Hepatitis C drugs and vascular event direct acting antiviral

-

1. Vol.:(0123456789)1 3 Internal and Emergency Medicine
https://doi.org/10.1007/s11739-018-1828-8
CE - SYSTEMATIC REVIEW
Direct‑acting antiviral drugs for chronic hepatitis C and risk of major
vascular events: a systematic review
Eleonora Tamborini Permunian
1
 · Lorenzo Gervaso
2
 · Victor Gerdes
3
 · Lorenzo Moja
4
 · Luigina Guasti
1,5
 ·
Alessandro Squizzato
1
Received: 28 October 2017 / Accepted: 8 March 2018
© SIMI 2018
Abstract
Direct-acting antiviral drugs (DAAs) were recently approved for treating hepatitis C virus-related chronic hepatitis. As
advanced chronic liver disease may predispose patients to thrombotic events, it is still uncertain whether DAAs may in-uence
the actual risk of major arterial and venous thrombotic events. We performed a systematic review to assess the incidence of
major vascular events in patients receiving DAAs for HCV chronic hepatitis during phase-III randomized controlled trials
(RCTs). Two reviewers identi ed studies through Pubmed database until October 2015. Reporting and incidence of any
vascular events were compared with reporting and incidence of major bleeding, anemia (a prespeci ed safety outcome) and
headache (a common non-prespeci ed safety outcome). 33 RCTs, encompassing 14,764 patients, were included. Only 13
(39%) and 4 (12%) RCTs provide data on any arterial or venous events, respectively. Occurrence of anemia and headache
is reported in all studies. Crude unweighted rate of major arterial events is 0.16% (95% CI 0.10–0.24) of the total included
population and 0.47% in those 13 RCTs reporting data. Crude unweighted rate of major venous events is 0.03% of the total
included population (95% CI 0.01–0.08) and 0.22% in those four RCTs reporting data. Crude unweighted rate of major
bleeding is 0.07% (95% CI 0.03–0.1). Incidence of thrombotic events in HCV patients receiving DAAs may be low, but an
incorrect estimation cannot be excluded.
Keywords Chronic hepatitis? Direct-acting antiviral drugs? Myocardial infarction? Underreporting? Venous
thromboembolism
Introduction
Chronic liver disease may predispose patients to thrombotic
events [1]. Portal hypertension and subsequent venous stasis,
a common condition of advanced liver disease, is consid-
ered the main predisposing factor to portal vein thrombosis.
Recently, other important factors have been recognized: the
prothrombotic state of chronic liver disease is also related to
a delicate haemostatic balance between reduced production
of procoagulant factors and platelets and decreased levels of
anticoagulants (such as protein C and antithrombin) in last
stages [2]. Therefore, patients with chronic liver disease may
have an increased risk of venous thromboembolism (VTE),
i.e. pulmonary embolism (PE) and deep vein thrombosis
(DVT) [1].
Numerous randomized controlled trials (RCTs) have been
published on the e cacy and safety of direct-acting antiviral
drugs (DAAs) for treating hepatitis C virus (HCV)-related
chronic hepatitis [3, 4]. A direct mechanism of these drugs
Electronic supplementary material The online version of this
article (https ://doi.org/10.1007/s1173 9-018-1828-8) contains
supplementary material, which is available to authorized users.
* Luigina Guasti
[email protected]
1
Department ofMedicine andSurgery, Research Centre
onThromboembolic Disorders andAntithrombotic
Therapies, University ofInsubria, Varese, Italy
2
Oncology Unit, IRCCS Fondazione Salvatore Maugeri,
University ofPavia, Pavia, Italy
3
Department ofInternal Medicine, MC Slotervaart,
Amsterdam, TheNetherlands
4
Unit ofClinical Epidemiology, I.R.C.C.S. Orthopedic
Institute Galeazzi, Milan, Italy
5
U.O. Medicina Interna 1, ASST Settelaghi, Viale Borri, 57,
21100Varese, Italy

2. Internal and Emergency Medicine1 3
on coagulation has not been demonstrated, and probably not
been investigated either. However, by modifying the in am-
mation cascade, they might indirectly play a role on throm-
bus formation and the possible regression of liver damage
related to their use might lead to a reduction of thrombotic
risk. As the number of patients included in DAAs’ RCTs is
very high, we performed a systematic review of phase III
RCTs comparing DAAs with standard treatment for HCV-
related chronic hepatitis, with the aim of determining the
incidence of major vascular events, i.e. myocardial infarction
(MI), ischemic stroke, pulmonary embolism and splanchnic
vein thrombosis.
Materials and methods
This systematic review was performed according to Pre-
ferred Reporting Items for Systematic reviews and Meta-
Analyses (PRISMA) guidelines [5 ].
Search strategy
Using PubMed database (up to 13th October 2015), we iden-
tied all published phase III RCTs that assessed e cacy
and safety of DAAs in patients with HCV chronic infection.
The search strategy was developed without any language
restriction. We used the following textwords: ‘direct-acting
antivirals’, ‘DAAs’, ‘daclatasvir’, ‘ledipasvir’, ‘ombitas-
vir’, ‘elbasvir’, ‘telaprevir’, ‘boceprevir’, ‘paritaprevir’,
‘simeprevir’, ‘faldaprevir’, ‘asunaprevir’, ‘vaniprevir’,
‘danoprevir’, ‘vedroprevir’, ‘sofosbuvir’, ‘mericitabine’,
‘valopecitabine’, ‘balapiravir’, ‘dasabuvir’, ‘beclabuvir’,
‘setrobuvir’, ‘grazoprevir’, ‘sovaprevir’, ‘narlaprevir’, ‘del-
deprevir?, ?neceprevir?, deleobuvir?, tegobuvir?, ?libuvir?;
with the following limits: ‘randomized controlled trial’,
‘humans’. We supplemented our search by manually review-
ing the reference list of all retrieved articles.
Study selection
Two reviewers (ETP, LG) performed study selection inde-
pendently, with disagreements solved through discussion
or by the opinion of a third reviewer (AS), if necessary.
Studies were considered potentially eligible for this sys-
tematic review if they met the following criteria: (i) studies
were phase III RCTs, (ii) e cacy and safety of DAAs was
assessed, (iii) enrolled patients had a chronic HCV infection.
Data extraction
Two reviewers independently extracted data on study (year
of publication, design), population characteristics (number
of patients, mean age, gender), type of intervention and
comparison, duration of treatment, vascular and bleeding
events, and follow-up duration. Any arterial ischemic and
venous thromboembolic events were collected. In particu-
lar, we planned to extract the following major vascular
events: MI, ischemic stroke, pulmonary embolism and
splanchnic vein thrombosis. We also recorded deaths from
unknown causes as they could have been related to a major
vascular event. Additionally, to verify whether major vas-
cular events are adequately reported, we collected data on
reporting of anemia, a prespecied safety outcome, and
headache, a common non-prespecied adverse event. We
selected anemia as safety outcome because it is a well-
known adverse event of ribavirin treatment [6 ]. Moreover,
anemia is relatively uncommon in the general population,
and its diagnosis is simple and objective. On the other
hand, headache was selected because it is biologically
unrelated to DAAs and standard HCV hepatitis treatment,
and it is actually quite common in the general population.
Finally, we collected data on major bleeding, as it is a
possible complication of HCV chronic hepatitis. Major
bleeding was dened according to International Society of
Thrombosis and Haemostasis (ISTH) criteria for non-sur-
gical patients [7 ]. When ISTH criteria were not applicable
because of lack of information about the bleeding event,
all events classied as ?life-threatening? or ?serious? were
reported as major. All outcomes occurring during active
treatment and during follow-up period were collected.
Risk of bias assessment
The same two reviewers assessed the methodological qual-
ity of selected studies on the basis of Cochrane criteria
[8]. The following items were evaluated: random sequence
generation (selection bias), allocation concealment (selec-
tion bias), blinding of participants and personnel (perfor-
mance bias), blinding of outcome assessment (detection
bias), incomplete outcome data (attrition bias). Each of
the items was judged as at ‘low’, ‘high’, or ‘unclear’ risk
of bias (where ‘low’ indicates that the study is less open
to bias). If the available information in published stud-
ies, protocols and supplementary data was partial or inad-
equate to assess the risk of bias, we assessed the risk of
bias as ‘unclear’.
We, therefore, performed a separate analysis to assess
risk of bias specic for adverse e ects both on conduct
and on reporting as proposed by the Cochrane Adverse
E ects Methods Group [9 ]. We evaluated the following
items, answering ‘yes’ or ‘no’ if present or not: method of
adverse event assessment description, exclusion of patients
from the adverse e ects analysis and presence of numeri-
cal data by intervention group (Table1).

3. Internal and Emergency Medicine 1 3
Table1   Risk of bias for adverse e ects [9]
Study Method of adverse event assessment description Exclusion of patients from
the adverse e ects analysis
Presence of numerical
data by intervention
group
ADVANCE [10] Yes
At all study visits (clinical examination and laboratory test-
ing), and spontaneously if reported between visits
No Yes
ALLY-2 [11] Yes
At all study visits (clinical examination and laboratory test-
ing), and spontaneously if reported between visits
No Yes
ASTRAL-1 [12] Yes
At all study visits (clinical examination and laboratory test-
ing), and spontaneously if reported between visits
No Yes
ASTRAL-2 [13] Yes
At all study visits (clinical examination and laboratory test-
ing), and spontaneously if reported between visits and after
the end of follow-up
No Yes
ASTRAL-3 [13] Yes
At all study visits (clinical examination and laboratory test-
ing), and spontaneously if reported between visits and after
the end of follow-up
No Yes
ASTRAL-4 [14] Yes
At all study visits (clinical examination and laboratory
testing), and spontaneously if reported between visits and
within 30days of the last study drug intake
No Yes
ATTAIN [15] Yes
At all study visits (clinical examination and laboratory test-
ing) and pre-specied questionnaires for some pre-specied
adverse events
No Yes
C-EDGE [16] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and up until 14days
after cessation of treatment
No Yes
CONCERTO-1 [17] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and up until 14days
after cessation of treatment
No Yes
CONCERTO-2 [18] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and up until 14days
after cessation of treatment and until the end of the post-
treatment follow up period
No Yes
DRAGON [19] No
Except for routine laboratory monitoring
No Yes
FISSION [20] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and until the end of the
4th week post-treatment follow up period
No Yes
FUSION [21] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and until the 4th week
post-treatment visit. After that visit investigators were not
obliged to actively seek adverse events, but have to report
them if they learn about any possible adverse events
No Yes
HALLMARK-DUAL [22] Yes
At all study visits (clinical examination and laboratory test-
ing) and through patient’s adverse events reports
No Yes
HCV RESPOND-2 [23] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study
No Yes

4. Internal and Emergency Medicine1 3
Table1  (continued)
Study Method of adverse event assessment description Exclusion of patients from
the adverse e ects analysis
Presence of numerical
data by intervention
group
ILLUMINATE [24] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and until the 4th week
post-treatment visit
No Yes
ION-1 [25] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and until the 4th week
post-treatment visit. After that visit investigators were not
obliged to actively seek adverse events, but have to report
them if they learn about any possible adverse events
No Yes
ION-2 [26] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and until the 4th week
post-treatment visit. After that visit investigators were not
obliged to actively seek adverse events, but have to report
them if they learn about any possible adverse events
No Yes
ION-3 [27] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and until the 4th week
post-treatment visit. After that visit investigators were not
obliged to actively seek adverse events, but have to report
them if they learn about any possible adverse events
No Yes
Mizokami etal. [28] Yes
Safety was assessed in all patients by physical examination
and by review of adverse events and blood and urine sam-
ples for clinical laboratory testing
No Yes
OPTIMIZE [29] Yes
Safety data were summarized from the date of rst intake
of study drug to the date of last intake of study drug plus
30days. Special search categories (SSCs) were created by
grouping AE terms representing similar medical concepts
from the same or di erent body systems to ensure that each
patient was counted only once
No Yes
PEARL II [30] Yes
Adverse event assessments were reported from the time of
study drug administration until 30days after the last dose
and were judged as mild, moderate, or severe; clinical labo-
ratory testing was performed at each study visit. Serious
AEs were recorded throughout the study
No Yes
PEARL III [31] Yes
The following safety evaluations were performed during the
study: adverse event monitoring and vital signs, physi-
cal examination, ECG, and laboratory tests assessments.
Adverse-event assessment and clinical laboratory testing
were performed at each study visit during treatment and in
the follow-up period. Adverse events were reported from
the time of study-treatment initiation until 30days after the
last dose. Data on serious adverse events were collected
throughout the study
No Yes

5. Internal and Emergency Medicine 1 3
Table1  (continued)
Study Method of adverse event assessment description Exclusion of patients from
the adverse e ects analysis
Presence of numerical
data by intervention
group
PEARL IV [31] Yes
The following safety evaluations were performed during the
study: adverse event monitoring and vital signs, physi-
cal examination, ECG, and laboratory tests assessments.
Adverse-event assessment and clinical laboratory testing
were performed at each study visit during treatment and in
the follow-up period. Adverse events were reported from
the time of study-treatment initiation until 30days after the
last dose. Data on serious adverse events were collected
throughout the study
No Yes
POSITRON [21] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study, and until the 4th week
post-treatment visit. After that visit investigators were not
obliged to actively seek adverse events, but have to report
them if they learn about any possible adverse events
No Yes
PROMISE [32] Not reported in main manuscript. Supplementary appendix
not available
No Yes
QUEST-1 [33] Yes
Adverse events were monitored throughout the trial. Blood
samples for biochemical and haematological analyses were
obtained at screening and during scheduled visits. Electro-
cardiogram, vital sign assessment, and physical examina-
tion were also done at scheduled visits. Data for fatigue and
productivity (including activity impairment and absen-
teeism) were gathered using the Fatigue Severity Score
(FSS) and the Work Productivity and Activity Impair-
ment questionnaire for hepatitis C (WPAI) at baseline and
throughout the study. The Centre for Epidemiologic Studies
Depression Scale (CES-D) questionnaire was used to assess
the e ect of treatment on depression, with data gathered at
baseline and all the way through the study. Perceived health
status and quality of life at baseline and during treatment
were assessed through the EuroQol 5-dimension question-
naire (EQ-5D). The ve health dimensions measured with
the EQ-5D were mobility, self-care, usual activities, pain or
discomfort, and anxiety or depression
No Yes
QUEST-2 [34] Yes
Adverse events were monitored throughout the trial. Blood
samples for biochemical and haematological analyses were
obtained at screening and during scheduled visits. Electro-
cardiogram, vital sign assessment, and physical examina-
tion were also done at scheduled visits. Data for fatigue and
productivity (including activity impairment and absen-
teeism) were gathered using the Fatigue Severity Score
(FSS) and the Work Productivity and Activity Impair-
ment questionnaire for hepatitis C (WPAI) at baseline and
throughout the study. The Centre for Epidemiologic Studies
Depression Scale (CES-D) questionnaire was used to assess
the e ect of treatment on depression, with data gathered at
baseline and all the way through the study. Perceived health
status and quality of life at baseline and during treatment
were assessed through the EuroQol 5-dimension question-
naire (EQ-5D). The ve health dimensions measured with
the EQ-5D were mobility, self-care, usual activities, pain or
discomfort, and anxiety or depression
No Yes

6. Internal and Emergency Medicine1 3
Completeness of outcome reporting
Reporting of any major vascular events, i.e. myocardial
infarction, ischemic stroke, pulmonary embolism and
splanchnic vein thrombosis, as pre-specified secondary
outcome and routine electrocardiogram (ECG) during study
period was judged as at ‘low risk of underreporting’.
Statistical analysis
Crude unweighted incidence rates of safety outcomes were
estimated for all included studies and studies reporting major
vascular events, and divided into intervention and control
groups. Clopper–Pearson was the method used to calculate
95% condence intervals of these unweighted incidences.
Di erence between proportions in interventions and controls
was calculated with Miettinen method reported as p value.
Results
From the electronic search strategy, we identied 64 poten-
tially relevant publications of which only 26 were included
[10, 15–39]. For the following reasons 38 papers were
Table1  (continued)
Study Method of adverse event assessment description Exclusion of patients from
the adverse e ects analysis
Presence of numerical
data by intervention
group
REALIZE [35] Yes
Data on all adverse events were collected throughout the
treatment period and at the safety follow-up assessment
4weeks after the administration of the last dose of a study
drug. Serious adverse events continued to be monitored
throughout the follow-up period. Blood samples for
biochemical and haematological analyses were obtained at
screening and during scheduled visits. Electrocardiogram,
vital sign assessment, and physical examination were also
done at scheduled visits. Data for fatigue and productiv-
ity were gathered using the Fatigue Severity Score (FSS)
and the Work Productivity questionnaire. Perceived health
status and quality of life at baseline and during treatment
were assessed through the EuroQol 5-dimension question-
naire (EQ-5D)
No Yes
SAPPHIRE II [36] Yes
The following safety evaluations were performed during the
study: adverse event monitoring and vital signs, physi-
cal examination, ECG, and laboratory tests assessments.
Adverse-event assessment and clinical laboratory testing
were performed at each study visit during treatment and in
the follow-up period. Adverse events were reported from
the time of study-treatment initiation until 30days after the
last dose. Data on serious adverse events were collected
throughout the study
No Yes
SPRINT-2 [37] Yes
At all study visits (clinical examination and laboratory test-
ing) and any time during the study
No Yes
TURQUOISE-II [38] The following safety evaluations were performed during the
study: adverse event monitoring and vital signs, physi-
cal examination, ECG, and laboratory tests assessments.
Adverse-event assessment and clinical laboratory testing
were performed at each study visit during treatment and in
the follow-up period. Adverse events were reported from
the time of study-treatment initiation until 30days after the
last dose. Data on serious adverse events were collected
throughout the study
No Yes
VALENCE [39] Yes
Safety was evaluated by assessment of clinical laboratory
tests, physical examinations, vital signs measurements, at
various time points during the study, and by the documen-
tation of AEs until the 4th week post-treatment visit
No Yes

7. Internal and Emergency Medicine 1 3
excluded: non-randomized trial (N = 11), commentary
(N = 2), subanalysis of already included RCTs (N = 11),
pooled analysis of already included RCTs (N = 7), post hoc
analysis (N = 1), HIV patients co-infected with HBV/HCV
(N = 5), phase IIb trial (N = 1) (references available on Sup-
plementary Data Appendix). Four extra papers were identi-
ed with additional search strategies [11– 14]).
A total of 33 phase III RCTs found in 30 publications
were included in this systematic review [10–39], encompass-
ing 14,801 patients randomized and 14,764 patients taking
at least one study drug (9204 in intervention group; 5560
in control group of which 2068 were controls with placebo,
not including a DAA), 37 patients were mainly excluded for
patients withdrawn before starting study drug. Randomiza-
tion was achieved in di erent ways. In 18 RCTs there was
a 1:1 randomization [10, 13–15, 18, 20, 21, 24–31, 37, 38].
Except for one study [19], in all others, randomization was
performed after stratication based on di erent characteris-
tics: HCV genotype/subgenotype or interleukin 28B geno-
type (genetic marker associated with treatment outcomes)
[10–12, 14–18, 20, 21, 23–27, 29, 31–34, 36–38], HCV
viral load [10, 35, 37], cirrhosis absence/presence or status
[11–13, 16, 20–22, 25, 26, 28, 29, 39], previous treatment
and type of response/nonresponse [13, 15, 23, 26, 28, 30,
35, 36, 38, 39], age ≥ 65years/<  65years [17, 18] and
self-reported race [24]. Baseline characteristic of included
RCTs are summarized in Table2. Median of the mean age
of included patients in each study is 53years (interquartile
range 50–55). Median of the percentage of males included
in each study is 59% (interquartile range 55–65). Median
treatment duration is 12weeks and median follow-up for
safety outcomes is 42weeks (interquartile range 24?72). Six
RCTs [17, 27, 30, 31, 36] include patients without cirrhosis
only, while among the other studies, the median of included
patients with cirrhosis is 20%. Of the six RCTs performed
among patients without liver cirrhosis, one enrolled only
Japanese patients [17]. Denition of absence of cirrhosis
as inclusion criteria is slightly di erent among these stud-
ies, but generally requires a liver biopsy performed within
24months prior to or during screening demonstrating the
absence of cirrhosis [e.g., a METAVIR Score (40  ≤ 3,
Ishak score  ≤ 4)] or a screening FibroTest score of  ≤ 0.72
and Aspartate Aminotransferase to Platelet Ratio Index
(APRI)  ≤ 2 or a screening FibroScan result of  < 9.6kPa [40].
Seven RCTs are in treatment naïve patients [22, 24, 25,
29, 33, 34, 37] and in one study the comparison is made
between previously treated and untreated patients [11]. In
fourteen RCTs the comparator is a placebo [10, 12, 16, 17,
21, 23, 30–36, 39], fourteen include the standard therapy
with peginterferon alfa-2a and ribavirin as part of inter-
vention treatment or comparison [10, 15, 17–20, 23, 24,
29, 32–35, 37] and twelve RCTs just ribavirin [13, 14, 21,
25–28, 30, 31, 38, 39]. In two RCTs the comparison is made
between di erent DAAs [15, 22] and in four studies di erent
duration/schemes of the same treatment are compared [18,
21, 28, 38]. All intervention treatment arms are combined
therapies: some of them are combination of DAAs [11– 14,
16, 22, 25–28], some others the addition of a DAA on stand-
ard therapy with peginterferon and ribavirin [10, 15, 17–19,
23, 24, 29, 32–35, 37] or ribavirin alone [20, 21, 39] and
some others combination of DAAs plus standard therapy
with peginterferon or ribavirin [30– 32, 36, 38]. In only 16
RCTs, control arms do not include use of a DAA [10, 12, 16,
17, 19–21, 23, 28, 32–37, 39].
Risk of bias and quality of reporting
Risk of bias and quality of reporting assessment are sum-
marized in Table3. None of the included studies report any
major vascular events as a predened outcome. In 20 stud-
ies [10– 14, 20, 24–29, 31–33, 35, 36, 38], ECG is regularly
tested during follow-up, as specied in the published manu-
script or in the protocol (if available) (Table3); in nine RCTs
[15–19, 21, 30, 34, 37
] this information is not reported in the
published manuscript, and in four RCTs [21, 23, 39], ECG
is clearly not performed during follow-up.
Arterial ischemic events
Thirteen (39%) RCTs [11– 14, 16, 19, 23, 26, 28, 35–37]
report a total of 23 probable arterial ischemic events for an overall crude unweighted rate of 0.16% of the total popu- lation (95% CI 0.10?0.24) (median follow-up 42weeks, interquartile range 24–72) and 0.47% in those 13 RCTs reporting arterial events (median follow up 44weeks, inter-
quartile range 19–63). Three arterial events occurred in patients receiving telaprevir plus peginterferon and riba- virin, and seven events in patients receiving boceprevir plus peginterferon and ribavirin. Seven RCTs report 14 coronary events [12, 14, 16, 23, 26, 28, 35, 37], of which
11 are dened as MI [12, 14, 16, 23, 28, 35, 37], 1 as
coronary artery occlusion [37], 1 as coronary artery dis- ease [37] and 1 as instable angina [26]. Seven of these MI occurred in intervention groups and four in control groups (two in placebo groups and two in control arms including DAAs), with an unweighted incidence of 0.07% of the total population (95% CI 0.04–0.1) (0.21% in those 13 RCTs reporting arterial events), divided into 0.08% in inter-
ventions (95% CI 0.04–0.2) and 0.07% in controls (95% CI 0.02–0.2) (0.09% considering the two events among placebo control groups only). Two RCTs report two cer-
ebrovascular ischemic events: one dened as fatal cerebral infarction [19] and one as transient ischemic attack [36], in both cases in intervention groups versus placebo (follow up, 16?72weeks). Five RCTs report a total of seven severe cardiac events dened as cardiac/cardiopulmonary arrests

8. Internal and Emergency Medicine1 3
Table2  Characteristics of the included studies StudyStudy designPopulationNo. of
patients rand -
omized
Mean ageInterventionComparator
ADVANCE NEJM, NCT00627926
[10]
RCT, DBHCV genotype 1 infection109549Telaprevir + PRPlacebo + PR
ALLY-2, NEJM 2015,
NCT02032888 [ 11]
RCT, OLHCV genotype 1, 2, 3, 4, 5 or 6
with HIV-1 co-infection
20351Daclatasvir + sofosbuvir in previ -
ously untreated patients
Daclatasvir + sofosbuvir in previ -
ously treated patients
ASTRAL-1, NEJM 2015,
NCT02201940 [ 12]
RCT, DBHCV genotype 1, 2, 4, 5 or 6. 19%
had cirrhosis (METAVIR score
F4)
70653.5Sofosbuvir + velpatasvirPlacebo
ASTRAL-2, NEJM 2015,
NCT02220998 [ 13]
RCT, OLHCV genotype 2 14% had cirrhosis,
and 14–15% had received unsuc -
cessful treatment for HCV
26957Sofosbuvir + velpatasvirSofosbuvir + ribavirin
ASTRAL-3, NEJM 2015,
NCT02201953 [ 13]
RCT, OLHCV genotype 3 29–30% had
cirrhosis and 26% had undergone
unsuccessful treatment
55849.5Sofosbuvir + velpatasvirSofosbuvir + ribavirin
ASTRAL-4, NEJM 2015,
NCT02201901 [ 14]
RCT, OLHCV genotype 1, 2, 3, 4, 5 or 6
with decompensated cirrhosis
(Child–Pugh B)
26858Sofosbuvir + velpatasvirSofosbuvir + velpatasvir + ribavirin
ATTAIN Lancet Infect Dis. 2015,
NCT01485991 [ 15]
RCT, DBDDHCV genotype 176351Simeprevir + PRTelaprevir + PR
C-EDGE Ann Intern Med 2015,
NCT02105467 [ 16]
RCT, BHCV genotype 1, 4, or 6 (22%
cirrhosis)
42154Grazoprevir + elbasvirPlacebo
CONCERTO-1 J Hepatol 2014,
NCT01292239 [ 17]
RCT, DBHCV genotype 1. Liver cirrhosis
excluded. Only Japanese
18855Simeprevir + PRPlacebo + PR
CONCERTO-2 J Gastroenterol
2014, NCT01288209 [ 18]
RCT, OLHCV genotype 1 infection. Only
Japanese
10860Simeprevir + PRSimeprevir +PR
DRAGON J Gastroent 2014,
NCT00996476 [ 19]
RCT, OLHCV genotype 1b. Only Japanese9254Simeprevir/PRPR
FISSION NEJM, NCT01641640
[20]
RCT, OLHCV genotype 2 or 3, treatment
naïve. 20% had cirrhosis
52748Sofosbuvir + ribavirinPR
FUSION NEJM, NCT01604850
[21]
RCT, BHCV chronic infection 34% had
cirrhosis
20254Sofosbuvir + ribavirinSofosbuvir + ribavirin, followed by
matching placebo
HALLMARK-DUAL Lancet 2014,
NCT01581203 [ 22]
RCT, DBDDHCV genotype 1b treatment naïve30755Daclatasvir + asunaprevirDaclatasvir/asunaprevir
HCV RESPOND-2 NEJM,
NCT00708500 [ 23]
RCT, DBHCV chronic infection. Non-
responders or relapsed
40352Boceprevir + PRPlacebo + PR
ILLUMINATE NEJM,
NCT00758043 [ 24]
RCT, OLHCV genotype 1, treatment naïve.
61 patients had cirrhosis
54051Telaprevir + PRTelaprevir + PR
ION-1 NEJM 2014, NCT01701401
[25]
RCT, OLHCV genotype 1 treatment naïve
(16% had cirrhosis and 67% had
HCV genotype 1a infection)
43452Ledipasvir + sofosbuvirLedipasvir + sofosbuvir +ribavirin

9. Internal and Emergency Medicine 1 3
Table2  (continued) StudyStudy designPopulationNo. of
patients rand -
omized
Mean ageInterventionComparator
ION-2 NEJM 2014, NCT01768286
[26]
RCT, OLHCV genotype 1 (20% had cirrho -
sis and 79% had HCV genotype
1a infection)
22056Ledipasvir + sofosbuvirLedipasvir + sofosbuvir +ribavirin
ION-3 NEJM 2014, NCT01851330
[27]
RCT, OLHCV genotype 1 without cirrhosis43152Ledipasvir + sofosbuvirLedipasvir + sofosbuvir + ribavirin
Mizokami etal. Lancet Infect Dis
2015, NCT01975675 [ 28]
RCT, OLHCV genotype 1. 22% had cirrho -
sis. Only Japanese
34159Ledipasvir + sofosbuvirRibavirin
OPTIMIZE Gastroenterology 2014;
NCT01241760 [ 29]
RCT, OLHCV genotype 1 treatment naïve.
14% had cirrhosis
74048Telaprevir + PRTelaprevir + PR
PEARL II Gastroenterology 2014,
NCT01674725 [ 30]
RCT, OLHCV genotype 1b infection, with -
out cirrhosis
17954ABT-450/ritonavir/ombitas -
vir + dasabuvir + ribavirin
ABT-450/ritonavir/ombitasvir + das-
abuvir
PEARL III NEJM 2014,
NCT01767116 [ 31]
RCT, DBHCV genotype 1b infection and no
cirrhosis
41950ABT-450/ritonavir/ombitas -
vir + dasabuvir + ribavirin
ABT-450/ritonavir/ombitasvir + das-
abuvir + ribavirin/placebo
PEARL IV NEJM 2014,
NCT01833533 [ 31]
RCT, DBHCV genotype 1a infection and no
cirrhosis
30550ABT-450/ritonavir/ombitas -
vir + dasabuvir + ribavirin
ABT-450/ritonavir/ombitasvir + das-
abuvir + ribavirin/placebo
POSITRON NEJM 2013,
NCT01542788 [ 21]
RCT, BHCV chronic infection28052Sofosbuvir + ribavirinPlacebo
PROMISE Gastroent 2014,
NCT01281839 [ 32]
RCT, DBHCV genotype 1. Liver cirrhosis
included
39352Simeprevir +PRPlacebo + PR
QUEST-1 Lancet 2014,
NCT01289782 [ 33]
RCT, DBHCV genotype 1 treatment naïve.
12% had cirrhosis (METAVIR
score F4)
39448Simeprevir + PRPlacebo + PR
QUEST-2 Lancet 2014,
NCT01290679 [ 34]
RCT, DBHCV genotype 1 treatment naïve.
Included cirrhosis (METAVIR
score F4)
39146Simeprevir + PRPlacebo + PR
REALIZE NEJM, NCT00703118
[35]
RCT, DBHCV genotype 1 25% had cirrhosis66351Telaprevir + PRPlacebo + PR
SAPPHIRE-II NEJM 2014,
NCT01715415 [ 36]
RCT, DBHCV genotype 1 without cirrhosis39553.5ABT-450/ritonavir–ombitasvir +
dasabuvir + ribavirin
Placebo
SPRINT-2 NEJM, NCT00705432
[37]
RCT, DBHCV chronic infection, treatment
naïve. 5% had cirrosi
109750PR + boceprevirPR
TURQUOISE-II NEJM 2014,
NCT01704755 [ 38]
RCT, OLHCV genotype 1 and Child–Pugh
class A cirrhosis
38057ABT-450/ritonavir–ombitasvir +
dasabuvir + ribavirin
ABT-450/ritonavir–ombitasvir +
dasabuvir + ribavirin
VALENCE NEJM 2014,
NCT01682720 [ 39]
RCT, DBHCV genotype 2 or 3, 21% had
cirrhosis
41950Sofosbuvir + ribavirinPlacebo
RCT randomized controlled trial, DBDD double blind double dummy, DB double blind, PR peginterferon alfa-2a and ribavirin, OL open label, IFN interferon

10. Internal and Emergency Medicine1 3
Table3   Risk of bias and quality of reporting assessment [8]
Study Random
sequence gen-
eration (selection
bias)
Allocation
concealment
(selection bias)
Blinding of
participants and
personnel (per-
formance bias)
Blinding of out-
come assessment
(detection bias)
Incomplete
outcome data
(attrition bias)
Arterial event
or VTE as pre-
specied second-
ary outcome
ECG
during
follow-up
ADVANCE [10]
☺ ☺ ☺ ☺ ☺ ☺
ALLY-2 [11]
☺ ☺ ☺ ☺ ☺ ☺
ASTRAL-1 [12]
☺ ☺ ☺ ☺ ☺ ☺
ASTRAL-2 [13]
☺ ☺ ☺ ☺
ASTRAL-3 [13]
☺ ☺ ☺ ☺
ASTRAL-4 [14]
☺ ☺ ☺ ☺
ATTAIN [15]
☺ ☺ ☺ ☺ ☺
?
C-EDGE [16]
☺ ☺ ☺ ☺ ☺
?
CONCERTO-1
[17]
☺ ☺ ☺ ☺
?
CONCERTO-2
[18]
?
☺
?
DRAGON [19]
☺ ☺ ☺
?
FISSION [20]
☺ ☺ ☺ ☺
FUSION [21]
☺ ☺ ☺
?
☺
HALLMARK-
DUAL [22]
☺ ☺ ☺ ☺ ☺
?
HCV
RESPOND-2 [23]
☺ ☺ ☺ ☺ ☺
ILLUMINATE
[24]
☺ ☺ ☺ ☺ ☺ ☺
ION-1 [25]
☺ ☺ ☺
ION-2 [26]
☺ ☺ ☺
ION-3 [27]
☺ ☺ ☺
Mizokami etal.
[28]
☺ ☺ ☺ ☺
OPTIMIZE [29]
☺ ☺ ☺ ☺
PEARL II [30]
☺ ☺ ☺
?
PEARL III [31]
☺ ☺ ☺ ☺ ☺ ☺
PEARL IV [31]
☺ ☺ ☺ ☺ ☺ ☺
POSITRON [21]
☺ ☺ ☺ ☺ ☺
PROMISE [32]
☺ ☺ ☺ ☺ ☺ ☺
QUEST-1 [33]
☺ ☺ ☺ ☺ ☺ ☺
QUEST-2 [34]
☺ ☺ ☺ ☺ ☺
?
REALIZE [35]
☺ ☺ ☺ ☺ ☺ ☺
SAPPHIRE II
[36]
☺ ☺ ☺ ☺ ☺ ☺

11. Internal and Emergency Medicine 1 3
[11, 13, 14, 28, 37], of which four were fatal [11, 13, 14,
28]. Three of these events occurred in interventions and
four in controls (two in placebo groups and two in control
arms including DAAs). Overall crude unweighted inci-
dence of these severe events is 0.05% of the total included
population (95% CI 0.02–0.1), 0.14% in those RCTs
reporting arterial events), 0.03% in interventions (95%
CI 0.001–0.1) and 0.07% in controls (95% CI 0.02–0.2)
(0.09% considering the two events among placebo control
groups only). Twenty RCTs (61%) do not report data on
arterial ischemic events [10, 15, 17, 18, 20–22, 24, 25,
27, 29–34, 38, 39]. We could not identify any signicant
di erence in patients’ characteristics between the studies
reporting arterial thrombotic events compared to those not
reporting. There is no signicant di erence between inci-
dence in intervention and controls.
Venous thromboembolic events
Four RCTs (12%) [11, 15, 21, 37] report data on at least one
major venous outcome (Table4). A total of ve VTE are
described: four pulmonary embolism [11, 15, 37], one of
which was fatal [15], and one portal vein thrombosis [21].
Two PE and the portal vein thrombosis occurred in interven-
tion groups [21, 37] while two PE, of which one was fatal,
occurred in control groups [11, 15], both of them including
DAAs. One venous event occurred in a patient receiving
telaprevir plus peginterferon and ribavirin, and two events in
patients receiving boceprevir plus peginterferon and ribavi-
rin. Overall crude unweighted rate of VTE is 0.03% (95% CI
0.01–0.08) of the total included population (median follow-
up 42weeks, interquartile range 24?72), divided into 0.03%
in interventions (95% CI 0.003–0.08) and 0.04% in controls
(95% CI 0.004–0.1) (no VTE among placebo control groups
only) and 0.22% in those four RCTs reporting VTE events
(median follow-up 54weeks, interquartile range 34?72). 29
RCTs (88%) do not report data on venous thromboembolic
events [10, 12–14, 16–36, 38, 39]. We could not identify any
signicant dierence in patients? characteristics between the
studies reporting VTE events compared to those not report-
ing. There is no signicant di erence between incidence in
intervention and controls.
Death from unknown causes
Six RCTs [10, 12, 13, 16, 20, 34] report seven deaths from
unknown causes. Five of these deaths occurred in the inter-
vention arm [10, 12, 16, 21, 34] and two in the compara-
tor arm of a single RCTs [13] that included DAAs. Crude
unweighted rate is 0.05% (95% CI 0.02–0.1) of the total
included population (median follow-up 42weeks, interquar-
tile range 24–72), divided into 0.05% of interventions (95%
CI 0.02–0.1) and 0.04% of controls (95% CI 0.004–0.1) (no
death from unknown cause among placebo control groups
only) and 0.19% in those six RCTs data [10, 12, 13, 16, 20,
34]. Pooling those RCTs in which data on death are reported
[10–24, 28, 29, 32–34, 36] and excluding those RCTs in
which data on death are not reported [25–27, 30, 31, 35,
37–39] the crude unweighed incidence is 0.07%. There is
no signicant di erence between incidence in intervention
and controls.
Combined outcomes
Overall, 11 MI, one ischemic stroke, four PE, and seven
cardiopulmonary arrests occurred for a total of 23 major
events. Thirteen of these events occurred in the interven-
tion arm [11– 14, 19, 23, 35, 37] and ten in the control arm
[11, 14–16, 28, 37], four in placebo control arms [16, 37]
and six in control arms including DAAs [11, 14, 15, 28].
Crude unweighted rate is 0.15% (95% CI 0.10–0.23) of
the total included population (median follow-up 42weeks,
interquartile range 24–72), divided into 0.14% of interven-
tions (95% CI 0.08–0.2) and 0.18% of controls (95% CI
0.09–0.3) (0.19% considering the four events among placebo
control groups only), and 0.42% in those 11 RCTs [11– 16,
19, 23, 28, 35, 37] reporting data on thrombotic events and
Table3  (continued)
Study Random
sequence gen-
eration (selection
bias)
Allocation
concealment
(selection bias)
Blinding of
participants and
personnel (per-
formance bias)
Blinding of out-
come assessment
(detection bias)
Incomplete
outcome data
(attrition bias)
Arterial event
or VTE as pre-
specied second-
ary outcome
ECG
during
follow-up
SPRINT-2 [37]
☺ ☺ ☺ ☺ ☺
?
TURQUOISE-II
[38]
☺ ☺ ☺
VALENCE [39] ?
☺ ☺ ☺ ☺
VTE venous thromboembolism
☺
low risk, ? unclear risk, high risk

12. Internal and Emergency Medicine1 3
Table4   Outcomes
Study Number of arterial
event or VTE events
Fatal arterial event or
VTE events
Death from unknown
cause
Major bleeding Safety Assessment
mean follow-up
(weeks)
ADVANCE [10] NR NR 1 (intervention group)0 52
ALLY-2 [11] 1 cardiac arrest (inter-
vention group)
1 PE (control group)
1 cardiac arrest (inter-
vention group)
0 0 36
ASTRAL-1 [12] 1 acute myocardial
infarction (interven-
tion group)
0 1 (intervention group)0 16
ASTRAL-2 [13] 1 cardiac arrest (inter-
vention group)
1 cardiac arrest (inter-
vention group)
0 0 16
ASTRAL-3 [13] 0 0 2 (control group)0 20
ASTRAL-4 [14] 1 AMI (intervention
group)
1 AMI (control group)
1 cardiopulmonary
arrest (control group)
1 cardiopulmonary
arrest (control group)
1 AMI (control group)
0 2 gastrointestinal
bleeding (interven-
tion group)
20
ATTAIN [15] 1 PE (control group)1 PE (control group)0 0 72
C-EDGE [16] 1 non fatal AMI (con-
trol group)
NR 1 (intervention group)1 retinal haemorrhage
(control group)
36
CONCERTO-1 [17] NR 0 0 0 72
CONCERTO-2 [18] NR 0 0 0 72
DRAGON [19] 1 cerebral infarction
(intervention group)
1 cerebral infarction
(intervention group)
0 0 60
FISSION [20] NR NR 1 (intervention group)0 42
FUSION [21] 1 portal vein throm-
bosis (intervention
group)
NR 0 1 oesophageal varices
haemorrhage (inter-
vention group)
28
HALLMARK-DUAL
[22]
NR 0 0 0 20
HCV RESPOND-2
[23]
1 AMI (intervention
group)
0 0 1 oesophageal varices
haemorrhage (inter-
vention group)
72
ILLUMINATE [24] NR NR 0 0 84
ION-1 [25] NR NR NR 0 42
ION-2 [26] 1 unstable angina
(intervention group)
NR NR 1 upper gastrointestinal
bleeding (interven-
tion group)
42
ION-3 [27] NR NR NR 0 32
Mizokami etal. [28] 1 cardiac arrest (con-
trol group)
1 AMI (control group)
1 (cardiac arrest)0 1 oesophageal varices
haemorrhage (inter-
vention group)
36
OPTIMIZE [29] NR NR 0 0 24
PEARL II [30] NR NR NR 0 16
PEARL III [31] NR NR NR 0 60
PEARL IV [31] NR NR NR 0 60
POSITRON [21] NR NR 0 0 24
PROMISE [32] NR NR 0 0 72
QUEST-1 [33] NR 0 0 0 72
QUEST-2 [34] NR NR 1 (intervention group)0 72
REALIZE [35] 3 AMI (intervention
group)
NR NR 0 72

13. Internal and Emergency Medicine 1 3
cardiopulmonary arrest (median follow-up 36weeks, inter-
quartile range 28–72).
Other safety outcome reporting
Seven RCTs [14, 16, 21, 22, 26, 28, 37] report a total of ten
major bleeding events, i.e. nine cases of gastrointestinal
bleeding (all occurred in intervention groups) and one reti-
nal haemorrhage (occurring in a placebo control group).
Crude unweighted rate is 0.07% (95% CI 0.03–0.1) for
the total included population (median follow-up 42weeks,
interquartile range 24–72), divided into 0.1% of interven-
tions (95% CI 0.04–0.2) and 0.02% in controls (95% CI
0.001–0.1) (p = 0.056) (0.05% considering the only event
among placebo control groups only) and 0.34% in those
seven RCTs reporting bleeding events (median follow-up
36weeks, interquartile range 32?57).
Anemia is described as a pre-specied outcome in all
manuscripts except for ve RTCs [16–19, 21]. Headache is
not described as a pre-specied outcome in any RCT. All
33 RCTs included in this review report data on anemia and
headache. Data on the incidence of anemia and headache
are available on Supplementary Data Appendix.
Discussion
To the best of our knowledge, this is the rst systematic
review aimed to evaluate the occurrence of arterial and
venous thrombotic events in patients with chronic hepati-
tis C receiving DAAs. Incidence of thrombotic events in
HCV patients receiving DAAs might be low, but published
data suggest that safety vascular prole of DAAs may be
incorrectly estimated due to a possible inadequate report-
ing of the real incidence of arterial ischemic and venous
thrombotic events in phase III RCTs.
Chronic liver disease has been claimed to increase
thrombotic risk by several mechanisms. In addition to
portal venous stasis secondary to portal hypertension and
a prothrombotic state due to an imbalance of procoagulant
and anticoagulant factors, hepatitis C virus has also been
associated with the presence of antiphospholipid antibod-
ies that may lead to an increased risk of both venous and
arterial thrombosis [41]. Recently, Enger and colleagues
published a matched cohort study aiming to estimate the
incidence of arterial and venous thromboembolic events
among patients with cirrhosis and HCV infection and
matched comparators [41]. The study includes 22,733
HCV-infected patients and 68,198 comparators, and
Table4  (continued)
Study Number of arterial
event or VTE events
Fatal arterial event or
VTE events
Death from unknown
cause
Major bleeding Safety Assessment
mean follow-up
(weeks)
SAPPHIRE II [36] 1 acute transient stroke
(cerebrovascular
accident) (interven-
tion group)
0 0 0 16
SPRINT-2 [37] 1 AMI (intervention
group)
1 AMI (control group)
1 cardiac arrest (con-
trol group)
1 cardiac arrest (inter-
vention group)
1 cardio-respiratory
arrest (control group)
1 coronary artery
disease (intervention
group)
1 coronary artery
occlusion (interven-
tion group)
1 arterial thrombosis
limb (intervention
group)
2 PE (intervention
group)
NR NR 1 gastrointestinal
haemorrhage (inter-
vention group)
1 hematemesis (inter-
vention group)
1 hemorrhoidal hemor-
rhage (intervention
group)
72
TURQUOISE-II [38] NR NR NR 0 66
VALENCE [39] NR NR NR 0 24
NR not reported, SAEs serious adverse events, PE pulmonary embolism, AMI acute myocardial infarction

14. Internal and Emergency Medicine1 3
15,158 cirrhosis patients and 45,473 comparators. The
incidence of any thromboembolic event is 233.4 events
per 10,000 person-years for the HCV cohort and 138.5 per
10,000 person-years for the comparators; the incidence of
any thromboembolic event is 561.1 per 10,000 person-
years for the cirrhosis patients and 249.7 per 10,000 per-
son-years for the controls [42].
In the Framingham study, low-risk men of similar age
show a risk of nearly 75 per 10,000 of a rst cardiovascu-
lar event in the rst 42weeks [43]. In our overall popu-
lation of adult patients, we estimate a crude unweighted
rate of arterial events of 15 per 10,000 during a median
follow-up of 42weeks. A recent update on epidemiology
of VTE [44] reports that the overall age-adjusted inci-
dence rate for men is 13 per 10,000 person-years. In our
overall population of adult patients, we estimate a crude-
unweighted rate of venous events of 3 per 10,000 during a
median follow-up of 42weeks. Moreover, only one portal
vein thrombosis is reported in more than 14,000 included
patients in this study.
In patients with advanced liver disease, the most fre-
quent and life-threatening site of bleeding is the gastro-
intestinal tract. In the included RCTs, major bleeding is
reported in only ten patients, nine bleeding events were
gastrointestinal. In cross-sectional studies, gastroesoph-
ageal varices are reported in 30–50% of patients with
cirrhosis [45]. As current evidence suggests that a third
of patients with documented oesophageal varices bleed
within 2years from the time of diagnosis [46, 47], and
among our population nearly one-fth of patients were cir-
rhotic, we may roughly estimate that at least 100 variceal
bleeding could have been reported in the median 42-week
follow-up of the included studies.
As DAAs have been also approved for treatment of HCV-
related cirrhotic patients with advanced disease, it is unex-
pected that both major vascular events and major (gastroin-
testinal) bleeding have not been considered as prespecied
secondary outcomes in phase III RCTs, in particular because
less clinically relevant outcomes, such as headache, have
been systematically reported. Even if it is possible to con-
clude that thrombotic events should not be underestimated
because other safety outcomes, i.e. anemia and headache,
have been reported correctly, all these ndings are unfor-
tunately in line with a previous review of the literature that
reveals that under-reporting of venous and arterial thrombo-
sis is a major shortcoming of RCTs, even when published in
major medical journals: indeed, venous and arterial throm-
bosis is not reported in 89 and 70% of published RCTs,
respectively [48].
The strengths of this study include the rigorous meth-
odological approach and the selection of all phase III RCTs
performed on the topic. Some limitations deserve further
comments. First, we did not contact the authors for providing
additional data, and safety outcomes may have been reg-
istered and not reported in the paper. Second, the crude
estimates are per denition imprecise, in particular when
assuming that studies in which no events were reported, no
events occurred. However, estimated incidence of safety
outcomes in those studies providing data still suggests pos-
sible inadequate reporting, Indeed, other possible statistical
approaches to correctly estimate reporting have limits when
most of the studies do not provide any data on VTE or AT
occurrence. In addition, imputing missing data for the stud-
ies in which no events are reported, appears to be problem-
atic for the current dataset because it involves most of the
included studies: this means that whatever incidence rate is
used to impute, this will dominate the result. Therefore, no
such analyses were performed.
Third, we included heterogeneous RCTs in terms of
enrolled population, intervention, control, comparator, and
follow-up. DAAs are very di erent. Particularly, rst gen-
eration DAAs (boceprevir and telaprevir) have many dif-
ferent known side e ects, especially if compared with sec-
ond generation DAAs. Moreover, duration of treatment is
di erent within the same generation of DAAs. In addition,
the main imbalance of coagulation deals with the presence
of liver cirrhosis, and most of the patients treated in the
considered studies are not patients with liver cirrhosis, but
only with chronic hepatitis. Every single study contains a
di erent percentage of cirrhotic patients (some less than
10%, others more than 50%). This may be a real important
bias. Although sub-analyses on di erent DAAs and cirrhotic
state are warranted, the lower number of reported thrombotic
events makes any further statistical estimate unsuitable.
Fourth, even after virus eradication, the cirrhotic condi-
tion still remains. It is still debatable if, when and how long
the cirrhosis may be reverted after the resolution of inam-
mation due to virus infection. Probably the time needed to
assess a real benet on vascular conditions (both bleeding
and thrombosis) need a longer evaluation time, considering
years after the sustained viral response. Phase IV studies will
certainly provide us additional data on this topic.
In conclusion, our study, including 14,764 patients, has
the purpose to investigate a possible role of DAAs in modi-
fying the thrombotic risk that plays such an important role
on natural history of HCV-related chronic hepatitis. How-
ever, our ndings based on published data, appear to be
inconclusive because of a possible inadequate reporting of
the real incidence of arterial ischemic and venous throm-
botic events in phase III RCTs evaluating these new drugs,
suggesting that safety prole of DAAs may be incorrectly
estimated, even though probably with an absolute and rela-
tive favorable trend.
In future studies on DAAs, investigators should system-
atically monitor and report the occurrence of any arterial
and venous event, to provide a better risk to benet balance

15. Internal and Emergency Medicine 1 3
of DAAs, given that the prevalence on HCV infection, at
least in Europe, has been recently estimated to amount of
5.6 million of patients, so there is expected to be a huge and
wide use of these agents [49].
Author contributions  ETP performed data collection and analysis,
manuscript writing and final approval. AS was involved in study
design, data interpretation, manuscript writing and nal approval. LG
performed data collection. VEAG, LM and LG performed data inter-
pretation, manuscript editing and nal approval.
Compliance with ethical standards 
Conflict of interest None.
Statement of human and animal rights  This article does not contain
any studies with human participants or animals performed by any of
the authors.
Informed consent None.
References
1. Galli L, Gerdes VE, Guasti L, Squizzato A (2014) Thrombosis
associated with viral hepatitis. J Clin Transl Hepatol 2(4):234–239
2. Tufano A, Di Capua M, Coppola A, Conca P, Cimino E, Cer-
bone AM etal (2012) The infectious burden in atherothrombosis.
Semin Thromb Hemost 38(5):515–523
3. Deval J, Symons JA, Beigelman L (2014) Inhibition of viral RNA
polymerases by nucleoside and nucleotide analogs: therapeutic
applications against positive-strand RNA viruses beyond hepatitis
C virus. Curr Opin Virol 9:1–7
4. Belousova V, Abd-Rabou AA, Mousa SA (2015) Recent advances
and future directions in the management of hepatitis C infections.
Pharmacol Ther 145:92–102
5. Shamseer L, Moher D, Clarke M, Ghersi D, Liberati A, Pet-
ticrew M, PRISMA-P Group etal (2015) Preferred report-
ing items for systematic review and meta-analysis proto-
cols (PRISMA-P) 2015: elaboration and explanation. BMJ
2(349):g7647
6. D’Avolio A, Cusato J, De Nicolò A, Allegra S, Di Perri G
(2016) Pharmacogenetics of ribavirin-induced anemia in HCV
patients. Pharmacogenomics 17(8):925–941
7. Schulman S, Kearon C, Subcommittee on Control of Anticoagu-
lation of the Scientic and Standardization Committee of the
International Society on Thrombosis and Haemostasis (2005)
Denition of major bleeding in clinical investigations of antihe-
mostatic medicinal products in non-surgical patients. J Thromb
Haemost 3(4):692–694
8. Higgins JP, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman
AD, Cochrane Bias Methods Group, Cochrane Statistical Meth-
ods Group etal (2011) The Cochrane Collaboration?s tool for
assessing risk of bias in randomised trials. BMJ 18(343):d5928
9. Loke YK, Price D, Herxheimer A (2011) Chapter14: Adverse
e ects. In: Higgins JPT, Green S (eds) Cochrane handbook
for systematic reviews of interventionsversion 5.1.0 (updated
March 2011). The Cochrane Collaboration, 2011. http://www.
handb ook.cochr ane.org
10. Jacobson IM, McHutchison JG, Dusheiko G, Di Bisceglie AM,
Reddy KR, Bzowej NH, ADVANCE Study Team etal (2011)
Telaprevir for previously untreated chronic hepatitis C virus
infection. N Engl J Med 364(25):2405–2416
11. Wyles DL, Ruane PJ, Sulkowski MS, Dieterich D, Luetkemeyer
A, Morgan TR, ALLY-2 Investigators etal (2015) Daclatasvir
plus sofosbuvir for HCV in patients coinfected with HIV-1. N
Engl J Med 373(8):714–725
12. Feld JJ, Jacobson IM, Hézode C, Asselah T, Ruane PJ, Gruener
N, ASTRAL-1 Investigators etal (2015) Sofosbuvir and vel-
patasvir for HCV genotype 1, 2, 4, 5, and 6 infection. N Engl J
Med 373(27):2599–2607
13. Foster GR, Afdhal N, Roberts SK, Bräu N, Gane EJ, Pianko S,
ASTRAL-2 Investigators, ASTRAL-3 Investigators etal (2015)
Sofosbuvir and velpatasvir for HCV genotype 2 and 3 infection.
N Engl J Med 373(27):2608–2617
14. Curry MP, O’Leary JG, Bzowej N, Muir AJ, Korenblat KM,
Fenkel JM, ASTRAL-4 Investigators etal (2015) Sofosbuvir
and velpatasvir for HCV in patients with decompensated cir-
rhosis. N Engl J Med 373(27):2618–2628
15. Reddy KR, Zeuzem S, Zoulim F, Weiland O, Horban A, Stanciu
C etal (2015) Simeprevir versus telaprevir with peginterferon
and ribavirin in previous null or partial responders with chronic
hepatitis C virus genotype 1 infection (ATTAIN): a randomised,
double-blind, non-inferiority phase 3 trial. Lancet Infect Dis
15(1):27–35
16. Zeuzem S, Ghalib R, Reddy KR, Pockros PJ, Ben Ari Z, Zhao
Y etal (2015) Grazoprevir-elbasvir combination therapy for
treatment-naive cirrhotic and noncirrhotic patients with chronic
hepatitis C virus genotype 1, 4, or 6 infection: a randomized
trial. Ann Intern Med 163(1):1–13
17. Hayashi N, Izumi N, Kumada H, Okanoue T, Tsubouchi H, Yat-
suhashi H etal (2014) Simeprevir with peginterferon/ribavirin
for treatment-naïve hepatitis C genotype 1 patients in Japan:
CONCERTO-1, a phase III trial. J Hepatol 61(2):219–227
18. Izumi N, Hayashi N, Kumada H, Okanoue T, Tsubouchi H, Yat-
suhashi H etal (2014) Once-daily simeprevir with peginter-
feron and ribavirin for treatment-experienced HCV genotype
1-infected patients in Japan: the CONCERTO-2 and CON-
CERTO-3 studies. J Gastroenterol 49(5):941–953
19. Hayashi N, Seto C, Kato M, Komada Y, Goto S (2014) Once-
daily simeprevir (TMC435) with peginterferon/ribavirin for
treatment-naïve hepatitis C genotype 1-infected patients in
Japan: the DRAGON study. J Gastroenterol 49(1):138–147
20. Lawitz E, Mangia A, Wyles D, Rodriguez-Torres M, Hassanein
T, Gordon SC etal (2013) Sofosbuvir for previously untreated
chronic hepatitis C infection. N Engl J Med 368(20):1878–1887
21. Jacobson IM, Gordon SC, Kowdley KV, Yoshida EM, Rodriguez-
Torres M, Sulkowski MS, POSITRON Study, FUSION Study etal
(2013) Sofosbuvir for hepatitis C genotype 2 or 3 in patients with-
out treatment options. N Engl J Med 368(20):1867–1877
22. Manns M, Pol S, Jacobson IM, Marcellin P, Gordon SC, Peng CY,
HALLMARK-DUAL Study Team etal (2014) All-oral daclatasvir
plus asunaprevir for hepatitis C virus genotype 1b: a multina-
tional, phase 3, multicohort study. Lancet 384(9954):1597–1605
23. Bacon BR, Gordon SC, Lawitz E, Marcellin P, Vierling JM,
Zeuzem S, HCV RESPOND-2 Investigators et al (2011)
Boceprevir for previously treated chronic HCV genotype 1 infec-
tion. N Engl J Med 364(13):1207–1217
24. Sherman KE, Flamm SL, Afdhal NH, Nelson DR, Sulkowski MS,
Everson GT, ILLUMINATE Study Team etal (2011) Response-
guided telaprevir combination treatment for hepatitis C virus
infection. N Engl J Med 365(11):1014–1024
25. Afdhal N, Zeuzem S, Kwo P, Chojkier M, Gitlin N, Puoti M,
ION-1 Investigators et al (2014) Ledipasvir and sofosbu-
vir for untreated HCV genotype 1 infection. N Engl J Med
370(20):1889–1898

16. Internal and Emergency Medicine1 3
26. Afdhal N, Reddy KR, Nelson DR, Lawitz E, Gordon SC, Schi 
E, ION-2 Investigators etal (2014) Ledipasvir and sofosbuvir
for previously treated HCV genotype 1 infection. N Engl J Med
370(16):1483–1493
27. Kowdley KV, Gordon SC, Reddy KR, Rossaro L, Bernstein DE,
Lawitz E, ION-3 Investigators etal (2014) Ledipasvir and sofos-
buvir for 8 or 12weeks for chronic HCV without cirrhosis. N Engl
J Med 370(20):1879–1888
28. Mizokami M, Yokosuka O, Takehara T, Sakamoto N, Korenaga
M, Mochizuki H etal (2015) Ledipasvir and sofosbuvir xed-dose
combination with and without ribavirin for 12weeks in treatment-
naive and previously treated Japanese patients with genotype 1
hepatitis C: an open-label, randomised, phase 3 trial. Lancet Infect
Dis 15(6):645–653
29. Buti M, Agarwal K, Horsmans Y, Sievert W, Janczewska E,
Zeuzem S etal (2014) Telaprevir twice daily is noninferior to
telaprevir every 8 hours for patients with chronic hepatitis C. Gas-
troenterology. 146(3):744.e3–753.e3
30. Andreone P, Colombo MG, Enejosa JV, Koksal I, Ferenci P,
Maieron A etal (2014) ABT-450, ritonavir, ombitasvir, and dasa-
buvir achieves 97% and 100% sustained virologic response with
or without ribavirin in treatment-experienced patients with HCV
genotype 1b infection. Gastroenterology. 147(2):359.e1–365.e1
31. Ferenci P, Bernstein D, Lalezari J, Cohen D, Luo Y, Cooper C,
PEARL-III Study, PEARL-IV Study etal (2014) ABT-450/r-
ombitasvir and dasabuvir with or without ribavirin for HCV. N
Engl J Med 370(21):1983–1992
32. Forns X, Lawitz E, Zeuzem S, Gane E, Bronowicki JP, Andreone
P etal (2014) Simeprevir with peginterferon and ribavirin leads
to high rates of SVR in patients with HCV genotype 1 who
relapsed after previous therapy: a phase 3 trial. Gastroenterology.
146(7):1669.e3–1679.e3
33. Jacobson IM, Dore GJ, Foster GR, Fried MW, Radu M, Rafalsky
VV etal (2014) Simeprevir with pegylated interferon alfa 2a plus
ribavirin in treatment-naive patients with chronic hepatitis C virus
genotype 1infection (QUEST-1): a phase 3, randomised, double-
blind, placebo-controlled trial. Lancet 384(9941):403–413
34. Manns M, Marcellin P, Poordad F, de Araujo ES, Buti M, Hors-
mans Y etal (2014) Simeprevir with pegylated interferon alfa
2a or 2b plus ribavirin in treatment-naive patients with chronic
hepatitis C virus genotype 1 infection (QUEST-2): a ran-
domised, double-blind, placebo-controlled phase 3 trial. Lancet
384(9941):414–426
35. Zeuzem S, Andreone P, Pol S, Lawitz E, Diago M, Roberts S,
REALIZE Study Team etal (2011) Telaprevir for retreatment of
HCV infection. N Engl J Med 364(25):2417–2428
36. Zeuzem S, Jacobson IM, Baykal T, Marinho RT, Poor-
dad F, Bourli?re M et al (2014) Retreatment of HCV with
ABT-450/r-ombitasvir and dasabuvir with ribavirin. N Engl J
Med 370(17):1604–1614
37. Poordad F, McCone J Jr, Bacon BR, Bruno S, Manns MP,
Sulkowski MS, SPRINT-2 Investigators etal (2011) Boceprevir
for untreated chronic HCV genotype 1 infection. N Engl J Med
364(13):1195–1206
38. Poordad F, Hezode C, Trinh R, Kowdley KV, Zeuzem S, Agarwal
K etal (2014) ABT-450/r-ombitasvir and dasabuvir with ribavirin
for hepatitis C with cirrhosis. N Engl J Med 370(21):1973–1982
39. Zeuzem S, Dusheiko GM, Salupere R, Mangia A, Flisiak
R, Hyland RH, VALENCE Investigators etal (2014) Sofos-
buvir and ribavirin in HCV genotypes 2 and 3. N Engl J Med
370(21):1993–2001
40. Bedossa P, Poynard T (1996) An algorithm for the grading of
activity in chronic hepatitis C. The METAVIR Cooperative Study
Group. Hepatology 24(2):289–293
41. Ambrosino P, Lupoli R, Tarantino P, Di Minno A, Tarantino L, Di
Minno MN (2015) Viral hepatitis and anti-phospholipid antibod-
ies positivity: a systematic review and meta-analysis. Dig Liver
Dis 47(6):478–487
42. Enger C, Forssen UM, Bennett D, Theodore D, Shantakumar S,
McAfee A (2014) Thromboembolic events among patients with
hepatitis C virus infection and cirrhosis: a matched-cohort study.
Adv Ther 31(8):891–903
43. D’Agostino RB Sr, Vasan RS, Pencina MJ, Wolf PA, Cobain M,
Massaro JM etal (2008) General cardiovascular risk prole for
use in primary care: the Framingham Heart Study. Circulation
117(6):743–753
44. Heit JA, Spencer FA, White RH (2016) The epidemiology of
venous thromboembolism. J Thromb Thrombolysis 41(1):3–14
45. Bosch J, Garcia-Pagan JC (2000) Complications of cirrhosis. 1.
Portal hypertension. J Hepatol 32(1 Suppl):141–156
46. Gores GJ, Wiesner RH, Dickson ER, Zinsmeister AR, Jorgensen
RA, Langworthy A (1989) Prospective evaluation of esophageal
varices in primary biliary cirrhosis: development, natural history,
and inuence on survival. Gastroenterology 96(6):1552?1559
47. Ferro D, Angelico F, Caldwell SH, Violi F (2012) Bleeding and
thrombosis in cirrhotic patients: what really matters? Dig Liver
Dis 44(4):275–279
48. Stuijver DJ, Romualdi E, van Zaane B, Bax L, Büller HR, Gerdes
VE etal (2015) Under-reporting of venous and arterial thrombotic
events in randomized clinical trials: a meta-analysis. Intern Emerg
Med 10(2):219–246
49. European Centre for Disease Prevention and Control (2016) Sys-
tematic review on hepatitis B and C prevalence in the EU/EEA.
ECDC, Stockholm