Good morning. My name is Jessa and I will be your conference operator today. At this time, I would like to welcome everyone to the Biogen Third Quarter 2019 Financial Results And Business update. Thank you. I would now like to turn the conference over to Mr.
Joe Mara President, Investor Relations. You may begin your conference.
Good morning, everyone, and welcome to Biogen's third quarter 2019 earnings conference call. On today's call, we will be discussing our Q3 results as well as an update on our Alzheimer's program aducanumab, including our plan to file in the U. S. Before we begin, I encourage everyone to go to the Investors section of biogen.com to find the earnings release and related financial tables including a reconciliation of the GAAP and Table 3 includes a reconciliation of our GAAP to non GAAP financial results. We believe non GAAP financial results better represent the ongoing economics of our business and reflect how we manage the business internally.
We've also posted slides on our website that follow the discussion related to this call. I would like to point out that we will be making forward looking statements, which are based on our current expectations and beliefs. These statements are subject to certain risks and uncertainties and our actual results may differ materially. I encourage you to consult the risk factors discussed in our SEC filings for additional detail. On today's call, I am joined by our Chief Executive Officer, Michelle Vinazzos, Doctor.
Al Sandrock, EVP Research And Development, and our Chief Medical Officer, Doctor. Samantha Bud Haberlein, Vice President, late stage clinical development and our CFO, Jeff Capello. Now, I will turn the call over to Michelle.
Thank you, Joe, and good morning, everyone. I will start by giving you an outline of this call since we're announcing both Q3 results and news on aducanumab. First, I will review the recent news on aducanumab. Next, I will provide the key highlights of a strong quarter. Al and Samantha will then provide additional details on aducanumab and progress across the rest of our pipeline.
Jeff will discuss our financial performance for Q3 and I will close before we open the call for questions. This is an important day as we are announcing that based on discussion with the FDA, we plan to submit a regulatory filing in the U. S. For aducanumab. If approved, aducanumab will become the first therapy to reduce clinical decline in Alzheimer's disease and the first therapy to show that removing amyloid beta can lead to better clinical outcomes.
This is an important milestones, providing hope for patients, physicians, caregivers and families around the world. It is also important to highlight that the path taken in the pursuit of discovering and developing breakthrough Treatments is not always direct and straightforward. As you know, in March, we announced our decision to discontinue the Phase III Immersion Engage Studies for aducanumab in early Alzheimer's disease based on a pre specified fertility analysis. In retrospect, the result of our futility analysis was incorrect. Based on what we know now, it is clear the pre specified futility criteria did not adequately anticipate the effect of all the variables in these trials.
So what happened? First, the decision to stop these trials relied on an earlier and smaller data set comprised only of patients who have the opportunity to complete 18 months of treatment as of December 26, 2018. At that time, the fertility analysis predicted that the trials were unlikely to meet the primary endpoint upon completion. Fertility analysis are common in lab studies and they use statistical modeling to attempt to predict the outcome of the studies based on a number of pre specified assumptions and criteria. There are multiple methodologies that can be used for futility analysis, and the methodology we used was a well accepted approach.
However, based on what we have learned, We know now that the futility analysis did not adequately account for the effect that the earlier enrollment in Engage had on patients' overhaul exposure to high dose aducanumab. 2nd, In the month following the discontinuation of these studies, our team has continued to analyze the vast set of clinical imaging and biomarker data that the studies have generated. In addition to further analysis of the data set, which informed the fertility analysis, we also gain access to and analyze additional data. Including data on patients who completed treatment after the cutoff date for the futility analysis, as well as data for patients who did not complete the full duration of the study. Once we became aware of the potential implication of this larger dataset we consulted with external advisors, followed by the FDA with a Type C meeting in June, as we began conducting further analysis.
3rd, the new analysis of the larger dataset which was conducted in consultation with the FDA showed that aducanumab had a dose dependent effect on the underlying pathology as measured by amyloid PET Imaging and reduced clinical decline in PACE with early Alzheimer's disease as measured by based on a second Type C meeting held with the FDA just yesterday, we believe this data support a regulatory filing. Beyond aducanumab, we are hopeful about the implication of these results for similar approaches targeting amyloid beta, including BAN2401. Looking forward, we plan new dialogue with regulatory authorities in international markets, including in Europe and Japan. Most importantly, we have a deep commitment to the With the support of the FDA, we aim to offer access to aducanumab as soon as possible to eligible patients previously enrolled in the Phase III studies, the long term extension of the Phase I PRIME B study, and the Phase II Evolve study. We will work towards this call with the regulatory authorities, principal investigators and the institutional review boards with a sense of urgency.
I know this development is unexpected for all our stakeholders. Including the many patients involved in our clinical studies and their families, our investigators, our investors and our employees. And I'm sure you have a lot of questions. We were also surprised when we initially application of this additional data, but our surprise quickly turned to the hope that we may be in the position to Alzheimer's patients, the first therapy to reduce clinical decline in this devastating disease. We are humbled and honored by this opportunity, and we are committed to following the science and doing the right Now, let me review some financial highlights from the third quarter.
Compared to the same period a year ago, Biogen delivered solid top line and bottom line growth. 3rd quarter revenues grew 5 percent to $3,600,000,000. 3rd quarter GAAP earnings per share grew 17% to $8.39 and non GAAP EPS grew 24 percent to $9.17. Now let me briefly review our progress against our strategic business priorities. First, we continue to demonstrate resilience in MS, and we remain focused on addressing the IP challenge 2nd, SPINRAZA revenues grew double digit both year over year and quarter over quarter.
3rd, our biosimilars business continued to grow, led by the recent launch of Imhaadi. 4th, we continue to progress our pipeline with a positive readout on VUMERITY and the addition of 2 new clinical programs. And 5th, Biogen has continued to be disciplined and focused on capital allocation. As we have demonstrated, we are committed to maximizing returns for our shareholders while continuing to bring innovative therapies to patients, something that demands a thoughtful approach towards all our investment over both the short term and the long term. Before turning the call over to Al, I would like to congratulate him on his new role as the head of R And D at Biogen.
I know I speak for the broader Biogen family in expressing my confidence in Al as a scientist as a clinician and as a leader. Al is taking the helm at the time when I believe Biogen R&D has never been stronger. And I believe he's well equipped to continue driving Biogen R&D forward.
Thank you, Michelle. Before diving in, let me first take a moment to say how excited I am about the opportunity to lead the R and D organization here at Biogen. I'm extremely proud of the team for that Biogen is uniquely positioned to bring breakthroughs in neuroscience and transform the lives of patients with neurological disease. As Michelle said, we were all surprised when we learned of the potential implications of the new analysis a larger dataset from the Phase III studies of aducanumab. Following discussions with external advisors and the FDA, We have now come to better understand what happened.
And even more importantly, with the positive implications of the larger data set may mean for patients physicians, and the broader scientific community. To start, I will briefly summarize our current understanding of the phase 3 data. Before I turn the call over to Samantha, we will describe in more detail the series of events and analyses that have led us to the current conclusions and status today. First, it is important to understand what the results that the results you will hear about today which we have analyzed in consultation with the FDA are based on a new analysis of a exposure to high dose aducanumab reduced clinical decline across multiple clinical endpoints. This reduction in clinical decline was statistically significant and emerge, and we believe that patients that data from patients who achieve efficient exposure to high dose aducanumab and engage support the findings of IMerge.
After consultation with the FDA, we believe that analysis where those who had enrolled early in the trials and those early enrolling patients had a lower average exposure to aducanumab in large part due to 2 protocol amendments that occurred sometime after the start of the trials. These 2 protocol amendments were put in place precise lead to enable the larger dataset available after trial cessation included more patients with sufficient exposure to high dose aducanumab. Moreover, differences between IMerge and Engage can mostly be accounted for by a greater level of exposure to high dose aducanumab in IMerge due to multiple factors, including the fact that ENGAGE started earlier and enrolled earlier than IMerge, meaning that fewer patients and engaged had sufficient exposure to high dose aducanumab, as well as other factors, including differences in the degree of dose suspension due to ARIA. Taken together, we believe that these data and the associated extensive analysis provide compelling evidence that aducanumab reduce as the otherwise devastating and inexorable clinical decline
Thank you, Al. Let me now describe in more detail how we got here. It's important to first understand the design of the studies and how that evolved over time. IMerge and Engage were phase 3 multicenter, randomized, double blind, placebo controlled parallel group studies designed to evaluate the efficacy and safety of aducanumab in early Alzheimer's disease The studies were identical in design. At full enrollment, IMerge included 1638 patients, and Engage included 1647 patients.
Based on the data from the phase 1b PRIME study of umab, we believed that higher doses of aducanumab may be associated with improved clinical outcomes. However, as the incidence of amyloid related imaging abnormality or aria for short, the most common adverse event associated with aducanumab also increased with aducanumab dose and occurred more often in APOE4 carriers than non carriers. The Phase III studies had a number of design elements such as titration, dose levels and management with MRI to mitigate and manage the risk of ARIA. In addition, dosing of aducanumab was stratified by APO E4 carrier status. The low dose was defined as 3 milligram per kilogram for APOE4 carriers and 6 milligram per kilogram for non carriers.
Whereas the high dose was initially defined as 6 milligram per kilogram for APOE4 carriers and 10 milligram kilogram for non 10 milligram per kilogram became available in August 2016. These data showed that the incidence of ARIA as well as discontinuations from treatment appeared to be reduced as compared to the fixed dose cohort. Following this analysis, the protocols for the ongoing phase 3 studies were amended such that the high dose Of note, in each study, approximately 2 thirds of enrolled patients were APOE 4 carriers, and the number of carriers was well balanced across each of the arms such that the maximum number of monthly in the 18 months spend their dosing and then restart and remain at a lower dose. Engage enrolled its 1st patient in August 2015, and IMerge began enrolling approximately 1 month later in September 2015. Two key amendments were made to the protocols during the conduct of the study.
In July of 2016, we amended the protocol to allow patients with ARIA who suspended dosing to resume aducanumab treatment at the originally assigned dose. And then in March of 2017, as I mentioned, The protocol was amended to increase the high dose for APOE4 carriers from 6 to 10 milligram per kilogram after titration. Taken together, these protocol amendments had the effect of allowing more patients to receive their target dose. Since enrollment in ENGAGE were impacted by the protocol amendments and received the 10 milligram per kilogram dosing of aducanumab. On March 21, 2019, we announced the termination of Emerge and Engage following the outcome of a pre specified futility analysis.
The futility analysis was based on efficacy data on all patients in the 2 trials who had enrolled early enough to have had the opportunity to have completed the 18 month study period 2018. Futility criteria, which were pre specified in the statistical analysis plan, would be met if both arms of both trials were predicted to have had less than 20% probability of being positive on the primary endpoint at the end of the study. To predict the future behavior for 2 trials that are identical heterogeneity between the studies. While Engage was not. We did not understand the drivers of these different results.
While we did know that the protocol amendment could have had differential effects data. Following the discontinuation of IMerge and Engage, additional data from these studies became available. Namely the 2066 patients who had the opportunity to complete the full 18 month study period by March 20, as well as all 3285 patients who had enrolled in the trials in what's known as the intent to treat or ITT population. Importantly, once we had applied the pre specified statistical analyses to the larger dataset We were excited to CDRS B scores are obtained through clinician interviews of patients and caregivers and assess 3 domains each of cognition and function namely memory, orientation, judgment and problem solving, community affairs, home and hobbies, and personal care. Specifically, in the intent to treat population, patients from the high dose aducanumab group showed a in clinical decline The corresponding result in the opportunity to complete or OTC dataset was also 23%.
With a nominal P value of 0.03. The low dose group also had less decline on the CDRS B than Placebo. However, the differences were smaller than in the high dose group and did not attain statistical significance. In IMerge, patients on high dose aducanumab also showed a reduction in clinical prespecified secondary endpoints. Specifically, we saw a 15% reduction in clinical decline as measured by MMSC, with a P value of 0.06, a 27% reduction in clinical decline, as measured by ADAS COGS13, with a P value of 0.01 and a 40% reduction in clinical decline as measured by ADCS ADL MCI with a P value of 0.001.
A small numeric advantage for low dose over placebo was observed on all endpoints, except for MMSC. MMC and ADES COGS are measures of cognitive performance in domains relevant to Alzheimer's disease, such as memory, orientation and language. ADCS ADL MCI on the other hand is a caregiver rated assessment of activities of daily living, including conducting personal finances, performing household chores, such as cleaning, shopping and doing laundry, and independently traveling out of the home. In Engage, the primary endpoint was not met, though the low dose showed results similar to the low dose of the merge. Results of the ITT and OTC analyses of the secondary endpoint in Engage showed no statistically significant differences in decline on MMSC, ADAS COGS13, or ADCS ADL NCI compared with Placebo.
However, small numeric advantages for the low dose of a placebo were served on these endpoints. In addition to the clinical data, we observed supporting data from multiple biomarkers Imaging of amyloid plaque deposition in both studies demonstrated that treatment with aducanumab resulted in a dose and time dependent reduction in amyloid plaque burden compared to Placebo with P values less than 0.001. The magnitude of reduction in the high doses impact on CSF biomarkers of tau pathology, a statistically significant reduction on CSF phospho tau levels was observed in IMerge and Engage with a dose proportional response in IMerge. Aducanumab produced a numeric reduction in CSF Total TOW levels in IMerge and Engage, with a dose proportional response in IMerge. Although the primary and secondary endpoints were not met in Engage, in post hoc analyses, the subset of patients who received sufficient exposure to 10 milligram per kilogram aducanumab.
In this case, at least 10 doses of 10 milligram per kilogram showed similar results to the comparable population from IMerge, in terms of both amyloid plaque reduction and reduced clinical decline on CDRS B. The safety and tolerability profile of aducanumab in Immersion Engage was consistent with previous studies of aducanumab with the most commonly reported adverse event being REE and headache. Although the underlying cause of REE is not well understood, it is likely that the MRI signal of REE is due to increased extracellular fluid. The majority of patients with REE did not experience symptoms during the REE episode, and REE episodes generally resolve within 4 to 16 weeks, typically without long term clinical sequelae. We plan to present further details on the new analysis of the larger data set from aducanumab at the CTAD meeting in December, which we plan to webcast.
So to summarize, First, across multiple clinical endpoints, the larger aducanumab data set demonstrated a statistically significant reduction of clinical decline in early Alzheimer's disease patients in IMerge, and we believe that data from a subset of ENGAGE support these findings. 2nd, exposure to high dose aducanumab was important for efficacy and differences in exposure to high dose aducanumab largely explain the different results between futility analysis and the new analysis of this larger data set as well we believe it is reasonable to submit a regulatory filing for aducanumab based on these data. As Michelle and Al mentioned, The path that brought us to today was long and complex. We will be working with our investigators to redose eligible patients who had previously participated in the aducanumab clinical trial, and we look forward to working with the FDA as well as regulators around the world to find a path to make the drug available to patients. Most importantly we envisage a future where physicians may finally have an option to offer patients to help
We believe that these positive results for aducanumab represent a turning point for patients, caregivers, physicians and scientists in the fight against Alzheimer's disease. More broadly, we believe these results represent an inflection point in neuroscience drug development and validate our core strategy. By demonstrating that removal of aggregated forms of amyloid beta can result in improved clinical outcomes We believe these results have positive implications for BAN2401, a distinct antibody that also targets aggregated amyloid beta that we are currently evaluating More generally, we believe these data may have positive implications for additional assets in our portfolio that target the causal pathobiology of neurodegenerative disease particularly those validated by human genetics. These include our tau directed assets for Alzheimer's disease and primary tauopathies, our alpha synuclein antibody for Parkinson's disease and our SOD1 and C9 ORF targeting antisense oligonucleotides for ALS ALS. Given our depth of expertise, our deep and interconnected neuroscience pipeline, including 9 additional readouts expected by theendofnextyear, We believe that Biogen is uniquely positioned to capture the opportunity in neuroscience and potentially deliver a suite of breakthrough therapies for diseases of the nervous system.
With that in mind, let me now review the highlights across the rest of our pipeline in the third quarter. Starting with MS, This quarter, we announced positive top line results from Evolve MS2, a Phase III study of VUMERITY or diroximel fumarate a novel oral the study's pre specified primary endpoint, the individual gastrointestinal symptom and impact scale with a P value of 0.0003. The proportion of patients who discontinued due to GI adverse events during the 5 week treatment period was 0.8% for VUMERITY and 4.8% for TECFIDERA. Of note, this continuation rate for TECFIDERA is similar to that observed in the phase 3 studies of TECFIDERA in which 4% of patients discontinued due to GI events. Earlier this month, Alkermes received a tentative approval from the FDA for BoomerITY we are working with Alkermes and the agency to secure final approval as quickly as possible.
Moving to Tysabri. Previous analysis of real world data from the U. S. Touch registry demonstrated that extended interval dosing of Tysabri reduced the risk of PML by between 78% 99% in this population. However, whether extended interval dosing preserves the efficacy profile of Tysabri remains an open question.
To this end, at last month's Xtrim's meeting, we presented data on the efficacy of extended interval dosing with Tysabri using data from the Tysabri observational program. After propensity score matching the results indicated that there was no significant difference in annualized relapse rate or risk of relapse between the two groups. And we recently completed enrollment for the phase 3b NOVA study of Tysabri, a 2 year randomized prospective trial that will directly compare the effectiveness of every 6 week dosing versus the approved every 4 week dosing regimen after at least 1 year of standard dosing. Also this quarter, the EMA updated the labels of Avinex and PLEGRITY to remove pregnancy contraindications and were clinically needed to allow use during pregnancy and breastfeeding in women with relapsing MS. With these updates, interferon beta therapies are the only MS therapies in Europe that can be considered for use in MS patients throughout the full course of pregnancy.
This is important given that women are diagnosed with MS at least 2 to 3 times more frequently than men and women are often affected during their childbearing years. Turning to our progress in neuromuscular disorders. At last month's Congress of the European pediatric neurological society, we presented new interim data from the open label shine extension study of SPINRAZA in children with later onset SMA. Now including data on patients up to fifteen years of age followed for up to 6 years, this analysis demonstrated that in stark contrast to the natural history of SMA, patients with Type 2 SMA treated with SPINRAZA demonstrated improvements in motor function as assessed by the Hammersmith functional motor scale expanded and patients with Type 3 SMA treated with SPINRAZA demonstrated stable Hammersmith scores and improvement in distance walked. No participants discontinued treatment due to adverse events and no new safety concerns were identified.
Taken together, these data again underscore the robust durable efficacy and well established longer term safety profile of SPINRAZA across a broad range of SMA patients. Also this quarter, we published data from the NURTURE study of SPINRAZA in presymptomatic infants in the journal Neuromuscular Disorders. Data from the NURTURE study showed that all patients treated with SPINRAZA were alive and achieved the ability to sit without support none required permanent ventilation, 92% achieved walking with assistance, and 88% achieved walking independently. Given SPINRAZA's well characterized safety profile, we recently announced the phase twothree devote study to evaluate whether higher doses of SPINRAZA can provide even greater efficacy than the currently approved Our review of PKPD data suggests that individuals with higher CSF concentrations of SPINRAZA achieved greater improvements in CHOP INTEND and motor milestones. As with any therapy that is developed to address high unmet need, companies who lead need to continually explore ways to optimize their treatment.
Building on the success of SPINRAZA, we aim to build a broader neuromuscular franchise including ALS. At the meeting of the northeast ALS consortium held earlier this month, we presented new data on neuro filament levels assessed in the phase 1two multiple ascending dose study of Tafersen our antisense oligonucleotide targeting SOD1 in patients with SOD1 ALS. As a reminder, previous data from this study showed that treatment with 100 milligrams of Tufersen was associated with a statistically significant reduction in CSF SOD1 protein levels and trends towards slowing of clinical decline as assessed by 3 independent measures relative to Placebo. This new analysis of the same population showed that baseline neuro filament levels in both the plasma and CSF were correlated with baseline disease activity and treatment with 100 milligrams of Tafersen reduced levels of neurofilament in both plasma and CSF These data further highlight the concordance across datasets generated in this study, including target engagement, clinical and neurofilament data and thus illustrate the potential for antisense oligonucleotides to target genetic drivers of neurodegenerative disease. In movement disorders, we continue to advance the phase 2 study of BIIB092 or gosuranumab an antibody targeting extracellular tau in PSP with data expected by the end of the year.
In this Phase 2, if this Phase 2 study is positive, we believe we would be in a position to file for regulatory approval. In Parkinson's disease, we continue to progress the phase 2 study of BIIB-fifty four and antibody targeting extracellular alpha synuclein We expect data from the 1 year placebo controlled period of this study, including data on safety as well as neuro imaging based assessment of striatal dopaminergic transporter density in the second half of next year. Also this quarter, as we work to build further depth in movement disorders, we dosed the 1st patient in the phase 1 study of BIIB094 and antisense oligonucleotide targeting Lucine rich repeat kinase 2 or LERC 2 in Parkinson's disease. Toxic gain of function mutations in LERC 2 constitute the most common genetic cause of Parkinson's disease representing approximately 5% of all Parkinson's disease cases. In addition to Lort 2's role in familial Parkinson's disease, Data from the literature suggests that Lort 2 gain of function may also contribute to the pathogenesis of sporadic Parkinson's disease.
As a result, this phase 1 study will include Parkinson's disease patients with or without verified mutations in LORC 2. Importantly, BIIB094 leverages the same RNAH mediated degradation mechanism utilized by Tofersen. With the addition of BIIB094 to our pipeline, we are now advancing ASOs targeting the most common genetic cause of Parkinson's disease 2 genetic causes of ALS and tau pathology, which underpins several primary and secondary tauopathies including Alzheimer's disease. And we aim to build further depth across on genetically validated targets and defined patient populations. In acute neurology, we continue to advance the Phase III study of BIIB093 or IV glabanklamide for cerebral edema caused by large hemispheric infarction or LHI.
As a reminder, BID-ninety three blocks SER-one TRIPM4 channels that are hypothesized to mediate brain edema following LHI. Given that these channels are also hypothesized to mediate expansion of hematoma and peri hematoma edema associated with brain contusion, This month, we dosed the 1st patient in a phase 2 study of BIIB093 for brain contusion. Approximately 280,000 patients are hospitalized due to head trauma annually in the United States. And we estimate that contusions occur in approximately 25% to 35% of these patients. There are no pharmaceutical agents approved to mitigate contusion expansion, which is associated with worsened clinical outcomes.
The primary objective of this new expansion in contusion volume over the course of a 96 hour infusion of BIIB 93 versus Placebo. Importantly, we believe that the shared pathophysiologic features of LHI and brain contusion exemplify the interconnectivity of neuroscience that we are leveraging as to pursue multiple indications for a given asset, particularly once we believe safety has been adequately established. And finally, in our Alzheimer's disease portfolio, this quarter, we completed enrollment in the phase 2 study of Gosuranumab in early Alzheimer's disease. And we continue to advance the phase 1 studies of BIIB076, a distinct anti tau antibody and VIB080 and ASO aimed at reducing the expression of tau in the central nervous system. Taken together with the positive results of aducanumab, we believe that no other company is better positioned to deliver breakthrough therapies for Alzheimer's disease.
Overall, this was a historic quarter for Biogen and for the patients, caregivers, physicians and scientists around the world who have been waiting decades for a therapy that can reduce the clinical decline of Alzheimer's disease. With that, I will turn the call over to Jeff.
Thanks, Al. Good morning, everyone. I'll now review our financial performance for the third quarter of 2019. As Michelle mentioned earlier, we had a strong financial performance in Q3 2019. Total revenues for the 3rd quarter grew 5% year over year, approximately $3,600,000,000, while GAAP earnings per share increased 17% and non GAAP earnings per share increased 24%.
Both compared to the prior year. Overall, our MS business delivered revenues of approximately $2,300,000,000 in the third quarter of this year, including OCREVUS royalties of approximately $188,000,000, growing 2% versus the prior year. Global MS revenues in Q3 2019 were stable versus the prior year without OCREVUS royalties and the total number of patients on our MS products globally continue to grow in the low single digits versus the prior year. U. S.
MS revenues in Q3 2019 were impacted by a decrease in channel inventory of approximately $30,000,000 compared to a decrease of approximately $5,000,000 in Q3 2018, and a decrease of approximately $30,000,000 in Q2 2019. Global third quarter TECFIDERA revenues increased 3% versus prior year. As TECFIDERA delivered strong global patient growth of approximately 8% year over year. In the U. S, Techvedere revenues were flat year over year as Techvedere share of total prescriptions remained relatively stable compared to the last couple of quarters.
Outside the U. S, TECFIDER performed very well again in Q3 2019 with continued volume increases across all large European markets and Japan versus the prior year, somewhat offset by pricing pressure in several European countries. Q3 global interference revenues, including both Avonix and Plegrady, decreased 10% versus Q3 2018, due to continued shift from the injectable platforms to oral or high efficacy therapies. SABRI worldwide revenues increased 3% versus the 3rd quarter of 2018. We were pleased with this growth into Sabri revenues as it continued to perform well in the high efficacy segment.
In the U. S, we grew to Saudi revenues 4% year over year, as we continue to maintain stable to Sabre share of total prescriptions compared to last couple of quarters. Outside the U. S, we grew to Sabre 2% driven by volume growth in several markets and favorable timing of shipments, somewhat offset by pricing pressures. Overall, we were pleased with the execution of our MS franchise and the continued strong performance of our MS business in the third quarter.
We remain focused on maintaining resilience and MS market leadership. Let me now move on to SPINRAZA. Global 3rd quarter spin revenue revenues increased 17% versus prior year and 12% versus the prior quarter to $547,000,000. In the third quarter, SPINRAZA achieved year over year and quarter over revenue growth both in the U. S.
And outside the U. S. Driven by continued patient growth across both mature and new markets and we now have approximately 9300 patients on SPINRAZA including the expanded q22 2019. The number of patients on therapy in the U. S.
Increased 3% as compared to the end of the second quarter of 2019. And importantly, we continue to make strong progress with adults, the largest patient segment, with the number of adults on SPINRAZA growing 8% compared to the prior quarter, which is the 3rd quarter in a row with upper single digit growth. As a reminder, Zolgensma is competing in a limited portion of the market, specifically the approximately 5% of SMA patients were under two years old. Within that segment, we have begun to see some impact on SPINRAZA performance. Outside the U.
S, revenues increased 27% versus Q2 2018 21% versus last quarter driven by strong performance in established markets such as the EU and Japan, as well as key markets in both Latin America and Asia Pacific. We were pleased Additionally, we recently dosed the 1st SPINRAZA patients in China. Overall, we were pleased to see continued patient growth across the larger mature markets and continued rapid uptake for more recently launched markets. Given our expected continued patient growth, and execution across many global markets, our established product profile and this significant market opportunity we remain optimistic about our SMA business. In our biosimilars business, we generated $184,000,000 this quarter growing 36% versus prior year, driven by IMRALDI.
We estimate that we now have approximately 180,000 patients on biosimilars in Europe. Total anti CD20 revenues in the 3rd quarter increased 16% versus the prior year, primarily driven by OCREVUS Royalties. Q3 OCREVUS royalties benefited by approximately $10,000,000 due to an adjustment related to prior periods. We continue to expect Rituxan revenues to be impacted by the entry 26% versus the prior year, driven by the decline in our manufacturing services revenues due to our divestiture of the Hilla Rock plant. Importantly, we continue to see geographic diversification of our revenue base, driven by growth in MS revenues outside the U S, the continued market expansion of SPINRAZA and our growing biosimilars business.
In the third quarter, approximately 41% of our product revenue from outside the U. S, versus approximately 37% in Q3 2018 and 32% in Q3 2017. We aim to continue capitalizing in global growth opportunities for both our current commercial portfolio and our pipeline of products. Let me now turn to the expense lines on the P and L. Q3 2019 gross margin was 88%, an improvement versus both prior year and prior quarter which were both 87% due to favorable cost of goods sold, product mix and higher margin contract manufacturing.
Q3 GAAP and non GAAP R and D expense were both 15% of revenue. R and D expense included approximately $58,000,000 in trial closure costs, for both elobescencestat and VG11. Q3 GAAP and non GAAP SG and A expense were both 15% of revenue. Q3 GAAP other expense was $27,000,000 and non GAAP other expense was $23,000,000. In the 3rd quarter, our GAAP tax rate was approximately 12 percent and our non GAAP tax rate was approximately 16%.
Compared to Q3 2018, our Q3 2019 GAAP tax rate benefited from the remaining amount realized from U. S. Corporate tax reform, a change in our tax profile in Q2 2019, and recently enacted tax reform in Switzerland. In the third quarter, we repurchased approximately 3,100,000 shares at an average price of $2.33 for total value of approximately $718,000,000. As of September 30, 2019, we had approximately $3,400,000,000 remaining under 20 19 share repurchase authorization, which now brings us to our diluting earnings per share.
In the third quarter, we booked GAAP EPS of $8.39, an increase of 17% versus the prior year, and non GAAP earnings per share of $9.17. A 24% increase versus the prior year. We generated approximately $1,700,000,000 of net cash flows from operations in Q3. We ended the quarter with approximately $6,300,000,000 in cash and marketable securities and $6,000,000,000 in debt. I'll now turn the call back over to Michelle for his closing comments.
Thank you, Geoff. To summarize, first, the positive clinical results for aducanumab positioned Biogen to potentially lead the fight against Alzheimer's disease. 2nd, this data validates Biogen's strategy to focus on an interconnected neuroscience pipeline and prioritization of targets supported by human genetics. 3rd, our base business continued to deliver solid performance in Q3 2019, driven by strong execution against our strategic priorities. Between now and the end of 2020, we expect continued progress as we aim to build a multi franchise portfolio, including 9 additional mid to late stage data readouts the expected launch of VUMERITY in the U.
S, and submitting the regulatory filing for aducanumab in the U. S, while continuing dialogue with regulatory authorities in international markets, including in Europe and in Japan. I am proud of the Biogen team for not being deterred by a history of disappointment in the pursuit of Alzheimer's therapies and more so for continuing their work of analyzing the clinical trial data with unprecedented focus and intensity even in light of an apparent futility results. This work reflects Biogen's steadfast examination to follow the science tackled the biggest challenges and do always the right thing for the patients. Finally, what is most important today is that in consultation with the FDA, we are excited to be moving ahead and preparing for regulatory filing for aducanumab on the ground of positive clinical from our Alzheimer's trials as quickly as possible.
This is a major step in the fight against Alzheimer's disease and an important inflection point for Biogen's neuroscience mission. We believe now more than ever that our core focus on neuroscience will enable us to maximize as well as for our shareholders. As the leader in neuroscience, we believe that no other company is better positioned to continue to deliver breakthrough therapies for diseases to build a multi franchise portfolio across our core and emerging growth areas we are inspired by the progress we have made in tackling Alzheimer's disease and the broader scientific implications of the positive clinical results for aducanumab. I would like to thank all the Biogen employees in particular those who have been working tirelessly on the aducanumab program and the many more who will contribute to this critical priority over time. I am incredibly grateful for all the patients, physicians and caregivers who have dedicated so much time and effort to our Alzheimer's clinical studies and advancing our understanding of this very complex disease.
I would like to thank the FDA for their We will now
Your first question comes from the line of Omar Rafatt from Evercore ISI. Please go ahead.
If I may, I only have a question on aducanumab, but it's got 3 parts. And given the significance of the news today, I would really appreciate if you could bear with us on it. So my my three parts are as follows. First, I'm and I'm not to correlate the 2, but Michael Ehlers is departure ahead of this data announcement. Just wanted to hear if those two things are related in any way or not.
Second, the implication in the data is that the high within sufficient exposure at the high dose, the second trial worked as well. But when we look at CDR some of the boxes, low dose actually looks more consistent than the high dose. And also for patients that did not make it to the large opportunity to complete data set. Those patients actually, especially in MMSC, more consistent than the patient that did have a sufficient exposure. So I guess I'm just trying to understand, how spot on is that finding on patients that had a sufficient exposure and those are the ones that drove efficacy.
Thank you so much.
Thank you, Umer. This is Michel. Mike decided to leave the company on his own. And, I cannot think thank him enough for his many contribution over the past three and a half years to Biogen. So thanks, Mike.
And at the same time, I'm extremely confident in Al's leadership as a clinician too, as a scientist, to take the helm at the time where the R and D portfolio never been as stronger. The team also and the capabilities.
Umer, this is Al Sandrock and I'll turn it over to Samantha. Later for, follow-up. But, look, your point is well taken. The low dose is consistent across Engage and IMerge. And that's because the, particularly the 2nd protocol amendment really affected the high dose arm of in the carriers.
So in the low dose arm, the protocol amendments had less of an effect. And I think that's one of the main reasons for the consistency in the results across the two studies. I'll turn it over to Samantha for follow-up.
Yes, that's correct. You'll also see that in the high dose for ENGAGE that we do have a partial response on ADAS COGS13 and the ADCS ADL MCI And so the potential, rate that you have there on CDR summer boxes and MMSCs that these are potentially less sensitive as endpoints So what Al said is that in that high dose group, we know that we have less doses than we had at the high dose than we had in IMerge And we also know that the studies were, to some degree, impacted by dose suspensions due to ARIA. So dosing is a complex, combination of duration, magnitude and no interruptions.
Thank you very much.
Your next question comes from the line of Phil Nadeau from Cowen and Company. Please go ahead.
Good morning. Thanks for taking my question. It's also, as you might imagine, on aducanumab. I guess in 2 parts. First, if you pulled the data from Engage and IMerge, would the pooled results still be positive on the primary endpoint?
And, kind of related to that, could you give us some sense of what the differences between Engage and IMerge were in exposures at those high doses. It seems like the trials didn't start on that far part. They're just a month. So, kind of when in response to the last question, you mentioned exposures rely on dose suspensions and whatnot. Can you give us some sense quantitatively of how different the patient populations at that high dose were an engaging emerge in terms of their exposure?
Thank you.
Certainly. So the first part that if we pulled the outcomes on engage and emerge at the high dose, essentially you'll get an intermediate effect, not much more complicated than that. But we are looking at these as stand alone studies as 2 independently identically designed studies. The second part of the question, which I've forgotten actually.
Difference is an engaged number?
Difference? Yes. So they started 1 month difference between the two studies, as I mentioned, and they remained different throughout the entirety of the studies in that initial 1 month at certain periods of the studies and particularly through the protocol amendments, was greater in the middle of the studies And more details around this will come at the presentation in CTAD.
Your next question comes from the line of Terence Flynn from Goldman Sachs. Please go ahead.
Hi, thanks for taking the question. Maybe two parts for me as well. I'm just wondering if you can share any additional commentary on the 2nd Type C meeting did FDA agree that a single positive trial could be sufficient for approval? Or is that likely a review question? And then can you give us the rates of Aria in the high dose arms of the 2 trials?
Thanks.
Hi, this is Al. It's generally our policy not to comment too much on the content of regulatory interactions. I will say though that, they thought it was reasonable for us to submit an application to, for approval. So that's the main, that was the upshot of the meeting.
And the second part of the question was on aria and high dose. And that was consistent in incidents, per the studies that we have previously reported. And we'll get more details on that at CTAD.
Your next question comes from the line of Jeff Porchas from Leerink. Please go ahead.
Thank you very much. And thanks for having Samantha on the call. It's very helpful. First, could you answer whether that the analysis that you presented and presented to the agency has been independently verified, what confirmation of both the statistics and the results do you have? Secondly, do you have any intention or plans for a confirmatory, a pivotal trial to supplement these 2 trials.
And then I hate to sort of push on the issue of the type, the FDA meeting, but Did the FDA see the full analysis or did the FDA just hear the company summary of the analysis? Thanks.
Let me start with the last question and I'll let Samantha answer the first couple. First of all, the FDA did see the full analysis of both studies. And I would also say that the only study we have planned right now is the redosing study. And any further study, we'll update you as soon as we plan 1.
Thanks, Al. Going to your first question, Geoff, have we had independent review. As we mentioned, one of the first steps that we undertook was to engage external advisors to help us review this data, and that did include independent statistical experts. But the data that we did take to the FDA as Al says was a full data set, which was an analysis of the blinded data, conducted in using the same statistical analysis plan at had originally planned for the end of the study. And the validity of the data set was the first thing that we analyzed together with the FDA.
To your second question on whether we are conducting another study, as we've mentioned, our next steps are twofold. One is the FDA indicated to us that it is reasonable for us to file these 2 studies and for us to go ahead and put together a redosing study for the patients who were in previously enrolled studies of aducanumab.
Your next question comes from the line of Michael Yee from Jefferies. Please go ahead.
Thanks. Thanks for the question. Appreciate it. Al or Samantha, I guess I just wanted to stand engage a little bit more specifically in the high dose you appropriately say that there was a slightly negative trend overall in the high dose but in the subgroup of, exposure patients, which is, I think about a third of it, they had a nice benefit. I guess the question is, how do you think about the patients you did not have enough exposure?
Did they drive a strong negative trend? Are those patients at harm? How do you think about that since that's a huge majority of the patients and how is that explainable given the difference in IMerge?
Thanks, Michael. I'll start. Look, I don't think, first of all, the there's a slight negative. I would say that that was just basically no effect. And then in the those who did not have the high dose I would not say that they had a negative effect.
In fact, in many ways, there was either a neutral or a positive effect, but I would point out this. You remember the 6 milligram per kilogram dose arm in prime that everybody was wondering about you always ask me questions about it. I remember, Michael. And we thought that was an outlier. Well, maybe that wasn't the outlier.
Maybe that was true and that the 3 milligram per kilogram that looked like it was trending was the outlier. So in other words, what I'm saying is that there's a very sort of sharp dose response, if you will. You have to get to high dose aducanumab and intermediate dosing at least in an 18 month trial is not enough.
Your next question comes from the line of Cory Kasimov from JP Morgan. Please go ahead.
Hey, good morning guys. Thanks for taking my question. I guess first just to ask Phil's question more directly, can you tell us yet how many patients got the 14 doses of of 10 mgs per kg in each study in the full data set. And then as a follow-up, did you see any difference in APO E4 carriers versus non carriers especially in the patients who completed after the futility cohort and would have had more exposure to the higher dose? Thanks.
Yes. So the first thing I want to mention is, in terms of the numbers of subjects who had the particular data that you're referring to more dose, more than 10 doses of 10 milligram per kilogram, there's more than a 10% difference between the two studies, but that's not the only parameter of difference that is important for And so that's a more complex calculation between the two studies.
2nd April versus
Oh, so your question regarding APOE carriers versus non carriers, the analyses that we've conducted today have been on the entire studies. And as we've mentioned for IMerge, we have a positive, we met the primary endpoint for the entire patient population and details of subgroups is something that we'll come to later, and we'll have details at CTAN.
Okay. Thank you.
Your next question comes from the line of Geoff Meacham from Bank of America. Please go ahead.
Hey guys, thanks for the question and all the detail on edutinumab. Al, just have a couple of regulatory type of questions all related. If half of a merge achieves significance at the high dose and of engaged achieved it. Is it you guys expectations that the prime study could count as one of the 2 pivotals? Second one is, does a conditional approval pending another successful phase 3?
Did that come up in the FDA discussion? And then third, Have you had any discussion with European regulators on the data? Thank you.
Let me start with the last question first. So we have just started to contact the European regulators and we haven't had any substantive discussion as of yet. And, in terms of the merchant engaged, We looked at Engage in totality as a positive study that stands on its own. And, remember, as Samantha said, we use pre specified primary and secondary outcomes. We didn't look at a subset.
We looked at all the patients And based on that, we believe the study met its primary endpoint and the secondary endpoints as well. I think that whether or not a single trial, can be approved. There are circumstances where an FDA can approve a drug based on a single study. It's up to them to determine what those circumstances are. And I'll just leave it at that.
And then I would say that Engage, we believe we showed the data, for example, in those who achieved sufficient exposure to 10 milligrams per kilogram, we do see evidence of efficacy. So I would say that emerge stands on its own, Engage has supportive evidence. And I would also say that prime is supportive. It's a well controlled you know, phase 1b, some might call it phase 2 trial. And we'll submit all the data.
Just to add there, Al, emerge is the study that primary. I get it wrong.
Yeah. I get it confused sometimes.
Your next question comes from the line of Brian Abrahams from RBC Capital Markets. Please go ahead.
The question, just a little bit more clarity on the Engage study. For the subgroups of patients, I guess for both Engage and IMerge with 10 interrupted receiving 10 interrupted, uninterrupted high doses. Can you talk about the baseline characteristics for the aducanumab versus placebo arms across both studies and how well balanced those were. And then can you maybe help us understand how feasible was it for patients? I guess once the protocol was protocols were amended, to remain on 10 uninterrupted doses, or is the lesson here that if you do need to temporarily discontinue for ARIA or whatever reason, you're probably best off not restarting the drug.
Thanks.
So, thanks. Thank you for the question. I want to point out that the analyses that we conducted in close consultation with the FDA around determining who in ENGAGE did have a response were exploratory analyses. And any time that you look at a subset of patients, you have very important questions in regards to whether they are balanced for all baseline characteristics and, as I mentioned, APOE 4 status. But those subgroups are exploratory in nature, and they help us understand that dosing is important for efficacy.
And in the context of an 18 month trial, one does tend to see that you need a certain number of doses for clinical benefit of aducanumab. However, that's not the same as one would anticipate in a real world situation where an individual is taking aducanumab for an extended period of time where a dose interruption would likely be of less significance.
Yes, I agree with Samantha. I think that, dose suspension in the context of an 18 month study was could be problematic because they didn't achieve enough of the high dose. But in clinical practice, we don't do 18 month treatment periods. We're going to treat patients for longer periods of time. And in that situation, I think dose suspension may be acceptable in some patients.
Your next question comes from the line of Matthew Harrison from Morgan Stanley. Please go ahead.
Great. Thanks for taking the question. I guess a follow-up and a sort of second question for me. So first one is, you've been talking about exposure and dose a lot. Could you just broadly comment on how many of these patients actually achieved all of the factors that you were looking for and how easily you think that will be the case in clinical practice And I guess the related question of that is this dose exposure curve that you're sort of talking about Al.
I mean, were there characteristics that were different where the kinetics the amyloid plaque reduction different, in the subgroup of patients with the achievement of tau or amyloid reductions were they significantly different? I'm wondering what you think is sort of biologically happened to account for this steep dose exposure curve that you're discussing?
These are good questions, Matthew. And we're still learning, as we look at the data, but I would say this, even an MCI patient, if you look at the amount of amyloid in the brain, it's tremendous. It took 20 years to build that much up. And in the context of an 18 month trial, you have to remove a large amount of amyloid. I think that's what distinguishes Tucanumab and BAN2401 is that we can it's safe enough to achieve the doses that allow us to remove a large amount of amyloid And if you don't remove a large effect.
We saw that in prime, for example, where we did have some amyloid lowering at 6 months, but we saw no difference and the clinical outcomes at 6 months. It was it took the 12 month time period to see starts to see an effect on clinical outcomes. So in addition to a large amount of amyloid being useful. I think you need to have a little bit of time for that biological activity to have an effect on clinical outcomes. That's what we see.
And I would say that if you look at the amyloid pet results that was on one of the slides, and those who had more than 10 doses of 10 milligrams, you can see that the SUVR score is very similar in Engage in that subgroup of patients and engage to the emerge total data set. So, and so again, what it says is that if you give enough of the high dose, you can achieve a certain amount of amyloid removal and that certain amount is what's required to see the reduction in clinical decline in an 18 month study.
Yes. Al, just to add to that on the question of numbers, on the graph that you've just referred to, you've got the N numbers. So they were 147 for IMerge and 116 for ENGAGE. In that CDR summer boxes analysis. But the question you ask of how many patients have the precise criteria where there aren't precise criteria, dose response is not binary.
And so given the levels dose, you'll have a different response. And it's a bit of a sliding scale. So we have that exploratory analysis that we disclosed to explained what it is we learned around the importance of dose, but there's no perfect number of doses that are required. It's not binary.
Your next question comes from the line of Ronny Gal from Bernstein. Please go ahead.
Taking the question. And I'm going to stay with aducanumab here. I'm just kind of struggling with the movement from the interim futility analysis to efficacy with relatively small number patient, just looking at the number of completed that you have here and then in IMerge, you moved from 803 patient, the futility announced to 980 patients in the treatment. So it's about 180 more patients. If we assume a third of those were on the high dose, 60 more, in your total number of folks that you have amyloid beta that you called got sufficient exposure is at the end of the trial, 127.
I kind of wonder if there's just a very small number of patients that drove the entire movement. If you can discuss a little bit that issue of, how many patients actually contribute to the difference between stopping the trial for futility and showing efficacy. Would be appreciated. And then I'm going to, if you don't mind, I'm going to throw my second one in and, and, it will be different, not another kind of objective for the variability. And do you have any way to protecting the highest dose Tysabri from biosimilars to the 1st generation products?
If you can discuss that at all, we appreciate it. Thanks.
Ronnie, this is Al. My head is swimming, even just with the first question, but So, I think first of all, you should remember that in a merge, even at the time of the futility analysis, That study was trending positive as Samantha said. And then we add those additional patients, and it didn't take that many now to then in the April data set, to see that it had met its primary endpoint. And then I would also say that we also looked at the patient who had not completed 18 months, all the rest of the patients, which is roughly half the patients, because we only looked at the first half, the first half of the enrollees for the futility. So it's a large number of patients that we ended up looking at.
And I remind you that result that you saw on that slide was all the patients in IMerge who had been randomized, the ITC populations and it was using the pre specified primary and secondary endpoints. And then I now forget the second question.
So before you jump into that, if you look at the slide that you had slide 22, the number of patient aducanumab that you have there is the number of patients received enough dose. The question some of my peers was how many patients had exposed? Are those the numbers that we're seeing on aducanumab on Slide 22 that are the numbers we should be thinking about?
So I just want to recap that Slide 22 was a piece of exploratory analysis. It is not the subset that we believe are supported. It's just a particular analysis to emphasize the point that there are subjects in Engage whom if they do have sufficient dosing do support the outcome of IMerge?
I would also say, Ronny, that the pet was done in a subset of patients too. So the numbers that you see on the left side, which is the amyloid bet, are from only the subset who got the pet imaging.
And the numbers on the right, I'm still not the complete set. This is just some sort of a subset.
That's correct.
I think the second question was around, high dose Tysabri?
So what we would say to that, Ryan, it's Jeff, is obviously, we'll wait and see kind of what happens with regard to the Phase III trials that are going on with regard to biosimilars. And we're supportive of biosurms coming into the market. We obviously have a biosimilar business. We'll just have to see how their products do and we'll deal with it when it comes.
All right. Probably have time for about 2 more questions.
Thank you. Your next question comes from the line of Jay Olson from Oppenheimer. Please go ahead.
Hi. First off, I want to congratulate you for hanging in there and delivering these aducanumab results today. This is very promising news for Alzheimer's patients and their families. And second of all, I want to thank you for taking my questions. Can you comment on the clinical meaningfulness of aducanumab efficacy profile?
And how does it line up with your target product profile in terms of improvements in cognition and function? Are there any gaps in the profile as you know it now? And how do you know if you've optimized the efficacy at the 10 mg per kg dose or would it make sense to test higher doses? Thank you.
Jay, I'll start and then Samantha will follow. We believe it is clinically meaningful. We've heard that anything above 20% is clinically meaningful as a neurologist. Being the 1st drug of its of its kind. We have no drugs right now that affect the clinical decline in Alzheimer's disease.
This would be the very first. So anything north of 20%, we believe clinically meaningful. And I would also add that, in clinical practice, I think that MCI patients will be, if approved, they'll enjoy the benefit of living a more independent life for longer periods of time. If you look at that AD, the activities of daily living, it's a 40% effect. And that's a caregiver assessment of whether or not the patients can live independently, can do their household chores, etcetera.
So that's all very clinically meaningful results.
There's a question around the dose, the 10 mgs.
Yes. So the question of whether we had achieved the correct dose, I think what we have learned clearly is that dose is very important, but that if individuals do receive 10 milligram per kilogram, then they do have an efficacious response I think the trials unfortunately were hampered by a number of operational and other implications that may that not enough patients got 10 milligram per kilogram. So we do believe that to be the correct dose.
Your last question comes from the line of Paul Matteis from Stifel. Please go ahead.
Great. Thanks for fitting me in. Really appreciate it. Within the high dose arm in the Engage study, can you talk about the magnitude of plaque reductions you observed in patients who titrated all the way up to the highest dose versus patients who are stocked at 6 mgs per kg. And I guess does the differential magnitude of plaque reduction in those patients that I'll tell same narrative you're seeing on the difference in clinical outcomes?
And then can you tell us anything else about other measures of function in the Engage subset of patients who titrate it all the way up to 10 mgs per kg? Thanks so much.
Thank you. So your first question in amyloid plaque reduction We do believe high dose is showing a lower reduction than in IMerge. And we do believe that that is a clear reflection of the lower doses that were achieved in that high dosing group in Engage. And the second question was
Other measures of function?
Yes, so the exploratory analysis that, we've demonstrated for you, we focused on the primary endpoint, and we do not have the same analyses for the functional endpoints, but you do have those results for the overall study where even in Engage, we do have some response on the functional scores albeit not statistically significant.
You. I'm going to turn it back to Michelle for some closing comments.
So thank you all for attending our Q3 call characterized by the go to file decision for decanumab with the U. S. FDA. But also with a solid performance for the for the patients first, but also for the shareholders. Have a good day.
Thank you. Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.