The role of bevacizumab in the treatment of glioblastoma
Roberto Jose Diaz1 · Sheikh Ali2 · Mehreen Gull Qadir3 · Macarena I. De La Fuente4 · Michael E. Ivan5 · Ricardo J. Komotar5

Received: 11 October 2016 / Accepted: 15 May 2017
© Springer Science+Business Media New York 2017

Abstract Bevacizumab has been used in patients with GBM as a salvage therapy since its approval in the United States for recurrent GBM in 2009. In order to review the therapeutic effect of bevacizumab in the primary and recur- rent clinical setting we have performed a systematic analy- sis of data from the published literature. Weighted median progression free survival and overall survival were calcu- lated and compared to standard therapy or other experi- mental therapies. A qualitative analysis of the limited stud- ies on health related quality of life and effects on steroid requirements was also undertaken. We found that the avail- able literature supports the use of bevacizumab for prolong- ing PFS and OS in the recurrent setting either alone or in combination with a cytotoxic agent (P < 0.05), but does not support its use in the primary setting (P > 0.05). The sur- vival advantage of bevacizumab compared to experimental therapy at recurrence is limited to 4 months. There is no

Electronic supplementary material The online version of this article (doi:10.1007/s11060-017-2477-x) contains supplementary material, which is available to authorized users.

* Roberto Jose Diaz [email protected]
1 Department of Neurology and Neurosurgery, Montreal Neurological Institute and Hospital, McGill University, 3801 Rue University, Montreal H3A 2B4, QC, Canada
2 College of Osteopathic Medicine, Nova Southeastern University, Fort Lauderdale, FL, USA
3 Department of Biological Sciences, Florida International University, Miami, FL, USA
4 Department of Neurology, University of Miami Miller School of Medicine, Miami, FL, USA
5 Department of Neurological Surgery, University of Miami Miller School of Medicine, Miami, FL, USA

additional benefit reported to date in health-related quality of life with the use of bevacizumab, although it may reduce steroid requirements. On average there is one side-effect event per patient and 74% of these events are grade 3 toxic- ity or higher. Further studies investigating the role of beva- cizumab in combination with cytotoxic agents at recurrence are awaited.
Keywords Avastin · Bevacizumab · VEGF · Glioblastoma · Chemotherapy · Survival

Introduction
Glioblastoma (GBM) is the most common malignant pri- mary brain tumor in adults. Despite advancements in GBM treatment, such as the Stupp protocol,[1] it remains an incurable disease. While an effective medical therapy remains elusive, surgical resection of recurrent tumor in a selected group of patients appears beneficial [2]. Multiple adjuvant therapies have been proposed at the time of GBM recurrence including cytotoxic chemotherapy, angiogenesis inhibitors, or immunotherapy [3]. The histology of GBM is characterized by marked microvascular proliferation and disruption of the blood brain barrier, which results in con- trast enhancement on MRI. This is associated with overex- pression of the vascular endothelial growth factor (VEGF- A) [4]. Bevacizumab is an IgG1 humanized monoclonal antibody that neutralizes the effect of VEGF-A [2]. The FDA has approved bevacizumab for use in the treatment of multiple cancers including colon (2004), lung (2006), kid- ney (2009), and cervix (2014) cancers. In 2009, the FDA approved the use of bevacizumab in treating adult patients diagnosed with recurrent GBM based on the success of two Phase 2 clinical trials [5, 6]. Subsequent clinical use of

bevacizumab has brought into question its utility in mono- therapy and its overall contribution to extending patient survival in the primary and recurrent setting. When beva- cizumab is used in combination with standard treatment in the primary setting, no additional overall survival (OS) benefit has been observed in two separate randomized con- trolled studies [4, 7]. In this systematic review, we sum- marize the current evidence for use of bevacizumab in the treatment of GBM in the primary or recurrent setting, high- lighting the effect size on progression free and OS. We also provide an estimate of the toxicities experienced by patients with GBM and the impact on quality of life.

Methods
Literature search
We searched PubMed for English language articles between January 2000 and April 2016 using the following phrases: “GBM AND Avastin” and “Astrocytoma Grade IV AND bevacizumab” and “GBM AND Avastin AND Quality of Life”. Articles were first evaluated for relevance based on whether GBM and bevacizumab were a focus. Laboratory studies or those that did not have full text available were excluded. All other articles were included for subsequent screening. The secondary screen of articles consisted of mining references for additional relevant studies and excluding case reports, commentaries, editorials, and prior reviews. Expert opinion on important studies published as abstracts was also considered. Only studies involving adults (≥18 years of age) were selected. In order to specify the type of tumor, only literature that specifically included patients with the diagnosis of GBM were used. Retrospec- tive and prospective studies, and phase I, II, and III clinical trials were selected for data analysis. A PRISMA flow chart with the number of articles screened is shown in Fig. 1. Tables describing the relevant studies were generated and studies were separated into use of bevacizumab in the pri- mary and recurrent setting. Effects on quality of life were assessed based on specific measures used in individual studies.
Data extraction
The studies included were analyzed based on histological confirmation of diagnosis, patient population, and phar- macological treatment. The studies were separated based on whether bevacizumab was single-agent or used in com- bination with other agents.. In addition, some articles dis- cussed GBM in addition to other high-grade gliomas. Data from GBM patients was extracted, and combined with the patient data from all other research articles that solely
discussed GBM. The primary outcome measures that were collected were median progression free survival (PFS) and median OS. A list of all toxicities reported for bevacizumab in patients with GBM was generated and a tally of patients affected as well total patients exposed was generated.
Statistical analysis
The median PFS and OS reported in each study and the number of patients in each study was tabulated. A weighted median and weighted standard deviation (SD) was calcu- lated for grouped studies using the weightedMedian and weightedSd functions in the matrixStats package available for R [8, 9]. The 95% confidence interval for the weighted median was calculated using the basic bootstrap method in R [9]. Significant differences were defined as no overlap in the bootstrap 95% confidence interval. DistillerSR Forest Plot Generator from Evidence Partners was used to create forest plots.

Results
A total of 52 relevant studies on the use of bevacizumab in patients with GBM were found in our literature search (Fig. 1). From these studies, 11 were retrospective and 9 were prospective case series. A total of 1 phase I, 22 phase II, and 4 phase III studies were evaluated. Six additional studies were included that discussed quality of life (QOL). Among all studies, only ten were randomized controlled trials. A total of 2163 patients were treated in these stud- ies. The median age of these patients ranged from 48 to 62 years.
Use at diagnosis
Bevacizumab has been used in the upfront setting in eight phase II and 2 phase III studies. Most commonly, beva- cizumab was used in addition to standard temozolomide based chemoradiotherapy. Table 1 shows the summary of all phase II/III studies involving the use of bevacizumab in the upfront setting [4, 7, 10–16]. The median PFS data and OS data is graphically represented in Figs. 2 and 3. The median PFS ranged from 31.7 to 61.7 weeks (weighted median PFS: 46.4 ± 7.0SD weeks, 95%CI 32.9–46.7) and the median OS ranged from 51.3 to 100 weeks (weighted median OS: 71.9 ± 11.0SD weeks, 95%CI 51.7–75.6). In the two phase III studies, the median PFS in patients receiving bevacizumab and stand- ard therapy was longer than in standard therapy alone by
4.4 months (P < 0.0001) in AVAGlio [7] and 3.4 months (P = 0.004) in RTOG0825 [4]. However, median OS was not different between treatment groups in both trials. In

Fig. 1 PRISMA flow chart for systematic review of bevacizumab use in the treatment of GBM

our analysis, the difference in weighted median PFS is statistically significant, but the weighted median OS is not statistically significant when comparing bevacizumab in combination with a cytotoxic chemotherapy versus Temozolomide alone in the primary setting [weighted median PFS: 46.4 ± 7.0SD weeks (95%CI 32.9–46.7)
versus 28.3 ± 7.3SD weeks (95%CI 19.4–33.4); weighted median OS: 71.9 ± 11.0SD weeks (95%CI 51.65–75.61)
versus 64.9 ± 10.0 weeks (95%CI 57.7–83.7)]. The sum- mary of studies used for determination of the weighted median PFS and OS for Temozolomide monotherapy in the primary setting is provided in Supplementary Table 1.
Use at recurrence
Only three studies reported the use of bevacizumab as monotherapy at recurrence, while 14 studies reported the use of bevacizumab combined with a secondary agent. Among the studies in which bevacizumab was used as monotherapy, one study involved the use of bevacizumab at first recurrence, one at first and second recurrence, and one did not discriminate the number of prior therapies. In eight of 17 studies in which bevacizumab was used at recurrence, the KPS of participating patients was ≥60%, while in nine of the 17 studies patients had a KPS ≥ 70%.

Table 1 Survival in patients treated with bevacizumab in combination and independently for newly diagnosed GBM

Author Year Treatment Phase # of Patients Median age
(range)

Lai et al. [12] 2011 Bev + TMZ II 70 57.4 (31.3–
75.8) 39:31 105.2 (NA) 59.1 (48.2– 85.2 (70–101)
Vredenburgh et al. [10] 2011 Bev + TMZ + iri- notecan II 75 55.6 (19–78) 45:30 100.8
(94.3–113) 61.7 (52.1– 92.1 (74.7–110)
Vredenburgh et al. [10] 2012 Bev + TMZ II 125 56.2 (19–80) 74:51 91.3 (NA) 60 (NA) 92.6 (81.7–
103.4)
Narayana 2012 Bev + TMZ II 51 54 (26–75) 31:20 69.5 (NA) 56.5 (NA) 99.9 (NA)
Hainsworth 2012 Bev + TMZ then II 68 59 (24–80) 39:29 73.9 (NA) 49.1 (40.4– 60.4 (53.8–NA)
et al. [14] Bev + everoli- 56.9)
mus
Hofland et al. [15]a 2014 Bev + TMZ II 32 62 (30–73) 21:11 134.7
(82.6–187) 33.5 (22.2–
44.3) 51.3 (35.6–
66.5)
Hofland et al. [15]a 2014 Bev + irinotecan II 31 59 (36–77) 18:13 134.7
(82.6–187) 31.7 (21.7–
40.4) 65.6 (41.7–
89.5)
Chinot et al. 2014 Bev + TMZ III 458 57 (20–84) 282:176 62.6 (NA) 46.1 (NA) 73 (NA)
Gilbert et al. [4] 2014 Bev III 320 57 (51–65)/59 (52–65)b 372:249 89.1 (0.43–
173) 46.5 68.2 (61.7–73)
Van Linde et al. [16] 2015 Bev + TMZ II 19 NA NA NA (NA) 41.7 (18.7– 69.5 (35.2–114)
Total patients:

Bevacizumab therapy in newly diagnosed GBM
M:F ratio Median F/U
(95% CI)
[weeks]
Median PFS (95% CI)
[weeks]

71.7)

69.5)
Median OS (95% CI)
[weeks]

et al. [13]

[7]

NA not available
aSame study, two sets of data
bArm 1 and arm 2, respectively

1249

(43.5–53)

62.6)

Thus, no patient requiring considerable assistance or medi- cal care, or confined to a bed or chair more than 50% of waking hours received bevacizumab at recurrence. Table 2 shows the summary of all phase II and III studies involv- ing the use of bevacizumab in the recurrent setting [5, 6, 17–27]. The median PFS data and OS data is graphically represented in Figs. 4 and 5. The median PFS in the mon- otherapy studies ranged from 13 to 18.3 weeks (weighted median PFS: 16.6 ± 2.2SD weeks, 95%CI 14.9–20.2) and the median OS ranged from 31 to 40 weeks (weighted median OS: 36.2 ± 3.8SD weeks, 95%CI 32.5–41.5). For
comparison, when GBM patients are treated with other therapies at recurrence the weighted median PFS and OS are 8.5 ± 2.3SD (95%CI 2.8–9.3) weeks and 22.4 ± 4.3SD
(95%CI 10.6–27.3) weeks respectively (Supplementary Table 1) [28–36]. When bevacizumab is used in combi- nation with a secondary agent at recurrence the median PFS ranges from 8 to 25.6 weeks (weighted median PFS:
18.2 ± 4.1SD weeks, 95%CI 13.9–20.5) and the median OS ranges from 15 to 44.6 weeks (weighted median OS:
39.5 ± 6.2SD weeks, 95%CI 39.5–44.8). The longest reported OS benefit for use of bevacizumab and a cytotoxic
chemotherapy at the time of GBM recurrence was for the combination of bevacizumab and lomustine (52.1 weeks median OS) [26 ]. In the only phase III study comparing bevacizumab monotherapy to combination therapy with lomustine (EORTC26101), median PFS was prolonged (4.2 vs. 1.5 months) with combination therapy but no difference in median OS was observed between treatment groups (9.1 vs. 8.6 months) [37] (Fig. 5).
Side effects
Among 503 patients with GBM treated in Phase II or III trials with bevacizumab at 10 mg/kg every 2 weeks as a single agent, a total of 563 side-effects were recorded. Table 3 lists all the described side effects of bevaci- zumab and their grade and frequency among treated GBM patients. A total of 417 events with Grade 3 or higher toxic- ity were recorded. The most frequent single side effect was hypertension (17%). The overall rate of leukopenia derived by combining reported lymphopenia, neutropenia, or leu- kopenia in Table 3 was 31.2% and this represented the most common hematological disturbance. The most common

Fig. 2 Forest plot demonstrating median PFS (MED PFS) for use of bevacizumab in the primary setting. POP study population, UCL upper 95% confidence limit, LCL lower 95% confidence limit

serious medical complication was a thromboembolic event (9.3%). Non-central nervous system hemorrhage was observed in 5.4% of all patients treated, but these were primarily grade 1 or 2 in severity. Central nervous system hemorrhage occurred in 0.4% of patients treated.
Effects on quality of life
The effect on health-related quality of life from adding bevacizumab to standard radiation and temozolomide chemotherapy in patients with GBM was investigated in the AVAglio trial [38]. A total of 450 treated patients who were randomized to treatment were followed for 1 year and their health related-quality of life was measured against that of patients randomized to standard therapy. This was done via a series of patient questionnaires at different intervals of treatment without the assistance of clinical investiga- tors. The trial demonstrated no adverse effect on HRQoL with the addition of bevacizumab to standard therapy. Fur- thermore, the decline in HRQoL correlated with tumor progression. No difference in quality of life measures was also observed by Herrlinger et al. via (QLQ) -C30 and QLQ-BN20 questionnaires when comparing patients with MGMT promoter non-methylated tumors treated with

bevacizumab and irinotecan versus standard therapy [39]. The RTOG 0825 trial was a randomized controlled trial in which patients received radiation with concurrent temozo- lomide and were then randomized to continue on bevaci- zumab or to receive placebo [4]. In that study, the EORTC QLQ-C30/BN20 questionnaire was used. The investigators observed a decline in neurocognitive function and quality of life measures of cognitive function, motor skills, and communication in the bevacizumab treated group com- pared to placebo [4]. The BELOB trial, which randomized patients to receive bevacizumab, lomustine, or a combina- tion of both, showed no difference in HRQoL (QLQ-C30 and QLQBCM20 questionnaires) among patients treated in any of the three arms; however, the trial was not powered to detect differences [40].

Discussion
In this study we have endeavored to generate a weighted estimate of the median PFS and median OS for the use of bevacizumab in the primary or recurrent setting when treating patients with GBM. Summarizing the multiple studies reviewed in this work, we highlight two clinically

Fig. 3 Forest plot demonstrating median OS (MED OS) for use of bevacizumab in the primary setting. POP study population, UCL upper 95% confidence limit, LCL lower 95% confidence limit

relevant findings. First, we have found that while there is a prolongation in median PFS with the use of bevacizumab in the primary setting, there is no significant difference in OS. These findings reflect well the results of the two ran- domized control trials AVAGlio and RTOG0825, which have as of yet not been published in full format. Secondly, in the setting of recurrent GBM, combination therapy with bevacizumab shows no benefit over the use of bevaci- zumab alone in patients with KPS of 60 or higher. While the overall analysis does not show a statistically significant advantage to combination therapy in general, synergistic interactions may still be observed for individual combina- tions. There is an OS advantage to using the combination of bevacizumab and lomustine over bevacizumab alone in the BELOB trial [26]. When using a weighted estimate and comparing to patients treated with other therapies at recur- rence, bevacizumab monotherapy and combination therapy both showed improved PFS and OS and the effect size was only slightly greater with combination therapy (delta PFS
9.7 weeks, delta OS 17.1 weeks) versus monotherapy (delta
PFS 8.1 weeks, delta OS 13.8 weeks). Thus, we cannot exclude a treatment effect for bevacizumab in the recur- rent setting when compared to other experimental cytotoxic therapy.
We recognize the limitations of using a weighted median estimate for comparing outcomes. Confounding factors include heterogeneity in the study populations, variabil- ity in the quality of studies, differences in treatment regi- mens, and variability in follow up. The weighted median is strongly influenced by the study size and thus results from larger studies will have a larger effect on the estimate. The use of other methods of evaluating outcomes from multi- ple studies, such as pooled median survival times and com- bined hazard ratios [41] were not possible due to missing or unavailable data even after contacting individual study investigators. Furthermore, we have sought to establish a measure that would be more easily interpreted by the clini- cian and patients when making treatment decisions.
One of the major hurdles in defining the clinical effect of bevacizumab in studies of recurrent GBM performed to date has been the lack of a control arm. The primary reason that there has not been a control arm in the phase II and III studies has been the lack of a standard chemo- therapy treatment regimen for recurrent GBM. Typically, patients with recurrent GBM will be offered treatment based on the extent of disease; the amount of tumor that can be resected at recurrence; the performance status; and the general medical condition. Other factors, such as access

Table 2 Survival in patients treated with bevacizumab in combination and independently for recurrent GBM

Author Year Treatment Phase # of patients Median Age
(Range)

Friedman et al. [6]a 2009 Bev + irinote- can II 82 57 (23–79) 57:25 8.7 (NA) 24.3 (19.1–
26.9) 37.8 (33.9–
47.4)
Verhoeff
et al. [17] 2010 Bev + TMZ I 15 50.9 (24.2– 10:5 NA (NA) 10.4 (NA) 15.7 (NA)
Hasselbach 2010 Bev + irinote- II 43 54 (23–70) 25:18 65.2 (30.4– 16 (13–20) 30 (23–37)

a. Bevacizumab combination therapy in recurrent GBM
M:F ratio Median F/U
(95% CI)
[weeks]
Median PFS (95% CI)
[weeks]
Median OS (95% CI)
[weeks]

et al. [18]

can + cetuxi- mab

68.7)

108.6)

Reardon et al. 2011 Bev + irinote- II 25 52 (19–76) 17:8 34.7 (27.2–48) 10 (7.8–15.6) 25.2 (17.4–
[21] can + carbo- 30.4)
platin
Reardon et al. 2012 Bev + irinote- II 40 51 (25.2–72.0) 30:10 66.9 (32.6–73) 25.6 (19.1– 36.1 (25.6–
[21] can + carbo- 36.1) 46.5)
Desjardins 2012 Bev + TMZ II 32 56 (25–80) 19:13 NA (NA) 15.8 (NA) 37 (NA)
Moller et al. [22] 2012 Bev + Irinote- can II 32 55 (31–72) NA NA (NA) 22.6 (13.5–
31.3) 34.3 (27.4–
41.7)
Lassen et al. [24] 2013 Bev + tem- II 13 48 (24–72) 5:8 NA (NA) 8 (NA) 15 (NA)
Soffietti et al. [23] 2014 Bev + fotemus- tine II 54 NA 35:19 40.4 (31.7–
45.6) 22.6 (16.5–
28.7) 39.5 (31.7–
44.8)
Lee et al. [25] 2015 Bev + panobi- nostat II 24 53 (22–66) 14:10 NA (NA) 21.7 (13-39.1) 39.1 (26.1–
82.6)
Taal et al. [26]b 2014 bev + lomus- tine II 52 NA NA NA (NA) 17.4 (13-34.8) 52.1 (34.8–
56.5)
Odia et al. [27] 2016 Bev + enzas- II 40 51 (25–73) 23:17 NA (NA) 8.7 (2.2–89.1) 32.6 (4.4-201.2)
Wick et al. [37] 2016 Bev + lomus- tine III 437 NA NA NA (NA) 18.25 (16.1–
18.7) 39.5 (35.2–
43.9)

Single-agent bev
acizumab in recurrent GBM Total patients:
Kreisl et al. 2009 Bev II 48 53 (21–69) 28:20 NA (NA) 16 (12–26) 31 (21–54)
Friedman et al. [6]a 2009 Bev II 85 54 (23–78) 58:27 8.7 (NA) 18.3 (12.6– 40 (35.6–46.5)
Taal et al. [26]b 2014 Bev II 50 NA NA NA (NA) 13 (13-17.4) 34.8 (26.1–
39.1)
Total patients:

Sathornsu- metee et al. [19]
2010 Bev + erlotinib II 25 52.4 (24.1–
70.4)
13:12 141.8 (141–
142.6)
18 (12.0–23.9) 44.6 (28.4–
68.7)

platin

et al. [20]

sirolimus

taurin

914

[5]

25.2)

NA not available
a,bSame study, two sets of data

183

to local phase II/III clinical trials can also affect patient dis- cussion and choice of treatment. Since no standard control arm can be defined in this population, trials have compared alternative regimens head-to-head. While this tells us a baseline comparison, it is not certain if the effective agent
is bevacizumab, which was added, or if there is a synergis- tic effect of combining bevacizumab with a cytotoxic agent.
Bevacizumab is routinely used either as monotherapy or in combination with cytotoxic agents at many clini- cal centers in North America for recurrent GBM. Both

Fig. 4 Forest plot demonstrating median PFS (MED PFS) for use of bevacizumab monotherapy versus combination therapy at GBM recurrence.
POP study population, UCL upper 95% confidence limit, LCL lower 95% confidence limit

monotherapy and combination therapy at recurrence showed longer PFS and OS compared to other experimen- tal agents, so that the clinical use as a therapy at recur- rence can be supported. Our analysis also showed there was a slight increase in median OS with combination therapy
versus monotherapy of 3.3 weeks; however, the difference was not statistically significant. Thus, monotherapy and combination therapy may have similar effect or the type of combinations used may be important determinants of the treatment effect. These results are concordant with the

Fig. 5 Forest plot demonstrating median OS (MED OS) for use of bevacizumab monotherapy versus combination therapy at GBM recurrence. POP study population, UCL upper 95% confidence limit, LCL lower 95% confidence limit

EORTC26101 study, which reported no difference in OS for bevacizumab monotherapy versus combination therapy with lomustine [37]. The results of NCT01860638, a phase

II randomized controlled trial comparing bevacizumab or placebo in addition to lomustine and standard of care in recurrent GBM, is expected to provide an answer as to the

Table 3 Side-effects of treatment with single-agent bevacizumab in all Phase II and Phase III trials in this review Complication Gr. 1 Gr. 1/2 Gr. 2 Gr. 3 Gr. 3/4 Gr. 4 Gr. 5 Total
number of events

% Total treated

Total # of patients treated = 503
Hypertension – 38 4 30 14 – – 86 17
Fatigue – 30 – 41 3 2 – 76 15.1
Lymphopenia – – – 51 2 15 – 68 13.5
Thrombocytopenia – – 2 29 – 33 – 64 12.7
Neutropenia/ neutropenic sepsis – – – 27 1 21 – 49 9.7
Thromboembolic event (both venous and arterial) – 2 – 19 5 20 1 47 9.3
Leukopenia – – 2 30 – 8 – 40 8.0
Non-CNS hemorrhage – 23 – 3 – – 1 27 5.4
Proteinuria 13 4 3 – – – – 20 4
Nausea and vomiting – 5 – 14 – – – 19 3.8
Wound healing comp./dehiscence – 3 – 6 2 1 – 12 2.4
Infection/wound infection/other infection – 7 – 3 – – – 10 2
Anemia – – – 5 – 3 – 8 1.6
Pulmonary/upper respiratory tract – 5 – 2 – – – 7 1.4
Convulsion – – – – 5 – – 5 1
Visceral perforation – – – 4 – 1 – 5 1
Hypokalemia – – – – 3 – – 3 0.6
Aphasia – – – – 3 – – 3 0.6
Hypophosphatemia – – 1 2 – – – 3 0.6
Thrombosis/DVT – – – – 2 – – 2 0.4
CNS hemorrhage – 2 – – – – – 2 0.4
Confusional state – 2 – – 2 – – 2 0.4
Diarrhea – 1 – – 1 – – 1 0.2
Pneumonia – 1 – – 1 – – 1 0.2
Pyramidal tract syndrome – 1 – – 1 – – 1 0.2
Somnolence – 1 – – 1 – – 1 0.2
Hepatic dysfunction – – – 1 – – – 1 0.2
Total 13 125 12 267 46 104 2 563
Gr toxicity grade, DVT deep venous thrombosis, CNS central nervous system

effect of bevacizumab in prolonging OS when used in com- bination therapy [42].
When considering the initiation of bevacizumab in patients with recurrent GBM it is important to weigh the potential survival benefits with the risk of severe side- effects. In our review of studies with GBM patients receiv- ing single agent bevacizumab (10 mg/kg every 2 weeks) either in the primary setting or at recurrence there was approximately 1 side-effect per patient on average (563 events in 503 patients) and 74% (417/563) of side-effects had a severity of Grade 3 or higher. The most common cummulative side effect was leukopenia (31.2%) followed by hypertension (17%). Serious complications such as CNS hemorrhage, thromboembolic events, or neutropenia occurred in 19.4% of all patients treated. These results are in keeping with an increased risk of developing hypertension
and fatal adverse events reported with the use of anti- VEGF in patients with non-small cell lung cancer [43, 44]. Detailed information regarding the adverse drug reactions seen in patients who have received bevacizumab is avail- able in the FDA drug monogram (http://www.accessdata. fda.gov/drugsatfda_docs/label/2015/125085s312lbl.pdf).
Despite the risks of therapy, bevacizumab use in the setting of recurrent GBM may be used clinically to reduce the side-effects of long-term steroid use in patients with recurrent GBM. In a retrospective comparative study, Nagpal et al. demonstrated a steroid sparing effect and also improvement in independent living using the inde- pendent living score (ILS) assessment when patients received bevacizumab at recurrence compared to other treatments [45]. After 2 months of therapy the median steroid dose had decreased to 0.4 mg in the bevacizumab

group and increased to 8 mg in the control group [45]. A non-randomized retrospective chart review of patients with recurrent GBM who received bevacizumab (n = 44) compared to other treatments (n = 79) demonstrated that 54% of patients treated with bevacizumab had a reduction in steroid dose compared to only 33% in the control group [46]. Patients treated with bevacizumab also maintained their functional status (Karnofsky Perfomance Score) for a longer period of time than patients in the control group [46]. While a reduction in steroid use is often seen when GBM patients are started on bevacizumab, the HRQoL data from randomized trials reported to date does not show any significant advantages overall. It is important to note that a qualitative review of the HRQoL data is not sufficient to conclude lack of a treatment effect. There- fore, we await for further QOL data collected from rand- omized phase III trials, which will provide greater clarity on the QOL effects of bevacizumab.

Conclusion
In current practice bevacizumab is used alone or in com- bination with cytotoxic chemotherapy in patients with good performance status at the time of GBM recurrence. There is no evidence in the literature to support the use of bevacizumab in patients with a KPS < 60% (ECOG/ WHO > 2). Our analysis is concordant with the current clinical use as there is an observed increased OS when patients with recurrent GBM are treated with bevaci- zumab alone or in combination with cytotoxic chemo- therapy compared to historical cytotoxic chemotherapy controls. Although there is an observable survival benefit with the use of bevacizumab at recurrence the absolute survival advantage is limited to 4 months. In addition to a potential effect on survival, some patients treated with bevacizumab may benefit from its steroid sparing effects. Importantly, the use of bevacizumab has not been shown to have a negative effect on quality of life in limited ran- domized studies. The benefits of bevacizumab come at a cost of significant side-effects and potential disqualifi- cation of the patient for the use of other investigational therapies.
Compliance with ethical standards

Conflict of interest All authors declare that they have no conflict of interest.
Ethical approval This article does not contain any studies with human participants performed by any of the authors.
References
⦁ Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO (2005) Radiotherapy plus concomitant and adjuvant temo- zolomide for glioblastoma. N Engl J Med 352(10):987–996. doi:⦁ 10.1056/NEJMoa043330
⦁ Quick J, Gessler F, Dutzmann S, Hattingen E, Harter PN, Weise LM, Franz K, Seifert V, Senft C (2014) Benefit of tumor resec- tion for recurrent glioblastoma. J Neurooncol 117(2):365–372. doi:⦁ 10.1007/s11060-014-1397-2
⦁ Seystahl K, Gramatzki D, Roth P, Weller M (2016) Pharmaco- therapies for the treatment of glioblastoma—current evidence and perspectives. Expert Opin Pharmacother 17(9):1259–1270. doi:⦁ 10.1080/14656566.2016.1176146
⦁ Gilbert MR, Dignam JJ, Armstrong TS, Wefel JS, Blumenthal DT, Vogelbaum MA, Colman H, Chakravarti A, Pugh S, Won M, Jeraj R, Brown PD, Jaeckle KA, Schiff D, Stieber VW, Brachman DG, Werner-Wasik M, Tremont-Lukats IW, Sulman EP, Aldape KD, Curran WJ Jr, Mehta MP (2014) A randomized trial of bevacizumab for newly diagnosed glioblastoma. N Engl J Med 370(8):699–708. doi:⦁ 10.1056/NEJMoa1308573
⦁ Kreisl TN, Kim L, Moore K, Duic P, Royce C, Stroud I, Gar- ren N, Mackey M, Butman JA, Camphausen K, Park J, Albert PS, Fine HA (2009) Phase II trial of single-agent bevacizumab followed by bevacizumab plus irinotecan at tumor progres- sion in recurrent glioblastoma. J Clin Oncol 27(5):740–745. doi:⦁ 10.1200/jco.2008.16.3055
⦁ Friedman HS, Prados MD, Wen PY, Mikkelsen T, Schiff D, Abrey LE, Yung WK, Paleologos N, Nicholas MK, Jensen R, Vredenburgh J, Huang J, Zheng M, Cloughesy T (2009) Beva- cizumab alone and in combination with irinotecan in recurrent glioblastoma. J Clin Oncol 27(28):4733–4740. doi:⦁ 10.1200/ ⦁ J⦁ CO.2008.19.8721
⦁ Chinot OL, Wick W, Mason W, Henriksson R, Saran F, Nishi- kawa R, Carpentier AF, Hoang-Xuan K, Kavan P, Cernea D, Brandes AA, Hilton M, Abrey L, Cloughesy T (2014) Bevaci- zumab plus radiotherapy-temozolomide for newly diagnosed glioblastoma. N Engl J Med 370(8):709–722. doi:⦁ 10.1056/ ⦁ NEJMoa1308345
⦁ Bengtsson H (2016) matrixStats: functions that apply to rows and Columns of matrices (and to Vectors). R package version
0.51.0 edn
⦁ Team RC (2013) R: a language and environment for statistical computing. R Foundation for Statistical Copmuting, Vienna
⦁ Vredenburgh JJ, Desjardins A, Kirkpatrick JP, Reardon DA, Peters KB, Herndon JE 2nd, Marcello J, Bailey L, Threatt S, Sampson J, Friedman A, Friedman HS (2012) Addition of beva- cizumab to standard radiation therapy and daily temozolomide is associated with minimal toxicity in newly diagnosed glioblas- toma multiforme. Int J Radiat Oncol Biol Phys 82(1):58–66. doi:⦁ 10.1016/j.ijrobp.2010.08.058
⦁ Vredenburgh JJ, Desjardins A, Reardon DA, Peters KB, Hern- don JE 2nd, Marcello J, Kirkpatrick JP, Sampson JH, Bailey L, Threatt S, Friedman AH, Bigner DD, Friedman HS (2011) The addition of bevacizumab to standard radiation therapy and temozolomide followed by bevacizumab, temozolomide, and irinotecan for newly diagnosed glioblastoma. Clin Cancer Res 17(12):4119–4124. doi:⦁ 10.1158/1078-0432.CCR-11-0120
⦁ Lai A, Tran A, Nghiemphu PL, Pope WB, Solis OE, Selch M, Filka E, Yong WH, Mischel PS, Liau LM, Phuphanich S, Black K, Peak S, Green RM, Spier CE, Kolevska T, Polikoff J, Fehren- bacher L, Elashoff R, Cloughesy T (2011) Phase II study of

bevacizumab plus temozolomide during and after radiation ther- apy for patients with newly diagnosed glioblastoma multiforme. J Clin Oncol 29(2):142–148. doi:10.1200/JCO.2010.30.2729
⦁ Narayana A, Gruber D, Kunnakkat S, Golfinos JG, Parker E, Raza S, Zagzag D, Eagan P, Gruber ML (2012) A clinical trial of bevacizumab, temozolomide, and radiation for newly diagnosed glioblastoma. J Neurosurg 116(2):341–345. doi:⦁ 10.3171/2011.9. ⦁ JNS11656
⦁ Hainsworth JD, Shih KC, Shepard GC, Tillinghast GW, Brinker BT, Spigel DR (2012) Phase II study of concurrent radiation therapy, temozolomide, and bevacizumab followed by bevaci- zumab/everolimus as first-line treatment for patients with glio- blastoma. Clin Adv Hematol Oncol 10 (4):240–246
⦁ Hofland KF, Hansen S, Sorensen M, Engelholm S, Schultz HP, Muhic A, Grunnet K, Ask A, Costa JC, Kristiansen C, Thom- sen C, Poulsen HS, Lassen U (2014) Neoadjuvant bevacizumab and irinotecan versus bevacizumab and temozolomide followed by concomitant chemoradiotherapy in newly diagnosed glio- blastoma multiforme: a randomized phase II study. Acta Oncol 53(7):939–944. doi:⦁ 10.3109/0284186X.2013.879607
⦁ van Linde ME, Verhoeff JJ, Richel DJ, van Furth WR, Reijneveld JC, Verheul HM, Stalpers LJ (2015) Bevacizumab in combina- tion with radiotherapy and temozolomide for patients with newly diagnosed glioblastoma multiforme. Oncologist 20(2):107–108. doi:⦁ 10.1634/theoncologist.2014-0418
⦁ Verhoeff JJ, Lavini C, van Linde ME, Stalpers LJ, Majoie CB, Reijneveld JC, van Furth WR, Richel DJ (2010) Bevacizumab and dose-intense temozolomide in recurrent high-grade glioma. Ann Oncol 21(8):1723–1727. doi:⦁ 10.1093/annonc/mdp591
⦁ Hasselbalch B, Lassen U, Hansen S, Holmberg M, Sorensen M, Kosteljanetz M, Broholm H, Stockhausen MT, Poulsen HS (2010) Cetuximab, bevacizumab, and irinotecan for patients with primary glioblastoma and progression after radiation therapy and temozolomide: a phase II trial. Neuro Oncol 12(5):508–516. doi:⦁ 10.1093/neuonc/nop063
⦁ Sathornsumetee S, Desjardins A, Vredenburgh JJ, McLendon RE, Marcello J, Herndon JE, Mathe A, Hamilton M, Rich JN, Norfleet JA, Gururangan S, Friedman HS, Reardon DA (2010) Phase II trial of bevacizumab and erlotinib in patients with recurrent malignant glioma. Neuro Oncol 12(12):1300–1310. doi:⦁ 10.1093/neuonc/noq099
⦁ Desjardins A, Reardon DA, Coan A, Marcello J, Herndon JE 2nd, Bailey L, Peters KB, Friedman HS, Vredenburgh JJ (2012) Bevacizumab and daily temozolomide for recurrent glioblas- toma. Cancer 118(5):1302–1312. doi:⦁ 10.1002/cncr.26381
⦁ Reardon DA, Desjardins A, Peters KB, Gururangan S, Sampson JH, McLendon RE, Herndon JE 2nd, Bulusu A, Threatt S, Fried- man AH, Vredenburgh JJ, Friedman HS (2012) Phase II study of carboplatin, irinotecan, and bevacizumab for bevacizumab naive, recurrent glioblastoma. J Neurooncol 107(1):155–164. doi:⦁ 10.1007/s11060-011-0722-2
⦁ Moller S, Grunnet K, Hansen S, Schultz H, Holmberg M, Sorensen M, Poulsen HS, Lassen U (2012) A phase II trial with bevacizumab and irinotecan for patients with primary brain tumors and progression after standard therapy. Acta Oncol 51(6):797–804. doi:⦁ 10.3109/0284186X.2012.681063
⦁ Soffietti R, Trevisan E, Bertero L, Cassoni P, Morra I, Fabrini MG, Pasqualetti F, Lolli I, Castiglione A, Ciccone G, Ruda R (2014) Bevacizumab and fotemustine for recurrent glioblas- toma: a phase II study of AINO (Italian Association of Neuro- Oncology). J Neurooncol 116(3):533–541. doi:⦁ 10.1007/ ⦁ s11060-013-1317-x
⦁ Lassen U, Sorensen M, Gaziel TB, Hasselbalch B, Poulsen HS (2013) Phase II study of bevacizumab and temsirolimus combi- nation therapy for recurrent glioblastoma multiforme. Anticancer Res 33(4):1657–1660
⦁ Lee EQ, Reardon DA, Schiff D, Drappatz J, Muzikansky A, Grimm SA, Norden AD, Nayak L, Beroukhim R, Rinne ML, Chi AS, Batchelor TT, Hempfling K, McCluskey C, Smith KH, Gaffey SC, Wrigley B, Ligon KL, Raizer JJ, Wen PY (2015) Phase II study of panobinostat in combination with bevacizumab for recurrent glioblastoma and anaplastic glioma. Neuro Oncol 17(6):862–867. doi:⦁ 10.1093/neuonc/nou350
⦁ Taal W, Oosterkamp HM, Walenkamp AM, Dubbink HJ, Beerepoot LV, Hanse MC, Buter J, Honkoop AH, Boerman D, de Vos FY, Dinjens WN, Enting RH, Taphoorn MJ, van den Berkmortel FW, Jansen RL, Brandsma D, Bromberg JE, van Heuvel I, Vernhout RM, van der Holt B, van den Bent MJ (2014) Single-agent bevacizumab or lomustine versus a com- bination of bevacizumab plus lomustine in patients with recur- rent glioblastoma (BELOB trial): a randomised controlled phase 2 trial. Lancet Oncol 15(9):943–953. doi:⦁ 10.1016/ ⦁ S1470-2045(14)70314-6
⦁ Odia Y, Iwamoto FM, Moustakas A, Fraum TJ, Salgado CA, Li A, Kreisl TN, Sul J, Butman JA, Fine HA (2016) A phase II trial of enzastaurin (LY317615) in combination with bevaci- zumab in adults with recurrent malignant gliomas. J Neurooncol 127(1):127–135. doi:⦁ 10.1007/s11060-015-2020-x
⦁ Fadul CE, Kingman LS, Meyer LP, Cole BF, Eskey CJ, Rho- des CH, Roberts DW, Newton HB, Pipas JM (2008) A phase II study of thalidomide and irinotecan for treatment of glioblas- toma multiforme. J Neurooncol 90(2):229–235. doi:⦁ 10.1007/ ⦁ s11060-008-9655-9
⦁ Weller M, Tabatabai G, Kastner B, Felsberg J, Steinbach JP, Wick A, Schnell O, Hau P, Herrlinger U, Sabel MC, Wirsch- ing HG, Ketter R, Bahr O, Platten M, Tonn JC, Schlegel U, Marosi C, Goldbrunner R, Stupp R, Homicsko K, Pichler J, Nikkhah G, Meixensberger J, Vajkoczy P, Kollias S, Husing J, Reifenberger G, Wick W, Group DS (2015) MGMT promoter methylation is a strong prognostic biomarker for benefit from dose-intensified temozolomide rechallenge in progressive glio- blastoma: the director trial. Clin Cancer Res 21(9):2057–2064. doi:⦁ 10.1158/1078-0432.CCR-14-2737
⦁ Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn U, Curschmann J, Janzer RC, Ludwin SK, Gorlia T, Allgeier A, Lacombe D, Cairncross JG, Eisenhauer E, Mirimanoff RO, European Organisation for R, Treatment of Cancer Brain T, Radiotherapy G, National Cancer Institute of Canada Clinical Trials G (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. New Engl J Med 352 (10):987– 996. doi:⦁ 10.1056/NEJMoa043330
⦁ Malmstrom A, Gronberg BH, Marosi C, Stupp R, Frappaz D, Schultz H, Abacioglu U, Tavelin B, Lhermitte B, Hegi ME, Rosell J, Henriksson R, Nordic Clinical Brain Tumour Study G (2012) Temozolomide versus standard 6-week radiotherapy ver- sus hypofractionated radiotherapy in patients older than 60 years with glioblastoma: the Nordic randomised, phase 3 trial. Lancet Oncol 13(9):916–926. doi:⦁ 10.1016/S1470-2045(12)70265-6
⦁ Gilbert MR, Wang M, Aldape KD, Stupp R, Hegi ME, Jaeckle KA, Armstrong TS, Wefel JS, Won M, Blumenthal DT, Mahajan A, Schultz CJ, Erridge S, Baumert B, Hopkins KI, Tzuk-Shina T, Brown PD, Chakravarti A, Curran WJ Jr, Mehta MP (2013) Dose-dense temozolomide for newly diagnosed glioblastoma: a randomized phase III clinical trial. J Clin Oncol 31(32):4085– 4091. doi:⦁ 10.1200/JCO.2013.49.6968
⦁ Rosenthal MA, Gruber ML, Glass J, Nirenberg A, Finlay J, Hochster H, Muggia FM (2000) Phase II study of combination taxol and estramustine phosphate in the treatment of recurrent glioblastoma multiforme. J Neurooncol 47(1):59–63
⦁ See SJ, Levin VA, Yung WK, Hess KR, Groves MD (2004) 13-cis-retinoic acid in the treatment of recurrent glioblastoma

multiforme. Neuro Oncol 6(3):253–258. doi:10.1215/ S1152851703000607
⦁ Groves MD, Puduvalli VK, Chang SM, Conrad CA, Gilbert MR, Tremont-Lukats IW, Liu TJ, Peterson P, Schiff D, Cloughesy TF, Wen PY, Greenberg H, Abrey LE, DeAngelis LM, Hess KR, Lamborn KR, Prados MD, Yung WK (2007) A North American brain tumor consortium (NABTC 99-04) phase II trial of temozolomide plus thalidomide for recurrent glioblas- toma multiforme. J Neurooncol 81(3):271–277. doi:⦁ 10.1007/ ⦁ s11060-006-9225-y
⦁ Puduvalli VK, Giglio P, Groves MD, Hess KR, Gilbert MR, Mahankali S, Jackson EF, Levin VA, Conrad CA, Hsu SH, Col- man H, de Groot JF, Ritterhouse MG, Ictech SE, Yung WK (2008) Phase II trial of irinotecan and thalidomide in adults with recurrent glioblastoma multiforme. Neuro Oncol 10(2):216–222. doi:⦁ 10.1215/15228517-2007-060
⦁ Wick W, Gorlia T, Bendszus M, Sahm F, Taal W, Taphoorn M, Domont J, Idbaih A, Campone M, Clement PM, Stupp R, Fabbro M, Le Rhun E, Dubois F, Klein M, Platten M, Weller M, Golfi- nopoulos V, Van Den Bent MJ (2016) EORTC 26101 phase III trial exploring the combination of bevacizumab and lomustine in patients with first progression of a glioblastoma. J Clin Oncol 34 (suppl; abstr 2001)
⦁ Taphoorn MJ, Henriksson R, Bottomley A, Cloughesy T, Wick W, Mason WP, Saran F, Nishikawa R, Hilton M, Theodore- Oklota C, Ravelo A, Chinot OL (2015) Health-related quality of life in a randomized phase III study of bevacizumab, temozolo- mide, and radiotherapy in newly diagnosed glioblastoma. J Clin Oncol 33(19):2166–2175. doi:⦁ 10.1200/JCO.2014.60.3217
⦁ Herrlinger U, Schafer N, Steinbach JP, Weyerbrock A, Hau P, Goldbrunner R, Friedrich F, Rohde V, Ringel F, Schlegel U, Sabel M, Ronellenfitsch MW, Uhl M, Maciaczyk J, Grau S, Schnell O, Hanel M, Krex D, Vajkoczy P, Gerlach R, Kortmann RD, Mehdorn M, Tuttenberg J, Mayer-Steinacker R, Fietkau R, Brehmer S, Mack F, Stuplich M, Kebir S, Kohnen R, Dunkl E, Leutgeb B, Proescholdt M, Pietsch T, Urbach H, Belka C, Stum- mer W, Glas M (2016) Bevacizumab plus irinotecan versus
temozolomide in newly diagnosed o6-Methylguanine-DNA methyltransferase nonmethylated glioblastoma: the randomized glarius trial. J Clin Oncol 34(14):1611–1619. doi:10.1200/ JCO.2015.63.4691
⦁ Dirven L, van den Bent MJ, Bottomley A, van der Meer N, van der Holt B, Vos MJ, Walenkamp AM, Beerepoot LV, Hanse MC, Reijneveld JC, Otten A, de Vos FY, Smits M, Bromberg JE, Taal W, Taphoorn MJ, Dutch Neuro-Oncology G (2015) The impact of bevacizumab on health-related quality of life in patients treated for recurrent glioblastoma: results of the randomised con- trolled phase 2 BELOB trial. Eur J Cancer 51(10):1321–1330. doi:⦁ 10.1016/j.ejca.2015.03.025
⦁ Zang J, Xiang C, He J (2013) Synthesis of median survival time in meta-analysis. Epidemiology 24(2):337–338. doi:⦁ 10.1097/ ⦁ EDE.0b013e318282a66c
⦁ Brandes AA, Mason W, Pichler J, Nowak AK, Gil M, Saran F, Revil C, Lutiger B, Carpentier AF (2014) Can bevacizumab pro- long survival for glioblastoma patients through multiple lines of therapy? Future Oncol 10(7):1137–1145. doi:⦁ 10.2217/fon.14.75
⦁ Gu B, Gao W, Chu H, Gao J, Fu Z, Ding H, Lv J, Wu Q (2016) Adverse events risk associated with anti-VEGFR agents in the treatment of advanced nonsmall-cell lung cancer: a meta- analysis. Medicine (Baltimore) 95(48):e3752. doi:⦁ 10.1097/ ⦁ MD.0000000000003752
⦁ Lai XX, Xu RA, Yu-Ping L, Yang H (2016) Risk of adverse events with bevacizumab addition to therapy in advanced non- small-cell lung cancer: a meta-analysis of randomized controlled trials. OncoTargets Ther 9:2421–2428. doi:⦁ 10.2147/OTT.S96156
⦁ Nagpal S, Harsh G, Recht L (2011) Bevacizumab improves qual- ity of life in patients with recurrent glioblastoma. Chemother Res Pract 2011:602812. doi:⦁ 10.1155/2011/602812
⦁ Nghiemphu PL, Liu W, Lee Y, Than T, Graham C, Lai A, Green RM, Pope WB, Liau LM, Mischel PS, Nelson SF, Elashoff R, Cloughesy TF (2009) Bevacizumab and chemotherapy for recur- rent glioblastoma: a single-institution experience. Neurology 72(14):1217–1222. doi:⦁ 10.1212/01.wnl.0000345668.03039.90