The FDA Gets Lost on its "Critical Path"

I have to admit it – even I am a little impressed with the continued good results from  Qnexa’s second year trial, Sequel. Subjects generally did not lose more weight, but they did an impressive job of sustaining the weight loss results of the one-year Equip trial. And apparently they did this without further complications of safety issues – but I do have to stress apparently, since we have only Vivus’ say-so to go by.

So Qnexa, whether it gets approved or not, is the harbinger of the future of weight loss drug development – what the FDA calls “combination products.” Qnexa is the first combination agent to be reviewed by the FDA. Contrave is up next, with a December Advisory Committee date.

There’s no need to belabor the urgency of the ever-growing rate of obesity in the American public. The U.S. is first among major nation is obesity rates, and it threatens to overwhelm the nation’s healthcare budget. Obesity has overtaken cigarette smoking as the number one health hazard & the leading cause of death. Healthcare costs related to obesity are already widely acknowledged to be 10% of all such costs – the most recent estimate is actually 17% – and growing. The Center for Disease Control just released an ominous study suggesting that one in three Americans could be diabetic by 2050.

The FDA’s Critical Path Initiative

Recognizing the developing urgency of the obesity epidemic, the FDA in 2003 and 2004 held Advisory Committee meetings on this topic. Both meetings reviewed the history of weight loss drugs, which highlighted the paucity of drugs, the checkered history of their approval & subsequent withdrawal, & their enduring lack of efficacy. In the face of this history, a prime motivation was to consider ways to encourage more drug development by the pharmaceutical industry.

In the 2003 meeting, Dr. Janet Woodcock, today’s current Director of the FDA’s Center for Drug Evaluation & Research, gave the meeting’s central presentation on the development of a Critical Path forward in obesity research and, in particular, to foster new product development.

There is this tremendous explosion of basic scientific knowledge. But

... we feel it hasn't been applied to actually the process of getting these

products developed.

This process, which is called the development process, is on the critical
path to getting these products to the patients. We think it is becoming
a serious bottleneck to delivery of new products as the basic science enables
more discovery and more candidates to be identified.

Dr. Woodcock stressed the importance of FDA guidance in achieving these ends:

FDA guidance documents . . . are known to foster innovation and
improve chances of success -- the existence of these guidance documents –
and they seem to. Where there is clarity in a development pathway, in
a critical path, it fosters attention and innovation and investment in
that product area. So, where developers see a clear path forward, they
will focus their innovation in that area.

FDA Assessment of the Current State of Obesity Agents

The 2004 meeting was convened specifically to discuss revisions to the original 1996 Guidance on weight loss products. Dr. David Orloff, then the Director of the DMEP, stressed, again, the importance of the FDA’s role in supporting industry in undertaking more obesity research & development:

Germane to our work here today, [former FDA Commissioner Mark McClellan]
charged the so-called therapeutic subgroup, which was led by our division,
to assess real or perceived barriers to obesity drug development and to
make recommendations on ways to encourage the development of new or
enhanced therapeutics for obesity.

Several reviews were presented at the 2004 meeting: of the regulatory history of weight loss agents (Dr. Eric Colman), an overview of the current situation (Dr. Frank Greenway), and a current & future perspective (Dr. Richard L. Atkinson). All of these reviews referenced the continuing modest efficacy achieved by virtually all obesity agents. Dr. Greenway concluded his overview of the current state of treatment:

In conclusion, all these drugs give between a 2 and 6.5 kg greater weight
lossmthan placebo in trials that last up to a year, and the amount of weight
loss appears to be medically significant. The weight loss between these
different drugs is not different statistically and the choice, therefore,
revolves around side effects.

In a slide presentation in 2003, Dr. Orloff had presented the following conclusion about the historical mean weight loss of obesity drugs: “Average placebo-subtracted weight loss of drugs evaluated to date [equals] 3-5% of baseline at year 1”

Dr. Orloff, in 2004, also speculated about the likely efficacy of future weight loss drugs:

Our experience … is with only modest efficacy of drugs that we have
approved and evaluated to date and ...[without the] anticipation,
frankly, that there is anything in the pipeline that is going to be
dramatically more effective than what we have seen already.

Given the accepted evidence of the continued modest efficacy of weight loss drugs, a prime consideration for the 2004 meeting was the development of guidance for the approval of “combination products” in the hope of increasing effectiveness. This “guidance” was incorporated into the 2007 revised Guidance on Weight Loss Products. As a first attempt at formulating standards of safety and efficacy for these products, perhaps I should not be overly critical. But this guidance is clearly inadequate, as I have detailed here.

The FDA gets lost on its Critical Path
However, it is not in its promotion of combination drugs that the FDA has gone astray. One aspect of the “wrong turn” the Agency has made is in forgetting what the “combination products” actually combine: individual single agents.

It seems to have forgotten that without novel drugs, based on new research, combination drugs will just keep combining old anergenic, stimulatory drugs like phentermine with haphazard, already-approved drugs that happen to demonstrate some weight loss properties. That field is pretty narrow – & the low-hanging fruit has already been picked. The FDA needs to encourage the development of new agents – that’s what its 2003 and 2004 meetings were designed to do.

In the 2003 advisory meeting, new developments in obesity research were hailed as promising innovative new treatments to come. Dr. Orloff’s slide presentation detailed the following new avenues for research, any of which could result in the development of effective new medications:

* Serotonin
* Leptin/CNTF
* Neuropeptide Y
* Peptide YY
* Ghrelin
* MC 3, MC 4
* MSH
* MCH
* CRH
* Urocortin
* Galanin
* H3
* CART
* Amylin
* Orexin
* CCK-A
* GLP-1
* Bombesin
* Beta 3 adrenergic system
* GH
* Cannabinoid receptor
* Topiramate
* Thyroid modulators

The research avenues were many & promising – but the very abundance of them augured something less positive. Apparently, weight is regulated by the human body in many different ways. The body’s weight was maintained by an evolutionary-designed abundance of feedback & homeostasis mechanisms. All of these avenues might provide just one small piece of the overall obesity puzzle. Given the multiplicity of mechanisms, however, even an effective intervention with one aspect would likely be overcome by the combination of all the other mechanisms.

So the “many avenues” of new research was a double-edged sword. Yes, it augured new drugs to come. But it also suggested that the weight loss from any one such agent was likely to continue to be modest, and likely to be short-term. Its effect would be swamped by the other biological mechanisms mandating a return to the body’s state of “normalcy.” It was this discouraging realization – that even future anti-obesity agents would be likely to provide only “modest” results, which stimulated the current interest in developing combination products.

The FDA’s mission was obvious: how could they encourage development of new obesity agents, so they can be utilized in creating safer and more effective combination agents?

Lorcaserin – the only novel obesity product under review

The only product of the new wave of obesity research is Arena Pharmaceuticals’ lorcaserin (brand name Lorqess).  It is the result of innovative research on 5HT2c receptors, and as such it is the only novel weight loss agent under FDA review.

Incredibly, the FDA seems intent on discrediting Lorqess based on what it has termed “marginal” efficacy, despite its clear achievement of the FDA’s categorical standard for efficacy. 

(Let me acknowledge here that lorcaserin may have unresolved safety issues. That is an entirely separate issue, & not the subject of this particular essay.)

Lorcaserin’s mean placebo-adjusted weight loss is 3.3% – clearly in line with the 3% – 5% historical average cited by Dr. Orloff.  True, the 3.3% mean efficacy does not satisfy the FDA’s first criterion for effectiveness. However, it clearly meets the FDA’s second, alternative criterion: the categorical guideline, requiring at least 35% of a trial’s drug arm to achieve the 5% weight loss standard; and that the proportion of the drug arm meeting this standard be “approximately double” the placebo arm.

A few more words about the 2004 meeting are in order here, because that meeting also discussed the categorical standard in some detail. One doctor at this meeting was extremely critical about using it as an acceptable standard for judging a drug’s efficacy.  Dr. Jules Hirsch argued:

I think the categorical thing, I would worry about that. I think
that is a snare and a delusion and any statistician knows that post
hoc you can always find a category that you can make significant.
You can pick between 8-11 percent, 3 percent, whatever it is. So,
if you do categorical things, they must be stated ante hoc and you
must examine the distribution, that is, the mathematical distribution
of the weight loss. . . one must be statistically extremely careful
about categorical things.

The critical thing to realize about Dr. Hirsch’s assessment is that he is criticizing the 1996 Guidance. In the 1996 Guidance, the categorical standard was very different than the current standard detailed above. It simply required that the proportion of drug arm subjects who achieved 5% weight loss be GREATER than the proportion of placebo arm subjects who did likewise; and that the difference be statistically significant. So by this guidance, a drug that had only, say, 10% of responders could be declared approvable, as long as still fewer placebo subjects achieved 5% weight loss. It was also not required that a Sponsor pre-specify the categorical standard as the trial’s endpoint. They could specify a mean weight loss of 5% as their endpoint; and then, if they failed to achieve it, they could “fall back” on the categorical standard to justify their agent’s efficacy.

Important in retrospect,at the 2004 meeting Dr. Eric Colman, Deputy Director of DMEP, not only endorsed the categorical criterion, but actually stipulated that it was not even necessary to pre-specify the categorical standard in a clinical trial:

Dr. Eric Colman (2004):

"I have seen companies choose one or the other [Mean Difference Criterion or Categorical Criterion] as the primary efficacy outcome. In some cases, if a company prespecifies that they will only use the mean difference between groups of 5 percent and they don't make it on that

but they have, as a secondary outcome, the categorical, and that does make it, then we would be inclined to consider that drug efficacious."

Dr. Susan Yanovski was the first of several doctors to express disagreement with Dr. Hirsch regarding the usefulness and appropriateness of the categorical criterion:

Regarding magnitude of effect, I am going to disagree with Dr. Hirsch,
my esteemed colleague, on not having categorical definitions. First

of all, I don't think any of the drugs we have today--I am not sure

any of them would meet the 5 percent across the board difference from

placebo. We are going to find responders and non-responders and I think
that having the categorical variables and allowing a drug to be approved
that may have a standard number of responders is important. Now,

the magnitude of responders versus non-responders in the studies that

I have seen is often pretty substantial. It might be 35 percent versus

17 percent in the placebo group. There is often a doubling of the

responders. I think it would be a good idea to actually quantify the
difference between placebo and active treatment groups in the percentage
of responders and non-responders, not just make it statistical significance.

As it turned out, Dr. Yanovski’s references were the standards actually incorporated into the revised 2007 Guidance: a minimum of 35% responders, which is “approximately double” the placebo group (“35 percent versus 17 percent in the placebo group”).Note that a 35% responder rate is considered "pretty substantial," in and of itself. The new 2007 categorical standard represented a much higher hurdle to overcome in establishing a drug's efficacy.

So how did Lorqess stack up against this guidance? Solidly – a full 47% of the drug arm achieved the 5% standard, versus only 22% of the placebo arm.

Yet how does the FDA view lorcaserin’s efficacy? Not a single FDA official appears to be able to get five simple words out of his mouth: “Lorcaserin meets the categorical standard.” He must insert “just” or “by a slim margin.” Dr. Colman summed up his opinion on lorcaserin’s efficacy, in the introductory Memorandum in the FDA briefing document, this way:

Therefore lorcaserin did not satisfy the guidance’s mean efficacy  criterion. However,
the lorcaserin 10 mg BID dose did, by a slim margin, satisfy the categorical efficacy
criterion.

Dr. Colman is denying basic arithmetic truth – 47% of responders is substantially higher than the required 35%. It is also plainly, mathematically, more than “double” 22%. What will he deny next: gravity?

This point cannot be made too often or too emphatically – lorcaserin met the categorical standard.  It met the standard solidly. It met the new, revised, much harder to achieve 2007 criterion. It met the criterion that Arena pre-specified in both its Phase III trials.

No one in the FDA has publicly challenged Dr. Colman in this false conclusion. He persisted in minimizing lorcaserin’s efficacy throughout the advisory panel, & his attitude was reinforced by Dr. Golden’s short, biased presentation on the subject. Their attitude was picked up & echoed by several panel members.

So extreme was Dr. Colman’s attitude that he all but said he would not consider lorcaserin approvable even if it had no safety issues. Is it any wonder, in the face of such intransigent casuistry, that many spectators are questioning the impartiality & integrity of the FDA in general, and Dr. Colman in particular?

The FDA appears to have “forgotten” the rationale for the categorical criterion. By taking lorcaserin, almost half the obese/overweight population would lose 5% or more of their excess weight – this is undeniably clinically significant.  

In addition, Dr. Frank Greenway, at the 2004 meeting, argued for the greater relevancy of the “completer” data:

The traditional way of analyzing these studies, as Susan suggested, has been the last observation carried forward [ITT-LOCF], and what that does is it dilutes the effect of the
drug because it assumes that the reason the people dropped out is because they didn't lose weight. Actually, what the physician treating a patient is interested in is ... what happens to the patient ... who stays in treatment, rather than the more public health perspective of this last observation carried forward, which looks at the entire group. If you treat everybody, what does the total group gain from this experience? So, from the way in which these medications are used, it is much more informative to me, as a clinician, to have the analysis of completers rather than the last observation carried forward.

Using the clinically relevant “completer” statistics, you can start to get an idea of how truly impressive lorcaserin’s weight loss can be. The average weight loss among all subjects who completed a year on lorcaserin was 17 pounds. Sixty-four percent (64%) of the subjects who completed a year on lorcaserin lost > 5% of their weight – the average weight loss for this 64% was 26 pounds. Over one-third of completers lost > 10% of their weight. For the top 25%, the average weight loss was 35 pounds.

This health-changing and life-changing degree of weight loss, which could be obtained by large subsets of the obese population, will be denied by a deranged or deeply biased FDA who insist on focusing on the 3.3% mean weight loss criterion, and who deny, contrary to rudimentary arithmetic, lorcaserin’s achievement of the categorical standard.

FDA Guidance is Central to the Critical Path

The purpose of FDA guidance bears reiteration. I think it’s appropriate to let Dr. Colman help out here. He concluded his address to the 2004 meeting on FDA regulation thus:

Dr. Eric Colman (2004):
Since the topic of today's discussion is the obesity guidance document,
I thought I would just provide a visual reminder of the goal ...of all [FDA] guidance
documents, and that is obviously, on the one hand, to facilitate industry's 
development of safe and effective drugs but, just as importantly, to provide
regulators with the best available evidence upon which to judge a new drug's
risk/benefit profile before the drug is approved.

It follows that,  when the new Guidance was crafted in 2007, the categorical standard was considered to be "best available evidence" upon which to judge efficacy. What changed by 2010?

Recall also the words of Dr. Janet Woodcock, in the central presentation of the 2003 advisory meeting:

FDA guidance documents … are known to foster innovation and improve
chances of success. Where there is clarity in a development pathway,
in a critical path, it fosters attention and innovation and investment
in that product area. So, where developers see a clear path forward,
they will focus their innovation in that area.

Dr. Woodcock specifically endorsed the importance of FDA guidance documents in providing this clarity -- clarity upon which industry could rely as they developed a new weight loss agent:.clarity that would prevent a biopharm from wasting hundreds of millions of dollars developing a product that would not meet the FDA's standards.

Eventually, if we have recommended a scientific approach, as we take these 
to our advisory committees or we publish guidance, we have a public process for 
discussion and then that becomes a public standard that developers can use. 

So they can pick up that standard, if they have a product, and they can just 

apply it. They don't have to develop it.They don't have to agonize over it. 

That is the standard they can use.We find that this really facilitates development.

Dr. Janet Woodcock is the Director of the Center for Drug Evaluation & Research, and the agency’s foremost proponent of the Critical Path. Where is Dr. Woodcock on this issue? Will she allow her subordinates to undermine the entire basis for industry trust in FDA Guidance, as well as her Critical Path initiative?

What chilling effect does she think this will have on industry as a whole? Dr. Woodcock cited the “bottleneck” in the obesity development cycle -- tremendous strides in obesity research are not being utilized to develop into usable products. What biopharm company will undertake novel product development in such treacherous waters -- which one will feel it can rely upon the Division of Metabolic/Endocrinology Products to apply its own Guidance fairly?

I would remind the FDA of Dr. Orloff's own assessment -- that future products, developed on the basis of new scientific advances, are likely to be as modest in efficacy as anything produced to date. I would remind them again of the wrong-headedness of comparing a single agent like lorcaserin to a double agent like Qnexa.

Given their own acknowledgment that virtually all obesity agents have produced mean weight loss of 3% – 5%; given their understanding that this modest efficacy is likely to characterize future as well as past drugs; given their stated commitment to standing by the standards they have themselves imposed – how can they reasonably deny lorcaserin’s acceptable efficacy?

This is an appraisal that the FDA needs to re-think and recant.