The news of failed Phase III efficacy studies on LibiGel (testosterone gel) being used as a treatment for female sexual dysfunction (FSD) was a surprise to many BioSante Pharmaceuticals (BPAX) investors. Failure announcement came after market closed on Dec 14, 2011 Wednesday with chatter already mentioning failure during the trading day. The news was so devastating that BPAX lost 16% on Wednesday followed by another 80% as soon as the news hit the wires in after hours. Antares Pharma (ATRS) wasn't spared in the devastation either due to the fact that ATRS was perceived to be the lesser-known but potentially-bigger beneficiary of a successful LibiGel launch. In fact, Antares not only stood to receive royalty and milestone payments from BioSante's North American launch of LibiGel, but also they retained the licensing rights of LibiGel for the rest of the world. According to the agreements Antares also owns worldwide manufacturing rights; so anywhere LibiGel is sold Antares would make money. Although BioSante had spent a fortune on conducting safety and efficacy studies Antares was entitled to use all that data at no charge, which can be used in seeking approvals for the rest of the world. In summary, BioSante stood to profit handsomely from an FDA approval of LibiGel while Antares was potentially a bigger beneficiary of LibiGel's success. In this article, I would like to present the perspective of a professional statistician on LibiGel's failed Phase III studies and discuss how it can be revived by re-designing the statistical qualities of the efficacy experiments through removing the placebo effect. Revival of the LibiGel's blockbuster potential is a clear catalyst for both BioSante and Antares Pharma.
Before delving into any further details it would be helpful to discuss the background on LibiGel and why it has such a blockbuster potential for BioSante and Antares. LibiGel is a testosterone gel for the treatment of women who suffer from female sexual dysfunction , specifically hypoactive sexual desire disorder (HSDD). The delivery mechanism of LibiGel is developed by Antares and known as ATD™ Gel Technology, which is a hydroalcoholic transdermal gel containing less than 50 percent ethanol and a patented mixture of penetration enhancers. The system is designed to simultaneously deliver the drug across the skin in a controlled and sustained fashion over a 24-hour period, and result in therapeutic plasma levels. The gel is crystal clear and cosmetically appealing, and dries readily after application to the abdomen, thighs, arms, or shoulders without leaving residues.
According to BioSante, HSDD affects millions of women in the US, especially those past menopause. In fact, studies have found that HSDD is more common than erectile dysfunction, which is a $2 Billion prescription market in the US. Since there is currently no FDA-approved product for HSDD, women usually are given off-label testosterone prescriptions. There are over 4 million off-label testosterone prescriptions given for the treatment of HSDD. BioSante's physician surveys found out that over 90% of such off-label prescriptions would be switched to LibiGel once FDA approval is received. BioSante predicted that the potential market for LibiGel is well over $2 Billion in the US alone. According to licensing agreements between Antares and BioSante, LibiGel sales would generate 4.5% royalty to Antares in addition to milestone payments. More importantly, Antares holds the rights for the rest of the world. Furthermore, if LibiGel were to receive FDA approval Antares can use that data free of charge to pursue rest of the world markets. It is also commonly believed that the market potential is even bigger since many women who suffer from HSDD choose not to seek medical attention, partly due to the fact that there is no FDA approved treatment. Needless to say, LibiGel is a clear blockbuster for both BioSante and Antares. Now, let me give you a statistician's look on LibiGel Phase II and Phase III clinical studies and discuss how to revive this blockbuster potential back to life.
BioSante communicated the results of Phase II in a solidly positive way: "In a Phase II trial, LibiGel significantly increased the number of satisfying sexual events in surgically menopausal women suffering from HSDD by 238 percent versus baseline (p<0.0001); this increase also was significant versus placebo (p<0.05)." All sounds well for laypeople even for professionals including statisticians. Although laypeople would be impressed with 238% increase, a typical statistician would be impressed with the small p-values of the clinical trial. After all, it is a statistical fact that smaller the p-value more statistically significant the result is. Certainly, the p-values communicated by BioSante pass the significance test from a scientific point of view, which usually sets the cutoff for statistical significance at p-value=0.05. Consequently, if Phase II results are statistically significant we should expect Phase III to generate even more significant results in favor of LibiGel as the test design is improved and optimum dosage is used in Phase III.
As a professional statistician I was well aware of the weak evidence from Phase II studies of LibiGel. Note that weak evidence didn't immediately mean Phase III would be a failure; rather it meant that the statistical evidence behind rosy Phase II results are not credible. Such results may repeat in Phase III or completely fail to show any difference between drug and placebo, which is exactly what happened in Phase III clinical studies of LibiGel. The weakness in BioSante's Phase II results lie in the small sample size. Some of the documents presented by BioSante included that tiny bit of information that Phase II results were based on a sample of size 46. Furthermore, this small sample size is used to test 150 mcg, 300 mcg, and 880 mcg dosage of the treatment as well as the placebo. Although Phase II achieved statistical significance it was a clear case of extremely small sample size where one should have less confidence that such rosy results can be duplicated in the future. Many other statisticians would argue that in small sample sizes significant p-values can only be achieved when the drug response is vastly different than the control response, which is true, because that's an academic and textbook oriented viewpoint. However, professional statisticians who work with real life experiments know that significant p-values from small sample sizes are unreliable since the statistical assumptions underlying the calculations can't be satisfied to a reasonable degree. In my professional experience, I have designed many statistical tests and also analyzed numerous tests designed by others. I have had many similar cases where a statistically significant observation from one test failed to reach statistical significance in a follow-up test when the first test suffered from small sample sizes.
Same set of facts can be communicated very differently; i.e., either with a positive spin or a negative one. Indeed, some investors felt that BioSante consistently misled the investors about the viability and effectiveness of LibiGel, and they filed a class action lawsuit following the failure of Phase III. Lead plaintiff Thomas Lauria claims: "Defendants...consistently misled investors about the commercial viability, effectiveness, and market potential for LibiGel. Defendants boasted about LibiGel's efficacy over placebo, and provided supposedly concrete 'data' regarding the drug's 'statistically significant' effect on increasing the 'number of satisfying sexual events' for women suffering from HSDD." Without commenting on the merits of this lawsuit I would like to point out one more time that the weak evidence from Phase II based on the small sample size doesn't imply a failure in Phase III; rather it gives less credibility that such stellar results are duplicable in a later trial. Let me explain this in more detail as I transition to reviving the blockbuster potential of LibiGel in a re-designed Phase III trial.
Although Phase II showed a 238% increase in sexually satisfying events, my conviction that Phase III would show something similar was weak based on the small sample size; however, I do expect to see something more moderate around 70-80% increase in Phase III. Since Phase III studies have much larger sample sizes such a difference in response should still reach statistical significance. Here is my reasoning. As some of you would remember, Procter & Gamble (PG) developed a testosterone patch, called Intrinsa, for the treatment of female sexual dysfunction back in 2004. Intrinsa patch had the exact same amount of testosterone as LibiGel. In P&G's studies over six months of surgically menopausal women, those who received a placebo said satisfying sexual activity increased by an average of 19%, vs. a 73% increase for Intrinsa patch users. In the placebo group, women had an average increase in satisfying sexual episodes of 0.5 per month, while those with patches had an average increase of 1.9 episodes per month. The patch was granted a license from the European Medicines Agency, and was available on Britain's National Health Service from March 2007. However, in December 2004 the US FDA 14-member advisory committee, plus voting consultants, for Reproductive Health Drugs unanimously rejected P&G's fast-track request for Intrinsa citing concerns about off-label use. FDA rejected Intrinsa for lack of safety data. FDA panel members strongly criticized the company funded clinical trials as "inadequate" to assess issues of long-term safety. Panel members also raised serious concerns about the risks of cardiovascular disease and breast cancer associated with a drug that appeared to offer women, on average, little more than one extra episode of sexual activity a month, compared with placebo. P&G didn't pursue the US approval of Intrinsa following this rejection; because of the time and costs involved in a large scale safety study. Since Intrinsa and LibiGel have the exact same active ingredient (300 mcg of testosterone) I expected that LibiGel Phase III trials to increase sexually satisfying events around 70-80%. BioSante's Phase III trials showed a 83% increase but the result was a statistical failure because 83% increase in treatment arm was no statistically different than the increase seen in the control arm. In layman's terms, women's sexual satisfaction increased similarly between LibiGel and placebo; making the difference between the two arms as statistically insignificant. Intrinsa showing statistical significance while LibiGel fails in Phase III trials clearly tells me that LibiGel's failure was a design issue rather than the treatment being ineffective. Motivated by this observation, here are the top 5 reasons why BioSante should fix the experimental design issue and re-conduct the Phase III trials in their pursuit of this blockbuster medication.
Top 5 reasons why BioSante should pursue LibiGel
1. BioSante already has the safety data, which Intrinsa was missing:
When BioSante announced that efficacy studies didn't meet the end-points they stated their intention to go forward with the safety part of the study, which completed enrollment of 3,656 women. The safety study is designed for a total of five years; however, BioSante could use the safety study data as part of a New Drug Application (NDA) submission after the last subject enrolled has completed 12 months of exposure to LibiGel or placebo. Confirming their intentions, BioSante continued the safety study until September 2012. BioSante ended the safety trial because a review of the data did not reveal any general or specific safety concerns. The company said patients who were treated with LibiGel were less likely to suffer cardiovascular problems like heart attacks, and they were diagnosed with breast cancer at a rate that was similar to the general population based on their ages. BioSante said the FDA asked it to report data from a study where patients used LibiGel for 12 months on average, and it said the subjects in the trial used the product for 24.5 months on average. The company ended the trial after a safety review by an independent monitoring committee.
With regards to Intrinsa, in a vote of 14 to three, the FDA committee agreed that Intrinsa's effect was clinically meaningful; so 300 mcg of testosterone received FDA's approval in terms of efficacy in treating HSDD. FDA was concerned in Intrinsa's safety profile in their review letter: "Areas of special concern regarding the use of testosterone in surgically menopausal women include the unknown risks of cardiovascular disease and breast malignancy potentially arising from chronic use of this hormone." Considering that LibiGel and Intrinsa are the same dose of the same hormone with the same pharmacokinetics; plus, no difference in blood concentration levels between the two methods of administration, BioSante has every reason to believe that properly designed Phase III trials should provide the missing piece of the puzzle. In summary BioSante has the safety part covered; now, they need to re-design the Phase III efficacy trials with special emphasis on removing the placebo effect.
2. Low-level testosterone is already approved in some parts of the world:
There is an approved 1% testosterone cream for women available in Australia. Andro-Feme® 1% testosterone cream has been available in Australia since 1998 and is currently the only available approved testosterone product for delivering testosterone to women in Australia. It is applied daily to the skin, and sometimes reduced or increased, depending on side effects and blood levels. As mentioned earlier, other approved testosterone product for HSDD is Intrinsa patch, which contains 300 mcg of testosterone and originally developed by P&G. Intrinsa was approved by European Medicines Agency in July 2006 for the treatment of HSDD in women who have had their uterus (womb) and both ovaries removed, and Warner acquired the product through its £3.1 billion purchase of P&G's pharma unit in 2009. However, Warner notified the European Commission in a letter dated Mar 2012 to voluntarily withdraw the marketing authorization citing commercial reasons. Consequently, on 25 May 2012 European Commission issued a letter withdrawing the marketing authorization for Intrinsa. The details about commercial prospects of Intrinsa weren't made available. One needs to note that Intrinsa was a patch and LibiGel is a gel formulation. Since the delivery mechanisms are different the side effects profiles would be different too. Furthermore, in the withdrawal request Warner didn't mention any issues with safety or efficacy; but rather commercial reasons. In Canada a testosterone enanthate for injection (Delatestryl) is government approved for the treatment of frigidity in women.
3. FDA knows about the off-label use of testosterone to treat HSDD in women:
According to IMS and primary research conducted by BioSante, there are over four million "off label" testosterone prescriptions written for the treatment of HSDD in 2009. According to American Society for Reproductive Medicine (ASRM) clinicians in the US are limited to off-label testosterone preparations since no testosterone product is FDA approved for use in women, while clinicians in Australia and some European Union countries can prescribe a testosterone cream for women. ASRM states that "the need for an approved testosterone product formulated for women is clear. Women in America are seeking out testosterone therapy, and gynecologists are supporting them by prescribing either compounded testosterone creams and troches or testosterone products that deliver doses appropriate for male testosterone replacement. Off-label use exposes women to the risks of supraphysiologic testosterone levels."
4. Truly blockbuster potential
Finally in June 2011 an FDA advisory committee to the division of Reproductive and Urologic Drug Products stated that HSDD is a significant medical condition for women. Note that this statement is 7 years after FDA rejected Intrinsa based on safety concerns. Considering 4 million off-label prescriptions in the US annually for the treatment of HSDD, the potential is truly blockbuster. In fact, 4 million prescriptions seem to be the tip of the iceberg in terms of the demand from women suffering from HSDD; because, many women prefer not to even talk to their gynecologist knowing that there is no FDA-approved product. Even in cases where they seek treatment, they have to convince the physician that asking for an off-label prescription is well-worth the risk which applies not only for the patient but also for the physician. According to a P&G's survey on female health, 30 million women in the US are naturally menopausal, 3 million are distressed by their lack of sexual desire, and 20% of 25 million women who are surgically menopausal are distressed. According to Leerink Swan the addressable postmenopausal patient population in the US is 6.5 million, which will only increase due to aging demographics. Note that LibiGel is patented until mid 2022, which leaves more than enough time to recover the costs of development.
5. BioSante's Phase III efficacy trials failed due to strong placebo effect
When a Phase III efficacy study fails to reach statistical significance there can be two reasons; either treatment arm didn't improve the condition, or the control arm (placebo) worked as great as the treatment arm; where in BioSante's case it is the latter. In the press release announcing the failure BioSante stated "there was an increase in the total number of satisfying sexual events of 3.87 from baseline (an increase of 83 percent) in the LibiGel group and in the placebo group there was an increase of 3.52 satisfying sexual events from baseline (an increase of 65 percent) for a p value of 0.698." Note that 83% increase is very much inline with the Intrinsa efficacy results of 74% which reached statistical significance due to weaker placebo effect in the Intrinsa trial. This shows that LibiGel delivered through Antares's ATD technology indeed did the job from a pharmokinetic point of view; but, patients in the placebo arm felt that their sexual lives were much improved, which is not quite surprising because there is considerable psychological part when it comes to sexual satisfaction. BioSante needs to re-design the efficacy trials by focusing on removing the placebo effect from the readings. Statisticians who design experiments to show efficacy of a medication where the improvement has to be measured from a patient response involving very subjective assessments (e.g., female sexual satisfaction) need to be very careful about how they design the trials such that potentially high rate of placebo response in these trials makes it difficult to demonstrate efficacy and end up to trial failure. For example, in assessing the efficacy of anti-depressant drugs researchers found out that a major reason for high placebo response rates is the inability of standardized rating scales to define a coherent group of 'drug responders'. They have developed alternative means of classifying patients in order to circumvent this problem and thereby improve the chances of trial success. Obviously, instead of using patient diaries BioSante needs to develop a more objective way of counting sexually satisfying events. Another method to remove the placebo effect is to run a small scale study where placebo responder profile is established for HSDD through statistical modeling; and then use that profiling model to screen out possible placebo responders from the real efficacy study. Finally, a third but more costly option is to run the screening on all patients in the study and remove the placebo responders in stage 1; and then continue on stage 2 with the patient population cleaned from the placebo responders. In this case stage 2 population can be randomized into treatment and control groups with the hope that placebo responders are already eliminated at stage 1; thus, yielding minimal placebo effect.
In this article, I presented a professional statistician's perspective on why BioSante's best course of action would be to re-conduct the Phase III efficacy trials with a special focus on removing the placebo effect from the readings as LibiGel has a clear blockbuster potential for both BioSante, which holds North American licensing rights; and for Antares, which holds manufacturing rights worldwide and licensing rights for the rest of the world on top of North American royalties and milestone payments. I am happy to see that BioSante already announced plans to initiate two new LibiGel Phase III efficacy trials based on an extensive analysis of previous efficacy data, consultation with key opinion leaders in female sexual dysfunction, testosterone therapy and placebo effects, as well as a meeting with the US FDA. I expect BioSante to agree on a Special Protocol Assessment (SPA) with the FDA and start the new trials in the near future.
Interestingly, the new trials can be conducted much effectively with help from a larger pharmaceutical partner with sufficient financial, technical, and regulatory expertise. Such a partnership would be very plausible for the bigger partner because BioSante already gathered the very costly and large safety data. Within this regard, Pfizer (PFE) can serve as that bigger partner, which I discussed previously as Pfizer being a potential suitor for Antares Pharma. A medication treating female sexual dysfunction would complement Pfizer's blockbuster drug Viagra which is used to treat male sexual dysfunction. Indeed, Pfizer is interested in providing a medication for female sexual dysfunction as evidenced by their eight years of work and tests involving 3,000 women to see if sildenafil citrate improves sexual dysfunction in women. Furthermore, in 2005, Pfizer funded a survey which showed 63 per cent of women had sexual dysfunction and that testosterone and Viagra might be helpful. Attempts from Pfizer to treat female sexual dysfunction are not restricted to these two projects. In 2010 Pfizer looked beyond sildenafil in treating female sexual dysfunction, the active ingredient in Viagra, to a new potion, a selective neutral endopeptidase inhibitor. Results published in the British Journal of Pharmacology show that the chemical increased vaginal and clitoral blood flow. Either through a partnership with Pfizer or otherwise, the future of LibiGel looks promising; and any progress towards that blockbuster potential are clear catalysts for both BioSante and Antares Pharma.