACM) iscaused by chronic and excessivealcohol intake (1–4). Although mod-
erate levels of alcohol consumption may
have some beneficial cardiovascular effects
(3,5), prolonged and excessive consumption
can lead to deleterious consequences
including cardiac arrhythmias and a dilated
cardiomyopathy (DCM) phenotype (2–5).
The pathophysiology of ACM is not fully un-
derstood, and the relationship between thedegree of alcohol exposure and severity of end-
organ damage is not simple (2,6). In particular, not
all individuals with high alcohol intake develop
ACM, and this heterogeneity in response indicates
differences in underlying susceptibility, likely both
genetic and environmental. However, studies of her-
itable contributors to ACM are currently limited
(7,8). Prognosis in ACM is poor, but is considered
more favorable than DCM generally, with recovery
observed in up to one-third of cases, especially
when alcohol intake is reduced (3). Current manage-
ment of ACM individuals includes cessation of
alcohol exposure, standard heart failure medications,
and prevention of sudden cardiac death (2,6,9). As
the genetic contribution to ACM is currently un-
known, familial evaluation is not part of ACM
management.SEE PAGE 2303DCM is estimated to affect up to 1 in 250 in-
dividuals (10) and has a significant genetic contribu-
tion, with truncation variants in the gene encoding
titin (TTNtv), a key sarcomeric protein, representing
the predominant genetic cause, seen in 10% to 20% of
cases (11–13). It is also recognized that up to 1% of the
general population carry a TTNtv, presenting a sig-
nificant challenge to interpretation (12). Genetic
and/or environmental factors likely underlie the
variable penetrance and expressivity. In line witha contributed equally to this work. This work was supported by th
istry of Economy and Competitiveness (SAF2015-71863-REDT), t
(SP/10/10/28431), the Medical Research Council, the National I
Research Unit based at Royal Brompton & Harefield NHS Found
Research Centre based at Imperial College London Healthcare NH
CVD-01), and a Health Innovation Challenge Fund award from th
ICF-R6-373). The CNIC is supported by the Ministry of Econom
, and is a Severo Ochoa Center of Excellence (SEV-2015-0505). Gra
titiveness are supported by the Plan Estatal de IþDþI 2013–2016—Eu
urope.” The Hospital Universitario Puerta de Hierro Majadahonda
eference Network for rare, low-prevalence, and complex diseases
e design, collection, analysis, or interpretation of the data or in t
s cofounder and a shareholder of Enleofen Bio. All other authors h
nts of this paper to disclose.
received January 5, 2018; revised manuscript received Februarythis, recent evidence has shown that over 10% of
patients with peripartum cardiomyopathy carry a
TTNtv, suggesting that in some patients, the DCM
phenotype results from a combination of pregnancy
with a genetic predisposing background (14).
Moreover, we have recently demonstrated that
TTNtv found in the general population are not
phenotypically silent (15); although the population
prevalence of TTNtv exceeds the prevalence of DCM,
careful phenotyping reveals differences in cardiac
volumes in subjects with and without TTNtv. Using a
rat model, we found essentially normal resting car-
diac function, but subclinical metabolic abnormalities
in TTNtv carriers and impaired cardiac physiology
under conditions of cardiac stress (15). Together,
these data suggest that TTNtv may predispose to
cardiomyopathy, with environmental factors modu-
lating penetrance and expressivity.
Here, we sought to evaluate genetic determinants
in the pathophysiology of ACM by characterizing ge-
netic variation in known DCM-causing genes in a
large ACM cohort. We sequenced 141 individuals with
ACM and compared these with healthy volunteers
(n ¼ 445), individuals with DCM (n ¼ 366), and
population-based variant frequency data (Exome
Aggregation Consortium [ExAC]; n ¼ 60,706). We
further evaluated the phenotypic effect of excessive
alcohol intake (below the levels required for a diag-
nosis of ACM) in the context of TTNtv in a wider
cohort of 716 DCM subjects.
METHODS
The study conformed to the ethical principles of the
Declaration of Helsinki and was approved by the
local institutional review board of Hospital Uni-
versitario Puerta de Hierro and a National Health
Service Health Research Authority Research Ethics
Committee. All patients provided written informed
consent.e Instituto de Salud Carlos III (ISCIII) (PI15/01551), the
he Wellcome Trust (107469/Z/15/Z), the British Heart
nstitute for Health Research (NIHR) Cardiovascular
ation Trust and Imperial College London, the NIHR
S Trust and Imperial College London, the Fondation
e Wellcome Trust and Department of Health, United
y, Industry and Competitiveness and the Pro CNIC
nts from ISCIII and the Spanish Ministry of Economy
ropean Regional Development Fund (FEDER) “A way
and Hospital Virgen de la Arrixaca are members of the
of the heart (ERN GUARD-Heart). The funders played
he decision to submit the manuscript for publication.
ave reported that they have no relationships relevant
14, 2018, accepted March 1, 2018.
TABLE 1 Clinical Characteristics of Patient Cohorts
ACM
(n ¼ 141)
DCM
(n ¼ 366)
Healthy Volunteer
(n ¼ 445)
Age at scan, yrs 53.2  10.0 56.0  13.6 40.8  13.5
Left ventricular ejection fraction (echo), % 26.5  9.3 — —
Left ventricular ejection fraction (CMR), % — 38.7  12.8 66.1  5.1
Left ventricular end-diastolic diameter (echo), mm 65.6  9.1 — —
Left ventricular end-diastolic volume (CMR), ml — 257.7  82.6 149.3  32.6
Males 138 (97.9) 255 (69.7) 201 (45.2)
Ethnicity (Caucasian) 141 (100.0) 366 (100.0) 445 (100.0)
Values are mean  SD or n (%).
ACM ¼ alcoholic cardiomyopathy; CMR ¼ cardiac magnetic resonance; DCM ¼ dilated cardiomyopathy.
J A C C V O L . 7 1 , N O . 2 0 , 2 0 1 8 Ware et al.
M A Y 2 2 , 2 0 1 8 : 2 2 9 3 – 3 0 2 Genetics and Alcohol Cardiotoxicity
2295ACM CASES. A total of 141 unrelated patients with
ACM (Table 1) were recruited for this study from 6
Spanish hospitals. ACM was defined as DCM with a
history of prolonged and heavy alcohol consumption:
that is, a self-reported history of alcohol intake of
>80 g/day over a period of at least 5 years (2,3,6),
with excess intake continuing up to no <3 months
before initial diagnosis of ACM, in combination with
DCM defined by established criteria of left ventricular
dilation and reduced ejection fraction in the absence
of coronary artery disease (invasive or computed to-
mography angiographic evidence of >50% stenosis in
any major epicardial coronary artery, or previous
percutaneous coronary intervention or coronary ar-
tery bypass grafting) or abnormal loading conditions
(uncontrolled hypertension or significant primary
valvular disease). Outcome information was collected
until last available follow-up, or at death or trans-
plantation, and follow-up time was truncated at 12
years. Although a specific program for alcohol
discontinuation was not provided, complete absti-
nence from alcohol was recommended to all ACM
patients. Endpoints were pre-specified as: 1) death or
cardiac transplantation; and 2) recovery defined as an
absolute increase in left ventricular ejection fraction
(LVEF) $10% to a final value of $40% (16). Survival
analyses measured time from diagnosis (first assess-
ment in heart failure clinic) to first event. Although
genotype, which defined groups for comparison in
survival analysis, was ascertained retrospectively,
both clinical care and outcome adjudication were
blinded to genotype.
DCM CASES. A total of 716 consecutive patients with
DCM confirmed by late gadolinium enhancement car-
diac magnetic resonance were prospectively enrolled
in the Royal Brompton Hospital Cardiovascular
Research Centre Biobank between 2009 and 2015 as
previously described (17). DCM was diagnosed based
on established criteria of left ventricular dilation and
reduced ejection fraction with reference to age- and
sex-adjusted nomograms (18) in the absence of known
coronary artery disease (defined as presence of sub-
endocardial late gadolinium enhancement suggestive
of previous myocardial infarction or >50% stenosis in
any major epicardial coronary artery or previous
percutaneous coronary intervention or coronary ar-
tery bypass grafting) or abnormal loading conditions
as for ACM. The complete 716 DCM cohort was evalu-
ated for phenotypic correlates of TTN genotype and
alcohol exposure (described in the following text), and
a subset of 366 unrelated cases that were matched
both technically and by ethnicity with the ACM cohort
were used for comparative genetic analysis.HEALTHY VOLUNTEERS. A total of 445 healthy
volunteers free from self-reported cardiovascular
disease or a family history of disease were recruited
prospectively via advertisement to the U.K. Digital
Heart Project at the MRC-LMS, Imperial College
London (15). All participants underwent clinical
assessment, including cardiac magnetic resonance, to
confirm the absence of cardiac disease.
NEXT-GENERATION SEQUENCING AND VARIANT
ANALYSIS. See the Online Methods in the Online
Appendix for full details on sequencing, variant
filtering, and annotation. In brief, sequencing was
carried out using the Illumina TruSight Cardio
Sequencing kit (San Diego, California) (19) or a
custom Agilent SureSelect XT target capture (Santa
Clara, California) with similar content and run on
Illumina platforms or Life Technologies 5500XL
(Waltham, Massachusetts). Rare (ExAC filtering allele
frequency [20] <8.4  105) protein-altering variants
were identified in genes and variant classes proven
to be robustly associated with DCM (Online Table 1).
In the case of titin, analysis was further restricted to
truncating variants in exons constitutively expressed
in the heart as described (12). Although the Illumina
TruSight Cardio sequencing kit captures 61 genes
purportedly associated with DCM (full gene list and
variants detected are given in Online Table 7), we
decided to be conservative and pre-specified a
focused analysis on 9 genes with the most robust
evidence of disease association (TTN, DSP, MYH7,
LMNA, TTNT2, TCAP, SNC5A, BAG3, and TNNC1) and
compared the prevalence of rare protein-altering
variants in subjects who were matched both
technically (TruSight Cardio panel and NextSeq
platform [both Illumina]) and by ethnicity (self-
reported Caucasian, confirmed by PCA analysis
[see Online Methods in the Online Appendix]). The
9 genes assessed are those with a demonstrated
excess of rare variation in DCM clinical cohorts over
FIGURE 1 Survival Analysis of ACM Cases According to Genotype
1.00
0.75
0.50
Ev
en
t-
Fr
ee
 S
ur
vi
va
l
0.25
0.00
0
0
122
14
94
10
67
9
48
9
30
6
18
5
14
3
p = 0.7
2 4
Number at Risk by Time
St
ra
ta
6
Years
8 10 12
2 4 6
Years
8 10 12
Strata Genotype Negative TTNtv Positive
Survival curves show freedom from composite primary endpoint (all-cause mortality or cardiac transplant) between ACM cases stratified by
genetic status: TTNtv positive (cases with a truncating variant in titin) or TTNtv negative. Event-free survival is measured from time of
diagnosis. There is no significant difference between groups. Curves are compared using the log-rank test. ACM ¼ alcoholic cardiomyopathy;
TTNtv ¼ titin truncating variant.
Ware et al. J A C C V O L . 7 1 , N O . 2 0 , 2 0 1 8
Genetics and Alcohol Cardiotoxicity M A Y 2 2 , 2 0 1 8 : 2 2 9 3 – 3 0 2
2296ExAC reference samples, for either truncating or
nontruncating variants (13,21).
EVALUATION OF ALCOHOL AS A PHENOTYPIC
MODIFIER IN DCM. We investigated the effect of
alcohol consumption on phenotype in DCM patients
using self-reported weekly consumption together
with a review of hospital and primary care medical
records for a history of alcohol excess prior to study
recruitment. No patients had a history of prolonged
heavy alcohol consumption for a diagnosis of ACM.
“Excess alcohol consumption” in DCM was defined as
a binary variable indicating a history of consumption
>21 U/week for men and >14 U/week for women (1 U
of alcohol ¼ 10 ml or 8 g of pure alcohol, an amount
the average adult metabolizes in 1 h) (22), levels
representing the “sensible limits” for alcohol con-
sumption from U.K. consensus medical advice (23)
from 1987 to 2016.
Univariable linear regression was performed to
identify variables predictive of LVEF at study
recruitment. LVEF was measured while blind to
genotype. Variables with p < 0.10 in univariableanalysis were included in a multivariable model,
which was then optimized by reverse stepwise se-
lection until only significant variables were
included. The pre-specified main analysis assessed
the significance of an interaction term between
TTNtv and “excess alcohol consumption” added to
this optimized multivariable model predicting
LVEF, to determine whether TTNtv and alcohol
consumption in combination have any additional
effect beyond the effects of TTNtv and alcohol
separately. A p value #0.05 was considered statis-
tically significant.
Statistical analyses were conducted in the R envi-
ronment, version 3.0 (R Foundation for Statistical
Computing, Vienna, Austria). All data and code
required to reproduce these analyses are available
online (24).
RESULTS
GENETIC CONTRIBUTION TO ACM. To investigate
the potential genetic contribution of cardiomyopathy
TABLE 2 Burden Analysis of Rare, Protein-Altering Variants in DCM-Related Genes Between Cohorts
ACM
(n ¼ 141)
DCM
(n ¼ 366)
Healthy Volunteer
(n ¼ 445)
*ACM vs.
DCM
*ACM vs. Healthy
Volunteer
*DCM vs. Healthy
Volunteer
All genes 19 (13.5) (7.8%–19.1%) 71 (19.4) (15.3%–23.4%) 13 (2.9) (1.4%–4.5%) 0.12 1.2  105 5.4  1015
TTNtv 14 (9.9) (5.0%–14.9%) 44 (12.0) (8.7%–15.4%) 3 (0.7) (0.0%–1.4%) 0.64 4.4  107 6.4  1012
Genes other than TTN 6 (4.3) (0.9%–7.6%) 28 (7.7) (4.9%–10.4%) 10 (2.2) (0.9%–3.6%) 0.23 0.23 0.00035
Values are n (%) (95% confidence interval). The number of individual cases with a rare protein-altering variant is shown. TTN variants are TTNtv only; other variants are as
described in Online Table 1. In ACM, 1 case had both a TTNtv and LMNA variant. In DCM, 1 case had both a TTNtv and a BAG3 variant. *Unadjusted p value (Fisher exact test).
TTNtv ¼ titin truncating variant; other abbreviations as in Table 1.
J A C C V O L . 7 1 , N O . 2 0 , 2 0 1 8 Ware et al.
M A Y 2 2 , 2 0 1 8 : 2 2 9 3 – 3 0 2 Genetics and Alcohol Cardiotoxicity
2297genes to ACM, we examined cases for the presence of
rare protein-altering variants in 9 genes known to
cause DCM that were selected according to their
previously reported excess of rare variants in DCM
compared with control subjects (21). The frequency of
variants was then compared between ACM cases,
technically and ethnically matched DCM cases, and
healthy volunteers (n ¼ 141, n ¼ 366, and n ¼ 445,
respectively) (cohort characteristics are shown in
Table 1). We identified 20 distinct variants in 19 ACM
cases involving 4 different genes (Table 2, Online
Tables 2 and 3A and 3B). The prevalence of variants
in ACM was significantly higher than in healthy vol-
unteers (13.5% of ACM cases carry at least 1 variant vs.
2.9% of healthy volunteers; p ¼ 0.000012), but not
statistically different from the prevalence in the DCM
cohort (19.4%; p ¼ 0.12). The rate in healthy volun-
teers was as expected for the general population
(Online Table 2). TTNtv accounted for the majority of
variants detected in ACM cases (9.9%) and were
found with a frequency similar to that seen in DCM
(12.0%; p ¼ 0.64), and significantly higher than inTABLE 3 Characteristics of ACM Cases With and Without Titin Trunca
TT
(n ¼
Alcohol, g/day 139.0
Age at initial clinical assessment, yrs 49.4
Initial left ventricular ejection fraction, % 25.1
Initial left ventricular end-diastolic diameter, mm 63.2
Male 13 (
Atrial fibrillation 5 (
Family history of cardiomyopathy 6 (4
Family history of sudden cardiac death 1 (
Outcomes
Mean follow up period, yrs 8.3
Death or transplant 3 (
Stable with recovery of left ventricular ejection fraction 7 (5
Stable without recovery of left ventricular ejection fraction 4 (2
Values are mean  SD, n (%), or n. Age, left ventricular ejection fraction, left ventricular e
*Unadjusted p values of TTNtv vs. genotype negative: Mann-Whitney U test for continu
test for survival (death or transplant).
TTNtv ¼ titin truncating variant.control subjects (0.7%; p ¼ 4.4  107). In line with
studies in DCM (10–12,15), TTNtv found in ACM were
in exons constitutively expressed in the heart and
distributed across the gene (Online Figure 1) with 13
of 14 being novel (i.e., absent from previous DCM
cases, healthy volunteers, and ExAC).
We identified 6 ACM cases with rare, protein-
altering variants in other DCM genes: 1 carrying a
BAG3 truncating variant previously reported in DCM
(25) and classified as pathogenic for DCM under cur-
rent variant interpretation guidelines from the
American College of Medical Genetics and Genomics
and the Association for Molecular Pathology (Online
Tables 3A and 3B) (26), 1 carrying a novel BAG3
missense variant, 1 carrying both a TTNtv and a novel
LMNA missense mutation, and 3 cases each carrying
different MYH7 variants.
There were no detectable differences in phenotype
or demographics between ACM cases with and
without TTNtv (Table 3, Online Table 4), with the
notable exception of family history of cardiomyopa-
thy. On follow-up (overall mean follow-up ¼ 5.9  5.2ting Variants
Ntv
14)
Genotype Negative
(n ¼ 122)
Other Variants
(n ¼ 5) *p Value
 68.7 136.0  50.1 122.0  34.6 0.85
 12.9 53.4  9.6 58.8  11.1 0.31
 10.7 26.5  9.1 30.4  10.5 0.35
 6.6 65.8  9.2 68.4  11.7 0.37
92.9) 120 (98.4) 5 (100.0) 0.28
35.7) 41 (33.6) 3 (60.0) 1.00
2.9) 9 (7.4) 1 (20.0) 0.0012
7.1) 12 (9.8) 0 (0.0) 1.00
14 120 5
 7.2 5.8  4.9 5.5  4.9 0.26
21.4) 19 (15.8) 3 (60.0) 0.96
0.0) 55 (45.8) 0 (0.0) 0.78
8.6) 46 (38.3) 2 (40.0) 0.57
nd-diastolic diameter, and atrial fibrillation taken at time of initial clinical assessment.
ous variables, Fisher exact test for categorical variables, and Cox proportional hazard
FIGURE 2 Alcohol and TTNtv Act in Combination, and
Together Are Associated With a Lower Baseline LVEF in
Patients With DCM
–15
TTNtv*Alcohol XS
TTNtv
Alcohol XS
Male Gender
LGE
Beta Blocker
A-Antag
–10
Effect on Baseline LVEF (%)
–5 0
Forest plot showing regression coefficient and 95% confidence
intervals from the multivariable linear regression model evalu-
ating the effects of TTNtv and excess alcohol consumption on
baseline LVEF. The effect on LVEF is shown as absolute differ-
ence in LVEF (% ¼ expressed as percentage of end-diastolic
volume) between groups. A-Antag ¼ aldosterone antagonist;
Alcohol XS ¼ excess alcohol consumption (binary variable
indicating consumption >21 U/week for men, >14 U/week
forwomen); LGE¼ late gadoliniumenhancement (indicative of
mid-wall fibrosis) on cardiovascular magnetic resonance;
LVEF ¼ left ventricular ejection fraction; TTNtv ¼ presence of
truncating variant in titin; TTNtv*Alcohol XS¼ interaction term
representing individuals with both a TTNtv and a history of
excess alcohol consumption.
Ware et al. J A C C V O L . 7 1 , N O . 2 0 , 2 0 1 8
Genetics and Alcohol Cardiotoxicity M A Y 2 2 , 2 0 1 8 : 2 2 9 3 – 3 0 2
2298years), TTNtv status did not predict outcomes after
reduction in alcohol intake and treatment for heart
failure, with approximately one-half of all ACM cases
showing LVEF recovery irrespective of TTNtv status
(Table 3), and no detectable difference in event-free
survival between the 2 groups (Figure 1).
ALCOHOL AS A PHENOTYPIC MODIFIER IN DCM.
Having established a genetic contribution to ACM in a
proportion of cases, primarily driven by TTNtv, we
investigated the interaction between TTNtv and
alcohol consumption in the context of DCM but in the
absence of prolonged and heavy alcohol consumption
meeting criteria for ACM. A total of 111 of 716 DCM
cases (15.5%) had a history of excess consumption
(above U.K. guidelines, see the Methods section).
These individuals were more likely to be male, and in
univariate analyses had modestly reduced LVEF(median: 37.0% vs. 41.0%; p ¼ 0.02) and right ven-
tricular ejection fraction (median: 48.0% vs. 54.0%;
p < 0.001) compared with DCM patients without a
history of excess alcohol consumption (Online
Table 5). A total of 83 DCM cases carried a TTNtv
(11.6%). The presence of a TTNtv alone did not predict
LVEF. In multivariable analysis accounting for co-
variate predictors of baseline LVEF, neither TTNtv
nor excess alcohol consumption were significant
predictors in isolation, but patients with both TTNtv
and excess alcohol consumption (n ¼ 13) had a sta-
tistically significant and biologically important
reduction in LVEF (Figure 2, Online Tables 6A and
6B), with an 8.7% absolute reduction (95% confi-
dence interval: 15.1 to 2.3; p ¼ 0.007) compared
with DCM with neither TTNtv nor excess alcohol
consumption.
DISCUSSION
This study demonstrates an important genetic pre-
disposition to ACM. We present a large series of ACM
patients genotyped for variants in 9 genes associated
with inherited DCM, and identified rare, protein-
altering variants in 19 of 141 ACM cases (13.5%), a
frequency significantly higher than that observed in
healthy volunteers (2.9%; p ¼ 0.000012) and popu-
lation controls (ExAC, 4.3%; p ¼ 0.0000059), but
similar to that seen in DCM cases (19.4%; p ¼ 0.12)
(Central Illustration). Our findings demonstrate that in
a proportion of ACM subjects, the disease has a ge-
netic etiology.
The data presented here indicate that patients with
alcohol-related cardiomyopathy should undergo a 3-
generation pedigree and should be considered for
familial evaluation, such as clinical screening and
genetic testing, to identify family members at risk for
developing DCM (in line with current practice for
idiopathic DCM).
An illustration of the utility of genetic management
in ACM is shown in Figure 3, where familial evaluation
identified several individuals with DCM, and molec-
ular genetic testing enabled informed genetic coun-
seling including reproductive advice. This reveals the
importance of recognizing genetic disease and famil-
ial assessment, although future work will be needed
to more fully understand the risk associated with ge-
netic variants found in the absence of overt familial
disease, to balance the costs and benefits associated
with genetic testing and clinical surveillance, and to
allow for fully informed genetic counseling.
We further identified a direct interaction between
TTNtv and alcohol consumption in the context of
CENTRAL ILLUSTRATION Alcohol Consumption and Genetic Background Act in Concert to
Determine Cardiac Phenotype
Prevalence of rare genetic variants
in cardiomyopathy genes:
LVEF according to titin truncating variant
(TTNtv) status and alcohol intake:
Alcoholic Cardiomyopathy
(ACM)
PHENOTYPE
Dilated Cardiomyopathy
(DCM)
ACM: 13.5%
Controls: 2.9%
no TTNtv or excess alcohol intake: 39.6 ± 12.2%
TTNtv only: 39.8 ± 13.2%
excess alcohol intake only: 37.8 ± 11.8%
TTNtv AND excess alcohol intake: 27.7 ± 12.7%PFisher = 0.000012
PMultivariate = 0.007
ALCOHOL + GENETICS
Ware, J.S. et al. J Am Coll Cardiol. 2018;71(20):2293–302.
ACM patients exhibit a higher prevalence of rare variants in DCM-associated genes than control subjects. In DCM patients, neither the
presence of a TTNtv nor excess alcohol consumption had a significant effect on baseline LVEF in isolation, but the combination was
associated with a significantly lower baseline LVEF. Values shown are absolute ejection fraction in each group. The p value is derived from
multivariate analysis. ACM ¼ alcoholic cardiomyopathy; DCM ¼ dilated cardiomyopathy; LVEF ¼ left ventricular ejection fraction;
TTNtv ¼ titin truncating variant.
J A C C V O L . 7 1 , N O . 2 0 , 2 0 1 8 Ware et al.
M A Y 2 2 , 2 0 1 8 : 2 2 9 3 – 3 0 2 Genetics and Alcohol Cardiotoxicity
2299typical DCM: cases with a TTNtv and excess alcohol
consumption have a markedly reduced LVEF
compared with those with low alcohol intake. Taken
together, these 2 lines of evidence both support a
model whereby alcohol and cardiac genotype interact,
contributing both to the development of ACM and to
disease severity in the context of DCM (Central
Illustration). Although we acknowledge that many
factors may contribute to the development of ACM, we
identified an illustrative family where alcohol abuse
and TTNtv were present in multiple relatives and
probably acted in conjunction to promote disease ex-
pressivity in certain family members. In this pedigree,
all 3 affected individuals both carried the TTNtv and
reported prolonged heavy alcohol consumption,
whereas 2 individuals who reported prolonged heavy
alcohol consumption without the TTNtv and 3 in-
dividuals with the TTNtv but without excess alcohol
consumption were all free from DCM (Figure 3).There is still much to understand. The molecular
mechanisms underlying ACM are not fully under-
stood, and this study only explores some of the ge-
netic factors that may influence susceptibility to
cardiomyopathy on exposure to alcohol. Although
there is strong evidence for an interaction between
alcohol and TTNtv, there is much more to learn about
the mechanisms underlying the variable penetrance
of TTNtv. In some families with DCM, TTNtv appear
highly penetrant and sufficient to cause disease in
isolation, but TTNtv are also seen in approximately
1% of the general population (12), a level well above
the prevalence of DCM and suggesting that other ge-
netic or environmental factors contribute to the car-
diomyopathic process (27).
The overall effect of alcohol on the occurrence of
DCM is also difficult to assess, but previous reports
have suggested that it may be involved in as many as
47% of cases (6), and a recent population-based study
FIGURE 3 Family Pedigrees Illustrating Coexistence of ACM and DCM and the Combined Effect of Excessive Alcohol Consumption and Genetic Background
I:1
II:1
III:1
VHD
SCA
DCM
MI?
III:2
IV:1
II:1
FAMILY 1016
FAMILY 978
TTN c.64453C>T; p.R21485X
TTN c.66816_66817delAT;
p.Leu22273Thrfs24
II:4
III:1
31y
III:2
29y
III:3
27y
III:4
17y
III:5
28y
III:6
27y
III:7
20y
III:8
31y
III:9
20y
III:10
11y
III:11
8y
II:5 II:6 II:8 II:9
51y 48y 46y
ACM
II:10 II:11
44y
ACM
II:7II:2
57y
ACM
II:3
53y
IV:2 IV:3 IV:4 IV:5 IV:6
44 y
Cancer
I:1
IV:7
O
IV:8
79y
I:2
+
+ –
+ + +––
IV:9 IV:10 IV:11 IV:12 IV:13 IV:14 IV:15
III:3 III:4 III:5 III:6
58y
N N
N N
N
N
N N NN
N N
N N
N N
56y 50y
DCM
46y 44y 42y
DCM
54y
ACM
CTx
III:7 III:8 III:9 III:10 III:11 III:12 III:13 III:14 III:15 III:16 III:17 III:18
II:2 II:3 II:4 II:5 II:6 II:7
+
+ +
+ +
+ –
+ –
+ +– *
*
* * * *
?
+
DCM
II:8 II:9
78 y
DCM
II:10
I:2
(Top) Family 978: coexistence of ACM and DCM. The proband (arrow) was diagnosed with ACM and underwent cardiac transplantation. When genetic and clinical
familial evaluation was performed, multiple individuals without excessive alcohol consumption were diagnosed with DCM and found to carry TTN truncating variants.
(Bottom) Family 1016: combined effect of excessive alcohol consumption and genetic background. The proband (arrow) was diagnosed with ACM at age 44 years and
was identified as carrying a TTNtv variant (TTN c.64453C>T; p.R21485X). One brother and 1 sister with prolonged heavy alcohol consumption (red asterisk) and TTNtv
also show ACM. Two family members with TTNtv but no regular alcohol intake, and 2 individuals with prolonged heavy alcohol consumption but without TTNtv, did
not show cardiac involvement. Standard pedigree notation is used: squares and circles indicate male and female subjects, respectively, a strike-through indicates a
deceased individual, an arrow indicates the proband in each family, and filled symbols indicate affected individuals with ACM or DCM. Symbols containing an N
represent individuals confirmed as unaffected. D/L symbols indicate genetic evaluation: D indicates carry TTNtv; L are noncarriers, oD are obligate carriers.
Red asterisks indicate cases with documented prolonged heavy alcohol consumption. CTx ¼ cardiac transplant; DCM ¼ dilated cardiomyopathy; MI ¼ myocardial
infarction; SCA ¼ sudden cardiac arrest; VHD ¼ valvular heart disease; other abbreviations as in Figure 1.
Ware et al. J A C C V O L . 7 1 , N O . 2 0 , 2 0 1 8
Genetics and Alcohol Cardiotoxicity M A Y 2 2 , 2 0 1 8 : 2 2 9 3 – 3 0 2
2300of >1.9 million U.K. individuals showed that 8.4%
recorded drinking above the recommended safe
levels (28). If this accurately reflects the proportion of
the population with above-recommended alcoholintake, then we see a significantly higher exposure in
our nonalcoholic DCM population (111 of 716 ¼ 15.5%,
pbinomial ¼ 5  1010). Together, these data suggest
that alcohol alone, as well as in combination with
J A C C V O L . 7 1 , N O . 2 0 , 2 0 1 8 Ware et al.
M A Y 2 2 , 2 0 1 8 : 2 2 9 3 – 3 0 2 Genetics and Alcohol Cardiotoxicity
2301genetic factors, may account for a substantial pro-
portion of disease risk.
Additional environmental factors that may act in
concert with TTNtv include viral myocarditis (29),
nutritional deficiencies (30), recreational drug use
(31), and certain drugs (32). Our data therefore also
have wider potential implications both for lifestyle
choices and for exploring the potential interaction of
genetics with other environmental factors.
ACM has a poor prognosis, although somewhat
better than DCM overall (2,3). In the ACM cohort
studied here, of the 139 cases with outcome data, 62
(44.6%) showed functional recovery following heart
failure therapy and reduction in alcohol, 52 (37.4%)
remained stable but without functional recovery, and
25 (17.9%) died or received a cardiac transplant, in
agreement with recent studies (3,16). We saw no dif-
ference between TTNtv and non-TTNtv cases with
respect to outcome, with equivalent proportions
showing improved cardiac function (50.0% vs. 45.8%,
respectively) (Table 3), indicating that the presence of
a TTNtv does not of itself preclude recovery. Func-
tional recovery in DCM resulting from TTNtv has been
previously reported both in severe end-stage failure
requiring LVAD support (33) and in milder cases
following medical therapy (34). Likewise, we
observed no difference in survival analysis (freedom
from death or cardiac transplantation) between
TTNtvþ and TTNtv groups.
STUDY LIMITATIONS. First, in the absence of a cohort
with prolonged and heavy alcohol consumption but
no cardiomyopathy, our comparison of ACM and
healthy volunteers cannot formally exclude the pos-
sibility that TTNtv are associated with increased
alcohol consumption, rather than the development of
ACM on exposure to alcohol. However, this would
seem highly unlikely and cannot explain the observed
interaction between excess alcohol consumption and
TTNtv as predictors of severity in an independent
DCM cohort. Second, one might postulate that the
individuals with coincident ACM and TTNtv simply
represent conventional familial DCM: because pro-
longed heavy alcohol consumption is not uncommon
in the population, a proportion of DCM cases will be
exposed; thus, the TTNtv could be the causative
driver, and the alcohol consumption a coincidental
bystander. However, the positive cardiac response on
reduction or cessation of alcohol points to anetiological role of alcohol in the disease process, and
the observed synergistic interaction between genetic
predisposition and environmental toxin in the DCM
cohort once again points to a biological interaction.
Third, although the association between aggre-
gated rare variation in this gene set and ACM can be
robustly interpreted as demonstrating an etiological
role, the interpretation of specific variants in indi-
vidual patients often remains uncertain. Improve-
ment in clinical variant interpretation would
substantially improve the utility of genetic testing in
cardiomyopathies more widely. We also restricted our
analysis to robustly validated DCM genes with a
published excess of rare variants in DCM compared
with control subjects. We acknowledge that rare
variants in other genes that might have a role in DCM
may make a further contribution to a genetic predis-
position to ACM.
Finally, self-reported alcohol consumption lacks
precision and is likely under-reported, which,
together with modest cohort size, limits our power to
detect modest effect sizes on phenotype and
outcome, to evaluate the contribution of genes that
are more rarely variant, and to fully dissect the in-
teractions between genetic and environmental
influences.
CONCLUSIONS
We have shown that TTNtv represent an important
genetic predisposition to ACM, and that the combi-
nation of TTNtv and excess alcohol consumption is
associated with worse LVEF in DCM patients. These
findings support a model whereby alcohol and other
environmental factors interact with genotype to
determine the cardiac phenotype. Furthermore,
based on our findings, familial evaluation and genetic
testing should be considered in patients presenting
with ACM.
ADDRESS FOR CORRESPONDENCE: Dr. Pablo
Garcia-Pavia, Department of Cardiology, Hospital
Universitario Puerta de Hierro, Manuel de Falla 2,
Majadahonda, Madrid 28222, Spain. E-mail:
pablogpavia@yahoo.es. OR Dr. Paul J. Barton, Car-
diovascular Research Centre, Royal Brompton and
Harefield NHS Foundation Trust and Imperial College
London, London SW3 6NP, United Kingdom. E-mail:
p.barton@imperial.ac.uk.
PERSPECTIVES
COMPETENCY IN MEDICAL KNOWLEDGE: Variants
in DCM-associated genes are more frequent in patients
with ACM than in the general population, and patients
with DCM and TTNtv who drink alcohol excessively are
more prone to decline in LVEF than those who drink less
or lack these genetic variants.
TRANSLATIONAL OUTLOOK: Further studies are
needed to understand how family history and genetic
testing can be used to identify patients at risk of devel-
oping ACM, and effectively employed in counseling and
other psychosocial interventions to reduce the incidence
of this form of DCM.
Ware et al. J A C C V O L . 7 1 , N O . 2 0 , 2 0 1 8
Genetics and Alcohol Cardiotoxicity M A Y 2 2 , 2 0 1 8 : 2 2 9 3 – 3 0 2
2302RE F E RENCE S1. Urbano-Marquez A, Estruch R, Navarro-Lopez F,
Grau JM, Mont L, Rubin E. The effects of alco-
holism on skeletal and cardiac muscle. N Engl J
Med 1989;320:409–15.
2. Fernandez-Sola J. Cardiovascular risks and
benefits of moderate and heavy alcohol con-
sumption. Nat Rev Cardiol 2015;12:576–87.
3. Guzzo-Merello G, Segovia J, Dominguez F, et al.
Natural history and prognostic factors in alcoholic
cardiomyopathy. J Am Coll Cardiol HF 2015;3:
78–86.
4. George A, Figueredo VM. Alcoholic cardiomy-
opathy: a review. J Card Fail 2011;17:844–9.
5. Xi B, Veeranki SP, Zhao M, Ma C, Yan Y, Mi J.
Relationship of alcohol consumption to all-
cause, cardiovascular, and cancer-related mor-
tality in U.S. adults. J Am Coll Cardiol 2017;70:
913–22.
6. Guzzo-Merello G, Cobo-Marcos M, Gallego-
Delgado M, Garcia-Pavia P. Alcoholic cardiomy-
opathy. World J Cardiol 2014;6:771–81.
7. Fernandez-Sola J, Nicolas JM, Oriola J, et al.
Angiotensin-converting enzymegene polymorphism
is associated with vulnerability to alcoholic cardio-
myopathy. Ann Intern Med 2002;137:321–6.
8. Teragaki M, Takeuchi K, Toda I, et al. Point
mutations in mitochondrial DNA of patients with
alcoholic cardiomyopathy. Heart Vessels 2000;15:
172–5.
9. Guzzo-Merello G, Dominguez F, Gonzalez-
Lopez E, et al. Malignant ventricular arrhythmias
in alcoholic cardiomyopathy. Int J Cardiol 2015;
199:99–105.
10. Hershberger RE, Hedges DJ, Morales A.
Dilated cardiomyopathy: the complexity of a
diverse genetic architecture. Nat Rev Cardiol 2013;
10:531–47.
11. Herman DS, Lam L, Taylor MR, et al. Trunca-
tions of titin causing dilated cardiomyopathy.
N Engl J Med 2012;366:619–28.
12. Roberts AM, Ware JS, Herman DS, et al. Inte-
grated allelic, transcriptional, and phenomic
dissection of the cardiac effects of titin trunca-
tions in health and disease. Sci Transl Med 2015;7:
270ra6.
13. Tayal U, Prasad S, Cook SA. Genetics and ge-
nomics of dilated cardiomyopathy and systolic
heart failure. Genome Med 2017;9:20.14. Ware JS, Li J, Mazaika E, et al. Shared ge-
netic predisposition in peripartum and dilated
cardiomyopathies. N Engl J Med 2016;374:
233–41.
15. Schafer S, de Marvao A, Adami E, et al. Titin-
truncating variants affect heart function in disease
cohorts and the general population. Nat Genet
2017;49:46–53.
16. Amor-Salamanca A, Guzzo-Merello G, Gonzá-
lez-López E, et al. Prognostic impact and pre-
dictors of ejection fraction recovery in patients
with alcoholic cardiomyopathy. Rev Esp Cardiol
2018 Feb 8 [E-pub ahead of print].
17. Tayal U, Newsome S, Buchan R, et al. Inte-
grated analysis of the phenotype and clinical
outcomes of Titin cardiomyopathy. J Am Coll
Cardiol 2017;70:2264–74.
18. Maceira AM, Prasad SK, Khan M, Pennell DJ.
Normalized left ventricular systolic and diastolic
function by steady state free precession cardio-
vascular magnetic resonance. J Cardiovasc Magn
Reson 2006;8:417–26.
19. Pua CJ, Bhalshankar J, Miao K, et al. Devel-
opment of a comprehensive sequencing assay for
inherited cardiac condition genes. J Cardiovasc
Transl Res 2016;9:3–11.
20. Whiffin N, Minikel E, Walsh R, et al. Using
high-resolution variant frequencies to empower
clinical genome interpretation. Genet Med 2017;
19:1151–8.
21. Walsh R, Thomson KL, Ware JS, et al. Reas-
sessment of Mendelian gene pathogenicity using
7,855 cardiomyopathy cases and 60,706 refer-
ence samples. Genet Med 2017;19:192–203.
22. Committee HoCSaT. Alcohol guidelines: Elev-
enth Report of Session 2010-12: UK Parliament,
2012.
23. Royal Colleges of Physicians, Psychiatrists and
General Practitioners. Alcohol and the Heart in
Perspective: Sensible Limits Reaffirmed. London:
Royal Colleges of Physicians, 1995.
24. ImperialCardioGenetics/ACM. Available at:
https://github.com/ImperialCardioGenetics/ACM.
Accessed April 15, 2018.
25. Chami N, Tadros R, Lemarbre F, et al.
Nonsense mutations in BAG3 are associated with
early-onset dilated cardiomyopathy in French Ca-
nadians. Can J Cardiol 2014;30:1655–61.26. Richards S, Aziz N, Bale S, et al. Standards and
guidelines for the interpretation of sequence
variants: a joint consensus recommendation of the
American College of Medical Genetics and Geno-
mics and the Association for Molecular Pathology.
Genet Med 2015;17:405–24.
27. Ware JS, Cook SA. Role of titin in cardiomy-
opathy: from DNA variants to patient stratifica-
tion. Nat Rev Cardiol 2018;15:241–52.
28. Bell S, Daskalopoulou M, Rapsomaniki E, et al.
Association between clinically recorded alcohol
consumption and initial presentation of 12 cardio-
vascular diseases: population based cohort study
using linked health records. BMJ 2017;356:j909.
29. Belkaya S, Kontorovich AR, Byun M, et al.
Autosomal recessive cardiomyopathy presenting
as acute myocarditis. J Am Coll Cardiol 2017;69:
1653–65.
30. Marinescu V, McCullough PA. Nutritional and
micronutrient determinants of idiopathic dilated
cardiomyopathy: diagnostic and therapeutic im-
plications. Expert Rev Cardiovasc Ther 2011;9:
1161–70.
31. Frustaci A, Russo MA, Morgante E, et al.
Oxidative myocardial damage in human cocaine-
related cardiomyopathy. Eur J Heart Fail 2015;17:
283–90.
32. Chang HM, Okwuosa TM, Scarabelli T, et al.
Cardiovascular complications of cancer therapy:
best practices in diagnosis, prevention, and man-
agement: part 2. J Am Coll Cardiol 2017;70:
2552–65.
33. Felkin LE, Walsh R, Ware JS, et al. Recovery of
cardiac function in cardiomyopathy caused by titin
truncation. JAMA Cardiol 2016;1:234–5.
34. Jansweijer JA, Nieuwhof K, Russo F, et al.
Truncating titin mutations are associated with a
mild and treatable form of dilated cardiomyopa-
thy. Eur J Heart Fail 2017;19:512–21.KEY WORDS alcohol, dilated
cardiomyopathy, genetics, titin, variantAPPENDIX For an expanded Methods section
as well as a supplemental figure and tables,
please see the online version of this paper.