Through a glass, Darkly – The FDA Perspective on Obesity Agents

It remains to be seen how this race will turn out. The FDA, in the final two weeks of October, issued a Complete Response Letter (CRL) to both Arena and Vivus Pharmaceuticals. The letters to both companies requested clarification on a number of safety issues, with the threat of more clinical trials if they could not satisfy the FDA.  The CRL issued to Arena additionally cited its "marginal efficacy" as being problematic in its approval.

Lorcaserin’s "marginal efficacy" was a canard continually promoted by the FDA in the lorcaserin briefing documents & in the lorcaserin advisory meeting.  Yet the FDA has written guidance that establishes the criteria by which a drug's efficacy should be judged – and lorcaserin clearly meets this guidance. This guidance also provides additional criteria that should be met by “combination products” such as Qnexa. The FDA, however, seems unable to correctly apply the clear language of its own guidance.  As a result, their views of both lorcaserin & Qnexa are seriously skewed.

I have taken the time to go back & review the FDA's purported views on efficacy.  I have reviewed not just the current 2007 Guidance for weight loss products, but the original 1996 Guidance and the transcripts from two Advisory Committees that the FDA held specifically on obesity drug development: one, their 2003 Advisory Committee on  the obesity epidemic & the Critical Path forward to develop new products ; and two, their 2004 Advisory Committee, which considered specific revisions to the 1996 guidance  The FDA appears to have selective memory of these events, and appears to apply their precepts haphazardly.

The current guidance offers two main alternative criteria to establish a drug's efficacy. The first is the mean weight loss standard, which requires mean placebo-subtracted weight loss of at least 5%. The second is the so-called "categorical standard," which requires demonstrating that the proportion of subjects who lose > 5% of their baseline weight is greater in the drug arm versus the placebo arm.

Frequently overlooked, however, is the additional guidance that is provided, for the first time, for combination products. The advisory committees just referred to, held in 2003 and 2004 to consider a "path forward" in obesity drug development, endorsed the necessity of trying combination products, combining two or more agents, in an attempt to procure greater efficay. To that end, the revised 2007 guidance made a first attempt at establishing efficacy and safety guidelines for such combination products.

Qnexa, as the preponderance of evidence will show, does not meet this guidance, and it should not be approved.  Its multiple serious risks outweigh the modest 3% benefit it provides in comparison to its constituent elements. This is the standard by which it should be judged. Lorcaserin, on the other hand, fully meets the FDA's guidance, & deserves approval once the FDA's illegitimate cancer concerns have been allayed.

The Double Standard, Part I
The FDA considered Qnexa's efficacy to be so solidly established that the FDA deputy director of the Division of Metabolic & Endocrinology Products (DMEP),  Dr. Eric Colman, opened the Qnexa Advisory Committee in July by taking it off the table as an issue for consideration:

In general, the FDA is in agreement with the company with respect to the

weight loss effects of the drug that's been proposed, trade name of Qnexa.

So you will see during the FDA presentation that the focus is primarily on

safety.

This "pass" which Dr. Colman gave Qnexa on efficacy was important, because Qnexa's efficacy, if examined, does not hold up to scrutiny.

In 2004, the FDA Advisory Committee met to discuss revisions to the 1996 Guidance. At this meeting, Dr. Colman provided an overview of the history of obesity agents, which emphasized the modest efficacy achieved by virtually all such agents.

In the face of this record, the advisory panel considered the future of obesity drug development, the obstacles the pharmaceutical industry faced in developing new & more effective agents, and how the FDA could encourage this development. The many new advances in obesity research foretold two developments in drug development: (1) since obesity seemed to involve so many biological processes & feedback loops, new research highlighted the near impossibility of finding any one simple solution; and (2) this research yet held out hope for many new avenues to exploit in drug development. The problem was, each individual “solution” to any specific piece of the “obesity puzzle” would likely be counterbalanced by some other biochemical process. This augured continued modest efficacy, and short-term weight loss that would tend to revert to “normal” by the body's homeostatic mechanisms.

Given these issues, the panel discussed how to encourage more research on combination products – whether by universities, the National Institute of Health (NIH), or the pharmaceutical industry. Given the size – and cost – of the trials necessary to provide sufficient evidence of efficacy and safety, it appeared industry must be encouraged to develop combination agents, with the opportunity to exploit these agents financially with new patents on the combinations themselves.

In a discussion of how to provide guidelines by which to assess the efficacy & safety of these new combination products, it soon became evident that many variables impinged on this issue. Should a combination be required to offer twice the efficacy of either of its constituents? Perhaps – but what if the combination offered a superior safety profile in addition to better efficacy? In that case, should not a lower efficacy be acceptable, as tradeoff for the better safety? And what if a combination agent attained similar efficacy, but at lower dosages than usually prescribed? And so on – to the point that Dr. David Orloff, Director of DMEP in 2004, decided it was too complicated to come to any resolutions at the panel meeting.

Given, perhaps, this complexity, the 2007 Guidance that eventually emerged did not provide very clear guidelines for combination products. On efficacy, the Guidance states:

   
            We have not defined a minimum difference in weight loss between a fixed-
            dose combination and its individual component products that should be
            achieved for the combination to be considered more efficacious than either
           of its components when used alone. However, a fixed-dose combination 
           that is associated with at least twice the weight loss observed with that of 
           each of the individual components will be viewed more favorably than
           combinations that do not achieve this degree of relative weight loss.

With regard to safety, the guideline is spelled out in the FDA Qnexa Advisory Committee briefing documents:
             Although not a requirement for approval of a fixed-dose combination
             product, the Division looks more favorably on combination products
             that exhibit a potentially meaningful improvement in the safety profile
             compared to the individual components.

So the guidance the FDA provides is not terribly clear, and is therefore amenable to interpretation.  Certainly, it did not meet either of these "preferred" guidelines, neither in efficacy nor in safety.

 “But they don’t have to,” says Vivus and the FDA. The Guidance just says it will be viewed more favorably.  That is true. However, if the weight loss was worse than the individual components, then that would obviously be viewed extremely unfavorably. It would not be approvable – it would negate the whole point of combining agents. 

Likewise, if the safety profile of the combination agent was worse than its constituent components, then that would also be viewed extremely unfavorably.  That may also negate the approvability of a combination product – unless, possibly, the efficacy was truly stellar.

The 2007 Guidance requests that a study trial be done which compares a combination product with both of its constituent components separately, as well as with placebo:

We recommend that the efficacy and safety of fixed-dose combinations be compared with the individual product components of the combination and placebo in phase 2 trials of sufficient duration to capture the maximal or near-maximal weight-management effects of the products

Vivus performed this comparison in its first Phase III clinical trial, OB-301. This trial showed that Qnexa High-Dose produced mean weight loss of 9.0 kg, compared to 6.1 kg  for its toparimate component and 5.8 kg for its phentermine component.

	PLACEBO	PHEN 7.5 MG	TPM 46 MG	QNEXA MID-DOSE	PHEN 15 MG	TPM 92 MG	QNEXA HIGH-DOSE
Mean % Weight Loss	-1.5%	-5.2%	-4.9%	-8.2	-5.8%	-6.1%	-9.0%
Placebo-adjusted Weight Loss	-----	-3.7%	-3.4%	-6.7%	-4.3%	-4.6%	-7.6%
Individual Component-Adjusted Weight Loss	-----	-----	-----	-3.0%*	-----	-----	-2.9%**

  *Qnexa at Mid-dose (Phen7.5/Tpm46) mean weight loss minus Phen 7.5 mean weight loss.

**Qnexa at High-dose (Phen15/Tpm92) mean weight loss minus Tpm 92 mean weight loss.

This table clearly shows how Qnexa compares to placebo versus how it compares to its individual components. Just as any weight loss agent must show that it provides a substantial increase in effectiveness as compared to placebo, a combination agent should additionally demonstrate substantial increase in effectiveness as compared with either of its individual components. As the FDA repeatedly emphasized in regard to lorcaserin, a 3% differential in efficacy is modest.  In the CRL, the FDA characterizes this difference as "minimal."

Why is lorcaserin being held to a higher standard than Qnexa?

The FDA reviewer concluded: “Treatment with PHEN/TPM [Qnexa] resulted in an additional 3% weight loss compared to the individual components alone.”   This degree of improvement in efficacy does not meet the FDA's preferred guidance of "at least twice the weight loss observed with that of each of the individual components."  More importantly, it also fails to meet the reasonable difference of 5% over each individual component – this 5% difference over components would be the natural corollary of the basic efficacy requirement which requires a 5% difference over placebo.

So although the  FDA has not yet set clear guidelines for the degree by which a combination agent should exceed the efficacy of either of its constituent elements,  nevertheless the cue should reasonably be taken from its efficacy guidance for weight loss drugs in general: a difference of 5% or greater compared to placebo. Since they require that the "efficacy and safety of fixed-dose combinations be compared with the individual product components of the combination and placebo,” it seems reasonable that the same 5% difference that applies to placebo should also apply with regard to the individual components.  That is a standard that Qnexa clearly fails.

But it’s the safety profile that really does not bear scrutiny.

Moving to the safety side of the equation, the FDA prefaces its safety review of the OB-301 data with its own statement of a desirable safety profile, then with two interesting hypotheses made by Vivus:

Although not a requirement for approval of a fixed-dose combination product, 
the Division looks more favorably on combination products that exhibit a 
potentially meaningful improvement in the safety profile compared to the 
individual components.

 In addition, the applicant has theorized that some of the expected side effects 
of the two drugs alone may be “mitigated by oppositional pharmacodynamic 
effects associated with the other component.” In particular, the cognitive slowing observed with topiramate treatment may be lessened with co-administration of phentermine.

This is followed by ten pages of complicated tables and lovely, colored charts -- and several bland statements of damning import.  These results might be best illustrated with a few simple charts, pulled from the chaotic detail of the data tables.

The safety issues most closely associated with phentermine and toparimate include psychiatric-related issues, cognitive-related issues, & cardiovascular issues. (These are three of the five explicitly labeled by the FDA as the relevant safety concerns – the other two are teratogenicity and metabolic acidosis.)

Percentage of  Adverse Events in OB-301, by treatment category.

Nervous System disorders

PLACEBO             24.8%

PHEN 7.5               22.0%

TOP 46                   38.7%

Qnexa MID-dose          40.6%

PHEN 15                20.4%

TOP 92                   36.4%

Qnexa HIGH-dose                   49.1%

Psychiatric disorders

PLACEBO             13.8%

PHEN 7.5               10.1%

TOP 46                   17.0%

Qnexa MID-dose      17.0%

PHEN 15                15.7%

TOP 92                   17.8%

Qnexa HIGH-dose                  25.0

Cardiac disorders

PLACEBO             0.9%

PHEN 7.5               1.8%

TOP 46                   1.9%

Qnexa MID-dose            2.8%

PHEN 15                0.0%

TOP 92                   0.9%

Qnexa HIGH-dose                                                             6.5%

These simple tables clearly illustrate that, for the safety issues most closely associated with phentermine and toparimate, the combination of these drugs in Qnexa creates a synergistically much worse result – especially in the Qnexa high-dose..

Vivus may argue that, although the differences look stark as set out in these tables, nevertheless many of these side effects are mild and frequently don't result in discontinuation on Qnexa. Though there may be some truth in that anticipated assertion, this table shows nevertheless that the safety profile of Qnexa does cause many more discontinuations than either placebo or phentermine or topiramate alone.

Adverse Events as Main Reason for Discontinuation of Study Drug:

PLACEBO       7.3%

                                                           

PHEN 7.5        9.2%   

TOP 46            7.4%

Q MID                        15%

PHEN 15         10.2%

TOP 92            16.8%

Q HIGH                      21.3%

The bottom line is this -- if an obese person was treated with phentermine or topiramate alone, even at the reduced dosages used in the Qnexa trials, they would achieve significant weight loss.  The modest incremental weight loss achieved by the combination of these agents in Qnexa is not justified by the additional safety concerns.

Lastly -- with regard to Vivus’ proposal that the safety issues would actually be “mitigated by oppositional pharmacodynamic effects associated with the other component”? The FDA reviewer provides no explicit conclusion – perhaps he hopes we’ll have forgotten Vivus’ awkward hypothesis by that time.  And Vivus’ even more hopeful conjecture – that the cognitive slowing associated with toparimate will be mitigated by phentermine?  Nope. Cognitive function was specifically demonstrated to be worse on Qnexa, at both dose levels, compared to phentermine or toparimate alone.

The Double Standard, Part II

The FDA recently rejected Arena Pharmaceuticals drug candidate lorcaserin, partly due to what they termed 'marginal efficacy."  This conclusion is truly puzzling, since lorcaserin easily met the required criterion. In the same 2004 Advisory Committee that debated the approvability of combination products,  other proposed revisions to the 1996 Guidance on weight loss products were considered & discussed. One such provision was the “categorical efficacy” guideline. Following discussion, the categorical efficacy criterion was judged to be relevant & important.

Importantly, however, the 2007 revision made this goal much harder to achieve.  In 1996, the standard only required that there be a difference in the proportion of subjects who achieved at least 5% weight loss, drug arm versus placebo arm, and that the difference is statistically significant. The new categorical criterion set much higher hurdles: at least 35% of the drug arm must achieve the goal; in addition, the percent of the drug arm achieving 5% must be “appoximately double” the percent of the placebo arm; and the difference must be statistically significant.

 Lorcaserin met this goal solidly -- 47% of the lorcaserin arm achieved the 5% weight loss goal.  Only 22% of the placebo arm met this goal.

Yet the FDA, primarily in the person of Dr. Eric Colman, has repeatedly characterized this achievement as minimal – Dr. Colman has all but said that lorcaserin should not be approved even with no safety concerns whatsoever. This is in direct contradiction to statements he made at the 2004 Advisory Committee, where he explicitly stated that the categorical standard would be considered evidence of efficacy – even if the sponsor did not pre-specify the categorical standard as the trial endpoint. Additionally, he made this statement before the categorical hurdle was raised.

Arena explicitly specified the categorical criterion as the endpoint of both of its clinical trials. Lorcaserin met the higher revised standard easily. And yet Dr. Colman persisted in characterizing its efficacy as “marginal.” It is hard to exaggerate the blatant bias displayed by his attitude.
 

It is also important to understand the context of lorcaserin in relation to all other weight loss agents that have been approved or under review in the past 20 years. Whether you look at drugs specifically designed for weight loss, like sibutramine, rimonabant, or orlistat; or antidepressants with anorectic properties like bupropion or fluoxetine; or anti-epileptics with the same properties, like toparimate or zonisamide – the result is the same in all cases.  All demonstrated similarly modest but clinically significant weight loss. Only amphetamines and amphetamine-like agents, such as phentermine, achieved better than 5% placebo-adjusted efficacy. (The weight loss indication has nevertheless been withdrawn from most such amphetamine agents, due to safety and abuse concerns, and the ones that retain it are restricted to short-term usage.)

In a review of all these drugs at the 2004 advisory meeting by Dr. Frank Greenway, he concluded that these drugs demonstrated:

 "between a 2 and 6.5 kg greater weight loss than 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."

Lorcaserin’s efficacy is right in line with that of all other such agents in the past 20 years.  It clearly meets the categorical efficacy standard. So why, all of a sudden, is this not sufficient?

Once again -- Why is lorcaserin being held to a higher standard?

Conclusions

The answer to that question can only be Qnexa -- misguided though that opinion has been shown to be. For the average obese person looking desperately for help, or for any average individual looking for any easy way to lose 10-20 pounds, it is not hard to understand the irresistible appeal of a drug that holds out the possibility of 37 pounds weight loss. What is very hard to understand, however, is why a supposedly independent, impartial agency like the FDA is being blinded by these numbers to the exclusion of safety concerns & common sense. The safety issues associated with phentermine & toparimate are exacerbated by their combination in Qnexa – this one point should determine the compound’s unapprovable status.

For any common sense approach,  comparing Qnexa and lorcaserin is like comparing apples to oranges. Qnexa is a combination product. Lorcaserin is a (novel) single agent. It is so completely inappropriate to compare a single agent to a double agent -- and hold the single agent “marginally” effective because its efficacy is less – that it staggers the mind that the FDA insists on doing it.

By the FDA's own guidelines, & by historical standards, lorcaserin is an eminently approvable drug, based on its efficacy.The safety issues for lorcaserin -- "drummed up" by a hostile FDA -- can be and will be resolved. At that point, lorcaserin should be approved without further demurral.

Qnexa, on the other hand, should be held to the same standards by which lorcaserin is approvable. It should be held to the additional, common-sense standard which is the primary efficacy standard for weight loss agents: Is the mean weight loss, individual component-subtracted, greater than 5%?   It is not -- it is only 3%. In addition, the safety profile for the combination of its components is clearly worse than either component individually.  This is a drug that completely deserves rejection by the FDA.