Description

Special Situation: BioDelivery Sciences International - BDSI

BDSI was a drug development company that has turned into a commercial pharma company and was just proving its profitability when COVID hit. BDSI crossed into commercial maturity and profitability while everyone was looking at vaccine companies and while doctor visits for non-COVID purposes dropped precipitously. 2020 was a transformational year for BDSI as they reached record levels in sales that drove the company into significant operating cash flow, but the price is still at 2018 levels due to concern over a patent case on BDSI’s biggest product. Trial has ended and we can now look at final briefs and court activity since the arguments ended. The case appears likely to result in no impact to BDSI’s sales and insiders are buying in advance of the formal court announcement, which will likely trigger a significant upward revaluation.

The Business

BDSI started as an unprofitable drug development company who licensed out their main drug, a long acting opiod (LAO) called Belbuca, and then eventually acquired the license back to benefit from the profits themselves when they realized how good the product was. Belbuca is an in-cheek strip that dissolves against the mucousal membrane inside the mouth, delivering buprenorphine - an opiod that is safer and harder to abuse than the more common ones in use today but which can’t survive pill ingestion very well [1]. BDSI replaced their CEO with someone chosen more for commercial pharma experience and transitioned into a profitable company.

BDSI has been doing very well selling Belbuca. Their earnings calls used to discuss the high price of capital, now they discuss their constantly rising cash balance and stock repurchases. Management likes to highlight gross margin improvements but, like many commercial pharmaceutical companies, salespeople are a necessary part of the delivery ecosystem.  Thus I prefer thinking about operating margin. The field reps have to be scaled up as prescriptions increase, although BDSI has less of this than - for example - surgical device makers who need to send field reps into actual surgeries to help surgeons. Operating margin has been improving with scale, sitting around 25% for the 40mm rev per quarter range (Q3 2020 to Q2 2021 - this also avoids thinking the $5mm ex-CEO payout is recurring).Both gross and operating margin show fairly uncomplicated behaviors exhibiting operating operating leverage and show the company has recently passed into profitable scale: 

Figure: A quick plot of gross and operating margin versus revenue for some recent data points. The fit and behavior show classic achievement of efficient scale. This is a good business, if it continues.

They have some minor other products, but most of their revenue comes from Belbuca. Here is recent revenue of the three main products, the only thing I’ve left out of overall revenues was a large one-off reserve release in 2017 that won’t be happening again and the only important take-away here is that their revenue is largely one product:

Figure: Revenue by product. Note that Bunavail has been discontinued (and they just recently licensed a new product with some of their free cash flow but it hasn’t launched yet). Belbuca is basically their business, without it they lack efficient scale. Belbuca is also the target of a lawsuit attempting to invalidate their patents so others can sell generics - that’s the special situation.

The one product is a good one! Belbuca is an in-cheek tab for delivering buprenorphine through the mucosal membrane. This is required because the medicine is neutralized very well by the liver - this both necessitates patch-style delivery and results in a safer profile for drug dosing. BDSI has slides emphasizing the lower risk of overdose and lower risk of respiratory problems. We reviewed the primary literature in medical journals and spoke with multiple doctors. We would say that BDSI’s promotional literature is a little guilty of cherry-picking - other research showed respiratory depression was possible, but agrees it is less of an issue than classic opioids. The doctors were 50/50 split on whether they had really heard of it before but after being given journals by us (NOT the promo materials) to look at, they indicated they would probably put Belbuca on their pain treatment escalation plan after Tramadol/Codeine but before going to Oxycodone or the like (read more on the Analgesic ladder here [2]). This suggests that, long term, Belbuca has an addressable market measured in billions. The only meaningful competing buprenorphine delivery system is a skin patch (Butrans) formerly made by Perdue and now in generics. Cheek patches are generally considered superior if the patient is able to use them correctly, as they avoid issues of skin contamination and damage. Belbuca is also available in a larger range of doses. We don’t need to figure out here what relative market share Belbuca will eventually achieve vs Butrans, as long as the market exists in significant size this idea is investable.

The special situation

The special situation is a patent case where a company called Alvogen is attempting to invalidate three of BDSI’s patents [3] which prevent competitors from bringing generic Belbuca to market. If Alvogen wins, BDSI will likely see margins on Belbuca drop in the next few years and, as shown above, Belbuca is the main engine of BDSI’s business. If BDSI wins they have the market to themselves until 2027 (at which point only Teva will be allowed to join in with generics for several years,[4]). Alvogen is trying to break the patent lock to compete with BDSI and Teva.

There are two important things in such a special situation: what can we intelligently predict about the outcome, and what outcome or uncertainty is priced in?

Evaluating the outcome

The main assertion of Alvogen is that the bioerodable mucoadhesive layers in the patents are and function the same as those in an earlier (“Tapolsky”) patent. Alvogen draws attention to BDSI’s earlier licensing of the Tapolsky patent and claim that the new material in the target patents would be obvious. This is a difficult road to travel, legally, especially when some of the patents have already survived previous inter-partes review (a patent board adversarial challenge) and previous litigation (Teva already tried to invalidate the ‘866 patent - although the case was eventually settled instead of ruled on the merits). Alvogen faces a particularly hard challenge here because there are three different patents asserted with 9 claims in total. Alvogen would have to achieve invalidation of all 9 claims to clear their path to manufacturing a generic.

Invalidating that many claims across 3 different patents which have already been tested in various ways would be highly unusual. Judges tend to default to leaving situations as they are. They also tend to like NOT undoing previous proceedings. From these properties of intellectual property litigation we can step into the documents to see if Alvogen rises to such standards. If you want to read just a few documents I recommend the Alvogen post trial opening brief and proposed findings of fact and then the BDSI response to proposed finding of fact. I have compiled an archive of some of the more important docs you can download at this link if I’ve done it right [5].

As for your friendly neighborhood author here, I  have ten years of recent patent litigation experience and know my way around this specific technical content well enough to understand what’s going on. My opinions ([6]) on the matter:

• Alvogen has a plausible argument in asserting that that the BDSI patents “BEM” bioerodable mucoadhesive is close in basic concept to the Tapolsky patent polymer elements (which he refers to as “bioerodable” and “mucoadhesive” many times, just not with the words next to each other or as an acronym).

• Alvogen has a plausible to questionable argument that practitioners might reasonably expect low pH to enhance buprenorphine dissolution. The lab results on low pH are countered by buprenorphine widely being considered unsuitable for oral delivery which undercuts arguments that someone would be inclined to combine the pH learning with oral dosing. The low pH results in the prior literature are mostly laboratory results and certainly not conceived of as a path for getting buprenorphine in through the mouth, nor do they conceive of dissolution in bioerodable polymers. Alvogen would have to convince the judge of all of their arguments and interpretations here to win.

• Alvogen’s argument falls apart where they attempt to completely ignore that the buprenorphine is already in solution inside the polymer in the Belbucca layers. Alvogen instead tries to pretend the buprenorphine is in powder form and must be hydrated by saliva in order to cause the “low pH environment.” This tortured reading, contrary to the physics of the polymer, is necessary to avoid dealing with the lab data and technical discussion in the patents and between BDSI and the patent office showing how their polymer system works. Any capable legal environment should recognize that the “buffered polymeric diffusion environment” in the patents under attack is, by itself, a large distinction over Tapolsky and not described by even the most charitable interpretation of Alvogen’s submitted preceding publications.

•  Alvogen has had notable embarrassing mistakes. In their arguments they accused BDSI of lying to the patent office but Alvogen had their documents mixed up [see DFF19], and after final arguments the judge smacked Alvogen down pretty hard - striking out whole sections of Alvogen’s brief as improper.

Ultimately, I’m comfortable taking a risk position on the outcome going in BDSI’s favor (which means at least one applicable patent claim left standing). One can never 100% predict the courts, but I consider this outcome highly probably if the case is allowed to proceed to ruling on the merits. (More likely would be a settlement with a nominal concession to Alvogen in return for BDSI not risking an unexpected ruling.)

(An Aside on Insiders)

Insiders seem to agree. They started buying in significant volume right around the time final arguments would have been prepared for the litigation. From Nov 2016 until april of 2021 (almost 5 years)  every transaction was a sell. Since then every transaction has been a buy and the share total bought within the last 4 months equals roughly the previous year’s worth of sales.

Valuation:

What outcome is priced in?

From the figure above we can see that Belbuca is their main revenue generator. Management likes to talk positively about Symproic but it’s a smaller market and while, yes, they were seeing some traction right before COVID it’s hard to say Symproic has any proven growth. Bunavail is discontinued, so it won’t contribute anymore. Thus, BDSI is all about Belbuca (which is why the litigation over Belbuca is taken so seriously).

Now we have to ask ourselves what growth rate to attribute to Belbuca. Let’s pull that data out and evaluate it closely:

Figure: We can see that Pre-COVID the product was growing at ~13% per year (look at that blue fit!), then when COVID hit growth slowed to 4%. This isn’t really a surprise - doctors don’t like prescribing painkillers without a visit and non-COVID doctor visits plummeted. IQVIA data shows pain practices still had 36% fewer patient visits as recent as Q2 2021 compared to pre-pandemic levels. So what growth level do we think it will be at when COVID is over? 

As of March 2021 numbers came out showing the overall LAO market was down overall 5.8% compared to end of year 2020. Conversely, throughout COVID Belbuca has continued to grow, taking significant market share - which is exactly what we would expect from a good product. Market share is estimated at 4.7% end of Q2 2021, versus 3.6% end of Q1 2020 just before the pandemic - that’s a 30% rise in market share during the pandemic. Belbuca’s “new RX share” (the percentage of new within-category prescriptions) is running 7.7% and climbing, well above the Belbuca absolute share which means the absolute share should continue to grow (these two metrics tend to converge over time). Belbuca’s market share gains appear to be coming mostly in the form of taking prescriptions from pill-based non-buprenorphin LAOs:

Figure: Some 2019 data that helps show that the increases in Belbuca share appears to be coming from Oxycontin, Fentanyl, and Morphine reductions. Belbuca as a mouth strip gives stable long term performance. Meanwhile the buprenorphine in Belbuca is abuse resistant: it can’t be crushed and even if a strip were dissolved and swallowed the “high first pass effect” from the liver (the same effect that makes a mucosal strip necessary) removes it from the bloodstream so fast there would be little effect.

We can evaluate the price to see what growth path is expected and what is priced in.

This isn’t a mystery novel so I’ll tell you the ending now. Reverse DCF shows that the current market valuation is pricing in continuation along or below the red (COVID) line, which I think only makes sense if interpreted as a large risk discount over the patent suit. Conversely, if we instead use a return to the blue line growth rate starting in 2022 [7] we get a stock value of $7.32 for a 90% up-side as of when I’m typing this, assuming nothing more than a return to trend.

Diving into the numbers

As of 2020 Q2 Belbuca total available market was about 900K Rx/month according to management - my model scenarios don’t depend on getting anywhere near this number so it isn’t very important except to know there is room to grow (which is what we validated with doctors). Meanwhile, Belbuca is currently running ~34K Rx/month. So the defining question here is: what will the growth look like in the future? I consider two waypoints. The first waypoint is for BDSI to double sales to meet their *current* new Rx share (note: that share is still growing), which drives revenue to $320mm per year and seems to be unchallenging given recent sales trends. The second waypoint I consider to be a likely limit on growth which is 30% of the overall LAO market or about 300K Rx per month ($1.5B in revenue). Right now Belbuca is basically the only LAO doing any marketing but once they start to reach the billions in revenue it seems likely a competitor will step up or generics (entering the market 2027) will be big so I’m going to assume $1.5B is a hard upper limit for this product right now.  

Down below are the detailed net present value calculations with an 8% discount rate and terminal growth at 2034 assuming terminal growth rate = cost of capital. This isn’t a spreadsheet driven idea where the outcome is dependent on tuning a variable somewhere, so I’m avoiding false precision with a very simple calculation. Debt is represented by actual payments straight from cash flow on the current debt schedule (i.e. assumes no refinancing). I put in Maintenance Capex / Investment of 3% of revenue which is actually a bit higher than their current annual cost of license/amortization of acquisition for their current products. Management has made clear they expect to spend some of the cash flow on new products licensed from others; we just got a look at the first deal and it looks 3% is a good representation of their operation and my modeling actually assumes that doesn’t create significant value. Sims use the share count from recent 10-K minus buybacks since then, although a higher number was also tested since the accounting for Series B preferred seems unclear [8], there was no significant change of results. Management current guidance is 170-180mm revenue for the year and low-med-high of those are used on the different cases. Again, missing revenue targets for this year has little to no ultimate impact on the opportunity here.

The most interesting thing I found was that doing a reverse DCF at these values returns an imputed growth rate at or below the during-COVID rate. This suggests to me that the market has a substantial legal risk discount baked in right now. This price level seems unsustainably low if Belbuca survives the litigation and sales recover at all from COVID.

“Low Case” is resuming an 8% trend long term growth rate in 2022. At this growth we don’t reach first waypoint (market share equals *current* new Rx share) until the end of 2029 and overall market share is still below 10% at 2034 when terminal value is hit. NPV per share is ~$5. That’s a 30% upside using what appear to be pretty conservative numbers..

“Mid Case” is a return to blue line 13% growth in 2022. This case sees the Rx share reach *current* new Rx share at the end of 2026, and Belbuca reaches about 15% LAO market share right at year end 2033 as we go into terminal value. The NPV for this case is $7.32.

“Rosy Case” assumes that some of their new marketing strategies (such as rolling out their successful “first start” program nationally) get traction and things go very well. It models 16% growth 2022-2024, then 12% until terminal value. This case sees the Rx share reach *current* new Rx share at the beginning of 2026 and has a NPV of $7.75. That’s more than double the share price as I write this.

Along the way they should have plenty of cash to pay off their debt (which is included in each model). If BDSI eventually stabilizes at 30% of the market for BELBUCA at a cautious 10x FCF that gives a market value of $3.4B and note that I’m leaving out growth of other products and a building portfolio as well as buybacks along the way. This looks like a very good long term bet at this price, but the special situation is more immediate.

[1] In technical terms is has “high first pass effect” which means the liver is very effective at nullifying buprenorphine

[2] https://www.ncbi.nlm.nih.gov/books/NBK554435/

[3] Technically, they are specifically targeting 9 claims across the three patents

[4] Due to a previous settlement on a Paragraph IV certifications suit

[5] Also, I tried to use publicly available versions and to scrub my personal PACER ID out of the ones I contributed, please let me know if you spot any of that remaining

[6] Not legal advice, personal opinion

[7]  COVID should be over by 2022, right?  Actually changing that in the models to 2023 doesn’t change a whole lot, the stock price is so low.

[8]  For shares outstanding I also tried 107mm vs 102mm (diluted equivalent reported after share buybacks since the last annual report) because I don’t like the way they are accounting for the remaining 443 shares of Preferred B shares outstanding - they convert at $1.80 equivalent on a $10,000 headline price but have some “deemed dividends” baked into them that seem a bit unclear so I tested a number that was certainly too high as well.

Catalyst

The main catalyst here will be the outcome of the litigation. If Belbuca survives still protected (even by one patent), or Alvogen agrees to a last-minute settlement, BDSI’s price is suddenly far too low. Now, I’m sure we’ve all been stuck holding stocks at prices that don’t make sense so we can’t assume it will instantly return to proper valuation. However, in this case there is a force to drive BDSI to proper valuation: growth return as COVID fades. Free cash flow is already piling up and management is already engaging in buybacks (and buying for their personal accounts)! So perhaps BDSI jumps immediately on litigation outcome, maybe it rises over the following year, most likely a bit of both - but given the situation this looks like perhaps around a 70-100% rise iin the next 12-24 months.