One hand on the prize

Paion is a Frankfurt-listed emerging specialty pharma company that develops products for anaesthesia and critical care. Its headquarters are in Aachen, Germany. Paion is focusing on lead programme remimazolam (CNS7056), which it acquired in 2008 through its purchase of CeNeS Pharma for €12m. Short-acting sedative remimazolam has potential in three indications: procedural sedation, general anaesthesia and intensive care unit (ICU) sedation. Paion completed a successful US Phase III study mid-year, with a second Phase III expected to complete enrolment in Q217. It has raised over €125m since flotation in 2005, including €46m at the IPO. Paion has licensed US rights to remimazolam to Cosmo Pharmaceuticals and is evaluating its commercialisation options in unpartnered territories including Europe and Japan, which could include appointing a salesforce in some regions and/or commercial partners.

Our sum-of-the-parts DCF valuation is €208m or €3.74 per share, a modest upgrade from €203m in our July report. Paion has sufficient funding to complete pivotal studies in the US and support operations to anticipated filing in H118. The current valuation excludes any potential dilution from the additional funding to file for approval in Japan or to initiate a new Phase III trial in general anaesthesia in Europe. This funding could come from a partner, an equity raise or, depending on timing, from proceeds of commercialisation of remimazolam in the US. We assume that Paion achieves average royalty rates of a 20% in Europe and Japan in line with the lowest tier in the Cosmo licence agreement. Paion has other options include negotiating a profit share, which may deliver a 35% margin.

The main sensitivity for Paion is the success or failure of clinical studies with lead product remimazolam. There is already a substantial body of advanced clinical evidence for remimazolam showing it has a good safety and efficacy profile in comparison to the standard of care. This broad evidence should support the product and the economic rationale, which will compete with established generic products. Other sensitivities include the usual regulatory, financial and partnering risks associated with a pharmaceutical company in the late stages of development, and preparing for commercialisation.

Operating expenses for 9M16 decreased to €20.6m vs €25.4m for 9M15, predominantly due to a lower R&D spend of €16.4m vs €20.9m in 9M15. Paion ended Q316 with €35.9m in net cash, which is expected to be sufficient to fund ongoing Phase III development of remimazolam, including preparation for filing for procedural sedation in the US. Thereafter, we expect that milestones from Cosmo for filing and approval would fund operations until Paion starts receiving royalties on US sales, as long as development goes according to plan. Additional funding or a partner would be required to support filing in Japan and renewed development of general anaesthesia in Europe. We rebalance our FY16 estimated costs, shifting €0.6m from R&D (now €25.4m) to SG&A (now €5.5m) based on guidance. We also allocate €6m of the €10m Cosmo upfront payment to deferred revenue in line with guidance, reducing FY16 revenue by €6m to €4.2m. Our end-FY16 cash estimate is €26.1m. We add €4.5m to costs in FY17 due to the slower recruitment in the bronchoscopy Phase III trial pushing more R&D costs into next year, taking end-2017 net cash to €14.4m from €18.5m.

Paion is at an advanced stage in developing remimazolam, an ultra-short-acting sedative, having recently successfully completed the first of two US pivotal studies, which could lead to it being approved in the US in H218/H119 in the lead indication procedural sedation. Restarting European studies in general anaesthesia is dependent on partnering or additional funding. Paion has licenced US rights to Cosmo Pharmaceuticals and is evaluating commercialisation options for unpartnered regions. The economic rationale for remimazolam focuses on faster induction and recovery from sedation and an improved safety profile compared to generic alternatives.

Paion granted Cosmo Pharmaceuticals an exclusive licence for the development and commercialisation of remimazolam in the US in June, soon after reporting positive results for the colonoscopy Phase III. The licence deal brings €10m upfront cash (received in July), an equity injection of €10m (€0.4m deferred), plus regulatory and sales milestones of €42.5m and a tiered royalty of 20-25%. The royalty rate can be adjusted downwards under certain conditions (believed to relate to product pricing) but not to below 15%. Cosmo will fund regulatory filing and commercialisation in the US, while Paion will fund the ongoing bronchoscopy Phase III.

End-September 2016 cash of €35.9m plus anticipated milestone payments for filing and approval of remimazolam the US should mean that Paion is fully funded to the point where it begins to earn royalty income in the US, as long as remimazolam development continues to progress as expected. This puts Paion on a strong footing as it assesses its options for further development of remimazolam in Japan and Europe.

Paion is in discussions with potential partners for remimazolam in Japan and Europe, and its strong financial position means it is not under pressure to do a deal in the near term, but can take its time to find the right partner and negotiate suitable financial terms.

Paion is focusing on developing remimazolam, an ultra-fast-acting intravenous (IV) sedative. Remimazolam could be a safer, faster alternative to approved sedatives and potentially carries a reduced risk of cardiac and respiratory depression, which is particularly significant for older and less healthy patients. Studies of remimazolam have shown it is suited for three indications requiring varying depths of sedation – general anaesthesia, procedural and ICU sedation – while maintaining the vital physiological and neurological functions of the patient.1 Paion has a global strategy in place through a combination of independently funded studies and partnered development and the lead indication per region varies depending on the most attractive business rationale. The characteristics of remimazolam compared to standard sedatives are shown in Exhibit 2.

The business rationale is based on an unmet need for safer and more controllable sedation supported by the potential for cost savings and reduction in the need for patient supervision during and after procedures compared to approved sedatives.

Paion reported positive results from the first of two US pivotal studies of remimazolam in procedural sedation; 91% of patients in the remimazolam arm achieved the primary outcome (completion of the colonoscopy procedure without rescue medication) vs 1.7% on placebo (Exhibit 4). The time from start of medication to start of procedure (induction time) was four minutes for remimazolam vs 19 minutes for midazolam, while time from end of procedure to fully alert was 7.2 minutes vs 15.7 minutes for remimazolam and midazolam, respectively. There were also fewer instances of hypotension in the remimazolam arm.

A key plank of the business case for remimazolam is that faster induction of and recovery from sedation compared with midazolam will allow increased patient throughput for colonoscopies and other procedures where conscious sedation is required.

The colonoscopy Phase III trial results give strong support to the business case with recovery time to full alertness 8.5 minutes faster for remimazolam than midazolam. However, the data on induction times are more difficult to interpret. The induction time in the midazolam arm of the Phase III trial was 19 minutes compared to 4.8 minutes in the midazolam arm of Paion’s previous Phase II trial (induction with remimazolam was 2.2-3.0 minutes faster than midazolam in the Phase II).

One possible explanation for the difference in midazolam induction times between the Phase II and Phase III trials is that the FDA-approved label for use of midazolam for procedural sedation recommends that each dose for the drug be injected slowly over two minutes, followed by a two-minute wait to assess depth of sedation before administering each top-up dose. Under this protocol each top-up dose would add four minutes to the time to sedation. The label says that a total dose greater than 5mg is not usually required to achieve sedation. Under this protocol, administering a starting dose of 1.75mg if midazolam plus three doses of 1mg to reach a total dose of 4.75mg to induce sedation would take 16 minutes. This time interval is similar to the 19 minutes reported in Paion’s Phase III trial.

However, it appears that the real-world use of midazolam does not follow the protocol outlined in the FDA-approved label. In each of the three studies listed below, for healthy adults under the age of 60 each dose of midazolam was administered as bolus injection, including the initial dose of ~2mg, with additional doses of medication administered at one- to three-minute intervals to achieve or maintain the desired level of sedation. In Paion’s Phase II trial, a higher initial dose of midazolam was used (2.5mg), allowing average time to procedure start to be even shorter than the published trials at 4.8 minutes. In addition, a higher dose of fentanyl was given prior to midazolam or remimazolam in the Phase II study (100µg) compared with in the Phase III trial (50-75µg).

DeWitt et al. (2008)2 reported a mean time to start of endoscopy procedure of 2.5 minutes for propofol and 6.4 minutes for midazolam plus meperidine (pethidine). A separate study by Sipe et al. (2002)3 reported mean time to sedation of 2.1 minutes for propofol and 7.1 minutes for midazolam plus meperidine. A third study by Ulmer et al. (2003)4 reported time to sedation of 2.1 minutes for propofol vs 6.1 minutes for midazolam plus fentanyl.

Average time to sedation for midazolam in the three studies above and in Paion’s Phase II trial was 6.1 minutes. The average time to sedation for propofol in the three published studies was 2.2 minutes.

We also examined the recovery times reported in published data where midazolam and propofol have been used for sedation in colonoscopies and similar procedures. In a Cochrane review (which did not report time to induce sedation) of the use of propofol for sedation during colonoscopy, the average recovery time in nine studies was 12 minutes for propofol-treated patients vs 31 minutes for traditional sedation (primarily midazolam, diazepam or other benzodiazepines combined with an opioid). For the two studies that used midazolam plus fentanyl, the combination used in the midazolam arm of Paion’s colonoscopy trial, the average recovery time was 30 minutes.

The average duration of the colonoscopy examination procedure reported in the Cochrane review was 18 minutes. Horiuchi et al. (2012)5 reported that the average colonoscopy procedure time was 14 minutes in a series of 2,100 patients in Japan.

Exhibit 5 compares the time taken from the start of sedation to the start of the colonoscopy procedure (induction) and the time to recover to full alertness for remimazolam, propofol and midazolam. For midazolam we have shown the data from the open-label arm of Paion’s Phase III and the average data from the Cochrane review and the published studies mentioned above. The second to last column combines the best-case outcomes for midazolam, recovery time from Paion’s trial with the induction time from published studies.

While acknowledging that there are limitations to indirect comparisons between clinical trials, Exhibit 5 suggests that combined induction and recovery time for remimazolam is comparable to that for propofol and substantially faster than for midazolam. Induction and recovery for remimazolam was 11 minutes faster than the best-case estimate for midazolam and 24-25 minutes faster than the two separate estimates for midazolam from Paion’s Phase III.

Exhibit 6 shows that even if we use the most conservative estimate of an 11-minute time saving with remimazolam, a clinician who performs four colonoscopies per day with midazolam could do one extra colonoscopy in the same or less time with remimazolam, while one who does 10 procedures with midazolam could do 13 in the same time with remimazolam.

In the US, moderate sedation (also known as conscious sedation) can be administered by an appropriately trained, licenced, registered nurse or physician assistant under the supervision of a non-anaesthesiologist sedation practitioner such as a licenced physician (eg gastroenterologist) who is specifically trained to administer moderate sedation.

However, eight US states (New York, Connecticut, Georgia, Maine, Nebraska, New Hampshire, South Carolina and West Virginia) restrict this use to sedation agents such as midazolam and exclude nurses from administering drugs such as propofol, which also be used to induce general anaesthesia. In these states propofol can only be administered by an anaesthesia professional such as an anaesthesiologist, a certified registered nurse anaesthetist (CRNA), or a physician trained in the administration of deep sedation.

The American society of Anaesthesiologists (ASA) has issued practice guidelines for sedation and analgesia by non-anaesthesiologists, and a statement on granting privileges for administration of moderate sedation to practitioners who are not anaesthesia professionals. The ASA used the following definitions:

Moderate sedation (conscious sedation) – a drug-induced depression of consciousness during which patients respond purposefully to verbal commands, either alone or accompanied by light tactile stimulation. No interventions are required to maintain a patent airway, and spontaneous ventilation is adequate. Cardiovascular function is usually maintained.

Deep sedation – a drug-induced depression of consciousness during which patients cannot be easily aroused but respond purposefully following repeated or painful stimulation. The ability to independently maintain ventilatory function may be impaired. Patients may require assistance in maintaining a patent airway, and spontaneous ventilation may be inadequate. Cardiovascular function is usually maintained.

In a separate position statement, the ASA said that because of the significant risk that patients who receive deep sedation may enter a state of general anaesthesia, non-anaesthesiologist physicians may neither delegate nor supervise the administration or monitoring of deep sedation by individuals who are not themselves qualified and trained to administer deep sedation, and the recognition of and rescue from general anaesthesia (ie practice nurses).

Benzodiazepine sedatives such as midazolam (and remimazolam if approved) are used in combination with an opioid such as fentanyl to induce moderate sedation and analgesia. Propofol can be used for moderate but is often used to induce deep sedation. Benzodiazepines, including remimazolam, have the advantage that the sedation and respiratory depression effects can be reversed with flumazenil, making these agents safer to use.

A study by the RAND Corporation6 estimated that 12.5 million colonoscopies and upper gastrointestinal (GI) endoscopy procedures were performed in the US in 2009. The study, which analysed claims data for 6.6 million patients, found that in 2009 27% of Medicare patients and 29% of commercially insured patients received anaesthesia services from an anaesthesiologist or nurse anaesthetist for their colonoscopy. The majority of anaesthesia services were delivered to low-risk patients with American Society of Anaesthesiologists (ASA) status level 1 or 2 (64% of Medicare patients and 84% of commercially insured patients where the use of anaesthesia services was considered to be potentially discretionary. They estimated that US$1.1bn of the US$1.3bn paid for anaesthesia providers for GI procedures in 2009 was for these low-risk patients, which highlights the potential savings that could be made if remimazolam was used in place of propofol, reducing the use of anaesthesiologists.

The percentage of procedures using anaesthesia providers varied across the US, with 48% in the Northeast, 38% in the South, 26% in the Midwest and 14% in the West.

The average payment per procedure for anaesthesia services was US$150 for Medicare patients and US$509 for commercially insured patients in 2009, according to the RAND Corporation report.

Average revenue $788 per colonoscopy at ASCs in 2012 VMG Health's 2012 Intellimarker report on the financial and operating results of ambulatory surgical centres (ASCs) in the US reported that in 2012 the gross charges billed per colonoscopy case were $3,610 and the net revenue per case was $788. The average discount of revenue received to the amount charged was 75%.

Looking just at Medicare fee-for-service patients our analysis of Medicare Provider and Utilisation data showed that in 2014 there were 3.9m claims for colonoscopies. The average physician charge submitted for the colonoscopy procedures was US$1,168 and the average Medicare payment was $208 per procedure.

There were also 0.56m claims for anaesthesiology services for lower intestine endoscopy, with an average submitted charge of US$907 and average Medicare payment of $101.

The American Gastroenterological Association (AGA) has developed models for bundling the entire cost of colonoscopy procedures, including the cost of sedation. The practice of bundling does not appear to be widespread at present, but with pressure growing to contain healthcare costs we expect the practice to become more common which would support uptake of remimazolam.

Remimazolam offers potential for considerable savings for practices that receive a bundled payment for provision of colonoscopy services, offering equivalent speed of sedation to propofol and a high level of safety without the cost of employing the services of anaesthesiologist.

Paion initiated a second US Phase III study of remimazolam in procedural sedation in June 2015, not long after the colonoscopy trial began, to fulfil the FDA’s requirement for two pivotal trials.

The bronchoscopy study was initially expected to be completed in 2016; however, recruitment has been moderate to date and full recruitment is now expected in Q217. Paion has introduced a number of measures to increase patient recruitment, and is in regular close contact with each trial site. The protocol has been amended to allow inclusion of patients taking benzodiazepines or opioids, which are often used to relieve anxiety and breathlessness in respiratory disease. The trial has been reduced from 460 to 420 patients by reducing the size of the midazolam arm, which does not affect the primary endpoint. The trial aims to treat 300 patients with remimazolam, 60 with placebo and 60 with midazolam (original design included 100 with midazolam). The comparison with midazolam is intended for pharmacoeconomic analysis and marketing purposes and is not required for approval of remimazolam.

The prospective double-blind, randomised study in bronchoscopy is recruiting patients across multiple US sites. Patients in the treatment arm will receive 5mg of IV remimazolam for sedation induction, and 2.5mg top-ups for sedation maintenance. The study specifically targets patients with pulmonary disease to demonstrate efficacy and safety in those that are high risk. In the study, patients will be randomised to receive remimazolam plus fentanyl, midazolam or placebo. The study endpoint is to conduct the bronchoscopy without requirement for any alternative sedation. As in the colonoscopy study, the secondary endpoints include speed of onset, time to alertness, full recovery and discharge.

Paion regained Japanese rights to remimazolam in November 2014 when partner Ono terminated the agreement that was struck in 2007. It has held a positive pre-NDA meeting with the Japanese regulator (PMDA) and received guidance that the drug manufactured in Europe is considered equivalent to the compound used in Japanese trials. The PMDA stated that the non-clinical and clinical data package was regarded as complete for filing for general anaesthesia.

Paion is in discussions with potential partners for remimazolam in Japan who would take the responsibility of submitting the marketing application. Alternatively, the company has the option to raise additional funding and file for approval in Japan itself, with the assistance of a contract research organisation (CRO) with appropriate expertise. This strategy has the advantage that a Japan filing dossier could also be used for additional filings in countries such as South Korea, which would attract a milestone payment from partner Hana Pharm. Depending on the timing of filing, we estimate that remimazolam could be approved in Japan during 2018.

The outcome of Ono’s Phase II/III trial of remimazolam in general anaesthesia reported in November 2013 was promising. The trial met its primary endpoint and showed remimazolam was 100% effective as a sedative in the 375-patient trial. The incidence rate of a drop in blood pressure was 35.3% and 34.7% of patients in the high- and low-dose remimazolam groups, vs 60% of patients in the propofol arm, and the difference between the remimazolam and midazolam groups was statistically significant. The data are consistent with earlier studies that indicated remimazolam’s good safety profile.

However, in August 2013 Ono discontinued a separate Phase II trial of remimazolam for sedation in ICUs. While all patients were sedated successfully and there were no significant unexpected adverse events, higher than expected plasma concentrations of remimazolam were observed in isolated cases after long-term treatment. The phenomenon of elevated remimazolam plasma concentrations could not be reproduced in preclinical studies or pharmacokinetic models. Further analysis by Paion has shown that such pharmacokinetic deviations are common for sedatives like midazolam and propofol in the ICU and are probably related to the underlying conditions of the patients in the ICU. Paion concluded that the maximum dose level has been identified for ICU sedation. Further development in ICU sedation is planned after the approval of the lead indications and when the required funds are available.

Paion is also assessing its options for the commercialisation of remimazolam in Europe after the European Phase III trial in general anaesthesia in cardiac surgery patients was discontinued in February 2016 due to slow recruitment. The company believes that an alternative trial design in general surgery patients could be completed successfully, but the initiation would require additional funding or a partner. If development in Europe is restarted, we would also expect the company to seek approval for use in procedural sedation in Europe, on the back of US Phase III trial data. The discontinuation of the European Phase III trial has allowed additional resources to be allocated to the development of remimazolam in the US.

Paion announced in February that the term of its US patent covering the crystalline form of remimazolam besylate had been extended by 4.3 years due to delays at the US patent office. This will extend US market exclusivity to late 2031 compared to 2027 in other major markets.

The US patent covers crystalline forms of the besylate salt of remimazolam, which Paion selected for use in the drug formulation due to its excellent stability profile. US market exclusivity may be further extended under the Hatch-Waxman Act following FDA approval, representing potential upside to our forecasts. The longer patent life dovetails with the company’s increased focus on the US development of remimazolam for procedural sedation following the discontinuation of the European general anaesthesia Phase III.

Paion has adopted a regional partnering strategy to accelerate remimazolam’s global development and provide marketing partners in each region. The results from pivotal studies in the US and Japan will be bridged to data from each region and could abbreviate clinical studies of remimazolam in these individual geographies. This strategy advances remimazolam’s global clinical status and market potential in a cost-effective way. Paion received upfront payments for each of its four regional partners and is eligible to receive milestone payments plus royalties. These partners will commercialise remimazolam in the respective regions.

In September Paion announced that its partner Yichang Humanwell has completed recruitment in its first Phase I study in China and is planning a Phase II trial in procedural sedation. In addition, a second Phase I study with continuous infusion is underway to prepare for a Phase II study in general anaesthesia.

Paion’s other clinical product, GGF2 (cimaglermin alfa) was being developed by Acorda to restore cardiac muscle function in heart failure patients. Paion out-licensed GGF2 at the preclinical stage to Acorda, which started a Phase Ib FDA trial in October 2013 to assess the safety, pharmacokinetics and tolerability of three dose levels of GGF2 in up to 30 patients. However, in June 2015 the company announced that it had stopped enrolment in the Phase Ib clinical trial of GGF2 as blood tests revealed a case of liver injury. Trials were subsequently placed on clinical hold by the FDA pending full analysis of liver interactions. The 22 patients who were dosed in the trial have completed 12-month follow up. In July 2016 the status of trial NCT01944683 on clinicaltrials.gov was change to completed without resuming patient recruitment. Acorda does not list GGF2 in its current development pipeline and in its 2015 annual report stated that it is “working with outside experts to better understand this liver effect and assess next steps for this programme”. In light of this, we have removed GGF2 from our valuation model for Paion pending further clarity on development plans.

A previous Phase Ia trial showed GGF2 was well tolerated and improved ventricular function in study participants.

Paion has two other products in its pipeline: M6G (perioperative pain) and human thrombomodulin, Solulin. Paion has halted development of both owing to the focus on remimazolam, although development of M6G may be resumed. Paion granted China development, manufacture and commercialisation rights for M6G to Yichang Pharma in September 2014 for an upfront payment plus milestones totalling €1.6m.

The key sensitivity is the outcome of clinical trials with remimazolam, notably the outcome of the Phase III bronchoscopy study in US. The Cosmo licence deal means that Paion should be fully funded until it begins to receive royalty income from US sales, but this is dependent on successful completion of the Phase III bronchoscopy trial and timely regulatory review and subsequent approval by the FDA. Any delay or clinical trial failure may mean that Paion needs to raise additional funds to support operations until a US market launch.

Paion has the option to file for approval in Japan in its own behalf with the assistance of a CRO, but this would likely mean the company would need to raise additional funds at some stage – appointing a partner in Japan who would fund the regulatory filing would be a lower risk approach. Similarly, resumption of development of remimazolam for general anaesthesia in Europe would require a partner or additional funding.

Our sum-of-the-parts DCF valuation is €208m, or €3.74 per share, a slight increase from €203m or €3.65 per share. This is based on the assumption that Paion forms post-approval commercialisation deals for remimazolam, which would yield a royalty rate of 20% in unpartnered regions. The summary of changes to our valuation includes:

In the US, our cost per procedure assumption for remimazolam is $40 and our peak sales estimate is $308m for the lead indication procedural sedation, assuming an addressable market of 35 million procedures a year.7 In Canada, our peak sales assumption is $42m, for seven million procedures at a revenue per procedure of $30, and we use a market penetration estimate of 22% in the US and Canada. The time savings over midazolam seen in the colonoscopy Phase III trial increase the likelihood that penetration could be higher and remimazolam could be adopted for additional conscious sedation indications; at this stage we have made only a modest revision to our market uptake assumptions.

Our peak sales assumption in Europe for the lead general anaesthesia indication is $175m, assuming a price of $25 per procedure and 35 million high-risk or class III/IV discharges in the OECD region per year. In Japan, our peak sales assumption is $75m at an average price of $25 and 20 million procedures a year (in general anaesthesia). We assume that the Japanese data package will be sufficient for filing in 2017. We assume a 15% market penetration in Europe and Japan given the apparent advantages over propofol for the high-risk patients.

The key catalysts in H216 include an update on recruitment progress in the bronchoscopy and confirming the filing timeline for Japan. Paion is evaluating its commercialisation options in unpartnered regions, which could lead us to adjust our royalty rate assumptions.

Operating expenses for 9M16 decreased to €20.6m vs €25.4m for 9M15, predominantly due to a lower R&D spend of €16.4m vs €20.9m in 9M15. Paion ended Q316 with €35.9m in net cash, which is expected to be sufficient to fund ongoing Phase III development of remimazolam, including preparation for filing for procedural sedation in the US. Thereafter, we expect milestones from Cosmo for filing and approval would fund operations until Paion starts receiving royalties on US sales, as long as development goes according to plan. Additional funding or a partner would be required to support filing in Japan and renewed development general anaesthesia in Europe. We rebalance our FY16 estimated costs, shifting €0.6m from R&D (now €25.4m) to SG&A (now €5.5m) based on guidance. We also allocate €6m of the €10m Cosmo upfront payment to deferred revenue in line with guidance, reducing FY16 revenue by €6m to €4.2m. Our end-FY16 cash estimate is €26.1m. We add €4.5m to costs in FY17 due to the slower recruitment in the bronchoscopy Phase III trial pushing more R&D costs into next year, taking end-of-year 2017 cash to €14.4m from €18.5m.

[Grab your reader’s attention with a great quote from the document or use this space to emphasize a key point. To place this text box anywhere on the page, just drag it.]