Antibiotics: When Science and Wishful Thinking Collide


Antibiotic resistance is a major concern confronting our health care system, and there is tremendous pressure on the Food and Drug Administration (FDA) to “do something” about it.  Unfortunately, the FDA is responding by approving drugs that are likely to do more harm than good.

FDA advisory committees are supposed to provide independent advice from experts across the country, but recent meetings have left observers wondering whether too many FDA advisory committee members are providing neither scientific nor independent advice, and whether the committee process itself is fundamentally flawed.  These concerns dovetail with essential questions about FDA objectivity and scientific judgment in its review of antibiotics.

For example, a few weeks ago, the FDA approved a new drug for multidrug-resistant tuberculosis (TB) — bedaquiline, to be sold under the name Sirturo — shortly after data reviewed by FDA scientists indicated a higher death rate for the new drug compared to the usual standard of care.  The FDA’s Anti-Infective Drugs Advisory Committee had publicly reviewed the TB antibiotic on November 28, 2012. (See Note 1)

The company’s clinical trial data compared the standard drug regimen to the new drug taken in conjunction with the standard drug regimen. This was a high priority, expedited review, so the company’s randomized, clinical trial wasn’t yet completed.  You can clearly see (as the committee could) that patients treated with bedaquiline in addition to standard therapy were 10 percent more likely to die than those receiving the standard therapy alone, and the survival gap between the new drug and standard therapies increased over time. (See Exhibit below, click to enlarge) The difference was statistically significant, and a 10 percent increased chance of dying is not something most doctors could justify to a patient when choosing an antibiotic.

At the FDA meeting, these risks were downplayed by the manufacturer, Janssen Therapeutics, whose spokesperson suggested that the higher deaths resulted from chance, not from the drug.  That seemed like an odd rationale, since the purpose of conducting this or any other controlled, randomized trial is to ensure that any statistically significant difference in mortality is due to the treatment, not chance. The FDA’s scientific staff had concluded that the new drug could be causing the larger number of deaths, and the Committee expressed their concerns.

Apparently worried about losing the support of the advisory committee, representatives from Janssen offered to immediately provide committee members with data from the longer-term follow-up of the same study, which had just been completed.  This offer seems reasonable but is against the FDA rules: all of the sponsor’s data to be discussed at the public meeting must be submitted 22 business days before the meeting, so that FDA scientists can review it and provide the data and the FDA’s written analyses to Committee members at least 14 business days before the meeting, and to the public 48 hours before the meeting.  The FDA rules are also reasonable, because FDA analyses sometimes disagree with the company’s methodology or interpretation, and advisory committee members rely on those FDA analyses for objective peer review.

The FDA rules were ignored at the November 28th meeting.  The Committee chair agreed to Janssen’s request, and FDA officials did too. Those new data, which had not been included in the FDA’s scientific analysis provided to the Committee, were more favorable to the drug than the data that the company had previously provided. Our notes from attending the meeting indicate that the death rate from the new drug was still higher but the difference was not statistically significant.  Regardless, the company’s enthusiastic presentation of the non-peer-reviewed data was accepted at face value by most committee members, who voted 11-7 that the new drug was safe. A few weeks later, the FDA approved the drug with great fanfare as the first new TB medication in 40 years.

Did the company’s new data objectively prove that this new TB drug is safe and effective?  It’s impossible to know, because the FDA has still not made those data available on the FDA web site or elsewhere for independent reviewers.  The law requires that advisory committee data be made public in advance of the meetings, so that anyone can review it.  In this case they were not public before the meeting, or before or since the FDA made its approval decision.

The desire to approve a new TB drug is understandable, and the FDA mandated that Sirturo carry a Black Box Warning regarding the higher death rate experienced by patients taking the drug. A Black Box Warning is an important safeguard, but does not justify approving an antibiotic under these circumstances.  Unfortunately, the questionable approval of bedaquiline is far from unique.

The day after the November 28th FDA meeting, the same advisory committee met to discuss telavancin, an antibiotic that is already on the market by the name of Vibativ for the treatment of complicated skin infections.  The drug has been found to cause birth defects in animals and potentially deadly kidney toxicity in humans. With many safer drugs available, telavancin has twice before been denied approval by the FDA for treating pneumonia.

These same adverse reactions remained concerns regarding the pneumonia trials submitted to the November advisory committee. However, unlike previous reviews for this same drug, the November advisory committee members were apparently overcome by wishful thinking.  They concluded that the hoped for but unproven potential benefit of telavancin for a small minority of patients was enough to warrant these high risks and recommended approval. As they had the previous day, the committee members speculated that the possible benefits to patients in severe need, such as those with MRSA (methicillin-resistant Staphylococcus aureus), outweighed the risk of death.  Unfortunately, the FDA scientific analysis of the evidence does not support that view.  Analysis of the company’s own studies showed that the sickest patients– those with diabetes, heart failure, and kidney failure– were most likely to die if they took telavancin compared to vancomycin.

These two committee meetings highlight a problem with the FDA’s accelerated approval process, and leave patients, consumers, and public health experts with growing questions about whether advisory committee meetings are merely a pretense that is being used by the FDA to justify decisions that are not based on science.

Question #1:  Why is a drug that lacks clear evidence of efficacy or safety the subject of the time-consuming and resource-intensive review of a public FDA advisory committee meeting?  Is it because the FDA is under so much pressure to approve new drugs, and especially antimicrobials, that they are afraid to make a decision without support from an outside advisory committee?

Question #2:   Why did the FDA allow data that were not vetted by their own scientists to be discussed and used as the basis of advisory committee recommendations?  A related question is whether this is only allowed for data that the company provides, but not for data provided by other reputable sources.  The most egregious comparison is the FDA’s advisory committee meeting on the birth control pill Yasmin, just a year before.  In that case, the well-respected former FDA Commissioner, Dr. David Kessler, had written a report that documented that Bayer covered up data describing the high risk of venous thromboembolisms (VTEs) among Yasmin patients, but the FDA would not allow Kessler’s report to be shared with the committee or discussed during the public meeting because it was submitted a few days before the meeting, which was “too late” to be included in the committee materials or FDA peer review.  In contrast, this time the Janssen data were allowed to be provided and discussed during the meeting itself.

Question #3:  Why are FDA advisory committee members recommending approval for drugs that are not proven safe or not proven effective?  Based on the one-sided enforcement of FDA rules described in #2, above, and the quick approval of bedaquiline just a few weeks afterward, one has to wonder.  The advisory committee recommendation in favor of telavancin further raises the question of whether advisory committee members are selected to recommend what someone at the FDA has already decided to do.

Feeling worried?  We are.  We don’t know what is going on behind closed doors at the FDA, but based on public meetings, policies regarding the analyses of safety data are vague and sloppy, with strong warning signs overlooked in the desire to push new drugs through the pipeline.  The FDA has clear guidelines for proving efficacy, but these standards are being loosened to accommodate surrogate endpoints that do not necessarily predict health outcomes. Science is being sacrificed for speed, and this puts patients at greater risk for exposure to drugs that are not proven safe.  We’ve given two recent examples but can provide dozens more.

The risks are very real.  The last two drugs considered for ventilator-acquired pneumonia, doripenem and tigecycline, were found to increase mortality in patients after they were approved for treatment of other infections.  The FDA denied approval to telavancin previously, but judging from their public meeting the FDA may now be bowing to pressure to approve more antimicrobials, regardless of the scientific evidence.

Even more disturbing, antibiotic drug development and review seem to be ingrained with indefensible scientific practices.  Drugs are assigned to priority review before Phase II and Phase III safety studies on humans are completed, and mortality is no longer considered an important outcome by the FDA. Instead, the FDA supports studies of drug efficacy based on test tube data rather than patient-centered outcomes, such as survival and better quality of life. This distinction is important to patients and their families, because an antimicrobial drug can kill more bacteria in a test tube and yet result in more rather than fewer deaths when used in patients.

Companies need to be held to higher standards when developing antibiotic drugs than the FDA currently requires. There is a lack of scientific rigor in clinical trials for drugs that are being recommended for approval.  Large numbers of patients who won’t benefit are being exposed to risky new therapies as part of clinical trials that are not always well designed and often do not measure the outcomes that matter to patients, such as survival. The hope of a benefit for a small group should not be reason enough to expose a large number of patients to increased harm.  We have seen how expediting drugs through the regulatory pipeline to reach these sick patients comes at a large cost in patients’ lives.

A few months ago, Congress passed a law that provides generous incentives for companies to develop new antimicrobial drugs, extending market exclusivity for five years.  Under this law, four drugs have already been given QIDP (Qualified Infectious Disease Products) status.  Another new regulation is being considered that would alter the review process to permit limited approval for drugs using another expedited process. As the FDA grants priority review and fast-track status, we need to ensure that they are based on solid clinical trials demonstrating drug efficacy and safety.

FDA guidelines need to do more to reduce the overuse of antibiotics. Overuse of one type of antibiotic drug can result in resistance to similar drugs, including even the newest drugs.  Pharmaceutical companies admit that use of older antimicrobials that were approved without appropriate studies is contributing to antibiotic resistance in critical pathogens and higher rates of mortality in patients. Based on their calculations, the rate of resistance to currently used antimicrobials will make drugs in development obsolete even before their patent exclusivity expires under QIDP. Instead of promoting the generation of new and improved drugs, the incentives being awarded by the FDA are restricting other companies from generating better drugs while approving those that have not been proven as effective or safe.

Something has to give. Public health advocates need to persuade the FDA to stick to the science when it makes approval decisions for anti-infective drugs. Approving drugs without proven safety, efficacy, and benefit (often with increased mortality in comparison to older drugs!) is only serving to put all of us at increased risk.

Note 1. Although “antibiotic” is the commonly used term, “antimicrobial” or “anti-infective” are more encompassing terms. All three are used interchangeably within this article.