AUTH/2705/3/14 - Roche v Merck Serono

Presentation of Erbitux clinical trial results in a press release

  • Received
    17 March 2014
  • Case number
    AUTH/2705/3/14
  • Applicable Code year
    2012
  • Completed
    10 October 2014
  • No breach Clause(s)
    7.4
  • Breach Clause(s)
    2, 7.2 (x5), 7.3 (x2), 7.10 (x2), 9.1, 10.2 and 22.2 (x2)
  • Sanctions applied
    Undertaking received
  • Additional sanctions
  • Appeal
    appeal by respondent
  • Review
    November 2014

Case Summary

​Roche complained about the way in which Merck Serono represented the results of the FIRE-3 AIO (Arbeitgemeinschaft Internistische Onkologie) clinical trial in a UK press release issued 28 September 2013 and also raised concerns about such data in unidentified Erbitux (cetuximab) promotional materials.

At that time Erbitux was indicated, inter alia, for the treatment of patients with epidermal growth factor receptor (EGFR)-expressing, KRAS wild-type metastatic colorectal cancer (mCRC) in combination with irinotecan-based chemotherapy, in first-line in combination with FOLFOX (folinic acid, fluorouracil and oxiplatin), and as a single agent in patients who had failed oxaliplatin and irinotecan-based therapy and who were intolerant to irinotecan.

Roche marketed Avastin (bevacizumab) which was indicated, inter alia, in combination with fluoropyrimidine-based chemotherapy for the treatment of adult patients with metastatic carcinoma of the colon or rectum.

The detailed response from Merck Serono is given below.

Roche explained that the FIRE-3 study trial evaluated the superiority of cetuximab plus combination chemotherapy, compared with bevacizumab plus combination chemotherapy in the first-line treatment of KRAS wild-type mCRC. The primary endpoint for the study was overall response rate. Secondary endpoints included progression-free survival and overall survival. Importantly it was not a treatment sequencing study as subsequent lines of treatment were not specified.

Roche stated that the primary analysis of the study, presented at the American Society of Clinical Oncology (ASCO) 2013, showed that the study failed to reach its primary endpoint. There was no significant difference in overall response rate (primary endpoint) between the two treatment arms. There was also no significant difference in progression-free survival between the two arms, but increased overall survival in the arm receiving cetuximab plus chemotherapy as first-line treatment (one of the secondary endpoints) was reported. The Kaplan-Meier curves of overall survival presented at ASCO 2013 showed that the lines, representing the different study arms, did not begin to separate until the 15-18 month time point. Given that the median time to first progression was approximately 10 months in both arms and the reported median duration of first-line treatment was significantly shorter than this in both arms, there would appear to be significant grounds to question the degree to which the first-line treatment was responsible for any overall survival difference demonstrated.

Roche noted that a second FIRE-3 analysis presented in July 2013 at the World Congress on Gastrointestinal Cancer, provided details of the second-line treatments administered to patients in the FIRE-3 trial. This analysis showed differences in the treatments received in the second-line setting by patients in the two arms. A further FIRE-3 analysis presented at the European Society for Medical Oncology (ESMO) European Cancer Congress (ECC), October 2013, showed the results of a pre-planned exploratory analysis of a sub-group of patients who were not only KRAS wild-type, but also NRAS wild-type (termed RAS wild-type). In that new sub-group of patients, the first-line cetuximab plus chemotherapy arm again failed to show a significant improvement over the bevacizumab plus chemotherapy arm in both overall response rate and progression free survival. However, the analysis showed a difference of 7.5 months in median overall survival between the two arms in favour of the group receiving cetuximab plus chemotherapy as their first-line regimen. As for the previous KRAS overall survival analysis, the Kaplan-Meier curves did not separate until well after completion of first-line treatments and first progression.

Merck Serono's press release about the FIRE-3 trial analysis after the ESMO-ECC congress was headed: 'Merck Serono's Erbitux Significantly Extends Survival by 7.5 Months in mCRC RAS Wild-Type Patients When Compared With Bevacizumab: New Analysis of FIRE-3 AlO Study'. Roche stated that the press release was the source material for at least one article in the medical press and similar messages were used in promotional materials in Ireland (with prescribing information stating it was for UK and Ireland) but was not sure if it was being used in the UK.

Roche alleged that the overall survival statement in the heading 'Merck Serono's Erbitux Significantly Extends Survival by 7.5 Months in mCRC RAS Wild- Type Patients When Compared with Bevacizumab: New Analysis of FIRE-3 AlO Study' was misleading because the FIRE-3 study failed to reach its primary endpoint of overall response rate. The heading was based on a sub-group analysis from this 'negative' phase III study. The fact that the study did not meet its primary endpoint was not prominently presented in the press release; it was only mentioned midway down the second page. Roche alleged a breach. Findings from secondary endpoints must be set within the context of the primary endpoints companies could not 'cherry pick' favourable findings.

The Panel noted that the press release was dated 28 September 2013 and thus the relevant Code was the Second 2012 Edition (amended) Code.

The Panel noted that the press release was headed 'Merck Serono's Erbitux Significantly Extends Survival by 7.5 Months in mCRC RAS Wild-Type Patients When Compared With Bevacizumab: New Analysis of FIRE-3 AIO Study', below the heading in slightly smaller text were two bullet points; 'New data from a pre-planned analysis of the FIRE-3 study show an increase of median overall survival (OS) from 25.6 to 33.1 months (p=0.011) in mCRC patients with RAS wild-type tumours receiving 1st line Erbitux plus FOLFIRI compared with patients receiving bevacizumab plus FOLFIRI' and 'In the group with any RAS mutations, patients who received Erbitux in 1st line reached a median OS of 20.3 months vs 20.6 months in the group that was treated with bevacizumab in 1st line (p=0.60)'.

Text beneath referred to the phase III head-tohead trial which showed a 'clinically relevant improvement from Erbitux (cetuximab) plus FOLFIRI vs bevacizumab plus FOLFIRI as first-line treatment in metastatic colorectal cancer (mCRC) in patients with RAS wild-type tumours'.

The Panel noted that the FIRE-3 study was a multicentre randomised phase III trial investigating 5-FU, folinic acid and irinotecan (FOLFIRI) plus cetuximab vs FOLFIRI plus bevacizumab in first-line treatment of mCRC. The study failed to meet its primary endpoint of overall response rate (ORR). Secondary endpoints included median progression free survival (PFS) and median overall survival.

The summary of the FIRE-3 study principal investigator's presentation at the European Cancer Congress stated 'OS was markedly superior (Δ = 7.5 months, HR 0.70) in all RAS wild-type patients receiving first-line therapy with cetuximab (p=0.011)'. The presentation concluded that upfront determination of RAS (KRAS and NRAS) mutation status appeared to be highly recommendable in patients with metastatic disease and concluded that 'Patients with all-RAS wild-type tumours have a clinically relevant survival benefit when first-line treatment with cetuximab is offered'.

The Panel disagreed with Merck Serono's decision that as the lack of difference in ORR and PFS had previously been reported in the ASCO press release and as there was no change in these endpoints it was not considered appropriate to include them in the heading. The Panel considered that the heading, 'Merck Serono's Erbitux Significantly Extends Survival to 7.5 Months in mCRC Wild-Type Patients When Compared with Bevacizumab: New Analysis of FIRE-3 AIO Study', was not a fair reflection of the overall data; it had not been placed within context of the study's primary outcome. The reference to the study's failure to meet its primary endpoint of objective response rate based on investigators' read in patients with KRAS EXON 2 wild-type tumours appeared in the third paragraph on page 2 and was insufficient to counter the heading. Insufficient information had been provided to enable the reader to properly assess how much weight to attach to the secondary endpoint findings. The heading was therefore misleading as alleged and the Panel ruled a breach of the Code. This ruling was upheld on appeal by Merck Serono.

Roche stated that the first bullet point: 'New data from a pre-planned analysis of the FIRE-3 study show an increase of median overall survival (OS) from 25.6 months to 33.1 months (p=0.011) …' was the result of a sub-group analysis from the negative phase III study. Further contextualisation outlined in the background section was critical for the audience to be able to understand the clinical relevance. The press release failed to set the finding clearly in the context of the overall study which failed to meet its primary endpoint. In addition, the word 'exploratory' was only used much later in the press release to describe that analysis. Roche alleged that this rendered the press release misleading.

Roche was concerned about the statistical validity of the analysis, as any sub-group analyses needed to be accounted for statistically to avoid bias from multiple analyses. It was acknowledgement later in the press release that the analysis was exploratory, this should have been reflected in the headlines/bullet points to avoid misleading the audience. In inter-company dialogue, Merck Serono was unable to comment on Roche's statistical concerns and directed Roche to the study sponsor. This had not reassured Roche that Merck Serono could sufficiently substantiate the data and Roche alleged a breach of the Code.

The Panel considered that its general comments above in relation to the heading of the press release were relevant here. The sub-group analyses had not been placed in context of the study's failure to achieve its primary endpoint. In addition, it was not clear at the outset that the data was from a preplanned exploratory analysis. The only reference to this was on the second page and there was no explanation that no confirmatory clinical conclusions could be drawn from such an analysis. In the Panel's view the press release invited the reader to draw such conclusions. Exploratory analyses should not be used as the basis for a robust comparison of medicines. The material should be sufficiently complete to enable the recipient to form their own opinion of the therapeutic value of the medicine. The Panel considered that the bullet point was misleading as alleged and ruled a breach of the Code. This ruling was upheld on appeal by Merck Serono.

The Panel noted Roche's allegation that Merck Serono was unable to substantiate the sub-group analysis. Merck Serono submitted that the bullet point in question was supported by the data presented at ESMO. However, the Panel noted that the ESMO presentation did not appear to cover statistical analysis of the sub-group although the abstract made it clear that the analysis was preplanned. The Panel however did not have any accompanying transcript.

The Panel noted Roche's allegation that the sub-group analysis needed to be accounted for statistically to avoid bias from multiple analyses. On balance and on this very narrow point the Panel ruled that the bullet point in question was not capable of substantiation. A breach of the Code was ruled. This ruling was appealed by Merck Serono.

The Appeal Board noted that this was clearly a complex area. As the FIRE-3 study had progressedit had started to become clear that patients with RAS wild-type mCRC responded better to therapy than those with RAS mutations. The analysis at issue in the press release involved only the RAS wild-type patients (n=342) and not the original ITT population (n=592). Although the Erbitux marketing authorisation had been restricted to patients with RAS wild-type mCRC, this was not the case when the press release was issued on 28 September 2013. In that regard the Appeal Board considered that only the data that was available on that date could be relied upon to substantiate the content of the press release.

The Appeal Board although concerned as to whether the analysis was sufficiently powered, considered that the bullet point was nonetheless factually correct and thus on balance, on this very narrow point, was capable of substantiation. No breach of the Code was ruled. The appeal on this point was successful.

Roche alleged that the second bullet point: 'In the group with any RAS mutations, patients who received Erbitux in 1st line reached a median OS of 20.3 months vs. 20.6 months in the group that was treated with bevacizumab in 1st line (p=0.60)' seemed to suggest that there was no difference between the arms with respect to overall survival in the sub-group of patients with RAS mutant mCRC. In Europe, cetuximab was not licensed in RAS mutant mCRC and was actually contraindicated in the treatment of RAS mutant mCRC with certain chemotherapy combinations. No such restriction applied to bevacizumab. The licence restriction, or indeed any of the licence particulars (eg should only be used for EGFR-expressing tumours) for cetuximab were not mentioned in the press release.

The comparison was actually based on a pooled analysis of two different populations of patients with RAS mutations. There was no information in the press release that these findings were based on pooling data from two different time points, using two different testing methods. In 2008, patients with mutations in the KRAS EXON 2 gene were no longer included in the licences for anti-EGFRs in Europe. As a result of this, the FIRE-3 trial was amended in 2008 to exclude recruitment of patients with KRAS MT gene in EXON 2. The analysis based on patients with RAS MT mCRC recruited into the trial after the protocol amendment reported a median overall survival of 16.4 months in the cetuximab arm and 20.6 months in the bevacizumab arm. With the press release only utilising the pooled analysis data set it appeared that there was no difference in overall survival between the treatment arms without clarification that cetuximab was unlicensed (or even contraindicated) in patients with RAS MT disease.

Roche was extremely concerned that the claim implied cetuximab had efficacy in a population for which it was unlicensed or contraindicated as it compared itself with a medicine that was licensed for use in that population. The statement, whilst factually accurate, did not provide balance, was misleading in itself and with respect to the safety profile of cetuximab and did not encourage rationale use of the medicine.

The Panel considered that the comparison was misleading as it was not clear that it was based on a pooled analysis of two different populations of patients with RAS mutations from two different time points. The Panel ruled breaches of the Code as it considered that the context of the comparison was not clear and it was therefore misleading.

The Panel disagreed with Merck Serono's submission that the comparison made no efficacy claims for cetuximab. The Panel considered that the overall survival comparison of cetuximab with bevacizumab in patients with any RAS mutations was misleading as it implied that like bevacizumab, cetuximab was licensed for the treatment of RAS mutant mCRC which was not so. It was only licensed for EGFR expressing RAS wild-type metastatic colorectal cancer. In the Panel's view the failure of Merck Serono to place the bullet point within the context of cetuximab's licensed indication and the failure to mention relevant contraindications was misleading and did not encourage the rational use of cetuximab and breaches of the Code were ruled. A breach was also ruled as the comparison was misleading.

The Panel noted Merck Serono's submission that the press release had been widely distributed to medical journals and health journalists. The Panel noted its rulings above in relation to the misleading statements made about Erbitux and considered that in relation to the matters discussed above the press release, which had been made available to the public, was not factual and had not presented information about Erbitux, a prescription only medicine, in a balanced way and a breach was ruled.

Roche alleged that the quotation on page 2: 'Such a prolongation is a paradigm shift in mCRC treatment since the introduction of monoclonal antibodies…' was misleading as it did not contextualise the subgroup analysis. In addition, whilst it reflected the views of the investigator, the discussant at ESMO strongly questioned it and recommended that based on the available data it was not a paradigm shift and the forthcoming results of CALGB (a forthcoming study evaluating the efficacy of first-line cetuximab vs first-line bevacizumab with a primary endpoint of overall survival) should be awaited to provide more insights into the outcomes of FIRE-3. Using words as strong as 'paradigm shift' in a press release was exaggerated and could raise unfounded hopes and Roche alleged breaches of the Code.

Overall, given the number and nature of its concerns and the very real risk to patient safety, Roche alleged that the press release and promotional materials failed to maintain high standards. Roche also alleged that such a concerted campaign based on misleading and unbalanced claims of this nature put patient safety at risk and brought the industry into disrepute in breach of Clause 2.

The Panel noted its comments and rulings at above with regard to the data from the FIRE-3 study showing a 7.5 month increase in median overall survival when using Erbitux plus FOLFIRI as compared with using bevacizumab plus FOLFIRI in metastatic colorectal cancer. The Panel considered that the quotation 'Such a prolongationis a paradigm shift in mCRC treatment since the introduction of monoclonal antibodies' was misleading as within the context of the median survival data it applied disproportionate weight to the results thereby exaggerating Erbitux's properties and consequently it did not encourage rational use. The Panel thus ruled breaches of the Code. The Panel noted its comments above regarding the provision of information to the public and similarly ruled a further breach of the Code. These rulings were upheld on appeal by Merck Serono.

The Panel considered that Merck Serono had failed to maintain high standards and ruled a breach in that regard.

The Panel noted that Clause 2 was used as a sign of particular censure and reserved for such use. The Panel noted that Roche had referred to patient safety. The Panel noted its rulings of breaches of the Code above. The Panel considered that it was very important that press releases about sensitive issues such as survival in cancer were fair, factual and not misleading. The press release had failed to reflect the study's primary endpoint and the product's licensed indications. In particular the headline claim about survival had been ruled in breach of the Code. The Panel considered that on balance the circumstances warranted such a ruling and a breach of the Clause 2 was ruled. This ruling was upheld on appeal by Merck Serono.