Clin Drug Investig (2014) 34:681–690
DOI 10.1007/s40261-014-0222-1
SYSTEMATIC REVIEW
Bevacizumab Increases the Risk of Severe Congestive Heart
Failure in Cancer Patients: An Up-to-Date Meta-Analysis
with a Focus on Different Subgroups
Wei-Xiang Qi • Shen Fu • Qing Zhang •
Xiao-Mao Guo
Published online: 6 August 2014
Ó Springer International Publishing Switzerland 2014
Abstract
Background and Objective Congestive heart failure
(CHF) risk with bevacizumab in breast cancer has been
previously investigated in a meta-analysis, but its incidence
and the risk of CHF in other tumor types remain unclear.
Thus, we performed this meta-analysis to gather current
data and evaluate the risk of CHF with bevacizumab in
cancer patients, with a focus on different subgroups.
Methods The databases of PubMed and abstracts pre-
sented at the American Society of Clinical Oncology up to
31 December 2013 were searched for relevant articles.
Statistical analyses were conducted to calculate the sum-
mary incidence, relative risk (RR), and 95 % confidence
intervals (CIs) by using either random-effects or fixed-
effect models according to the heterogeneity of included
studies.
Results A total of 16,962 patients from 19 RCTs were
included. The use of bevacizumab significantly increased
Electronic supplementary material The online version of this
article (doi:10.1007/s40261-014-0222-1) contains supplementary
material, which is available to authorized users.
W.-X. Qi
S. Fu (&)
Q. Zhang
X.-M. Guo
Department of Radiation Oncology, Fudan University Shanghai
Cancer Center, 270 Dong’An Road, Shanghai 200032, China
e-mail: fushen2014@sina.com
W.-X. Qi
e-mail: qiweixiang1113@163.com
Q. Zhang
e-mail: chenlingyan0302@163.com
X.-M. Guo
e-mail: qwxzlk@163.com
W.-X. Qi
S. Fu
Q. Zhang
Department of Radiation Oncology, Shanghai Proton and Heavy
Ion Center (SPHIC), 4365 Kang Xin Road, Shanghai, China
the risk of developing high-grade CHF in cancer patients
(RR 1.98, 95 % CI 1.30–3.02, p = 0.002), but not for all-
grade CHF (RR 1.14, 95 % CI 0.87–1.49, p = 0.33). Risk
might vary with bevacizumab dose and tumor types. RRs
for patients receiving bevacizumab at 5 and 2.5 mg/kg/
week were 2.25 (95 % CI 1.43–3.56) and 1.00 (95 % CI
0.33–3.05), respectively. High risks were observed in
patients with breast cancer (RR 2.43, 95 % CI 1.48–3.98),
renal cell cancer (RR 3.66, 95 % CI 0.41–33.01), and
glioblastoma (RR 4.90, 95 % CI 0.24–102.39). Addition-
ally, bevacizumab in combination with taxanes signifi-
cantly increased the risk of high-grade CHF (RR 2.15,
95 % CI 1.09–4.25, p = 0.027).
Conclusions Bevacizumab treatment significantly
increases the risk of developing high-grade CHF in cancer
patients. The risk may vary with bevacizumab dose and
tumor types. Clinicians should be aware of the risks of
CHF with the administration of this drug in cancer patients.
1 Introduction
Bevacizumab, a humanized monoclonal immunoglobulin
G1 antibody that targets vascular endothelial growth factor
(VEGF), is the first angiogenesis inhibitor widely used in
the treatment of various solid tumors. Currently, bev-
acizumab has been approved by the US FDA for the
treatment of metastatic colorectal cancer [1, 2], metastatic
renal cell cancer [3], non-small-cell lung cancer [4], met-
astatic human epidermal growth factor receptor-2 (HER-
2)-negative breast cancer [5], glioblastoma [6], and ovarian
cancer [7–9]. In addition, its efficacy in other malignancies
such as castration-resistant prostate cancer [10], gastric
cancer [11], and cervical cancer [12] has also been inves-
tigated, though the clinical benefits from bevacizumab are
682
limited. Congestive heart failure (CHF) has been reported
as a rare but potentially life-threatening adverse event with
bevacizumab in cancer patients [13]. In a previous meta-
analysis conducted by Choueiri et al. [14], a nearly fivefold
increase in the risk of CHF with bevacizumab in breast
cancer patients was observed. However, only 3,784
patients from five randomized controlled trials (RCTs) of
breast cancer are included for analysis [3, 15–18]; thus, the
power to investigate the risk of CHF with bevacizumab is
limited. In addition, two of the five included trials are
conference presentations; thus, the follow-up time for these
trials is limited and the safety data are immature. There-
fore, the incidence and risk of CHF with bevacizumab in
breast cancer hasn’t been well-defined. Additionally, the
incidence and risk of CHF in other tumor types remains
unknown, and the potential risk factors of CHF with bev-
acizumab, such as tumor types, treatment duration, and
concomitant drugs, haven’t been systematically investi-
gated. As a result, we conducted this up-to-date meta-
analysis of all available RCTs that particularly focused on
different subgroups in order to characterize which patient
populations are at particular high risk for CHF from
bevacizumab.
2 Materials and Methods
2.1 Data Sources
We conducted an independent review of citations from
PubMed between 1 January 2000 and 31 December 2013.
Key words were bevacizumab, avastin, and cancer. The
search was limited to prospective randomized controlled
clinical trials. The search strategy also used text terms such
as congestive heart failure, cardiotoxicity, angiogenesis
inhibitors, and vascular endothelial growth factor to iden-
tify relevant information. We also performed independent
searches using Web of Science databases between 1 Jan-
uary 2000 and 31 December 2013 to ensure that no clinical
trials were overlooked. A subsequent search was run in
February 2014 to identify any relevant publications added
to the databases since December 2013. Additionally, we
searched the clinical trial registration website http://www.
ClinicalTrials.gov to obtain information on the registered
prospective trials. We also searched abstracts and virtual
meeting presentations from the American Society of
Clinical Oncology (ASCO; http://www.asco.org/ASCO)
conferences that took place between January 2004 and
December 2013. We also reviewed FDA submission doc-
uments, the updated manufacturer’s package insert, and
pertinent review articles. The reference lists of identified
articles were examined for additional publications. Each
publication was reviewed and in cases of duplicate
W.-X. Qi et al.
publication only the most complete, recent, and updated
report of the clinical trial was included in the meta-
analysis.
2.2 Study Selection
The primary goal of our study was to determine the overall
incidence of CHF associated with bevacizumab and
establish the association between treatments with bev-
acizumab and the risk of developing CHF. Thus, only
prospective RCTs that directly compared cancer patients
treated with and without bevacizumab were included for
analysis. Phase I and single-arm phase II trials were
excluded due to lack of control groups. Clinical trials that
met the following criteria were included: (1) prospective
randomized controlled phase II or III trials involving can-
cer patients; (2) participants assigned to treatment with or
without bevacizumab in addition to concurrent chemo-
therapy and/or biological agent; and (3) available data
regarding events or incidence of CHF and sample size.
2.3 Data Extraction and Clinical Endpoint
Data extraction was conducted independently by two
investigators (W.-X.Q. and S.F.), and any discrepancy
between the reviewers was resolved by consensus. For each
study, the following information was extracted: first
author’s name, year of publication, phase of trials, number
of enrolled subjects, treatment arms, number of patients in
treatment and controlled groups, underlying malignancy,
median age, median treatment duration, median progres-
sion-free survival, adverse outcomes of interest (CHF), and
bevacizumab dosage. The following adverse outcomes
were considered as CHF events and were included in the
analyses: left ventricular ejection fraction decline or dys-
function, CHF (not specified), and cardiomyopathy. All-
and high-grade (Cgrade 3) adverse events, as assessed and
recorded according to the National Cancer Institute’s
common terminology criteria (NCI-CTC, version 2 or 3;
http://ctep.cancer.gov), which has been widely used in
cancer clinical trials, were extracted for analysis. In the
event a study reported high-grade but not low-grade CHF
events, no assumption of all-grade incidence was made.
2.4 Statistical Analysis
For the calculation of incidence, the number of patients
with CHF in the bevacizumab group and the total number
of patients receiving bevacizumab were extracted from the
selected clinical trials; the proportion of patients with CHF
and the 95 % confidence interval (CI) were derived for
each study. To calculate relative risk (RR), patients
assigned to bevacizumab were compared only with those
CHF Associated with Bevacizumab
assigned to control treatment in the same trial. For one
study that reported zero events in the treatment or control
arm, we applied the classic half-integer correction to cal-
culate the RR and variance [19]. We explored a potential
dose-effect relationship by further dividing bevacizumab
therapy into low dose (5 or 7.5 mg/kg per dose per sche-
dule, which is equivalent to 2.5 mg/kg/week) and high
dose (10 or 15 mg/kg per dose per schedule, which is
equivalent to 5 mg/kg/week). We also conducted the fol-
lowing prespecified subgroup analyses to find the potential
risk factor of CHF: tumor types, concomitant drugs, med-
ian treatment duration, and quality of trials. To assess the
stability of results, sensitivity analysis was performed by
sequential omission of individual studies. Additionally, to
test whether effect sizes were moderated by differences in
length of treatment, we carried out meta-regressions with
difference in median length of experimental treatments
(expressed in months) as the predictor and RR as depen-
dent variable. Between-study heterogeneity was estimated
using the Chi-square (v2)-based Q statistic [20]. Hetero-
geneity was considered statistically significant when Phet-
erogeneity \0.1. If heterogeneity existed, the pooled estimate
calculated based on the random-effects model was reported
using the DerSimonian and Laird method [21]. In the
absence of heterogeneity, the pooled estimate calculated
based on the fixed-effects model was reported using the
inverse variance method. A statistical test with a p value
less than 0.05 was considered significant. The presence of
publication bias was evaluated by using the Begg and
Egger tests [22]. The Jadad scale was used to assess the
quality of included trials based on the reporting of the
studies’ methods and results [23]. All statistical analyses
were performed by using version 2 of the Comprehensive
MetaAnalysis program (Biostat, Englewood, NJ, USA) and
Open Meta-Analyst software version 4.16.12 (Tufts Uni-
versity, Boston, MA, USA).
3 Results
3.1 Search Results
Our search yielded 236 clinical studies relevant to bev-
acizumab. After excluding review articles, phase I studies,
single-arm phase II trials, case reports, editorials, letters,
commentaries, meta-analyses, and systematic reviews
(Fig. 1), we selected 18 RCTs, including 17 phase III trials
and one phase II trial (Table 2). One additional conference
abstract was identified as a result of searching the ASCO
conference database [7]. Finally, a total of 16,962 patients
from 19 RCTs were included for analysis. The character-
istics of patients and studies are listed in Table 1. Most
patients had good performance status and adequate
683
Fig. 1 Selection process for randomized controlled trials included in
the meta-analysis. CHF congestive heart failure
hematologic, hepatic, and renal function. Patients were
generally excluded from the individual trials if they had a
history of clinically significant cardiovascular disease or
dysrhythmias, prolongation of the QT interval, or uncon-
trolled hypertension. Underlying malignancies included
breast cancer [5, 15–18, 24–27] (nine trials), gastrointes-
tinal cancer [11, 28, 29] (three trials), gynecological cancer
[7–9] (three trials), renal cell cancer [3, 30] (two trials),
prostate cancer [10](one trial), and glioblastoma [31] (one
trial).
3.2 Overall Incidence of Congestive Heart Failure
(CHF)
For calculating the incidence of all-grade CHF, a total of
5,090 patients from nine RCTs who received bevacizumab
were included for analysis. CHF, regardless of grade,
occurred in 122 patients, representing a total incidence of
0.9 % (95 % CI 0.4–2.3; Fig. 2a). For high-grade CHF, a
total of 8,862 patients from 19 RCTs who received bev-
acizumab were included. Using a random-effects model,
the summary incidence of high-grade CHF (0.9 %, 95 %
CI 0.5–1.4; Fig. 2b) was comparable with that of all-grade
CHF. One possible explanation for this finding was that the
included studies tended only to report grade 3 or higher
events; thus, the total number of all-grade CHF events
might be under-reported.
3.3 Relative Risk (RR) of CHF
To determine the specific contribution of bevacizumab to
the development of CHF, and to exclude the effect of any
confounding factors, we calculated the overall RR of CHF
from these RCTs. For calculating the all-grade RR of CHF,
Table 1 Baseline characteristics of 19 randomized controlled trials included in the meta-analysis (n = 16,962)
684

Study/year/phase Histology Patients Treatment arm Median Median treatment Median PFS/ Median OS No. for No. of high- Reported
enrolled age duration (months) TTP (months) (months) analysis grade CHF events
(years)

Kabbinavar et al. CRC 209 BEV 2.5 mg/kg/wk ? FU/ 71.3 7.2 9.2 16.6 104 0 Heart failure
[29]/2005/II LV
Placebo ? FU/LV 70.7 5.4 5.5 12.9 105 1
Miller et al. [5]/ Breast cancer 462 BEV 5 mg/kg/ 51 NR 4.86 15.1 229 5 CHF
2005/III wk ? capecitabine
Capecitabine 52 NR 4.17 14.5 215 1
Escudier et al. [3]/ RCC 649 BEV 5 mg/kg/ 61 9.7 10.2 NR 337 1 CHF
2007/III wk ? interferon-a
Placebo ? interferon-a 60 5.1 5.4 19.8 304 0
Miller et al. [16]/ Breast cancer 722 BEV 5 mg/kg/wk ? PTX 56 7.1 11.3 25.6 365 8 CHF
2007/III PTX 55 5.1 5.8 24.8 346 1
Rini et al. [30]/ RCC 732 BEV 5 mg/kg/ 61 5.7 8.5 18.3 362 2 LV
2010/III wk ? interferon-a dysfunction
Interferon-a 62 2.9 5.2 17.4 347 0
Miles et al. [15]/ Breast cancer 736 BEV 2.5 mg/kg/wk ? TXT 54 NR 9 30.8 250 3 CHF
2010/III BEV 5 mg/kg/wk ? TXT 55 NR 10.1 30.2 247 0
Placebo ? TXT 55 NR 8.2 31.9 233 0
Brufsky et al. Breast cancer 648 BEV 5 mg/kg/ 55 6 7.2 18 458 4 LV
[18]/2011/III wk ? chemotherapy dysfunction
Placebo ? chemotherapy 55 4 5.1 16.4 221 0
Ohtsu [11] et al./ Gastric cancer 774 BEV 2.5 mg/kg/ 58 6.8 6.7 12.1 386 2 CHF
2011/III wk ? DDP ? CAP
Placebo ? DDP ? CAP 59 5.8 5.3 10.1 381 1
Perren et al. [9]/ Ovarian cancer 1,528 BEV 2.5 mg/kg/ 57 NR 19 NR 745 2 CHF
2011/III wk ? PTX ? CBP
PTX ? CBP 57 NR 17.3 NR 753 3
Robert et al. [17]/ Breast cancer 1,237 BEV 5 mg/kg/wk ? CAP 56 NR 8.6 NR 404 6 LV
2011/III Placebo ? CAP 57 NR 5.7 NR 201 1 dysfunction
BEV 5 mg/kg/wk ? taxanes 55 NR 9.2 NR 203 5
Placebo ? taxanes 55 NR 8 NR 102 0
BEV 5 mg/kg/ 55 NR 9.2 NR 210 13
wk ? anthracycline
Placebo ? anthracyclines 55 NR 8 NR 100 6
Aghajanian et al. Gastrointestinal 484 BEV 5 mg/kg/ 60 8.4 12.4 33.3 247 3 LV
[8]/2012/III cancer wk ? GEM ? DDP dysfunction/
Placebo ? GEM ? DDP 61 7 8.4 35.2 233 2 CHF
W.-X. Qi et al.
Table 1 continued
Study/year/phase Histology Patients Treatment arm Median Median treatment Median PFS/ Median OS No. for No. of high- Reported
enrolled age duration (months) TTP (months) (months) analysis grade CHF events
(years)

Bear et al. [27]/ Breast cancer 1,206 BEV 5 mg/kg/ NR NR NR NR 595 8 LV systolic
2012/III wk ? chemotherapy dysfunction
Chemotherapy NR NR NR NR 596 1
Kelly et al. [10]/ CRPC 1,050 BEV 5 mg/kg/ 68.8 5.6 9.9 22.6 504 0 LV systolic
2012/III wk ? TXT ? PDN dysfunction
Placebo ? TXT ? PDN 69.3 5.6 7.5 21.5 505 1
CHF Associated with Bevacizumab

von Minckwitz Breast cancer 1,948 BEV 5 mg/kg/ 49 NR NR NR 956 2 CHF

et al. [26]/2012/ wk ? chemotherapy
III chemotherapy 48 NR NR NR 969 0
Pujade-Lauraine Ovarian cancer 361 BEV 5 mg/kg/ NR NR 6.7 NR 179 1 CHF
et al. [7]/2012/ wk ? chemotherapy
III Chemotherapy NR NR 3.4 NR 182 1
Gianni et al. [24]/ Breast cancer 424 BEV 5 mg/kg/ 53 11.7 16.5 NR 215 11 Cardiac
2013/III wk ? TXT ? trastuzumab events
TXT ? trastuzumab 55 NR 13.7 NR 206 6
Cameron et al. Breast cancer 2,591 BEV 5 mg/kg/ 50 NR NR NR 1,271 7 LV
[25]/2013/III wk ? chemotherapy dysfunction
Chemotherapy 50 NR NR NR 1,288 3
Cunningham et al. CRC 280 BEV 2.5 mg/kg/wk ? CAP 76 5.8 9.1 20.7 134 0 CHF
[28]/2013/III CAP 77 4.2 5.1 16.8 136 1
Chinot et al. [31]/ Glioblastoma 921 BEV 5 mg/kg/ 57 NR 10.6 16.8 461 2 CHF
2014/III wk ? TMZ ? RT
Placebo ? TMZ ? RT 56 NR 6.2 16.7 450 0
BEV bevacizumab, CAP capecitabine, CBP carboplatin, CHF congestive heart failure, CRC colorectal cancer, CRPC castration-resistant prostate cancer, DDP cisplatin, FU/LV fluorouracil plus
leucovorin, GEM gemcitabine, LV left ventricular, NR not reported, OS overall survival, PDN prednisone, PFS progression-free survival, PTX paclitaxel, RCC renal cell cancer, RT
radiotherapy, TMZ temozolomide, TTP time to progression, TXT docetaxel
685
686
Fig. 2 Incidence of a all- and b high-grade congestive heart failure associated with bevacizumab
nine RCTs, representing 10,148 patients, were included.
Among the 5,090 patients treated with bevacizumab, 122
presented with all-grade CHF (incidence of 0.9 %; 95 %
CI 0.4–2.3), whereas 98 of 5,058 patients in controlled
groups (incidence of 0.5 %; 95 % CI 0.1–2.0) had an any-
grade CHF event. This conferred an overall RR of devel-
oping all-grade CHF of 1.14 (95 % CI 0.87–1.49,
p = 0.33; Fig. 3a). No significant heterogeneity was
observed in the RR analysis of all-grade CHF events
(Q = 12.42; p = 0.13; I2 = 35.6 %). Considering only
high-grade CHF events, 85 of 8,862 patients treated with
bevacizumab (incidence of 0.9 %; 95 % CI 0.5–1.4) and 29
of 7,873 patients in controlled arms (incidence of 0.5 %;
95 % CI 0.3–0.8) experienced a high-grade CHF event.
This conferred an overall RR of 1.98 (95 % CI 1.30–3.02,
p = 0.002; Fig. 3b). No significant heterogeneity was
observed in the RR analysis of high-grade CHF events
(Q = 11.02; p = 0.89; I2 = 0.0 %). We also performed
sensitivity analysis to examine the stability and reliability
of pooled RRs by sequential omission of individual studies.
The results indicated that the significance estimate of
pooled RRs was not significantly influenced by omitting
any single study (Electronic Supplementary Material Fig-
ure 1). Additionally, a meta-regression analysis was carried
W.-X. Qi et al.
out to test whether the RR of CHF varied as a function of
difference in the length of the experimental treatments. The
result indicated that the RR of CHF didn’t vary with
treatment duration (p = 0.75; Electronic Supplementary
Material Figure 2).
3.4 Subgroup Analysis for RR of High-Grade CHF
To determine whether the observed increase in RRs of
developing high-grade CHF was the result of confounding
bias, we performed subgroup analyses according to the
underlying malignancy, bevacizumab dosage, median
treatment duration, concomitant drugs, and quality of trials.
When stratified by tumor type, a significantly increased risk
of high-grade CHF was observed in breast cancer (RR 2.43,
95 % CI 1.48–3.98, p \ 0.001; Table 2), and a non-sig-
nificantly increased risk of high-grade CHF was observed
in renal cell cancer (RR 3.66, 95 % CI 0.41–33.01,
p = 0.25) and glioblastoma (RR 4.90, 95 % CI
0.24–102.39, p = 0.31), while no increased risk of high-
grade CHF was observed in gastrointestinal cancer (RR
0.77, 95 % CI 0.15–4.00, p = 0.76), gynecological cancer
(RR 0.98, 95 % CI 0.31–3.10, p = 0.98), and prostate
cancer (RR 0.33, 95 % CI 0.014–8.20, p = 0.50).
CHF Associated with Bevacizumab
Fig. 3 Relative risk of a all- and b high-grade congestive heart failure associated with bevacizumab versus control
However, clinicians should be cautious when interpreting
these results due to the limited RCTs of specific types of
cancer included for the RR calculation.
There were two approved doses for bevacizumab:
2.5 mg/week (low dose) and 5.0 mg/week (high dose).
Therefore, we attempted to look at the incidence and RR
stratified by bevacizumab. Our results showed that the risk
of high-grade CHF with bevacizumab seemed to be dose
dependent. The RR of high-grade CHF with low-dose
bevacizumab was 1.00 (95 % CI 0.33–3.05, p = 0.99;
Table 2), while the RR for the high-dose bevacizumab was
2.25 (95 % CI 1.43–3.56, p \ 0.001; Table 2). We also
carried out a subgroup risk analysis stratified according to
concomitant drug with bevacizumab. Our results showed
that bevacizumab in conjunction with taxanes (RR 2.15,
95 % CI 1.09–4.25, p = 0.027) significantly increased the
risk of high-grade CHF when compared with controls,
while a non-significantly increased risk of severe CHF was
observed when used in combination with interferon-a (RR
3.66, 95 % CI 0.41–33.01, p = 0.25), fluoropyrimidines
(RR 1.82, 95 % CI 0.60–5.52, p = 0.29), anthracyclines
(RR 1.19, 95 % CI 0.48–2.90, p = 0.71), and gemcitabine
(RR 1.41, 95 % CI 0.24–8.22, p = 0.70). The risk of high-
grade CHF might be related to the length of treatment
687
duration with bevacizumab. Our results demonstrated that
the risk of CHF in long bevacizumab treatment was com-
parable with that in short bevacizumab treatment (RR 1.91
vs. 1.52; Table 2). Additionally, we performed a subgroup
analysis according to quality of trials (double-blinded vs.
open-label). Patients from open-label trials had an RR of
2.23 (95 % CI 1.19–4.21, p = 0.013), while patients from
placebo-controlled double-blind studies had an RR of 1.80
(95 % CI 1.02–3.16, p = 0.043; Table 2).
3.5 Publication Bias
No publication bias was detected for the primary endpoint
of this study (RR of high-grade CHF) by either the Egger
or the Begg test (RR of severe CHF: Begg’s test p = 0.75;
Egger’s test p = 0.72).
4 Discussion
CHF is a rare but potentially life-threatening complication
of VEGF signaling pathway inhibitors, and concerns have
arisen regarding the risk of CHF with the use of these
drugs. Sunitinib, a small-molecule receptor tyrosine kinase
688 W.-X. Qi et al.

Table 2 Incidence and relative risk of high-grade congestive heart failure with bevacizumab according to prespecified subgroups
Subgroup No. of studies Severe CHF no./total no. Incidence (95 % CI) [%] RR (95 % CI) p value
Bevacizumab Control

Overall 19 85/8,862 29/7,873 0.9 (0.5–1.4) 1.98 (1.30–3.02) 0.002

Tumor types
Breast cancer 9 72/5,403 19/4,477 1.4 (0.8–2.4) 2.43 (1.48–3.98) \0.001
Renal cell cancer 2 3/699 0/651 0.4 (0.1–1.4) 3.66 (0.41–33.01) 0.25
Glioblastoma 1 2/461 0/450 0.4 (0.1–1.7) 4.90 (0.24–102.39) 0.31
Gastrointestinal cancer 3 2/624 3/622 0.5 (0.2–1.5) 0.77 (0.15–4.00) 0.76
Gynecological cancer 3 6/1,171 6/1,168 0.6 (0.2–1.6) 0.98 (0.31–3.10) 0.98
CRPC 1 0/504 1/505 0 0.33 (0.014–8.20) 0.50
Bevacizumab dosea
Low dose: 2.5 mg/kg/wk 5 7/1,619 6/1,608 0.6 (0.3–1.2) 1.00 (0.33–3.05) 0.99
High dose: 5.0 mg/kg/wk 15 78/7,233 23/6,498 1.0 (0.6–1.4) 2.25 (1.43–3.56) \0.001
Concomitant therapy
Taxanes 7 37/3,124 12/2,741 1.3 (0.6–2.8) 2.15 (1.09–4.25) 0.027
Interferon-a 2 3/699 0/651 0.4 (0.1–1.4) 3.66 (0.41–33.01) 0.25
Fluoropyrimidines 5 13/1,257 5/1,038 1.3 (0.7–2.3) 1.82 (0.60–5.52) 0.29
Anthracyclines 2 15/1,166 6/1,069 1.2 (0.1–26.8) 1.19 (0.48–2.90) 0.71
Gemcitabine 1 3/247 2/233 1.2 (0.4–3.7) 1.41 (0.24–8.22) 0.70
Treatment duration
C7.1 months 5 23/1,268 10/1,194 1.7 (0.7–4.1) 1.91 (0.91–4.03) 0.089
\7.1 months 5 8/1,844 3/1,590 0.6 (0.3–1.1) 1.52 (0.42–5.53) 0.53
Not reported 9 54/5,750 16/5,089 0.8 (0.4–1.6) 2.12 (1.21–3.70) 0.008
Quality of trials
Double-blinded trials 10 46/5,082 15/4,123 0.7 (0.4–1.4) 1.80 (1.02–3.16) 0.043
Open-label trials 9 39/3,780 14/3,750 1.1 (0.5–2.2) 2.23 (1.19–4.21) 0.013
CHF congestive heart failure, CRPC castration-resistant prostate cancer, RR relative risk
a
One trial was a three-arm study containing low-dose and high-dose bevacizumab treatment

inhibitor targeting the VEGF receptor and several other
kinases, is associated with an increased risk of CHF in
cancer patients [32, 33], and the risk of CHF with bev-
acizumab in breast cancer has also been investigated in a
previous meta-analysis conducted by Choueiri et al. [14].
However, our meta-analysis differs in several aspects from
prior study: we analyze the risk of CHF in a broader
spectrum of tumors by using data from all published
studies and we were able to include more randomized
controlled studies than that previous meta-analysis. In
addition, we performed subgroup analysis to identify the
potential risk factors for bevacizumab-related CHF.
Our study includes a total of 16,962 patients from 19
RCTs. Despite the low incidence of high-grade CHF with
bevacizumab (0.9 %, 95 % CI 0.4–2.3), our results show
that the use of bevacizumab significantly increases the risk
of developing high-grade CHF (RR 1.98, 95 % CI
1.30–3.02, p = 0.002), but not all-grade CHF (RR 1.14,
95 % CI 0.87–1.49, p = 0.33), in cancer patients. Sensi-
tivity analysis indicates that the significance estimate of
pooled RRs of high-grade CHF is stable and reliable. We
also performed a meta-regression analysis to test whether
the RR of severe CHF varies as a function of difference in
the length of the experimental treatments. The result indi-
cates that the RR of CHF doesn’t vary with treatment
duration.
Risk factors for CHF associated with bevacizumab are
poorly understood. As such, we then carried out a subgroup
analysis to identify potential risk factors for bevacizumab
related-CHF. When stratified by tumor types, a twofold
increased risk of severe CHF with bevacizumab is observed
in breast cancer, which is lower than previously reported
(RR 2.43 vs. 4.74) [14]. The high risk of CHF with breast
cancer may be due to prior or concomitant exposure to
other cardiotoxic medications in the neoadjuvant, adjuvant,
or metastatic settings. Patients with renal cell cancer in our
study also have a high risk of severe CHF; the potential
reason for this finding is that patients with renal cell cancer
have usually undergone nephrectomy, and might have an
underlying decreased clearance of bevacizumab. The use of
CHF Associated with Bevacizumab
high-dose bevacizumab in the trials that include patients
with renal cell cancer would also confer a higher risk of
CHF. The potentially higher RR of bevacizumab-related
CHF seen in patients with glioblastoma is puzzling, but
may be due to the unique biology of glioblastoma. Inter-
estingly, a relatively low risk of severe CHF is observed in
gastrointestinal cancer, gynecological cancer, and castra-
tion-resistant prostate cancer. However, caution should be
taken when analyzing this subgroup analysis because of
limited RCTs of specific cancer types included for analysis.
When stratifying trials according to their concomitant
drugs, we found that the use of bevacizumab significantly
increases the risk of high-grade CHF when used in con-
junction with taxanes, and a higher risk of CHF is also
observed in trials using bevacizumab in combination with
interferon-a. We also performed a subgroup analysis to
investigate whether the risk of high-grade CHF with bev-
acizumab is dose dependent. Our results showed that the
RR of CHF at a dose of 5 mg/kg/week seems greater than
at 2.5 mg/kg/week. Additionally, no significant differences
in the risk of CHF with bevacizumab are found according
to bevacizumab treatment duration and the quality of trials.
Based on our results, we could conclude that bevacizumab-
containing therapy significantly increases the risk of high-
grade CHF in comparison with controls. High bevacizumab
dose is a potential risk factor for CHF with bevacizumab.
Clinicians should be cautious when adding bevacizumab to
the treatment regimen of breast cancer, renal cell cancer,
and glioblastoma. A close monitoring for bevacizumab-
related CHF is recommended during treatment.
Despite the size of this meta-analysis of almost 17,000
patients, several limitations needed to be considered. First,
this is a meta-analysis at study level, and individual patient
information is not available. Therefore, confounding vari-
ables at the patient level, such as co-morbidities and prior
exposure to cardiotoxic agents, could not be incorporated
into the analysis. Secondly, our analyses are based on trials
primarily designed to demonstrate efficacy. Therefore, the
sample size and time horizon of individual RCTs may not
have been sufficient to detect CHF. Also, misclassification
may occur because the diagnosis of CHF is complicated
and may often appear as a cardiomyopathy diagnosis. We
include cardiomyopathy with CHF to help minimize these
potential reporting errors. Thirdly, not all of the studies
report low-grade CHF events because the included studies
tend only to report grade 3 or higher events; therefore, the
total number of all-grade CHF events is likely to be under-
reported, and the calculated incidence from this analysis
may in fact be an under-representation of all-grade CHF.
Fourthly, all of these studies exclude patients with poor
renal, hematological, and hepatic function at study entry,
so the overall incidence and risk may be higher in medical
practice. Fifthly, our meta-analysis pools trials with
689
heterogeneous tumor types, bevacizumab dosage, treatment
regimens, periods of study conduct, and treatment line, and
all of these factors would increase the clinical heteroge-
neity among included trials, which also makes the inter-
pretation of a meta-analysis more problematic, although we
perform subgroup and sensitivity analyses for key differ-
ences. Finally, as in all meta-analyses, our results may be
biased as a result of potential publication bias. However, a
funnel plot evaluation for the primary endpoint does not
indicate publication bias.
5 Conclusion
Our study has shown that the use of bevacizumab almost
doubles the risk of high-grade CHF when compared with
controls. The risk might be dependent on bevacizumab
dose, with an increased risk in patients receiving high
doses. The risk might also vary with tumor types, with high
risks seen in patients with breast cancer, renal cell cancer,
and glioblastoma. Overall, however, the clear benefits of
bevacizumab therapy are likely to outweigh the risks in
most patients. In the appropriate clinical scenario, the use
of bevacizumab remains justified in its approved
indications.
Acknowledgments We are indebted to the authors of the primary
studies, as without their contributions, this work would have been
impossible.
Conflicts of interest statement All authors declare that they have
no potential conflicts of interests.
References
1. Hurwitz H, Fehrenbacher L, Novotny W, Cartwright T, Hains-
worth J, Heim W, et al. Bevacizumab plus irinotecan, fluoro-
uracil, and leucovorin for metastatic colorectal cancer. N Engl J
Med. 2004;350(23):2335–42.
2. Hurwitz HI, Fehrenbacher L, Hainsworth JD, Heim W, Berlin J,
Holmgren E, et al. Bevacizumab in combination with fluorouracil
and leucovorin: an active regimen for first-line metastatic colo-
rectal cancer. J Clin Oncol. 2005;23(15):3502–8.
3. Escudier B, Pluzanska A, Koralewski P, Ravaud A, Bracarda S,
Szczylik C, et al. Bevacizumab plus interferon alfa-2a for treat-
ment of metastatic renal cell carcinoma: a randomised, double-
blind phase III trial. Lancet. 2007;370(9605):2103–11.
4. Sandler A, Gray R, Perry MC, Brahmer J, Schiller JH, Dowlati A,
et al. Paclitaxel-carboplatin alone or with bevacizumab for non-
small-cell lung cancer. N Engl J Med. 2006;355(24):2542–50.
5. Miller KD, Chap LI, Holmes FA, Cobleigh MA, Marcom PK,
Fehrenbacher L, et al. Randomized phase III trial of capecitabine
compared with bevacizumab plus capecitabine in patients with
previously treated metastatic breast cancer. J Clin Oncol.
2005;23(4):792–9.
690
6. Friedman HS, Prados MD, Wen PY, Mikkelsen T, Schiff D,
Abrey LE, et al. Bevacizumab alone and in combination with
irinotecan in recurrent glioblastoma. J Clin Oncol.
2009;27(28):4733–40.
7. Pujade-Lauraine E, Hilpert F, Weber B, Reuss A, Poveda A,
Kristensen G, et al. AURELIA: A randomized phase III trial
evaluating bevacizumab (BEV) plus chemotherapy (CT) for
platinum (PT)-resistant recurrent ovarian cancer (OC). J Clin
Oncol 2012;30:(suppl; abstr LBA5002).
8. Aghajanian C, Blank SV, Goff BA, Judson PL, Teneriello MG,
Husain A, et al. OCEANS: a randomized, double-blind, placebo-
controlled phase III trial of chemotherapy with or without bev-
acizumab in patients with platinum-sensitive recurrent epithelial
ovarian, primary peritoneal, or fallopian tube cancer. J Clin
Oncol. 2012;30(17):2039–45.
9. Perren TJ, Swart AM, Pfisterer J, Ledermann JA, Pujade-Lau-
raine E, Kristensen G, et al. A phase 3 trial of bevacizumab in
ovarian cancer. N Engl J Med. 2011;365(26):2484–96.
10. Kelly WK, Halabi S, Carducci M, George D, Mahoney JF, Sta-
dler WM, et al. Randomized, double-blind, placebo-controlled
phase III trial comparing docetaxel and prednisone with or
without bevacizumab in men with metastatic castration-resistant
prostate cancer: CALGB 90401. J Clin Oncol.
2012;30(13):1534–40.
11. Ohtsu A, Shah MA, Van Cutsem E, Rha SY, Sawaki A, Park SR,
et al. Bevacizumab in combination with chemotherapy as first-
line therapy in advanced gastric cancer: a randomized, double-
blind, placebo-controlled phase III study. J Clin Oncol.
2011;29(30):3968–76.
12. Tewari KS, Sill MW, Long HJ 3rd, Penson RT, Huang H, Ra-
mondetta LM, et al. Improved survival with bevacizumab in
advanced cervical cancer. N Engl J Med. 2014;370(8):734–43.
13. des Guetz G, Uzzan B, Chouahnia K, Morere JF. Cardiovascular
toxicity of anti-angiogenic drugs. Target Oncol.
2011;6(4):197–202.
14. Choueiri TK, Mayer EL, Je Y, Rosenberg JE, Nguyen PL, Azzi
GR, et al. Congestive heart failure risk in patients with breast
cancer treated with bevacizumab. J Clin Oncol.
2011;29(6):632–8.
15. Miles DW, Chan A, Dirix LY, Cortes J, Pivot X, Tomczak P,
et al. Phase III study of bevacizumab plus docetaxel compared
with placebo plus docetaxel for the first-line treatment of human
epidermal growth factor receptor 2-negative metastatic breast
cancer. J Clin Oncol. 2010;28(20):3239–47.
16. Miller K, Wang M, Gralow J, Dickler M, Cobleigh M, Perez EA,
et al. Paclitaxel plus bevacizumab versus paclitaxel alone for
metastatic breast cancer. N Engl J Med. 2007;357(26):2666–76.
17. Robert NJ, Dieras V, Glaspy J, Brufsky AM, Bondarenko I, Li-
patov ON, et al. RIBBON-1: randomized, double-blind, placebo-
controlled, phase III trial of chemotherapy with or without bev-
acizumab for first-line treatment of human epidermal growth
factor receptor 2-negative, locally recurrent or metastatic breast
cancer. J Clin Oncol. 2011;29(10):1252–60.
18. Brufsky AM, Hurvitz S, Perez E, Swamy R, Valero V, O’Neill V,
et al. RIBBON-2: a randomized, double-blind, placebo-con-
trolled, phase III trial evaluating the efficacy and safety of bev-
acizumab in combination with chemotherapy for second-line
treatment of human epidermal growth factor receptor 2-negative
metastatic breast cancer. J Clin Oncol. 2011;29(32):4286–93.
W.-X. Qi et al.
19. Choueiri TK, Schutz FA, Je Y, Rosenberg JE, Bellmunt J. Risk of
arterial thromboembolic events with sunitinib and sorafenib: a
systematic review and meta-analysis of clinical trials. J Clin
Oncol. 2010;28(13):2280–5.
20. Zintzaras E, Ioannidis JP. Heterogeneity testing in meta-analysis
of genome searches. Genet Epidemiol. 2005;28(2):123–37.
21. DerSimonian R, Laird N. Meta-analysis in clinical trials. Con-
trolled Clin Trials. 1986;7(3):177–88.
22. Sterne JA, Gavaghan D, Egger M. Publication and related bias in
meta-analysis: power of statistical tests and prevalence in the
literature. J Clin Epidemiol. 2000;53(11):1119–29.
23. Moher D, Pham B, Jones A, Cook DJ, Jadad AR, Moher M, et al.
Does quality of reports of randomised trials affect estimates of
intervention efficacy reported in meta-analyses? Lancet.
1998;352(9128):609–13.
24. Gianni L, Romieu GH, Lichinitser M, Serrano SV, Mansutti M,
Pivot X, et al. AVEREL: a randomized phase III Trial evaluating
bevacizumab in combination with docetaxel and trastuzumab as
first-line therapy for HER2-positive locally recurrent/metastatic
breast cancer. J Clin Oncol. 2013;31(14):1719–25.
25. Cameron D, Brown J, Dent R, Jackisch C, Mackey J, Pivot X,
et al. Adjuvant bevacizumab-containing therapy in triple-negative
breast cancer (BEATRICE): primary results of a randomised,
phase 3 trial. Lancet Oncol. 2013;14(10):933–42.
26. von Minckwitz G, Eidtmann H, Rezai M, Fasching PA, Tesch H,
Eggemann H, et al. Neoadjuvant chemotherapy and bevacizumab
for HER2-negative breast cancer. N Engl J Med.
2012;366(4):299–309.
27. Bear HD, Tang G, Rastogi P, Geyer CE Jr, Robidoux A, Atkins
JN, et al. Bevacizumab added to neoadjuvant chemotherapy for
breast cancer. N Engl J Med. 2012;366(4):310–20.
28. Cunningham D, Lang I, Marcuello E, Lorusso V, Ocvirk J, Shin
DB, et al. Bevacizumab plus capecitabine versus capecitabine
alone in elderly patients with previously untreated metastatic
colorectal cancer (AVEX): an open-label, randomised phase 3
trial. Lancet Oncol. 2013;14(11):1077–85.
29. Kabbinavar FF, Schulz J, McCleod M, Patel T, Hamm JT, Hecht
JR, et al. Addition of bevacizumab to bolus fluorouracil and
leucovorin in first-line metastatic colorectal cancer: results of a
randomized phase II trial. J Clin Oncol. 2005;23(16):3697–705.
30. Rini BI, Halabi S, Rosenberg JE, Stadler WM, Vaena DA, Archer
L, et al. Phase III trial of bevacizumab plus interferon alfa versus
interferon alfa monotherapy in patients with metastatic renal cell
carcinoma: final results of CALGB 90206. J Clin Oncol.
2010;28(13):2137–43.
31. Chinot OL, Wick W, Mason W, Henriksson R, Saran F, Nis-
hikawa R, et al. Bevacizumab plus radiotherapy-temozolomide
for newly diagnosed glioblastoma. N Engl J Med.
2014;370(8):709–22.
32. Richards CJ, Je Y, Schutz FA, Heng DY, Dallabrida SM, Moslehi
JJ, et al. Incidence and risk of congestive heart failure in patients
with renal and nonrenal cell carcinoma treated with sunitinib.
J Clin Oncol. 2011;29(25):3450–6.
33. Qi WX, Shen Z, Tang LN, Yao Y. Congestive heart failure risk in
cancer patients treated with VEGFR-TKIs: a systematic review
and meta-analysis of 36 clinical trials. Br J Clin Pharmacol. Epub
2014 Mar 25. doi:10.1111/bcp.12387.