Pharma Mergers and Acquisitions: No-win or a No-Lose Propositions

This seems to be the season for mergers and acquisitions in Pharma.  What is the motivation for these transactions.  One possibility is that if structured correctly, a merger or acquisition can be a no-lose proposition for accounting purposes and a no-win proposition for stockholders.

How can this happen.  Consider Company A which is slowly losing patent protections for its products and cannot get replacement products approved quickly enough to replace them.  Given the massive profitability of many pharma products and the enormous landscape of potentially successful development projects, Company A must be relatively poorly managed.  However, Company A can find Company B that has income producing products and acquire it.  What is the cost to Company A?  The cost is minimal because the money Company A spends to acquire Company B is counted as an investment.  Company B’s assets are simply a replacement for the money Company A spent.  Over time, Company A will have to depreciate these assets but this depreciation can often be spread over many years.  So, the cost of acquisition is low from an accounting point of view.  Just a lot of lawyer payments.

If the cost is low, what’s the benefit.  Well, that’s easy, the net income of Company B is added to the income of Company A.  So, profits go up automatically.  In fact, profits can go up even more if Company A fires as many people in Company B as possible.  Of course the firing should concentrate on people who don’t add immediately to the current income of Company B.  This maximizes the amount of Company B’s income that is available to add to Company A’s bottom line.  Who will be the first targets for firing?  The R&D group is prime.  R&D doesn’t contribute to current profits.  R&D generates future profits and a good R&D group can be extremely valuable.  But, in the short term, R&D is a cost.  So, getting rid of R&D makes short-term sense.  It makes many mergers and acquisitions a no-lose proposition, at least on paper.

But what about reality?  Firing all the people responsible for generating future profits means there will be no future profits.  All the value locked in the R&D organization is lost.  So, the value of Company B is greatly reduced.  However, this loss is probably not recognized for accounting purposes.  The value of Company B is still carried at close to its cost.

Thus, an acquisition can, for accounting purposes, be a no-lose transaction.  However, over time, the profits from Company B must decline because their future revenues are limited by the loss of R&D.   So, in due course, the slow depreciation of Company B that must be recognized year after year, will overtake the rapidly declining revenues.  This is when reality starts to set in.  Now Company A may have to start recognizing losses from the acquisition.

What to do?  Another acquisition will fill the bill.  Of course, it must be larger because the income from newly acquired Company C will need to cover the losses from depreciation of Company B in addition to the depreciation of Company C.  So, the acquisitions must get bigger and bigger.

That’s just what we are seeing.  Larger and larger acquisitions but sluggish growth or declining profits, except for a few years following the acquisition.

I’m not sure how something like this will end.  But, he size of acquisitions cannot grow to infinity.

Changes at Merck

As I have written in this blog for the last couple of years and talked about for many years, the path to success in the pharmaceutical industry is clear and relatively straightforward.   The problem is that many managers find it difficult to tread this path because, it requires a different management style than many other industries where they may have had their prior training and experience.  Also, the massive profits from a single drug approval can delay the need to think about the long-term future and foster short-term and shortsighted management.

These challenges, which have yet to be fully overcome by many pharma companies, most notably the bigger pharma companies, have led to the current state of crisis in the industry.  For many companies, sales and profits are falling as patents expire on old, very profitable products, and companies, bound by shortsighted, unrealistic management techniques, are unable to replace them with new innovative products.

When crisis strikes, companies typically respond in one of two ways.  The most common is to try to minimize blame on management by asserting that existing management approaches were sound, but just not imposed strongly enough.  Thus, the rigid, bureaucracy-based management system is intensified, intensifying the problems.  A second approach is the more obvious one.  Admit that the current system is not working, fire the people responsible and hire new people who have a history of success in bringing new innovative products to market (yes this does still happen and the people who do it are well known).

Merck is now in the midst of a long crisis phase.  Last year Merck announced a workforce reduction of 20%.  This likely represents a massive loss of experienced personnel and a big blow to morale among the remaining people.  The question is, can Merck quickly rebuild a more productive organization.  The new head of R&D, Roger M Perlmutter, is a former Merck employee suggesting that he may have trouble changing the corporate culture that led to the problems now plaguing Merck.  So, it may be that Merck is still unable to create a culture of innovation and success and is merely intensifying the old system that led to the problem in the first place.

The final sentence of the New York Times article in the link above is not encouraging: “The company said it would also increase its focus on diabetes, acute hospital care and vaccines.”  Unfortunately, great opportunities in the pharma business do not generally present themselves in pre-specified therapeutic areas and it is, in my opinion foolish to restrict ones’ horizons.  This can lead to missing great opportunities elsewhere.

Another negative sign is Merck’s decision to create an innovation hub in the San Francisco Bay area in order to “actively seeking external scientific innovation”.  This sounds like Merck is to some extent giving up on innovation within the Company.  That it may be partially outsourcing its most critical function.  Outsourcing has been a consistent trend over my career and it has paralleled the decline of the industry.  Of course, association does not prove causation.  But, common sense tells us that a company should not outsource its most critical function to people and organizations that do not share company goals and incentives.

We shall see as time goes on.  A good manager should be able to start submitting successful NDAs within three years of taking over.  So, tune in to Merck in 2016.  If they don’t have at least one or two NDA’s successfully submitted, we will have our answer.

Related articles

• Merck plots drug R&D hub in Bay Area (bizjournals.com)

Indian Patent Decision – A Spur to Innovation

A court in India recently made news by rejecting the patent on a polymorph of Gleevec, an anti-cancer drug.  A New York Times article cited the decision as a potential impediment to innovation in the pharmaceutical industry because it could reduce short-term profits for Novartis, the company that sells Gleevec and “the brand name pharmaceutical companies contend the profits they reap are essential to their ability to develop and manufacture innovative medicines”.

In my experience, this kind of interpretation of anything that effects the short-term profitability of pharma companies is very common.  But, does it stand up to rigorous scrutiny?  Allowing companies to extend the effective patent lives of their existing products could actually reduce their incentive to innovate.  This is because investing in short-term strategies to extend the lives of existing products could be a more prudent investment than the high-risks associated with innovation of important advances in medical therapy.

The Novartis patent in question is for a novel “polymorph”.  The concept of a polymorph is fairly easy to understand.  When a drug is made in solid form its molecules can arrange themselves in different patterns depending on exactly how the solid form of the drug is made.  Sometimes the molecules line up in very orderly arrays, sometimes they are all jumbled together in a disordered pile.  As part of assuring the quality of drug products, regulators require that the polymorph in a particular product always be the same.  This is required because polymorphs can differ in the rate and extent of absorption into the body when a drug is administered.  So, it is important that the polymorph be controlled to avoid potential differences in the rate and extent of absorption when the drug is administered.

But, what if two polymorphs have exactly the same rate and extent of absorption?  Should a company be allowed to patent a new polymorph and keep all other products off the market just because the molecules are arranged in a different pattern but the drug still has the same biologic effects?  This is equivalent to putting the drug in a differently colored pill.  It looks different but behaves the same when given to patients.  Does this kind of change justify 20 years of monopoly?  The common sense answer is “no”.  Patents should only cover real innovations.  Granting long monopolies for trivial accomplishments will tend to remove the pressure to make real innovations.

Thus, the Indian court decision is likely to spur innovation.  One cannot criticize Novartis for pursuing their patent case.  It was a good business decision in the existing business environment.  But, the court has changed the environment.  Perhaps going forward, money that might be used to pursue trivial patents and their resulting legal cases, will be channeled toward discovering innovative new therapies and the Indian decision will result in a long-term increase in investments in real innovations resulting in long-term benefits to pharma companies and patients.

Some Good News

The trend of falling R&D productivity in the pharmaceutical industry seems to be reversing.  As shown in the following Figure, the number of new chemical entity approvals in 2011 and 2012 have increased above those for the last 7 years.

The number of new molecular entity approvals by year. “New Drugs” was published in the Summer of 2009 

I published “New Drugs” in the Summer of 2009 to help spread understanding of the drug development methods I was taught and that I found lead to a high rate of success.  This was one of many efforts to improve the efficiency of the drug development process in recent years.  As detailed in other blog entries, both the NIH and some non-profit organizations have recently started programs that seek to improve the efficiency of drug development

We will know in a year or two if this increased number of approvals is a real reversal of the prior low number of approvals or results in an increased number of product withdrawals, suggesting the increase was caused by excessive FDA leniency (in my opinion the latter is unlikely).

My expectation is that this increase in drug approvals is due to a real increase in R&D productivity and that this productivity will accelerate over the years to come.  I hope that my book contributed in some small way to what looks like the beginning of an exciting new trend.  If, as I expect, this trend is real, it represents a real opportunity for investors, pharmaceutical companies and, most importantly, patients.

Alzheimer’s Tragedy (Continued)

 

As expected, a second shoe has fallen.  Eli Lilly has reported disappointing results of two studies of solanezumab, an experimental treatment for Alzheimer’s disease.  The studies were named “Expedition1” and “Expedition2”.

I’ve expected this kind of result for many years and it has slowly become a mainstream belief in the investment community and to some extent in the academic community, especially since the recent failure of the similar product from Pfizer/Elan.

In spite of the widespread skepticism about the Lilly study it still seems to me that the news reports of the results of the solanezumab study do not emphasize how definitive the failure was.  The following quote taken from the Lilly press release is instructive.

“The EXPEDITION1 study did not meet co-primary cognitive and functional endpoints in the overall mild-to-moderate patient population; however, pre-specified secondary subgroup analyses in patients with mild Alzheimer’s disease showed a statistically significant reduction in cognitive decline. Based on those results, Lilly modified the statistical analysis plan (SAP) for EXPEDITION2 prior to database lock to specify a single primary endpoint of cognition in the mild patient population. This revised primary endpoint did not achieve statistical significance.”

What can we infer from this quote?  First, we learn that the primary analysis of the Expedition1 study failed.  That is, when the data from all patients were analyzed, they showed no statistically significant benefit from solanezumab treatment .  This means that from the regulatory point of view, the study failed.  However, a pre-specified sub analysis is reported to have shown statistically significant benefit.  Based on this observation the statistical plan for the second study was changed prior to locking the database.  However, the Press Release reports that in spite of this change, the second study also failed to show a statistically significant benefit.  How is this possible?  We can’t the reason in this particular case. However, this kind of outcome often happens because there are many pre-specified secondary analyses.  Thus, there is a possibility that one of them will show statistical significance purely by chance.  This is especially true if the analysis of the secondary endpoints is not corrected for the multiplicity of statistical comparisons and if the level of statistical significance is low.  Neither issue is addressed in the Press Release.  The fact that the second study failed in spite of the change in statistical plan suggests that the success in the first study was indeed due to chance rather than reflecting a real effect of solanezumab on Alzheimer’s disease.  These issues will become clearer if the full results of the two studies are published in a peer-reveiwed scientific journal.

There is one other speculation one can make.  Lilly decided to dramatically change the statistical analysis of the second study based on the results of the first.  This suggests that the results of the first study were so poor that Lilly felt there was no hope the second study would show efficacy unless the statistical analysis was changed dramatically.  This suggests that the first study didn’t just fail to show a statistically significant benefit of solanezumab but rather that is showed essentially no benefit (or that it showed the drug made patients worse).

Until these issues are clarified, I believe one must seriously question the justification for further investment in solanezumab as a treatment for Alzheimer’s disease.  As mentioned in a prior post, it is possible to show statistically significant benefit of Alzheimer’s treatments in much smaller studies, if the efficacy is there.  My team has done so in a number of studies and so have other teams.

Thus, the tragedy continues.  Money is poured into Alzheimer’s drugs that have a very low probability of success, while products with much better prospects are essentially ignored.

Like all bubbles, this one will pop sooner or later.  Let’s hope that when it does, companies don’t give up developing treatments for Alzheimer’s entirely.  There are a number of outstanding opportunities for developing Alzheimer’s therapeutics for much less money, in much less time and with much less risk.  Click to see one.

 

Alzheimer’s Tragedy

Alzheimer’s disease is a tragic condition.  It slowly robs a person of their most precious possession, consciousness.  However, these days, that’s not the only tragic aspect of Alzheimer’s disease.  There is an ongoing tragic misallocation of research resources.

As long as 10 years ago I was warning investors that anti-amyloid antibodies were a high-risk approach to the treatment of Alzheimer’s disease.  This played out in spades when the active immunization product produced serious and sometimes lethal side effects.  The high risks were confirmed by the early results with bapineuzumab that showed no statistically significant benefit on primary endpoints.  I was told that an investor who had consulted me on the likely outcome of that study did quite well on a short sale.

Further evidence against the amyloid antibody approach came from a different kind of therapy; the Eli Lilly gamma secretase inhibitor.  Gamma secretase inhibitors block the production of beta amyloid rather than binding to and theoretically removing beta amyloid like bapineuzumab.  Nonetheless, the gamma secretase inhibitor not only didn’t improve cognition in AD patients, it made their cognition worse.

In light of these negative findings, one would think that enthusiasm for anti-amyoid therapies would be tempered.  However, that’s not what happened.  In fact, a number of big pharma companies have invested what will probably end up being billions of more dollars on antibodies against beta amyloid.

The first of what should be a number of study results were recently described here.  Bapineuzumab failed to show any evidence of efficacy in patients with a genetic abnormality associated with a higher than average risk of Alzheimer’s disease.  This was supposed to be the subgroup of AD patients who would respond to bapineuzumab.  This is another nail in the coffin for this therapeutic approach.  However, the implications of these study results go beyond that.

One has to question the judgement of those who have committed huge sums of drug development money to such a high risk approach to the treatment of AD when much lower-risk approaches are evident.

Results of other similar antibody studies will likely come out over the next few months.  Assuming, as everyone now does, that these studies also fail, it will be important to yield to what will be overwhelming data and switch drug development efforts to approaches with stronger scientific support, of which there are a number.

Some will propose doing even longer, more expensive and time-consuming clinical studies.  But remember, therapies that are highly effective (and that’s what we want, right?), don’t require very large, long and expensive trials to show efficacy.  The benefit of highly effective therapies can be shown in small short studies.  Large clinical trials are required only to rule out rare side effects.

We were able to show donepezil worked in a study with about 50 patients per group studied for a few months.

The failure of bapineuzumab and the expected failure of the remaining similar drugs may actually be good news for AD patients.  Hopefully, it will result in the redirection of development money to products with real chances for success and will force pharmaceutical managements to make future allocations of research dollars in a more considered and responsible way.  That would also be good news for pharmaceutical investors.

I hope companies don’t give up on AD therapeutics as a results of negative findings with the anti-amyloid antibodies.  These were very high risk from the beginning.  There are a number of much lower risk approaches available that could mean the difference between life and death for patients and could offer rich financial rewards for companies with the foresight to take pursue them.

Bye Bye Pfi

In my book, published about 2 years ago, I outlined why generally accepted accounting practices (GAAP), at least as commonly implemented,  often don’t capture the reality of pharma company profits and losses in a common sense way.  For example, if a company lays off its entire drug development team, it is counted as a cost savings (net of severance payments of course) since expenses are cut.  However, an efficient and effective development team is often one of the most valuable company assets, and its dispersal, could reasonably be considered a large loss of future revenues and profits.  To my knowledge, this kind of loss is never recognized in GAAP financial statements until revenues actually start to fall which may take significant time.  Similarly, an NDA approval converts a potential product into a real one and often represents a significant increase in enterprise value.  This may be recognized by the stock market through an increase of share price.  But, again, to my knowledge, it is not recognized in GAAP accounting unless the rights to the product are sold and the value received as cash or other payment.

Naturally, mergers and acquisitions provide ample opportunities for distortions due to these quirks in GAAP accounting especially since they often involve massive R&D layoffs.

It has been my feeling that for at least the last 10 or 15 years, these distortions have been at the heart of accounting at many pharmaceutical companies and for this reason, these companies would ultimately disappear as ongoing businesses because GAAP accounting didn’t accurately reflect business realities.  Indeed many pharma companies, e.g. Schering Plough, Rorer, Squibb, Warner Lambert, etc., have disappeared.  Just as I had expected.

Now there is speculation that it may happen to the biggest of them all.  Pfizer has  has engaged in extensive merger activity, associated with massive layoffs for many years now.  And, according to a recent article on Bloomberg, it looks like this strategy may be reaching the end of the GAAP accounting road.  The unthinkable may happen.  Pfizer may split up and the giant may be no more.

I have no way of knowing if the speculation in the Bloomberg article will prove to be accurate.  However, the fact that such speculation exists at all, suggests that the concerns summarized in my book are well founded.

 

Long Drug Approval Times for Central Nervous System Drugs

I saw an interesting news article reporting that the average approval time for Central Nervous System (CNS) drugs is significantly longer than for other classes of drugs.  The article is apparently based on a subscription report by the Tufts Center for the Study of Drug Development.  The news article reports that the mean time for drugs to complete the clinical development process was 102.2 months (almost 9 years) and the mean time for approval was 20.3 months (almost 2 years) for a total development plus approval time of 122.5 months or over 10 years.  The average development time for CNS drugs was 40% longer than the average for non-CNS drugs and the average approval time was 13% longer than for non-CNS drugs.  What isn’t reported in the news article was any measure of variability around these average values.  Perhaps, the variability is given in the original research report.

What is interesting to me is that the period studied by the Tufts researchers includes the date when donepezil hydrochloride was approved and most of the time it was in development.  I selected donepezil for development and lead the development effort.  Unlike most of the large companies studied in the Tufts report, I had to assemble the development team as part of building a new subsidiary company to handle the project.  We had to prepare our own standard operating procedures, case report forms, buy computers, etc.  Thus, there was a lot of additional work required to get the company infrastructure in place.  However, as I recall we were able to complete the development in 62 months (compared with the average of 102.2 reported by Tufts) and as I recall donepezil was approved 7 months following the NDA submission (as compared with Tufts average of 20.3 months).  Thus, the entire process of development and approval took less than 5 years as compared with over 10 years for the Tufts average.

Now, one might assume that donepezil was a special case and this was indeed true.  But, rather than being easier than average, donepzil was more difficult for a number of reasons.  For example, at the time the development program was begun no other Alzheimer’s disease treatments had ever been approved.  Thus, the requirements for FDA approval were in flux and the endpoints for proving efficacy were, in one case very difficult to measure consistently.  There were also no successful clinical trials upon which to base reasonable power calculations. So, in my opinion, donepezil could reasonably have taken longer to develop than an average CNS drug.

What explains the difference between the time and effort to achieve the approval of donepezil hydrochloride and the average CNS drug?  The book explains the methods we used.  We assembled an extraordinary group of development experts and pooled our accumulated knowledge to implement the methods described in the book.  Little of these methods were my own invention.  I learned most of them from my colleagues and my mentors at ER Squibb and Co where I had my first pharmaceutical company job.  I suspect my colleagues also learned a great deal at their former companies because drug approvals were far more efficient in those days.

Alas, these methods have been lost in most companies.  The book is my attempt to revive them. Ten-year drug development timelines are not a necessity.  They are, in general, evidence of disorganization and/or sloth.  The FDA is required to produce its review in 10 months.  There is no reason the final approval should be delayed much beyond that deadline.  A longer review time is generally evidence of a lack of ability of the drug developers to communicate effectively with the FDA.  Of course, the drug developers blame FDA and the FDA personnel are prohibited from defending themselves by confidentiality requirements.  So, in general the public hears only half the story.  But, in my experience, the vast majority of FDA concerns are well justified and in accord with their published Guidance documents.  So, a good development group can anticipate them and by addressing them in advance, achieve a rapid approval.

The Beginning of a New Era

Today is a landmark day in the pharmaceutical business.  It is the day FDA approved generic atorvastatin (Lipitor®).  Atorvastatin was the largest selling drug in the world but it will likely lose that distinction as generic sales replace Lipitor sales.  In the short term, the loss of sales of the brand product is often painful to the brand company, especially once the 6-month exclusivity for the first-to-file generic runs out and additional generic products enter the market and increase competition.  But in the longer run, the loss of patent protection for an important product spurs companies to develop innovative new products to replace the lost sales and may also have a similar effect on other companies facing immanent patent expirations.  Thus, in due course, patent expirations can be healthy by encouraging brand pharmaceutical companies to fulfill their core social function, to bring innovative new medicines to the market.  With new drug approvals beginning to tick up, perhaps, we are seeing the beginning of such a turnaround in R&D productivity.  Time will tell.

Finding New Drug Candidates with High Potentials for Success

With all of the advances in biology basic science in the last 20 years it is tempting to conclude that drug discovery and development can be reduced to a relatively methodical process of identifying a new receptor or pathway, analyzing it in detail through the many powerful genetic, biochemical and analytical methods we now possess and designing new therapeutics though computer assisted drug design.  While this approach is practical in many cases, it is not the only approach to finding new drugs; nor, I believe, is it the most effective in many cases.  Because biology is very complex,  it can be very difficult or impossible to accurately predict the impact of a very novel agent on the whole human organism based purely on computer models.

However, this does not mean that discovery of entirely new therapeutics is inherently risky.  We can often provide effective therapies without understanding the pathophysiology of a disorder in detail.  For example, vaccines were in widespread use long before the immune system was understood at any level of detail.  Antibiotics were in widespread use long before their mechanisms of action were understood.

These therapeutics were discovered by chance when thoughtful observers integrated seemingly disparate observations to come up with a novel approach to therapy.  Although this is not the only way, nor even the best way to discover novel therapeutics, it is a way that has been successful many times in the past.  I would suggest it is getting short shrift today.  In addition to determining the crystal structures of newly identified therapeutic targets, we should be screening the scientific literature and folk medical literature for clues to new approaches to therapies.  After all, minoxidil (Rogaine®) was originally approved as an anti-hypertensive drug that was commercially unsuccessful because it caused excessive hair growth.  Eventually, someone had the reasonable idea to develop it as a treatment for hair loss with a potential side effect of hypotension (not a real problem since it is given topically as a treatment for hair loss).  There are other examples as well.

I have a few projects like this that I am currently working on but I am sure that there are many more waiting to be discovered.

Perhaps, it would be possible to develop a computer program that would scan the scientific literature looking for observations that could be the basis for novel therapeutics or new therapeutic indications for existing drugs?