[Call Started Abruptly] Analyst Call on the GSK Fourth Quarter 2018 Results. During the presentation your lines will remain on listen only [Operator Instructions]. I will now hand you over to Sarah Elton-Farr, Head of Investor Relations, who will introduce today's session.

Thank you. Good morning and good afternoon. Thank you for joining us for our full year 2018 results, which were issued earlier today. You should have received our press release and can view the presentation on GSK's Web site. For those not able to view the webcast, slides that accompany today's call are located on the Investors section of the GSK Web site. Before we begin, please refer to Slide 2 of our presentation for our cautionary statements. Our speakers today are Chief Executive Officer, Emma Walmsley; Simon Dingemans, Chief Financial Officer; and Dr. Hal Barron, Chief Scientific Officer and President of R&D. We have a broader team available for Q&A. We request that you ask only a maximum of two questions so that everyone has a chance to participate. Our presentation will last for approximately 45 minutes, slightly longer than usual to get Hal time to update you on our R&D progress. And with that, I'll hand the call over to Emma.

Thank you, Seth. Before we take you through our 2018 achievements, as this is the last quarter that Simon is going to representing GSK, I would really like to take the opportunity to reiterate my sincere thanks and appreciation to him for all he has done for the company over the last eight years. Our CFO designate, Iain Mackay, is also on the call today just in listening mode at this time. And he will be enrolled in April, so you will hear more from him on our Q1 call in May. I'm absolutely delighted to welcome him to our team. In 2018, we have made good progress across the group with improvements in sales, the group operating margin, earnings per share and cash flow. Group sales growth of 5% in CER terms reflected an increase in sales in all three of our global businesses, with the particularly performance in vaccines. The pharma business continues to shift its portfolio shape with excellent new launch growth and although consumer had a slower quarter, we remain confident and excited about the outlook for this business. Group operating margins this year were up 50 basis points on a CER basis. On a total basis, earnings per share more than doubled to 73.7 pence and adjusted earnings per share were up 12% CER. Our free cash flow position continues to improve, and we are particularly pleased with the underlying improvement in our cash flow. For the year, free cash flow was almost £5.7 billion, up 63% in actual terms versus last year. Today, we declared a dividend in respect of the fourth quarter of £0.23, resulting in a total dividend for 2018 of £0.80. When I became CEO of GSK in 2017, I laid out my three long-term priorities for the company: innovation, performance…

Thank you, Emma. I’m delighted to be presenting to you such a strong set of results, my last as CFO after 32 quarters. I'm sure you all have saved some particularly challenging questions to round everything off, so I'm looking forward to answering those later on. Overall, the group's results for the year ahead of the top end of our guidance and demonstrate continued operational execution of our key strategic objectives with strong performance in all three businesses. Our earnings release provides an extensive amount of information, so our focus on major points, our expectations for 2019 and important comparatives to make for your models. As usual, my comments today will be on a constant currency basis except where I specify otherwise, and I will cover both total and adjusted results. Starting with the headline results, group sales up 5% to 30.8 billion, total EPS more than doubled to £73.7 and adjusted EPS were up 12% to £219.4. Total operating profit was 5.5 billion, up 43% and showed strong progression on 2017. Higher charges for the revaluation of acquisition related liabilities, principally the BCCL were more than offset by stronger operating performance, lower restructuring costs, lower asset impairment charges and a favorable comparison with the charges taken in 2017 related to U.S tax reform with 0.7 billion. Adjusted operating profit grew at 6% with operating margin of 50 basis points, driven by margin growth in vaccines and consumer healthcare. Pharmacy's gross operating profit was flat with operating margin impacted by continued investments in our new products and a weaker gross margin in the face of ongoing pricing pressures. Free cash flow delivery was significantly stronger at 5.7 billion, up 2.2 billion, reflecting continued focus on cash conversion throughout the group with particular progress this year on working capital management. We…

Thank you, Simon. Q2 I set out our new R&D approach based on science, times technology, times culture. We made a commitment at that time to be much more transparent with you about the decisions we are taking and the progress we are making for regular update. This is the first of those updates and I'm very pleased with the advances we made to the portfolio in the last six months. I believe our pipeline is now more focused on our most promising assets, allowing us to accelerate them while terminate those, which have less potential. Eight assets have made encouraging progress, which I will describe in a moment. Overall, we have significantly strengthened our oncology portfolio. Since Q2, we have added three new internally generated assets to the portfolio and through business development added five, four from TESARO acquisition and one from the strategic alliance with Merck to jointly develop and commercialize M7824, resulting in a doubling of the size of our oncology clinical stage portfolio from 8 to 16. On culture we have made a number of key leadership appointments, most recently with Chris Corsica joining us from Boehringer Ingelheim to head up our newly created development organization, as well as introducing a new more robust governance model, which has about October 1st is up and running and I think going very well. In addition, we are in the process of redesigning the discovery performance units and we will establish a much smaller number of research units aligned with our focus on immunology and genetically validated targets. This is all helping us to create a culture of smart decision-making, single point of accountability and importantly focus. Given the limited amount of time we have today, I’m going to focus on our pipeline and defer taking about the progress…

Thank you very much, Hal. So in summary, 2018 has been a year of significant progress in terms of operational performance, reshaping of the portfolio and developments of the pipeline. In 2019, we will be building on this with continued execution of our priorities of innovation, performance and trust with an ongoing focus on the pipeline to provide a clear pathway to the creation of two exceptional businesses. And across GSK worldwide, we are all very committed to this tremendous opportunity to create substantial value for shareholders, patients and consumers. So with that, operator, I think a bit ahead of our 45 minute promise, the team is ready to take questions now.

Thank you. Your first question comes from Richard Parkes from Deutsche Bank. Please go ahead.

The first one is just for, Hal, on the M7824 deal. Obviously, since the disappointment with Zydol, it seems like the industry has moved away from more, time to move away from making big decisions based on signals in single trials with the next generation I-O agents. Obviously, with M7824, it looks like Glaxo is not necessarily following that trend given that's being advanced into pivotal studies. I just wondered if you could help us understand. You've obviously seen a lot more data what makes you confident to do that. Or is this just a calculated risk? The second question is on the outlook for HIV. When I look at NBRx volumes in the U.S., it look like those declined on an absolute basis by 20% to 30% since the launch of Biktarvy. But obviously the drug is only just launching now in Europe. I wondered if you could help us understand whether we should expect more or less impact to European new patients share of Biktarvy launches. Thank you.

So we will come to, Hal, first and then to David. And just to say aside from the scientific commentary that Hal will add. Just a reminder the construction of this deal is very heavily stage gated for GSK, but will be based on data. But Hal perhaps you'd like to comment on that questions further.

Let me walk you through why we’re pretty convinced this is a smart risk to take. When you look at what advances have been made in cancer therapeutics by inhibiting PD1 and PDL1, it's really been very transformational. But it's important to remember that the vast majority, actually about 75% of the patients either don't benefit or would relapse after therapy. And so there is a clear need to find new agents, either given in combination with or that can compete with these agents to provide patients with greater options. When you look at the facet, it's very unique. It not only combines the IgG backbone of a PDL1 inhibitor, but has this other component that allows it to be basically a receptor trap to bind the three e isoforms of the TGF beta ligands. And the preclinical data is pretty compelling here in terms of TGF data playing a role in tumor progression and particularly want PD1, PDL1 resistance at the tumor level, because of the suppressive effects that TGF beta has been seen in the tumor microenvironment. And so we think there is a very unique first in class novel mechanism agent, was very, very interesting. Now, you combine that with the very unique situation, which is that they had for us to review almost 700 patients treated in various phase I setting were for signal seeking. And an in fact in that finding, four different diseases that I mentioned earlier where there is clear evidence of activity; so based on that preclinical biology and the data generated in those four diseases and particularly the data generated in the second line lung setting where the response rates really did appear to be superior to those historically seen with PD1 and similar PD1 inhibition in some of those settings. So you are right that we don’t have randomized control trials. But if that data that that was exciting enough to initiate the phase II randomized controlled trial against pembro and as Emma said, the deal is structured in a way that gives us confidence that this was a smart risk to take, and hopefully for patients just one that being the superior therapy.

Thanks very much Hal. And so David, some specific questions around the dynamics of NBRx. I mean, I would just also like to repeat that we do expect of these business, our HIV business, to continue to be a key growth driver for GSK, primarily because of the bet that we are making on two drug regimens. And we are very excited about all the -- hopefully, both approvals and the further data that’s going to come through this year. So David, over to you.

Yes, thanks Richard. As Emma and Simon said, we do expect it to be a meaningful growth driver in 2019 and going forward. Although, the dynamics of that growth I think will vary a little bit across the world. So in the U.S., the Tivicay and Triumeq business is basically flat over the last few months. We are getting some penetration with Tivicay, particularly into new patients. But there is some switching of both products. So overall, we are broadly flat at around 35,000, 36,000 scripts a week and our share is roughly just over 27, or around £0.27 of the core in SDR market. And NBRx is also pretty flat. So going forward in the U.S., the growth will come from ongoing momentum with Juluca. And we've actually seen a pretty decent pickup in Juluca in the last quarter, in part I think driven by quite a favorable reaction to the two year SWORD data that was presented. so Juluca beginning to build quite strong momentum. And of course importantly, the launch once we get approval of dolutegravir and lamivudine, which we see as the key growth driver going forward. I think away from the U.S. and the European and international businesses, we expect the growth actually has been more broad based across the whole dolutegravir portfolio, really building on the momentum we have seen this year. So the European business, dolutegravir was up 17% and growing very nicely and international business up 35% very strong performances in places like Brazil, Japan and so forth. And also just remember that this year 2018 we had a reasonable drag predominantly from the genericization ongoing genericization of Kivexa/Epzicom, it's about £150 million and that will be less going forward.

The next question comes from the line of Tim Anderson from Wolfe. Please go ahead.

If I could go back to the M7824, just a couple of questions. When do you expect you will have the first registrational data with that compound? I'm assuming that the phase II study that you referenced that's randomized, it's not registrational. And also can you talk about dose limiting toxicities with compound? And then second question on dolutegravir. I think in March, we are supposed to get the updated results from the Tsepamo study, looking at the possible side effect of neural tube defects. I'm wondering if you can tell us any update on what you think that may show or if there is anything new to share on that.

David, briefly on neural tube update, and we'll go back to Hal.

So as you say, Tim, that study is being run on to tying to get to the bottom of this issue. We're obviously need to wait and see when that data comes out, so study not run by us but actually run by the NIH. All I can say at this point is time has gone on. We have seen I think no new cases of neural tube defects and obviously more babies being born. So it looks less and less of a meaningful signal but obviously we need to see the definitive data when it comes out.

First, we’re not really giving timelines on when we will have data that can be submitted for registration. That said, let me be clear that registrational studies in oncology can sometimes take different forms. And depending on the efficacy observed, there is always opportunities to think through novel strategies.

And I would just add, Tim, if you take a step back and look at the scale of the opportunity, I mean, pembro reported £7 billion, Opdivo was close to £7 billion, I think, in the high sixes. So, in terms of a relatively modest down payment, this gives us an opportunity potentially to disrupt this market.

Thanks. Next question -- do you want to add another part now?

Tim, I just didn't get a chance to answer the dose-limiting tox part of the question. I think the profile, from an immuno-oncology perspective, we don't see any new immune-related side effects. There is the skin findings with acanthosis, some skin disorders that we think are very manageable. But, overall, we don't see that being a limiting force in the development program.

Thank you. Your next question is from the line of Emmanuel Papadakis of Barclays. Please go ahead.

It's Emmanuel Papadakis from Barclays. Maybe one on Shingrix. Simon, you were kind enough to provide a little more clarity on, I think, a significant volume expansion in 2019. You'd previously alluded to some uncertainty as to the pace of step-up to that mid-high teens target. And I think you also now specified it will be high-teens. If you could perhaps give us a little bit of further clarity on what kind of volume expansion we should expect in 2019 that should be very well-received. Also, any comment on implication for margins? It looks like you're already at that mid-30s target, any thoughts on that. And then maybe just a quick one on Respiratory. You'd previously alluded to potential spillover, so to speak, of Advair generic pricing impact in the broader space, particularly for Breo. Could you just let us know what's embedded in your current guidance, in terms of broader pricing risk for the Ellipta franchise? Many thanks.

So I'll come to Simon in a minute. I'll just make a few comments on guidance. There's no new update to the Shingrix margins. And, obviously, the guidance range at the moment is about the impact of an Advair generic across our ICS, which we've always flagged, and we still need to know the pricing and the supply rate there. But Simon can make some more comment on those. And just to add on Shingrix capacity build, first, we're obviously absolutely delighted with the launch trajectory of this vaccine and see it as being a meaningful contributor to growth for, hopefully, years ahead as we pursue not only fully serving the markets we're in but also, eventually, geographic expansion. We're also very pleased to mobilize very effectively across our supply chain, both in Europe and the U.S., to increase supply through the second half of last year, and so are confirming high-teens over the next two to three years. And you'll have heard Simon mention, so I shall reiterate, we're looking to continue the momentum that we were able to establish in the second half but we're not going to guide specifically for the number for '19. Obviously, we'll update you more as the year goes on. Simon, do you want to add?

I think on the Respiratory side, my remarks focused on Breo because the ICS/LABA category is where we see the main pressure points. And I know you've asked us a number of times on the impact on some of the other products. Clearly, there is a broader pricing dynamic that the respiratory sector is dealing with but I think the particular Advair -- knock on wood -- just will be largely restricted to ICS/LABA.

Thank you. Your next question is from Graham Parry of Bank of America. Please go ahead.

Thanks for taking my questions and say well to Simon and thanks for working with you and also welcome to Ian. And then first question on guidance. So, could you just help us understand the assumptions baked into guidance for the rate of Advair generic decline? I think consensus is running at about 60% at the moment. Is that broadly in line with your internal planning guidance? And also the timing of the Pfizer consumer deal is assumed in the guidance. I think you said second half '19. Are you assuming a full second half of that deal being in action and some dilution from it? And then, secondly, if you could help us to understand your views on the HHS rebate/safe harbor rule proposal that came out last week, GSK's potential exposure to this, and what comments GSK would be submitting back to the administration, either individually or via pharma. Thanks.

Thanks very much. So, I'll come to Simon in a moment on the guidance questions and assumptions. I mean, broadly speaking, in terms of what was said about safe harbor, we support the administration's approach, which is about bringing more transparency to the kind of pricing value chain and continue to encourage, thereby, responsible pricing and, most importantly, passing on the discounts that manufacturers provide to patients so that out-of-pocket can be impacted. Obviously, we're digesting what's come through and we're looking to collaborate, as ever, with the administration on participating in next steps. But we're broadly supportive. So, Simon, do you want to comment on the guidance question?

So, clearly, at this point, there's a pretty wide range of outcomes. But if you look at the various analogs, then you would expect to see most of the decline to the endpoint we previously indicated during the course of 2019, given we're sitting at the beginning of February. And I think as we've also said before, I think unlike a sort of conventional tablet-type generic, which would normally lose about 80% in the first year, you would probably expect less than that. But it depends very heavily on what supply they've got and we don't know that yet and we don't know, really, until they start to signal. But I'll just remind you, we said back in 2015, we would expect to end 2020 with £200 million to £300 million of sales of Advair. And if you assume most of it goes in 2019, hopefully that gives you a reasonable range. Around consumer, I think you should assume later in the second half rather than earlier in the second half and that's why we're expecting a broadly neutral impact in '19 but we gear up to the first full year being 2020 and where you'll see the impact from the synergies beginning to kick in. So, we don't know precisely yet. It's quite a complex regulatory process to go through. But it'll be toward the end of the year.

We'll update you more as we go. Thanks, Simon. Thanks, Graham. Next question, please.

Your next question comes from the line of Keyur Parekh from Goldman Sachs. Please go ahead.

Can I have two questions, please? First, on HIV, Emma, I think you said you continue seeing this being a long-term growth driver for the company. More specifically, as you've thought about 2019 guidance, does it incorporate any HIV growth for 2019 and, if so, can you give us a flavor for what you expect that growth to be? And then, secondly, for Hal, on the ICOS compound, your appendix slide shows that you kind of got the data for the pembro combination in-house for the combination therapy. Can you give us a flavor for what the data is? I'm surprised you haven't -- if it's positive, why it hasn't been moved to Phase 3 as yet. And when do we actually see the data from that? Thank you.

So, we'll come to Hal in a second. Just to reiterate, we do think that HIV is a growth driver for the company. You've seen a lot of activity from us in terms of two-drug regimens and a lot more to happen this year. That is where we expect the growth to come from, also to reiterate David's comments. And we do expect to see growth in '19. Obviously, that's going to be at a slower rate because the business is bigger and it's got a lot more competitive near-term. But we're looking forward to that 3TC, we hope, approval and building that portfolio of two-drug regimens looking forward. Hal?

So as you heard me say, I think we have encouraging clinical data and we'll be presenting that at a meeting, likely in the second half of this year. We continue to enroll patients in the study at a nice clip and we'll be learning more about which indications we think are most appropriate to pursue and who to combine it with pembro and potentially other agents to optimize the impact it has on patients.

Thank you. Our next question is from James Gordon of J.P. Morgan. Please go ahead.

The question on M7824 is just where is the company or where is Hal most excited about the prospects of the product? So, is it in patients where PD-1 therapies have really worked well and this is going to work even better? So, a synergy angle. In that case, you are bullish about the head-to-head with Keytruda in PD-L1 high patients. Or is the excitement about using it in a broader population where PD-1 monotherapy is less successful and this could sensitize? And is that why the £500 million milestone, you're not yet committed to paying that because you're seeing lots of risk around actually showing you're better in high PD-L1 patients and the opportunity is really about going forward, although not necessarily being better?

Thanks, James. Hal?

Well, I think that with these kinds of molecules, you have to really let the data tell you how to develop it. And the data to date suggests that the response rates in the second-line lung cancer setting were better than historically seen with pembro. And that gives us -- and particularly in the PD-L1 high. So, the design of the program is, as you say, the latter example, where we're looking to go head-to-head in the PD-L1 high, because we believe that the mechanism is that for -- because this molecule, and it's important to remember this, the molecule has the PD-L1 backbone so it can work like a PD-1 inhibitor. But at the same time, the PD-L1 -- because PD-L1 is expressed on tumor cells -- this actually enables the combined trap PD-L1 construct to be targeted to the cancer cell. So, when the cells become resistant through TGF-β, we think that this will prevent that by inhibiting the TGF-β locally. So, we will work where the, potentially, PD-1/PD-L1 work but be more effective by both having the TGF-β there, as well as preventing the resistance. Now, of course, that's all pre-clinical and hypotheses but that's what the data to date would suggest, that we can actually work where PD-L1 inhibitors or PD-1 inhibitors work but better because of avoiding the resistance mechanism that appears to emerge. Maybe, as we get more data, we'll find that it can expand even further beyond that and that may be what we see in some of the other indications where the activity does seem to be in places where PD-1 inhibition or PD-L1 inhibition hasn't previously been very robust. So, it's possible that both opportunities are pursued but the lung cancer opportunity is one, as I described.

Thank you. Our next question is from Mark Purcell from Morgan Stanley. Please go ahead.

Thanks very much for taking my questions. On HIV, there's a lot of focus on NBRx but that's a market that's only about 7% of total prescriptions. So, I'd be interested in any comments you can give in terms of how to make that population more dynamic, expand the amount of switching that's going on in the marketplace with your dual strategy, where you could ultimately get much more significant market share gains going forward. That's the first question. And then, secondly, on the '165 assay, GM-CSF. When you spoke last, Hal, you discussed a bit of uncertainty around the dosing of that assay and how it's optimally dosed at. It sounds like there's been some resolution or there'll be work over the next few months to resolve those questions. So, I'd be very interested to understand how you're dosing it in pivotals and the discussions you've had with regulators around dose, whether you can give us some clarity there. And then just linking a very quick one, on the Merck relationship, on TGF-β, I'm just wondering how broad that can become going forward, giving that the ATM/ATR assets would fit very nicely with your park strategy and moving, ultimately, into earlier stages of cancer. Thanks very much.

Thanks very much, Mark. We'll let Hal pick up your two-and-three-quarter questions. But, first, over to David, with the sort of slight caveat from me that there's only so much we're going to declare on our competitive approach to driving switch. But, David?

Well, as we said, and we've said this many times, we see the major part of our growth coming from two-drug regimens. And I think, in particular, hopefully starting this year, dolutegravir and lamivudine, which is absolutely an opportunity in naïve patients and that's where the GEMINI data was studied. But I think, also, in switch patients as well. So, let me just make a few comments on that. I mean, I don't need to go through the 48-week GEMINI data. You've seen that. I think, following the presentation at IAS, we've actually had a pretty strong reaction from HGPs in the U.S. and around the world. In general, it exceeded their expectations and I think, particularly, in two areas. The fact that the efficacy was maintained and was so strong at the higher viral loads has definitely been important. And the fact that we saw no resistance at all at 48 weeks has also resonated. And so there's a pretty active debate going on with HGPs, thinking about exactly which patients they should prescribe it in. I think there's an important point here, which is the 48-week is really just the start of the dolutegravir/lamivudine story. We run the GEMINI studies on through two years and then three years and we've seen, from the Juluca update, the two-year data is important. So, that data in the middle of the year, I think, will be important. And we are also investing now very heavily in switch studies, as demonstrated on the slide, the TANGO study and the SALSA study, and in a whole range of further studies around overall patient quality of life and a whole lot of technicalities around DNA archiving and getting really to the bottom of resistance. So, there's a lot going on there and while this is a conservative market and it will take time to build this story, we are ever more confident in the potential of two-drug regimes, in particular, dolutegravir/lamivudine.

Thanks, David. Hal?

Thank you, Mark, for two thoughtful questions. First, we're not going to really disclose discussions we have with regulators but, in the spirit of transparency, I wanted to show you what the design will likely evolve to and that we've made the commitment to move to Phase 3. As you rightly point out, we highlighted one or two aspects of the Phase 2 program that we thought could be optimized. The first was to make sure that our primary endpoint in the study is ACR 20 and that that's done in the setting of compared to placebo. So, that's the new design. It's kind of harder to see. It's a little smaller on the slide but that's the primary endpoint. We also have, on the left side, that the 180mg every other week dose that was used in the Phase 2 study appeared between Weeks 12-24 to be suboptimal. And what we were hoping to do is design a study that would use something close to that but on a weekly basis rather than an every other week basis. And as you can see in the design now, we have 150mg weekly being given beyond the 12 weeks. So, that actually will give us an increased exposure and a higher dose in that period of time when we, in the previous study, in Phase 2, kind of saw a flattening off, and maybe even a diminution, of the treatment effect over that time period. So, that's the increased exposure that we were alluding to hoping to get and, in fact, have designed into it. So, that, I think, hopefully answers your question about why we're optimistic about both the design and the dose there. We've had limited, virtually no, discussions with Merck regarding their DDR programs, the ATM and ATR, as you mentioned. But let's just say we find the opportunity to look at medicines that could be synergistic with PARPs in a very interesting way and there are many that are emerging and just stay tuned to see how we're going to be approaching that for the future.

Thank you. Our next question is from Kerry Holford of Exane BNP Paribas. Please go ahead.

Yes, two questions from me, please. Firstly, on COGS, you talked about the increased price pressure in Respiratory and now also established vaccines. Is the latter there, is that a new issue from the end of this year? And can you talk more about these highlights. Just trying to understand whether that weaker gross margin in Q4 is something that we should expect to continue into 2019 and beyond. And then, secondly, on M&A and in licensing, clearly, we've seen a flurry of recent deals so I just want to understand your flexibility and your appetite to do more from here. And in the context of some cash constraints here, what is your appetite to do more potential divestments of non-core Pharma assets. Within your Pharma business, we've seen you quite active in Consumer. I wonder if there's more you could do now on the Pharma side.

Thanks, Kerry. So, I'll take the second question and then I'll ask Simon to comment on the gross margin and COGS dynamics. So, first of all, as you've noticed, we were reasonably busy through the last quarter, in terms of our business development. And our No. 1 focus is to make sure we deliver the value from those deals beyond the Consumer side or, indeed, the Pharma side. That said, I was extremely clear in July 2017 that our No. 1 priority is the strengthening of the pipeline. I'm pleased with the progress but BD will continue to be a key part of that. And I think the Merck alliance that we announced yesterday is exactly the kind of thing that we want to continue to do, whether it be on assets or technology platforms, and there will be cases of us looking for creative business development. There will probably also be examples of us looking to out-license things in the portfolio as well, a bit to your secondary comment, which is will we continue to review the portfolio and making sure we're allocating our capital as intelligently as we possibly can. And, yes, we will. But our No. 1 priority is to up the value from the various deals that we've done. So, with that, Simon?

Kerry, on COGS, I think as we've talked about for some time, we are seeing some pressure at the gross margin, given the pricing environment, particularly in the Pharma business, and we're now seeing some more of that in the established and older part of the vaccines portfolio, which have a higher degree of exposure to some of the tender business and GABI type contracts. And that one's a bit more visible in the back half of last year. But I think those trends will continue. It's one of the reasons why we're putting a lot of focus and effort into restructuring the supply chain to deliver some efficiencies to offset those pressures. I think Q4, specifically, you shouldn't read straight into 2019 because it had a number of specifics in there, particularly, the Relenza tender and some one-off sales of products that we've already sold, so we're contract manufacturing for third parties, which is a pretty low margin business but we delivered quite a heavy load of around that total of £80 million that I referred to. So, it was a particular factor in why there was such a sharp step-up in Q4.

So, I think we have one last question. I really hope it's the last question for Simon. So, with that, please, over to the last question.

Thank you. That question comes from Steve McGarry of HSBC. Please go ahead.

Hi, thanks for taking the question. Unfortunately, Simon, it's not a financial one.

You're not upsetting me at all.

Just on M7824 and head-to-heads in non-small cell lung cancer versus Keytruda, what would encourage you to develop that drug more broadly? Does it have to be better than Keytruda or is not inferior or equivalent enough? And then following on from that, if you look elsewhere in the industry, you've got Keytruda in trials and 900 studies, Opdivo with 900 studies, and they consume the majority of R&D at those companies. Although it would be a great problem to have if M7824 was superior to Keytruda, how big could the clinical program and the R&D spend become at that point in time? Thanks.

Thanks very much, Steve. So, I'll hand that on to Hal. I think the main point to underline is your "good problem to have" point. One of the things that Hal has brought in with tremendous discipline -- he referred to it when he talked about new governance -- is really looking at the efficiency frontier across our R&D spend and the assets that we want to bet on and where we can get the biggest kind of returns. And so we are particularly disciplined about that. But you'll also remember that Simon said in his outlook for '19 that we do expect a meaningful uptick in our R&D spend. Whether that be the continuing bet on our internal assets that we've accelerated, like BCMA, or, indeed, backing the TESARO teams and assets too. But the key is to make sure we are also dropping off and cancelling things that we don't think come high enough up that efficiency curve. So, Hal laid out his "what's in, what's out, what's accelerating, what's adding" chart today and that's something that we will, each six months, make sure we update you on to see what the progress is. So, Hal, do you want to come back on the M7824 question?

Thanks, Steve. I think I'll turn it over to Simon.

It would be a short answer.

Okay. So, let me try to tackle the first one and then sort of reflect on the second one. I think the study is designed as a superiority trial, just to be clear, in Phase 2. Maybe your point is if we don't achieve superiority, would there be an opportunity to move forward with something that had similar effects. I'd say two things. First, one thing you have to be careful about in these Phase 2 studies is they're usually powered and the primary endpoints are usually on response rate, which doesn't always track to PFS and OS and we've seen that with IO agents in the past for a number of reasons. So, I think that one has to both consider the effect on response rates but also look at the data as it relates to PFS and OS, although it will be very underpowered. But I want to point out that the philosophy, the sort of vision for our immuno-oncology group, is to really develop transformational medicine. So, I think that our focus is going to be on having benefit beyond Keytruda. It's a wonderful drug. It's done transformative things for patients and we think we can do better and this is one of what we believe are many smart bets we are taking to see if we can have superior therapies for patients with lung cancers and others. How big is big and how big to go? It really is all dependent on the data. And I really think that we have a number of programs where you could ask the same questions. Should the data read out in a very profoundly positive way, they could result in lots of opportunities for us to do development in patients to benefit. But these are, as was mentioned, high-risk, high-reward. And that's why we're doing a number of these. We think it's a smart bet and I really hope this is the problem we have to face. So, more as we can unravel data soon.

Thanks very much, Hal. So, with that, I will reiterate my last public thanks to Simon. Thank you all for joining and we look forward to updating you through the year, a year that we hope will build on the good momentum of '18, be very focused on delivering operational performance, and planning for the delivery of value against our various deals, and, particularly, a highly effective integration under Brian's leadership at the Pfizer joint venture, and, most of all, updating you on our progress on R&D. Thank you very much.

This presentation has now ended.