FDA Panel Split on Telavancin for Nosocomial Pneumonia

Miriam E. Tucker

November 30, 2012

SILVER SPRING, Maryland — A federal advisory panel has given qualified support for the use of telavancin in the treatment of patients with nosocomial pneumonia (NP) resulting from susceptible isolates of gram-positive organisms.

The US Food and Drug Administration (FDA) Anti-Infective Drugs Advisory Committee members split the vote, with 9 voting no and 6 yes, in answering whether they believe overall that the data support the efficacy and safety of the proposed indication for telavancin, but they subsequently voted 13 to 2 in favor of more limited use of the drug in certain seriously ill patients with NP for whom other alternatives are not suitable.

Telavancin is an intravenous lipoglycopeptide drug that has bactericidal activity against gram-positive bacteria. Manufactured under the name Vibativ by Theravance Inc, it has been licensed by the FDA since September 2009 for the treatment of complicated skin and skin structure infections. In October of that year, it was approved for the treatment of NP in the European Union.

However, in the United States, the FDA has twice rejected Theravance's new drug application for the NP infection indication. The first time, in November 2009, the FDA cited concerns about missing mortality data and whether the participants in the company's studies met guideline-based criteria for NP (chest X-ray plus 2 clinical features).

In November 2010, the FDA published new guidelines for the development of drugs to treat NP, which recommended noninferiority in 28-day all-cause mortality as a study endpoint rather than noninferiority in test-of-cure to an active comparator drug (vancomycin, in this case). It also recommended that companies conduct at least 2 randomized controlled studies.

Theravance's original new drug application for telavancin was based on 2 randomized, double-blind, active-controlled, parallel-group, multinational trials of identical design. The first, study 0015, was conducted in 22 countries and enrolled 761 patients, including 31% from the United States. The other, study 0019, enrolled 771 patients in 29 countries, with 14% from the United States. Patients were randomly assigned to receive either telavancin or vancomycin treatment for 7 to 10 days.

The test-of-cure endpoint, assessed by physician determination of the resolution of clinical signs and symptoms of NP, was noninferior to vancomycin overall, with an overall cure rate of about 60% with both, and telavancin was superior to vancomycin in certain analyses, including those with NP caused by Staphylococcus aureus, said Steven Barriere, PharmD, from Theravance.

However, after the new guideline was published, the FDA denied telavancin's NP indication a second time because only study 0019, and not study 0015, met the newly specified endpoint criteria of a 10% noninferiority margin in all-cause mortality for the patients who had at least 1 baseline gram-positive pathogen.

Difficulties Interpreting the Data

A major problem for the advisory panel in analyzing the data was that although the 2 trials were designed to be identical, they differed significantly in certain baseline characteristics, including the proportion of patients with a history of diabetes (31% in study 0015 vs 21% in study 0019), chronic renal failure (9% vs 4%, respectively), and baseline creatinine clearance less than 50 mL/minute (36% vs 27%), according to FDA medical reviewer Benjamin Lorenz, MD.

Mathai Mammen, MD, PhD, from Theravance, presented pooled data for the 2 trials. All-cause mortality at 28 days was 24% with telavancin vs 22% for vancomycin, which is a statistically insignificant difference. However, according to Dr. Lorenz, noninferiority was not demonstrated in study 0015, which showed a trend toward increased all-cause mortality with telavancin.

In Dr. Mammen's analysis, the increased mortality was primarily seen among patients with creatinine clearance values less than 30 mL/minute, and he said the product label would include guidance for using serum creatinine levels to guide therapy. (The label currently includes dose adjustments for patients with creatinine clearance values lower than 50 mL/minute.)

The committee struggled with the statistical interpretation of much of the efficacy and safety data, given the limitations of studies using both types of endpoints. The FDA had turned away from the use of the test-of-cure endpoint because of the inability to precisely determine efficacy for active comparators such as vancomycin, which must be used for ethical reasons, said Edward Cox, MD, director of the Office of Antimicrobial Products at the FDA.

Noninferiority for all-cause mortality was chosen instead because it is a more definitive endpoint for which there are data in the population being studied. However, that endpoint is also problematic in very sick populations with high overall mortality rates. Use of other endpoints is being explored, Dr. Cox said.

Support for Use With Restrictions

Most panel members agreed that telavancin should only be approved for NP caused by methicillin-resistant S aureus and not methicillin-susceptible S aureus or Streptococcus pneumonia, for which other agents are available. Several members also advised limiting the drug's use to patients with creatinine clearance levels above 30 or 50 mL/minute.

Matthew Goetz, MD, chief of infectious diseases at the Veterans Affairs Greater Los Angeles Healthcare System in California, said he voted no to the first question primarily out of concern for the renal data. "As a clinician I can't predict whose creatinine clearance will decrease over time," he said. He subsequently voted yes for more limited use.

Judith Voynow, MD, professor of pediatrics in the Division of Pulmonary Medicine at Duke University in Durham, North Carolina, voted yes to both questions with the caveat that it should be indicated for staph and not strep infections, and with the renal status restrictions. "These are very sick patients, mortality rates are high, and I thought that under the right circumstances this may be a useful drug in this population."

Peter Katona, MD, clinical professor of medicine at the David Geffen School of Medicine at the University of California, Los Angeles, said he voted no on the first question because he was disturbed by the fact that the FDA's rules had changed midstream and that the mortality data were therefore derived post hoc.

"I feel very strongly that when you start a study and you have predesignated endpoints, you really don't want to deviate from that if you can at all help it. Once you open the window to deviate from what you originally intended to do, it opens up a Pandora's box of things that may not serve us well," he said. However, he also voted yes for more limited use.

Wallace Kemper Alston, MD, professor of medicine in the Infectious Disease Unit at the University of Vermont, Burlington, voted yes to both questions. "Of all these applications I've listened to over the years, I think this one ended up the most muddled, but I don't think that's necessarily the fault of telavancin. It speaks to the fact that this is a very heterogeneous, very difficult diagnosis to make. It's very hard to assess a cure."

He concluded, "In my heart of hearts, I think telavancin would probably cure [methicillin-resistant S aureus] pneumonia at least as well as vancomycin and that the problems we're encountering have to do with the unknowns surrounding trial design for nosocomial pneumonia and not the drug itself."

Participants of FDA advisory committee meetings are vetted for conflicts of interest and waivers are granted to participants with conflicts if necessary. No waivers were granted for this meeting.