Mircera Reviewers Turned To Top-Level FDA Briefing With CRP Concerns

Unlike trials of other ESAs, Roche’s Mircera studies excluded patients with elevated levels of C-reactive protein, a marker for inflammation. In a move reflective of growing use of interdisciplinary and cross-division expertise within FDA, the review division sought center-level “regulatory briefing” to determine if CRP screening produced a database that was not characteristic of the market. The briefing found the rate of exclusion due to CRP screening insufficient to delay approval of Mircera

The brevity of the Mircera label's mention of C-reactive protein levels as exclusion criteria in the clinical trials of Roche's long-acting erythropoeisis stimulating agent belies the extent of the review team's concern over the CRP screening - a concern which culminated in a "regulatory briefing" that drew top-level panelists from five CDER offices.

Reviewer concern centered on Roche's use of active screening for CRP values in the Mircera (methoxy polyethylene glycol-epoetin beta) Phase II and Phase III trials. High levels of CRP are correlated with inflammation and infection.

Development programs for other ESAs did not screen prospective subjects to exclude patients with high CRP values.

According to the minutes of the top-level regulatory briefing, held March 16, 2007, the reviewing division (Medical Imaging and Hematology Products) "is concerned the use of active CRP screening may have resulted in a database insufficiently characteristic of the market population."

Regulatory briefings can be called when the review team requests center-level input on an issue. Typically, the topic is targeted and specific, unlike the extensive questions and approvability votes seen in advisory committee meetings. Examples of issues that have resulted in regulatory briefings include the potential interaction of Purdue's Palladone with alcohol, pediatric use of Sanofi-Aventis' Arava (leflunomide), and visual and skin effects of Pfizer's Lyrica (pregabalin).

The Mircera regulatory briefing was convened specifically to discuss the implication of CRP screening for "the sufficiency of the safety database" (For a look at the regulatory briefing participants, see chart p. 39).

[Editor's Note: This is the second of two installments of Pharmaceutical Approvals Monthly's analysis of FDA's review of Mircera. For more, including a look at the review against the backdrop of the changing ESA landscape, development chronology and reviewers, see the August 2008 issue.]

The division posed only two questions to the 30-plus FDAers assembled: is the exclusion of patients with elevated CRP "an important limitation of the Mircera safety database?" and is that limitation "sufficient to preclude licensure until additional safety data verify the product's safety?"

As is often the case, the input of high-level FDA staff seemed to moderate reviewer concerns. The Mircera regulatory briefing panel found the CRP concerns to be manageable with labeling and a post-marketing study; similarly, the Lyrica regulatory briefing alleviated the primary reviewers' instinct to delay approval of the Pfizer drug ( ¹ (Also see "Lyrica Reviewer Concerns On Vision, Skin Led To Consults With Top FDA Staff" - Pink Sheet, 21 Aug, 2006.) , p. 28).

Potential Clinical Consequences

Elevated CRP has been recognized as a biomarker for detection of chronic renal failure patients at risk for morbidity and mortality over the past few years, as well as a marker for ESA "hypo-responsiveness," the briefing minutes note.

But the reviewers were especially concerned that the other marketed ESAs do not have CRP data in labeling.

Roche compared CRP data from Mircera and pooled epoetin (Amgen's Epogen and J&J's Procrit ) and darbepoetin (Amgen's Aranesp ) studies, finding that "higher baseline CRP levels (>10 mg/L) correlated with death and more composite cardiovascular events (death, congestive heart failure, non-fatal myocardial infarction or stroke), in both Mircera and comparator groups with similar effects."

Because the approved ESAs' trial programs did not screen for elevated CRP, the products are used in clinical practice without regard to patients' CRP levels. Clinical reviewer John Lee highlighted the possibility of harm if Mircera were prescribed as if the patient population was equivalent to that for epoetin and darbepoetin, since the Mircera study population was different.

Roche's decision to screen for elevated CRP levels "systematically excluded from Phase II and Phase III studies some patients who may receive Mircera in clinical practice," Lee said.

Infection or inflammation, conditions associated with elevated CRP, "may increase the risk of serious adverse events, particularly in patients requiring hemodialysis," Lee noted.

"In comparison with no therapy, ESA therapy in inflammation may increase the risk for serious thrombotic events," the clinical reviewer cautioned.

Lee's concerns have yet to be tested in actual clinical practice, as patent litigation with Amgen has kept Mircera off the U.S. market ( ² (Also see "Mircera Review Provided FDA With Forum To Suggest Ideal ESA Studies" - Pink Sheet, 15 Aug, 2008.) , p. 28).

Roche's Case For CRP Screening

Roche's rationale for screening for CRP was that it was a means of eliminating patients that under-respond from the clinical trial database in order to optimize the assessment of Mircera's treatment effect, according to a background document prepared for the regulatory briefing.

"Amgen excluded patients with systemic inflammation/infection from their studies of darbepoetin. Roche regards active CRP screening as accomplishing the same goal," the briefing document states. Roche's decision to exclude patients based on active CRP screening was intended to be a more efficient form of excluding patients than Amgen's.

"Roche maintains that exclusion of patients based upon the presence of 'active inflammatory disease' or 'systemic infection' is tantamount to screening with CRP," the briefing document explains, adding that "no data are available to verify this contention."

Roche argued that Amgen's "stringent eligibility criteria" in darbepoetin trials "may have also eliminated 'hypo-responders' due to the exclusion of patients with: systemic infection, active inflammatory disease, malignancy, [or] hepatic enzymes >2x ULN."

Roche said that the CRP exclusion was designed with a cut-off based on the distribution of CRP levels in CRF patients on dialysis, so that "the added laboratory exclusion did not 'significantly alter the overall exclusion rate,'" the briefing document reported. "The number of patients excluded from participation in the phase safety database studies was a relatively small fraction of those screened."

"Overall, only approximately 3 percent of CRF patients were excluded from Mircera confirmatory clinical studies because of elevated CRP," Roche found, translating to "82 potential patients from the final database size of 2,737 patients."

The briefing document poses a worst-case scenario analysis, assuming that all 54 of the 82 patients who received Mircera died and all 28 patients receiving the comparator survived. In this analysis, Mircera would show a 3 percent death outcome vs. 0 percent. However, looking at deaths plus estimated deaths, the mortality rate rises to 10.3 percent of Mircera and 6.8 percent of comparator patients, according to the backgrounder.

"Given the overall Mircera safety pattern and assuming extreme outcomes for fatalities as a result of CRP screening, the sufficiency of the Mircera safety database appears unclear," the briefing paper says.

The 3 Percent Solution

However, "the panel did not think the CRP limitation of the database was a reason not to approve the application," the regulatory briefing minutes state.

The exclusion rate in particular was convincing. The panel noted that only 3 percent of patients were excluded by the CRP screening, "so that they could not have affected results to any significant effect had they been included," the minutes report.

"Moreover, dividing patients by CRP</= 10 and CRP > 10 showed no tendency for death or CV events to be more affected by Mircera than reference drug in the higher CRP group," the minutes add.

"The panel therefore concluded that the inclusion of the extra 2 percent data regarding CRP will not make a difference and that CRP exclusion is not an important limitation of the Mircera database," the minutes report.

The panel did feel that the CRP screening data should be in the product label. Final labeling states in the Clinical Studies section that in all studies, patients were "without evidence of infection or inflammation as determined by history and laboratory data, including C-reactive protein. … A CRP value above the threshold led to the exclusion of no more than 3 percent of the screened patients."

There was brief discussion at the regulatory briefing of additional analyses that could assist in better understanding the effect of CRP screening in development of the safety database. The panel suggested that the sponsor should do a study to determine drug effect on patients with high CRP as a post-marketing commitment.

The regulatory briefing also suggested that the agency could learn from the Mircera CRP screening case, particularly in the context of addressing exclusion criteria in internal documents.

"Regarding FDA development of reviewer tools and guidances, the panel noted that the CRP concern illustrated the importance of eligibility criteria having important review considerations," the minutes state. "Hence, the guidances/tools should highlight this potential."

Falling Cross-Divisional Barriers

The Mircera regulatory briefing is consistent with the rising profile of interdisciplinary assessment of difficult review issues at FDA, using expertise across divisional and disciplinary boundaries.

Mircera, for example, was also assessed by the Interdisciplinary Review Team for QT Studies. An oncology office consult was sought given the cancer reviewers' experience with the other ESAs, all of which have chemotherapy-induced anemia indications. (The Mircera BLA did not seek a CIA claim; development was terminated after a Phase II study was stopped early due to excess mortality.)

More cross- or extra-divisional input can be expected in coming drug reviews. Unlike the dedicated staff of specialized groups like the Controlled Substances Staff or the SEALD team, these ventures draw from multiple FDA offices. At the 2008 Drug Information Association annual meeting, the agency described the QT team as an example of an interdisciplinary approach that builds consistency and knowledge ( ³ (Also see "Relistor Review Offers Glimpse Into FDA’s Evolving Approach To QT Studies" - Pink Sheet, 15 Jul, 2008.) , p. 37).

The Executive Carcinogenicity Assessment Committee is a long-standing example of FDA's use of interdisciplinary teams to serve as routine consultants on topical issues. Additionally, an internal FDA Pediatric Review Committee was established in 2007 as a condition of the FDA Amendments Act; the PeRC acts as a consulting body to review divisions on pediatric data. The Interdisciplinary Pharmacogenomics Review Group reviews voluntary genomic data submissions and other submissions upon request. Plus, the Clinical Hold/Refusal-To-File committee has been restructured to play a greater role (⁴ (Also see "FDA Clinical Hold And Refusal-To-File Actions To Be Reviewed By Internal Cmte." - Pink Sheet, 28 Jan, 2008.), p. 29).

What Did The Review Division Take Home From The Briefing?

The division appears to have incorporated the regulatory briefing's conclusions into the final approval smoothly.

Nonetheless, Lee identified "potentially inadequate representation of the intended treatment population" in clinical trials as a "major residual safety concern" in his second-cycle review, dated Nov. 7, 2007, one week before Mircera's approval. Lee did recommend approval of Mircera, but was forceful in communicating his perspective that more extensive studies were needed to address emerging ESA class concerns ( ⁵ (Also see "Mircera Review Provided FDA With Forum To Suggest Ideal ESA Studies" - Pink Sheet, 15 Aug, 2008.) , p. 28).

A joint decision memo completed Nov. 9, 2007, by Acting DMIHP Director Dwaine Rieves and Office of Oncology Products Director Richard Pazdur is somewhat more sanguine. "Analyses of adverse outcomes categorized by C-reactive protein support the sponsor's conclusion that the active screening program did not adversely impact the safety database," the joint memo states.

At the internal regulatory briefing, "the consensus was that active CRP screening did not importantly compromise the Mircera safety database" because "the upper limit (30 mg/L) was relatively high" and "a small number of subjects were excluded ([approximately] 3 percent)."

The CRP analyses and data "support the reasonableness of the submitted safety database," Rieves and Pazdur conclude. "Nevertheless," they stress, "a post-marketing commitment is proposed to address this subject."

The Nov. 14, 2007, approval letter requires two clinical post-marketing studies: a pediatric study and a Phase IV open-label randomized controlled trial with a primary endpoint assessing all-cause mortality and cardiovascular morbidity, to address broad questions about cardiovascular safety of conventional ESA regimens for CKD-related anemia.

As a secondary analysis, the Phase IV trial will also evaluate the safety of Mircera in patients with high CRP concentrations; the directions in the approval letter are for the study to "enroll patients with a broad range of C-reactive protein blood concentrations."

- Bridget Silverman ([email protected])