healthmedicinet

Background and history of events

Bevacizumab is a recombinant humanised monoclonal antibody to vascular endothelial
growth factor (VEGF) marketed by Genentech as Avastin®. Bevacizumab inhibits the binding
of VEGF to its receptors on the surface of endothelial cells, thereby reducing the
vascularisation of cancerous tumours and inhibiting their growth.

In 2004, bevacizumab was first approved by the FDA as treatment for metastatic colorectal
cancer. It then gained approval for other indications including non-small cell lung
cancer, renal cell carcinoma, and glioblastoma 3]. In February 2008, bevacizumab was approved for first-line treatment of metastatic
breast cancer under the FDA’s accelerated approval program (Table 1). The approval was based on the results from a single study, the E2100 trial, published
in the New England Journal of Medicine (NEJM) journal in 2007 4]. This trial showed an increase in median progression-free survival in women with
advanced breast cancer treated with bevacizumab in combination with paclitaxel but
no increase in overall survival compared to those treated with paclitaxel alone 4] (Table 2). Reprints of the NEJM article were actively used by Genentech, the makers of bevacizumab,
to promote the drug to physicians 5] and Genentech estimated that about 9000 breast cancer patients in the United States
had been treated with ‘off-label’ bevacizumab (i.e. outside marketing authorisation)
at the time of the FDA’s approval in 2008 6].

Table 1. History of regulatory events

The official review of bevacizumab for metastatic breast cancer by ODAC was less positive
than the NEJM paper. It highlighted several methodological shortcomings of the E2100
trial, including the use of progression-free survival as an endpoint and the lack
of blinding 7]. Progression-free survival has not been shown convincingly to be an appropriate surrogate
endpoint for breast cancer or to be predictive of overall survival 8]. The lack of blinding meant that knowledge of which drug(s) individual patients were
receiving could have influenced judgements of their responses to treatment. An independent,
blinded review of radiological and clinical data of all patients in the E2100 trial
was therefore required by the FDA. Although this confirmed that the addition of bevacizumab
to paclitaxel resulted in a statistically significant improvement in progression-free
survival (hazard ratio 0.48, 95 % CI 0.39 to 0.61; P??0.0001) 9], the estimate of the magnitude of the effect lacked reliability because of incomplete
data (10 % of patients), loss to follow-up (34 % patients) and lack of consistency
in determination of radiologic disease progression 7]. ODAC also highlighted important safety issues with bevacizumab including a 20.2 %
increase in grade 3–5 toxicity (including hypertension, sensory neuropathy, thromboembolism,
gastrointestinal perforation, haemorrhage) and 1.7 % incidence of treatment-related
death in the bevacizumab plus paclitaxel arm compared to 0 % the paclitaxel arm 7], 10].

Based on all this evidence, ODAC, which comprised seven experts in the field of oncology
or statistics, and two consumer representatives, voted five to four in an open voting
process not to recommend approval on the question of whether the data provided were
‘sufficient to establish a favourable risk/benefit analysis for the use of bevacizumab
plus paclitaxel for first-line treatment of patients with metastatic breast cancer” 11].

Despite the ODAC vote, the FDA granted approval contingent on the results of additional
studies in February 2008. This decision generated great interest for both financial
and health reasons. It markedly benefited the drug maker’s potential market and sales
revenue, and Genentech’s stock rose by more than 8 % in after-hours trading 12]. The FDA decision was also perceived as clinically controversial. Some health professionals
and a patient advocacy group, the National Breast Cancer Coalition Fund, expressed
concern about a lowering of the FDA standards for medicine approval 12]. Other experts welcomed the decision because they believed that it was ‘a matter of time before a survival benefit is documented’ 6].

Almost two and a half years later, in July 2010, the ODAC convened to re-evaluate
the approval and examined the results of two additional clinical trials 13], 14] (Table 2) and voted 12 to one to recommend removing the indication of bevacizumab for metastatic
breast cancer 15]. The two new studies did not demonstrate a difference in overall survival and showed
smaller improvement in progression-free survival than in the original E2100 trial.
None of the studies demonstrated an improvement in quality of life and all showed
an increased risk of serious adverse effects including gastrointestinal perforation
and severe bleeding (Table 3). The overall proportion of treatment-related deaths was similar (1.8 %) in both
bevacizumab and control groups 16].

Table 3. Pooled safety results of bevacizumab trials for first-line treatment of advanced breast
cancer (E2100, AVADO, RIBBON-1 y

In December 2010, the FDA announced it would start withdrawing the indication. Around
17,000 women with advanced breast cancer were being treated with bevacizumab at that
time, and financial analysts estimated that a withdrawal of FDA approval for breast
cancer could cost Genentech nearly $1 billion in sales based on previously projected
figures 17]. Genentech filed an opposition petition requesting an administrative hearing.

In June 2011, there was a two-day hearing which involved ODAC experts, experts nominated
by Genentech and permitted members of the public to provide oral testimony. Electronic
or written comments on the FDA’s proposal to withdraw approval were also invited 18].

Around 450 public submissions were sent to the FDA during the hearing period, mostly
by consumers asking the FDA to maintain the indication as their perception was the
drug had benefited themselves or close friends and family 19]. In the first part of the hearing, 35 members of the public provided their views.
They included survivors of advanced breast cancer and representatives from consumer
organisations including ‘Facing Our Risk of Cancer Empowered’(FORCE, representing
people living with a genetic mutation or hereditary cancer risk), the Abigail Alliance
for Better Access to Developmental Drugs, the Ovarian Cancer National Alliance, breastcancer.org,
Marti Nelson Cancer Foundation, Colon Cancer Alliance, Kidney Cancer Association,
Cancer Support Community. Most consumer groups urged the FDA to keep bevacizumab available
20]. Survivors ascribed their survival and current quality of life to bevacizumab and
called themselves ‘super-responders’. There was no comment from the public on the possibility of predicting in advance
how responders might be distinguished from non-responders or whether responders may
constitute a minority of women. This question was acknowledged as unresolved during
the following scientific discussions of ODAC as no discriminatory biological or genetic
marker for predicting the efficacy of bevacizumab had been identified in the clinical
trials.

Concerns on the safety of bevacizumab were rarely raised by members of the public
during the hearing, and when they were it was usually to say that the adverse effects
of bevacizumab were mild or manageable. Only one woman from SHARE leaders, a group
of cancer survivors, said that ‘for every woman here testifying, there are other women who we know – a member of our
group who bled out of every orifice of her body…another woman…who had a brain haemorrhage.
So those people don’t come to testify’. Among the few consumer voices that supported bevacizumab withdrawal was that of
Christine Brunswick, the vice president of the National Breast Cancer Coalition and
breast cancer survivor who stated that ‘This decision can’t be driven by anecdotes. It must be driven by science’. At the end of the hearing, ODAC voted 6–0 in favour of removing the indication
and the FDA finally withdrew its approval in November 2011, three and a half years
after the initial conditional approval.

The FDA withdrawal process prompted reactions from patients, consumer organisations,
health professionals, healthcare policy makers and the pharmaceutical industry, and
the decision generated major media coverage in newspapers and on television. Some
patients were very upset when the vote was announced with one woman accusing the FDA
committee of ‘killing 17,000 women with one vote’ 21]. But criticism was not based solely on patients’ beliefs in bevacizumab’s benefits.
Freedom of Access to Medicines, a project from The Abigail Alliance that lobbies for
wider access to developmental cancer drugs, questioned the right of the FDA ‘to make a decision that should be left to a woman and her doctor’ 22]. The FDA was also accused of taking cost into consideration 22].

Some consumer organisations feared that insurers would not pay for the cost of bevacizumab
after the approval withdrawal. Susan G. Komen from the Cure, the US largest breast
cancer organization, declared that “as a patient advocacy organization, we want to ensure that women who are successfully
using Avastin today continue to have access to the drug, and that their treatment
be covered by third-party payers” 23]. Medicines without FDA regulatory approval may still be prescribed and used by patients
‘off-label’, i.e. for indications outside the marketing authorizations, however, the
extent to which public or private health insurers may cover off-label use may vary
substantially 24]. Insurers initially appeared to respond to these consumer concerns. The day after
the FDA hearing, the Centers for Medicare and Medicaid Services (CMS), which determines
how products are reimbursed under Medicare, the largest public health insurance in
the US, announced that they would ‘continue to cover the drug for breast cancer as long as doctors are prescribing it’ 25]. The guideline committee of the National Comprehensive Cancer Network (NCNN), a consortium
of 21 cancer centers, whose guidelines are one of the main compendia used by American
health insurers to inform their coverage decisions 26], voted 24 to 0 with one abstention to continue recommending bevacizumab in combination
with paclitaxel as an appropriate treatment option for metastatic breast cancer for
the reason that it ‘improves time to progression and response rates but does not improve overall survival’27]. It was noted that nine of the 27 members of the Breast Cancer Panel in charge of
the NCNN guidelines on management of breast cancer received financial support from
Genentech, the company manufacturing bevacizumab 28]. The extent to which this had an impact, if any, on the reluctance of the NCCN to
change their guidelines in response to the FDA withdrawal of the indication is unknown.
Three months later, however, Blue Shield of California became the first large insurance
company to announce it would end payments for bevacizumab in the management of advanced
breast cancer 29].

Health professional and medical organisation responses

The FDA withdrawal process was heavily debated in medical journals. An editorial in
the Journal of Clinical Oncology supported the FDA’s decision, stating that “the outcomes were arguably not clinically compelling” 30]. Ralph D’Agostino, a professor of mathematics who served on ODAC at the time of FDA’s
initial approval of bevacizumab, argued in the New England Journal of Medicine that progression-free survival was not an acceptable primary endpoint for approval
of first-line therapies in cancer and ‘if its use becomes standard for accelerated or even final approval, it will be more
difficult, if not impossible, to obtain solid data on overall survival’ 31]. An oncologist who served as an ODAC member at the FDA hearing that considered the
withdrawal decision declared ‘we did not want people to be hurt by a drug that does not work that well. We do not
want to provide false hope’ 32]. In contrast, Dr Milton Wolf, a radiologist, wrote an article for the conservative
Washington Times entitled ‘The FDA’s one-man death panel’ and claimed that the ‘FDA denies Americans access to life-saving drugs.’ He described the cost and complexity of the FDA’s processes as ‘regulatory barriers’ that impede the access to innovative medicines in the US.

The opinions of health professionals on the FDA withdrawal decision were examined
in an study conducted after ODAC’s recommendation to revoke approval of bevacizumab
between September and December 2010 33]. This email survey included 564 participants from across the world, mostly medical
oncologists. About 50 % of the respondents said that, if the DFA cancelled approval,
they would use bevacizumab in an off-label indication, mainly in patients with triple-receptor
negative breast cancer. A small majority (52 %) agreed with the FDA decision to withdraw
the indication on the ground that the benefits shown in the two additional studies
with bevacizumab were not of the same amount observed in the initial E2100 trial but
48 % believed this was not a valid reason. Another study examined trends in use of
bevacizumab for breast cancer in 122 oncology practices involving 570 oncologists
in the US 34]. It found that the use declined by 37 % between May 2010 (just prior to the ODAC
meeting revoking approval) and November 2010 (just prior to the start of the FDA withdrawal
process) and by 63 % just prior to the FDA official withdrawal notice without concomitant
changes to clinical guidelines or insurers’ coverage policies that may explain these
trends.

Media coverage

A study of 359 articles published in North American newspapers about bevacizumab and
breast cancer before, during and after the FDA approval period showed that, prior
to the FDA approval, the reports tended to present bevacizumab positively: 82 % of
articles noted efficacy and only 23 and 24 % reported the lack of efficacy or side
effects respectively 35]. The proportions of positive headline tone (36 % before approval, 18 % during approval,
9 % after approval, p?=?0.0002) and positive article tone (42 %, 19 %, 15 %, p??0.0001) declined with each study period. Industry representatives were more likely
to be quoted prior to the approval than at later times (33 %, 23 %, 11 %, p?=?0.014).