Memorandum

1Q18-2Q18 Diabetes and Obesity Industry Roundup – GLP-1s (+25% YOY), CGM (+60% YOY), SGLT-2s (+31% YOY) Drive ~70% of Growth; Insulin Flat – October 15, 2018

Executive Highlights

  • Overall diabetes industry: In 1Q18, the diabetes drug and device industry grew 13% YOY to a remarkable $13 billion; in 2Q18, the industry climbed 6% YOY to $13.3 billion. The primary growth drivers remained GLP-1 agonists (29%/34% share of growth), CGM (13%/27%), and SGLT-2 inhibitors (19%/19%) in 1Q18 and 2Q18, respectively. Pricing pressure remains acute in the US, where drug and device sales were roughly flat for 1Q18 (+3% YOY to $6 billion in 1Q18; -3% YOY to $5.9 billion in 2Q18). By contrast, OUS sales rose 17% YOY to $6.2 billion in 1Q18 and 8% YOY to $6.4 billion in 2Q18.

  • Diabetes therapy: The overall insulin market climbed 4% YOY to $5.2 billion in 1Q18 and fell 3% YOY to $5.1 billion in 2Q18.

    • By geography, insulin sales were actually lower in the US than outside the US during 1H18, the first time we can remember this happening.

    • Sales were stronger in the rapid-acting segment, climbing 8% YOY in 1Q18 and 6% YOY in 2Q18 for pooled revenue of $1.7 billion in each quarter. That’s pretty strong growth for a product where use is said by some to be slowing among those with type 2 – Dr. Anne Peters doesn’t really even prescribe mealtime insulin for her type 2 patients, she has emphasized (or not many of them!). By contrast, however, we believe that rapid acting insulin use in type 1 is likely increasing as the volume of patients is increasing and pump use also continues to grow.

    • In contrast, basal insulins grew a modest 2% YOY in 1Q18 to $2.4 billion and dropped 5% YOY in 2Q18 to $2.5 billion, compared to a higher base.

    • In the middle ground, human insulin grew 4% YOY to $746 million in 1Q18 and fell 5% YOY in 2Q18 to $739 million to stay mostly flat YOY.

    • Basal insulin/GLP-1 agonist fixed ratio combinations, a growing class, posted a combined $147 million in sales for 1H18, nearly tripling YOY from $53 million in 1H17 (the class’ first two quarters on the market).

  • The GLP-1 agonist class continued skyward, rising 30% YOY to $1.9 billion in 1Q18 and then 20% YOY to $1.9 billion in 2Q18. Sequentially, the market fell 1% in 1Q18 and grew a modest 2% in 2Q18, as the class struggles to climb from an ever-higher base.

    • Novo Nordisk’s Victoza maintained market leader status with 47% of total GLP-1 revenue in 2Q18 ($898 million) and 53% in 1Q18 ($989 million), growing 4% YOY in 1Q18 and falling slightly by 1% YOY in 2Q18

    • Lilly’s Trulicity is not far behind: The once-weekly captured 41% value share in 2Q18 ($780 million sales) and 36% in 1Q18 ($678 million), growing 82% YOY in 1Q18 and 62% YOY in 2Q18

  • The SGLT-2 inhibitor class surged 41% YOY in 1Q18 and 21% YOY in 2Q18 to just over $1 billion in sales for each quarter but was flat sequentially with a 2% decline followed by 4% growth. At the close of 2H18, Jardiance continued to lead the class with 43% value share ($445 million) followed by Farxiga at 33% ($340 million) – both are offsetting Invokana’s fall to 21% value share ($215 million), down from 34% in 2Q17. Newcomer Steglatro rounded things out at 4% value share ($37 million).

  • DPP-4 inhibitors achieved sales of nearly $5 billion in 1H18, showing 5% YOY growth. They were relatively flat sequentially at $2.4 and $2.5 billion in revenue.

  • Diabetes technology: CGM had a truly breakout half-year, growing faster from a higher base – we estimate global sales were ~$561 million in 1Q18 (+55% YOY) and a record-smashing ~$652 million in 2Q18 (+64% YOY). This marked the first-ever half-year crossing $1 billion in CGM sales (sales were actually $1.2 billion!), an acceleration from 1H17’s ~50% YOY growth, and places CGM nearly on par with the size of the global insulin pump industry.

  • In contrast to CGM, BGM struggles continue, with pooled global BGM-only revenue for Roche, J&J, and Abbott totaling an estimated ~$985 million in 1Q18 (+2% YOY) and ~$1.0 billion in 2Q18 (-4% YOY). It may not be too long before CGM sales eclipse BGM sales. The US BGM market in particular is showing no sign of recovery, with seven consecutive quarters of pooled YOY declines.

  • Insulin pumps/delivery devices of $1.3 billion rose ~20% YOY in both 1Q18 (~$679 million) and in 2Q18 (~$630 million), marking three consecutive quarters of positive YOY growth for the industry. Following last year’s 4% decline, the growth here is encouraging – we believe automated insulin delivery (and anticipation thereof) is driving the faster growth. With the obvious exception of Animas, we estimate every company saw growth in both 1Q18 and 2Q18. While Medtronic has carried the field in 2018, with an ~70% market share by sales and supplying ~73% of 1H18 pump industry growth, Insulet and Tandem are more than holding their own. We expect a resurgence in pump growth now that automated insulin delivery is only growing – it’s the killer app. We also expect new pump entrants like Beta Bionics and Bigfoot Biomedical to growth the market more than steal share, starting in earnest around 2021.

  • Obesity: The branded obesity drug market continues to struggle commercially with the exception of Novo Nordisk’s Saxenda; in particular, 1H18 saw Orexigen (maker of Contrave) file bankruptcy and later be acquired by Nalpropion Pharmaceuticals. Holding the last reported quarterly revenue for Contrave ($27 million) steady, the obesity market climbed 43% YOY to $174 million in 1Q18, then 27% YOY to $185 million in 2Q18. Excluding Novo Nordisk’s Saxenda, growth slowed to 5% and 17% YOY for 1Q18 and 2Q18, respectively ($47 million/quarter).

    Based on financial results from the nearly 30 public companies that we regularly track, this diabetes/obesity industry roundup provides a bird’s eye view of trends in the first half of 2018, including a deep dive into each major drug class and device category.

    This report is divided into four sections: (i) overall industry performance; (ii) diabetes therapy; (iii) diabetes technology; and (iv) obesity. Within each section, we analyze sales performance, share of growth, and provide graphs with current and historical trends. Please note that in many cases, the data we’ve used for our analysis reflects our best estimates, since a number of companies don’t disclose financial information in great detail (or in the case of BI, at all, except once a year). We’ve made the basis of our assumptions clear in the text below, and we recognize that a number of the estimates may be meaningfully different from the “actual” that is not known! If you have opinions on how an estimate could be improved, please don’t hesitate to reach out and let us know.

    Table of Contents 

    Overall Industry Performance

    Diabetes Industry Sales Climb ~9% YOY, Totaling $13 Billion (1Q18, +13% YOY) and $13.3 Billion (2Q18, +6% YOY); GLP-1 Agonists, CGM, and SGLT-2 Inhibitors Drive ~70% of Growth in 1H18

    Diabetes Industry Summary for 1Q18 and 2Q18 

    Drug or Device Class

    Total Sales in 2Q18 (1Q18)

    YOY Growth in 2Q18 (1Q18)

    Share of Industry Growth in 2Q18 (1Q18)

    Share of Total Industry Revenue in 2Q18 (1Q18)

    Insulin (Basal + Rapid-Acting + Human)

    $5.1 ($5.2)

    -3% (+4%)

    0% (12%)

    39% (40%)

    DPP-4 Inhibitors

    $2.5 ($2.4)

    +1% (+9%)

    4% (13%)

    19% (18%)

    GLP-1 Agonists

    $1.9 ($1.9)

    +20% (+30%)

    34% (29%)

    14% (14%)

    SGLT-2 Inhibitors

    $1.0 ($1.0)

    +21% (+41%)

    19% (19%)

    8% (8%)

    Basal Insulin/GLP-1 Combos

    $0.08 ($0.07)

    +140% (+252%)

    5% (3%)

    <1% (<1%)

    BGM

    $1.0 ($1.0)

    -4% (+2%)

    0% (1%)

    8% (8%)

    Insulin Pumps

    $0.63 ($0.68)

    +20% (+20%)

    11% (8%)

    5% (5%)

    CGM

    $0.65 ($0.56)

    +64% (+55%)

    27% (13%)

    5% (4%)

    Other (BD)

    $0.28 ($0.27)

    +5% (+10%)

    1% (2%)

    2% (2%)

    Total Diabetes Drug + Device Industry

    $13.3 ($13.0)

    +6% (+13%)

    --

    100% (100%)

    • By our calculation, the diabetes drug and device industry grew 13% YOY in 1Q18 to hit $13 billion, before growing 6% YOY in 2Q18 to $13.3 billion. Sequentially, sales actually fell 3% in 1Q18 from an all-time-high of $13.4 billion in 4Q17; in 2Q18, total revenue climbed 2% sequentially. Also by our calculation, we estimate that the US diabetes market was flat, posting 3% YOY growth to $6 billion in 1Q18 and a 3% YOY drop to $5.9 billion in 2Q18. In strong contrast, OUS sales were very strong, climbing 17% YOY to $6.2 billion in 1Q18 and 8% YOY to $6.4 billion in 2Q18. As ever, these trends can be attributed to the relatively more competitive atmosphere for diabetes in the US relative to the rest of the world (margins have been higher in the US historically, providing more room to fall), and we also imagine that strong growth in so-called emerging (non-European, non-Japanese) markets is boosting ex-US sales even further.
    • Particularly among pharmaceutical companies, pricing pressure is the ever-recurring explanation for muted growth stateside, particularly in insulin and also among DPP-4 inhibitor and even SGLT-2 inhibitor manufacturers. In the US, manufacturers have been driven to offer greater rebates to PBMs in exchange for formulary positioning, but there are also more (in our view) positive factors contributing to lower realized price in the US. These include patient savings/discount programs and, in some cases, greater segment mix in favor of patients on Medicaid – which although technically translates to lower profitability also means, importantly, that those most in need are gaining access to necessary medications. We hope the CEOs know how much they have the opportunity to help people in the largest public health crisis of our time.

    • As has become the norm, GLP-1 agonists, CGM, and SGLT-2 inhibitors continue to drive the lion’s share of growth in the diabetes drug and device industry. In 1Q18 (also by our calculation), GLP-1s accounted for 29% of growth, followed by SGLT-2s at 19% and CGM at 13%. Additionally, in 1Q18, DPP-4s contributed 13% of growth (particularly high due to an easy comparison), insulin 12% of growth, and insulin pumps 8% of growth, though our understanding is that not all this movement reflects of sustained growth. In 2Q18, GLP-1s drove an even more impressive 34% of growth, followed by 27% from CGM and 19% from SGLT-2s – totaling to 80% of total market growth, a much more heavily-skewed quarter than 1Q18 when these three classes drove 61% of growth (see the full breakdown above).

      • Compared to 1H17, CGM and SGLT-2s are driving a relatively similar share of industry growth, but the share driven by GLP-1s has fallen. In 1H17, CGM was responsible for 18% and 20% of diabetes industry growth in 1Q17 and 2Q17, respectively. Similarly, SGLT-2s drove 16% and 22% of industry-wide growth in those quarters, respectively, so the contribution of this class has also held steady. In contrast, the relative contribution of GLP-1s to growth has fallen from 55% in 1Q17 and 57% in 2Q17 to 29% in 1Q18 and 34% in 2Q18. To be sure, GLP-1s are still the most significant growth driver in diabetes overall, but the class’ dominance in this one regard seems to be waning slightly. We take a deep dive on GLP-1s below, but suffice to say these agents are now growing from a much higher base and hovering around $2 billion in quarterly sales (a milestone they’ve yet to break – could 2018 be the year?). The arrival of Novo Nordisk’s Ozempic, we think, will spur on class growth in coming quarters, so we wouldn’t be surprised to see the class gain a new set of legs. 

    Values represent 2Q18 estimated sales

    Diabetes Therapy

    Insulin

    Insulin Sales Roughly Flat (+1% YOY) at $5.1 Billion; Growth in Rapid-Acting Class (+7% YOY to $1.7 Billion) from an Easy Comparison in early 2017 Balances Modest Reduction in Basal Insulin Class (-2% YOY to $2.5 Billion)

    • The overall injectable insulin market climbed 4% YOY to $5.2 billion in 1Q18 (from $5.0 billion in 1Q17) before falling 3% YOY to $5.1 billion in 2Q18 (from $5.3 billion in 2Q17). Sequentially, insulin sales fell 3% in 1Q18 from $5.4 billion in 4Q17, then fell another 2% sequentially in 2Q18. This includes basal insulin analogs, rapid-acting analogs, and human insulin products from Novo Nordisk, Sanofi, and Lilly; this number does not include BI’s share of Basaglar sales nor does it include GLP-1/insulin combos. As shown above in the graph for the entire market, global insulin sales have fluctuated between $5 billion and $5.5 billion quarterly, since peaking at $5.9 billion in 4Q14; overall, this 1H18 performance is nothing out of the ordinary. In 1H18, rapid-acting insulins gave a stronger performance, which offset a modest decline in the basal insulin market. In the pooled rapid-acting market, sales climbed 8% YOY in 1Q18 and 6% YOY in 2Q18 for pooled revenue of $1.7 billion in each quarter; sequentially, the class grew 4% in 1Q18 and 3% in 2Q18. By contrast, pooled sales of major branded basal insulins grew a modest 2% YOY in 1Q18 to $2.4 billion and declined 5% YOY in 2Q18 to $2.5 billion; sequentially, basal insulins fell a notable 9% in 1Q18 (from $2.7 billion in 4Q17) and grew only 1% sequentially in 2Q18. Hovering somewhere in-between is the human insulin market, which grew 4% YOY to $746 million in 1Q18 (from $716 million in 1Q17), then fell 5% YOY in 2Q18 to $739 million (from $776 million in 2Q18) to stay mostly flat YOY. Sequentially, human insulins fell 6% in 1Q18 (from $790 million in 4Q17) then dropped another 1% in 2Q18. See below for an individual deep dive into the basal, rapid-acting, and human insulin categories, as well as a separate analysis of fixed-ratio basal insulin/GLP-1 agonist combinations.

      • At 48%, basal insulin sales continue to comprise a plurality of insulin market revenues; rapid-acting insulin sales make up another 39% of the insulin market, following by human insulin at 14%. These proportions have been mostly consistent over time. For the past eight quarters, human insulin has comprised 14%-15% of the insulin market, basal insulin 48%-51%, and rapid-acting insulin 36%-40%, with no real changes in trends over time. Broadly, we expect basal insulin volume to expand in sales while average pricing will likely continue to decline; we expect rapid acting insulin volume of sales to grow more slowly as the population of people with type 2 and a longer duration of diabetes grows but as it takes them longer to move to basal insulin (they still will do so whoever and will likely stay alive on basal insulin longer and lifespans continue).

      • As of 2Q18, Novo Nordisk held 47% of the global injectable insulin market by value with $2.4 billion in sales, while Sanofi held 28% with $1.4 billion and Lilly held 26% with $1.3 billion. These shares have changed considerably over the past couple of years, a trend driven mainly by Sanofi’s loss of exclusivity on Lantus. For reference, Novo Nordisk held a comparable 46% of the insulin market in 1Q17 with $2.3 billion in revenue, but Lilly held a much smaller 21% ($1.1 billion) and Sanofi a larger 33% ($1.6 billion). Since then, Lilly’s value share has grown every quarter, while Sanofi’s has mostly declined. Overall, this shift can be attributed to strong growth from Lilly’s biosimilar basal insulin Basaglar, and even mild but notable growth from Humalog in recent quarters due to formulary gains in the US. Specifically, Lilly’s insulin portfolio grew 20% YOY in 1Q18 and 17% YOY in 2Q18. Novo Nordisk’s insulin business has remained stable on the strength of next-gen basal Tresiba and addition of Fiasp, posting 4% YOY growth in 1Q18 balanced by a 3% YOY loss in 2Q18. However, Sanofi’s entire diabetes portfolio has seen reductions driven by the decline of Lantus in the US, somewhat balanced by next-gen basal Toujeo; Sanofi’s insulin frachise posted an 8% YOY reduction in 1Q18 followed by a 15% YOY reduction in 2Q18.

    • Insulin continues to comprise the bulk of all diabetes drug and device sales. Considering diabetes drugs only, insulin sales of $5.1 billion in 2Q18 were roughly on-par with combined sales of $5.4 billion for the three other major drug classes (SGLT-2 inhibitors, GLP-1 agonist, and DPP-4 inhibitors). The insulin class overall did drive 12% of growth in the overall diabetes drug and device market in 1Q18 (0% in 2Q18) on the strength of 4% YOY revenue growth, but generally speaking insulins have not contributed sustained growth to the diabetes market in years. In 1Q18, insulin products comprised 40% of total, global industry revenue ($5.2 billion of $13.0 billion); in 2Q18, insulin made up 39% of global diabetes sales ($5.1 billion of $13.3 billion). This does, so far, reflect a drop from 2017 when insulin comprised 43% of total global revenue ($21 billion of $49 million). There’s no doubt that insulin’s market share has been slowly shrinking; our records show that, back in 2013, insulin made up almost 50% of overall industry revenue. Of course, this is due not so much to losses in insulin revenue as the growth of other, more advanced therapy and technology classes and products (GLP-1s, SGLT-2s, CGM) – we do also believe that there has been far more pricing pressure for insulin than for newer classes.

    • Indeed, to be sure, insulin continues to face substantial commercial challenges, particularly in terms of pricing pressure and profitability. Looking at overall insulin market changes in the US vs. ex-US geographies reveals that this dynamic is largely driven by the US: Insulin sales in the US fell by 5% YOY in 1Q18 to $2.5 billion, then dropped 9% YOY in 2Q18, also to $2.5 billion. In contrast, ex-US sales actually climbed 9% YOY in 1Q18, to $2.6 billion and grew 1% YOY in 2Q81 to $2.6 billion. In fact, according to our records and calculations, 1Q18 was the first quarter (at least in semi-recent history) that ex-US insulin sales outpaced US insulin sales – at least in part reflecting just how meaningful US pricing pressure has become. Overall, US insulin sales have been on the decline since 1Q15, and while manufacturers often cite particularly strong effects in the rapid-acting insulin segment, both basal and prandial insulins alike face these challenges – this is particularly true since Lantus loss patent exclusivity in 2016. Because rapid-acting insulins are seen as largely interchangeable within the given class, manufacturers are often driven to offer steep rebates in exchange for formulary access. For example, New Lilly CEO Mr. Dave Ricks said at JPM 2017 that Humalog and NovoLog are “already on the floor” in terms of pricing. Novo Nordisk CSO Dr. Mads Thomsen told the Wall Street Journal late last year that insulin profitability is at an all-time low, specifically pointing to US pricing pressure and the (flawed) perception that new insulins offer only marginal rather than disruptive benefit. Barring a significant change to the US healthcare and drug pricing system, we don’t expect US pricing pressure to abate any time soon though we note that next-generation modern insulin (Tresiba and Toujeo) along with insulin/GLP-1 combos continue to receive rave reviews comparably.

    Basal Insulin Market Steady at ~$2.5 billion in Quarterly Sales; +2% YOY in 1Q18, -5% YOY in 2Q18; Next-Gen Tresiba + Toujeo Experience First YOY Drop; Basaglar Thrives; Pricing Pressure Abounds

    • Pooled sales of major branded basal insulins grew a modest 2% YOY in 1Q18 to $2.4 billion (from a base of $2.3 billion in 1Q17), and declined 5% YOY in 2Q18 to $2.5 billion (from $2.6 billion in 2Q17). This class has been reaching pooled sales of ~$2.5 billion per quarter and ~$5 billion per half for the last several years (see our industry roundups from 2015, 2016, and 2017); 1H18 market revenue of $4.9 billion fell 1% YOY from $5 billion in 1H17. Basal insulin sales tell a story of competing forces: consistent losses for first-generation products (Sanofi’s Lantus and Novo Nordisk’s Levemir) offset by rising revenue from next-generation products (Novo Nordisk’s Tresiba and Sanofi’s Toujeo) and from biosimilar glargine (Lilly/BI’s Basaglar). Sanofi’s Lantus (insulin glargine U100) business has been struggling since 2015, with consistent YOY revenue decline since 3Q15. The first half of 2018 was no exception to this trend, as Lantus sales dropped 26% YOY in both Q1 and Q2. This reflects increasing competition from next-gens as well as from Basaglar as well as continued pricing pressure. An additional headwind for Lantus came with major US formulary exclusions in 2017 and 2018 (CVS Health, UnitedHealthcare, and Medicare Part D), which were a boon for Lilly/BI’s Basaglar franchise and for Novo Nordisk’s Tresiba (and presumably Levemir franchises, now both preferred by Medicare Part D over Lantus and Toujeo). While next-gen basal insulins are a bright spot in this market, we note that for the first time in 2Q18, pooled Tresiba + Toujeo sales actually declined YOY (-2% to $559 million). In 1Q18, the next-gen class grew 27% YOY but fell 5% sequentially to $533 million. Thus, only Basaglar contributed meaningfully to class growth in 1H18. We see this as unequivocal evidence that pricing pressure is affecting each and every insulin product on the market.

      • Notably, pricing pressure extends to mealtime insulin therapies as well – in fact, we’ve heard from multiple industry experts that pricing pressure is most severe in the rapid-acting insulin category because PBMs/payers view these products as highly interchangeable (at least in the basal insulin category, there is more consensus around the advantages of Tresiba or Toujeo over Levemir or Lantus). In line with this, basal insulin continued to reflect 48% of the overall insulin market in 2Q18, on par with 47% in 1Q18, 48% in 1Q17, and 48% in 1Q16. Basal insulins are now tied with DPP-4 inhibitors in terms of quarterly revenue at $2.5 billion in 2Q18.

    Basal Insulin Market vs. DPP-4 Inhibitor Market (1Q06-2Q18)

      • The present analysis covers basal analogs old and new, including Sanofi’s Lantus (insulin glargine U100), Novo Nordisk’s Levemir (insulin detemir), Sanofi’s Toujeo (insulin glargine U300), Novo Nordisk’s Tresiba (insulin degludec), and Lilly/BI’s Basaglar (biosimilar insulin glargine). As a caveat, we have estimated total Basaglar sales by doubling Lilly’s reported product revenue ($202 million in 2Q18), since BI’s share is not publicly disclosed. We can only speculate that the two companies split worldwide Basaglar revenue 50/50 but this is conjecture. This estimated total for Basaglar sales has been factored into pooled revenue numbers and whole class analysis.

    Basal Insulin Market (1Q06-2Q18)

    • By volume (see slide 8 in Novo Nordisk’s 2Q18 roadshow presentation), Lantus holds a 39% share (TRx) of total basal insulin prescriptions in the US (a steep drop from its 50% US TRx in 2Q17). Levemir follows with 23% US TRx (steady YOY), Tresiba holds 13% (up 5% YOY), Basaglar holds 10% (up 6% YOY), and Toujeo holds 8% (down 1% YOY). These trends make sense in light of each product’s formulary status. For 2018, the Medicare Part D formulary lists Tresiba, Levemir, and Basaglar on a lower tier (i.e. lower patient co-pay) than Toujeo and Lantus. Sanofi’s basal products have been excluded by CVS Health and UnitedHealthcare since 2017; both PBMs prefer Basaglar, which is really too bad for patients since so many prefer next-generation basal insulins and for many, these insulins are safer due to less hypoglycemia. We should note that this does not mean that  Novo Nordisk or Lilly/BI are immune from the negatives related to pricing pressure, because we know these companies are paying high rebates to PBMs/payers in order to maintain the formulary status of Tresiba/Levemir and Basaglar, respectively. Indeed, there continues to be effectively zero transparency around the price-setting and rebating practices of PBMs and we believe they are having the last laugh – we do not yet understand much at all about the value they are providing but it’s our sense their revenue continues to increase (though at a slower pace since pricing increases have slowed). Novo Nordisk leaves 6.5% of the market unaccounted for on its slide, which we suspect reflects human/regular insulin; 6% was also unaccounted for in 2Q17, which suggests that this segment isn’t meaningfully changing. Dr. Irl Hirsch and others suggest that perhaps human insulin sales should increase as it is the most affordable insulin; it’s also the least stable, which is important for many patients.

    • Oddly and very surprisingly, next-gen basal insulins experienced a first-ever YOY decline in 2Q18, with combined Tresiba + Toujeo sales falling 2% YOY to $559 million. Until this point, the class had consistently posted double-digit YOY growth since Toujeo and Tresiba both launched. In 1Q18, Tresiba and Toujeo combined for $533 million in global revenue, rising 27% YOY but dropping 5% sequentially, from $419 million in 1Q17 and $561 million in 4Q17. We saw a more disheartening picture in 2Q18, as Tresiba revenue dipped 11% YOY, Toujeo revenue climbed only 3% YOY (falling 30% YOY in the US), and pooled sales fell 2% YOY to $559 million; we should note that the base for this comparison was an all-time-high of $569 million in 2Q17, but even so, we would certainly have expected to see continued positive growth for products that are still relatively new and that offer such incredible value. After all, the Tresiba label now includes data from DEVOTE showing 40% risk reduction for severe hypoglycemia compared to Lantus – we continue to believe that it would also be valuable to see SWITCH data, which also reflected data for type 1 patients. Toujeo has impressive backing from real-world evidence pointing to hypoglycemia benefits, and importantly, both these drugs have a flatter PK/PD profile vs. first-gen Lantus or Levemir. In our view, Tresiba and Toujeo are leaps and bounds above the basal insulins that came before them (particularly Levemir), and we want to see these advanced products in more patient hands. We were reassured to see 5% YOY US volume growth for Tresiba in 2Q18 (above) despite the slight financial downturn in 2Q18. Together, Tresiba and Toujeo held 21% US TRx within the basal insulin class in 2Q18 vs. only 17% at the same time last year. We imagine the next-generation basal insulin class will see growth again in future quarters, but we’re not uncrossing our fingers until we see additional sales/volume growth.

      • The sluggish performance from next-generation insulins in 1H18 is additional evidence that pricing pressure is ubiquitous across the insulin category. This becomes especially apparent when examining disparities between US and OUS sales for basal insulin. For example, Lantus’ OUS sales have been relatively steady over the past four years and surpassed US sales for the first time ever in 1Q18 (though this was driven by US losses, not OUS gains). In 2Q18, OUS Lantus revenue was $570 million (-9% YOY) vs. $471 million stateside (-39% YOY). Similarly, OUS revenue from Toujeo has outpaced US revenue, growing 49% YOY in 2Q18 to $153 million as US sales plummeted 30% YOY to $100 million. We fully believe that US pricing pressure is what’s preventing meaningful class growth for basal insulins (and for rapid-acting insulins, too) -we do also believe that lower prices for next-gen insulins would help more patients gain access. As we understand it, there have been pricing drops for Tresiba overseas, which is a positive for patients since these prices were too high to start.

      • In 2Q18, Tresiba once again led this two-member class with 55% market share by value ($306 million). Toujeo accounted for the remaining 45% ($253 million). We saw similar dynamics in 1Q18, when Tresiba captured 54% of the market by value and Toujeo captured 46%. While it’s becoming more and more common to see comparisons between these two drugs (most recently, we saw new data at EASD from Sanofi’s BRIGHT trial comparing degludec and glargine U300 head-to-head), we think it’s important to emphasize that both are strong therapeutic options for people with diabetes and far stronger than Lantus, Levemir, and Basaglar for many patients. And to be clear, we see more than enough room to grow the next-gen basal insulin class as a whole, so that both Tresiba and Toujeo are commercially successful.

    • Lantus and Levemir drove decline in the basal insulin market in 1H18. Pooled sales of these two first-gen products fell 16% YOY in 1Q18 to $1.6 billion (from $1.9 billion in 1Q17) and fell 21% YOY in 2Q18 to $1.5 billion (from $1.9 billion in 2Q17). Looking at each franchise individually, Lantus dropped 26% YOY in both 1Q18 and 2Q18, driven by US losses due to formulary exclusions and biosimilar competition from Basaglar. Levemir, by comparison, fell 31% YOY in 1Q18 and fell 16% YOY in 2Q18 from a much lower base – certainly a negative sign.

    • Basaglar continued its success story in 1H18, with sales tripling YOY in 1Q18 to $332 million and more than doubling YOY in 2Q18 to $404 million. As a reminder, we’ve estimated Basaglar’s total revenue by doubling Lilly’s reported sales ($166 million in Q1 and $202 million in Q2). BI does not report revenue publicly, so we assume for the purposes of this report that the two partners split Basaglar sales 50/50. Importantly, our calculations do not affect the YOY or sequential changes in Basaglar sales. An underwhelming 8% sequential increase for the biosimilar in 1Q18 was more than made up for with a 22% sequential surge in 2Q18. Lilly has repeatedly cited pricing pressure as a challenge for its insulin business, arguing that high rebates within commercial channels offset the positive effects of lower-tier formulary placement. Higher rebates to PBMs/payers implies lower realized price for Lilly, and we suspect this is a dominant reason for Basaglar’s single-digit sequential growth in 4Q17 (+6%) and 1Q18. Again, we can’t talk about an insulin therapy – even one that’s soaring commercially – without acknowledging the persistent obstacle of pricing pressure.

      • Interestingly, Basaglar reflects 16% of the basal insulin market by value by our calculations, but only holds 10% TRx (as of June 2018). We wouldn’t expect value share to exceed volume share for a “discount product,” so to speak. Are real-world patients seeing the promised 15%-20% discount on Basaglar vs. Lantus? A couple caveats here: First, US TRx does not necessarily match global TRx. Second, we’ve estimated total Basaglar sales by doubling Lilly’s recorded revenue, which may not be accurate. It’s possible we’re overestimating total Basaglar sales and therefore are also overestimating the biosimilar’s share of the market by value. This analysis is speculative and indeed should be considered “illustrative.”

      • FDA held a public hearing on its biosimilar action plan in September, and we hope the initiative streamlines availability of new biosimilar insulins. As context, both Merck’s Lusduna Nexvue and Mylan/Biocon’s Semglee have experienced significant delays in launch due to ongoing patent disputes with Sanofi and indeed, Merck said today it would no longer proceed with plans to commercialize Lusduna – obviously positive for Sanofi and Novo Nordisk and even Lilly – and disappointing for patients looking for another lower priced option.

        • Notably, FDA commissioner Dr. Scott Gottlieb alluded to movement on biosimilars in his prepared remarks, stating “we’re also taking new steps to communicate with patients, payers, and providers to improve the understanding of biosimilar and interchangeable products. And we’ll act where appropriate to deter the gaming of FDA requirements to unfairly delay competition among biologics. The Federal Trade Commission is a vital partner in this work and we look forward to continued coordination with them to address anti-competitive behavior in the drugs and biologics marketplace.”

        • As a reminder, Basaglar was tentatively approved by FDA in 2014, but only reached US pharmacy shelves in December 2016 after Lilly/BI signed an agreement with Sanofi to pay royalties and delay US launch. We look forward to a market with multiple biosimilar insulins available, because prior experience with the generic marketplace suggests that at least two generics are needed before patients see meaningful cost-savings. And as we know all too well, patients with diabetes need help affording their insulin – we are sure there is more to come on the Merck decision. Our confidence in the Mylan biosimilar remains low at present.

      • We do see some manufacturers trying to create change on the drug pricing front, including Lilly through its new insulin helpline, a resource designed to help patients navigate the system as it currently stands.

    Rapid-Acting Insulin Sales Grow ~7% YOY in 1H18 Despite Pricing Pressure and GLP-1/SGLT-2 Competition; Faster-Acting Fiasp Arrives on the Scene – New Ultra-Rapid-Acting Category Emerging

    • The rapid-acting insulin market climbed 8% YOY in 1Q18 and 6% YOY in 2Q18 for pooled revenue of $1.7 billion in each quarter. Sequentially, the class grew 4% in 1Q18 and 3% in 2Q18. This is a pretty positive performance, especially when considering the intense pricing pressure that affects all products in the mealtime insulin category; management from Novo Nordisk, Sanofi, and Lilly have all explained how rebating is higher for rapid-acting insulin than for any other diabetes drug class because PBMs/payers view all rapid-acting insulins as more or less interchangeable, and thus design more exclusive formularies. Mealtime insulins are also facing increasing competition from GLP-1 agonists and SGLT-2 inhibitors, therapies that address postprandial glucose excursions without raising hypoglycemia risk. Given this context, a 7% YOY jump in class-wide sales from 1H17 ($3.2 billion) to 1H18 ($3.3 billion) is a good sign. However, we should note that pooled sales fell 8% YOY in 1Q17 and dipped 1% YOY in 2Q17; in general, the rapid-acting insulin market has been fluctuating, and we’re not confident that 1H18 performance signals a true return to growth.

      • This analysis covers Novo Nordisk’s NovoLog (insulin aspart) and Fiasp (faster-acting insulin aspart), Lilly’s Humalog (insulin lispro), and Sanofi’s Apidra (insulin glulisine). Humalog and NovoLog continue to lead the class: Humalog posted $770 million in 2Q18 and $792 million in 1Q18, reflecting 47% of total market revenue in both quarters, while NovoLog posted $752 million in 2Q18 and $776 million in 1Q18, reflecting 46% of total market revenue in both quarters. Notably, Humalog drove class growth in 2Q18, rising 14% YOY while NovoLog dropped 6% YOY. Lilly management cited lower Medicaid utilization (i.e. more favorable segment mix) and higher realized price as reasons for the boost in Humalog sales – this is troubling from a patient access perspective, as it implies that Medicaid beneficiaries are having a harder time securing their Humalog and might be forced to switch to another bolus insulin. It’s also yet another sign of how pricing pressure is plaguing the insulin market, creating a lose-lose scenario for patients (who face higher out-of-pocket costs and restricted choice) and for manufacturers (who take a loss on net price and have less to invest in R&D, given large rebates paid to the middlemen, which we see as largely a waste of money).

      • Turning to the other players in this market, Sanofi’s Apidra held 6% of total class revenue in 2Q18 with $107 million sales (-1% YOY). The newly launched Fiasp held 1% of class revenue in 2Q18 ($21 million, +65% sequentially) and 1Q18 ($14 million). Read more on Fiasp immediately below the graph.

    Rapid-Acting Insulin Market (1Q06-2Q18)

    • We expect Fiasp sales to take off in future quarters and drive growth for this class. As a reminder, 1Q18 was the first quarter that Fiasp was on the US market – and it was a partial quarter at that, since the drug became available in US pharmacies in February 2018 (after a rollout in Canada and Europe in mid-2017). As the first next-generation mealtime insulin on the market, Fiasp offers meaningful improvements for patients in terms of flexible dosing (can be taken up to 20 minutes after the start of a meal), convenience, and reduced hypoglycemia risk. Moreover, Fiasp is priced at parity with NovoLog, which should help ensure broader access, and should certainly make this a popular option among patients (a better patient experience at the same price point). Of course, because of how complex (often convoluted) drug pricing is in the US, it’s hard to discern how many patients are really able to access Fiasp at the same rate as NovoLog – we hope to learn more about this as Novo Nordisk continues Fiasp’s roll out in the states. Kelly’s insurance company Aetna has refused to let her use Fiasp despite the pricing and we imagine Aetna gets a far better deal on Novolog and Humalog. That said, patient enthusiasm seems to be building around Fiasp. Market research firm dQ&A conducted a survey in spring 2017 that found 23% of respondents taking rapid-acting insulin (n=2,312) said they would “definitely” switch to Fiasp if it was suggested by their doctor and reimbursement was similar to their current mealtime insulin. An additional 55% said they would “likely” switch. There is certainly a significant demand for these agents – it’s now a matter of working around the numerous existing barriers in the system to ensure that this interest/demand is met.

    • Pricing pressure remains a significant challenge for the rapid-acting insulin class. On nearly every earnings call from an insulin company in 1Q18 and 2Q18, we heard management lament pricing pressure in the way of higher PBM/payer rebates as the primary headwind preventing growth for the insulin market as a whole. As we understand it, these pricing pressures disproportionately affect the rapid-acting insulin class: PBMs and payers view mealtime insulins as essentially interchangeable, leading to exclusive contracts that beget higher rebates from manufacturers in order to ensure preferred formulary positioning. Unfortunately, we don’t see this issue resolving anytime in the near future, which puts a damper on projected growth rates for the class. That said, we urge all stakeholders to continue working (hopefully, collaboratively) toward a solution, because insulin pricing is a problem that needs to be addressed – we can’t imagine PBMs have so much power!  As step one, it’s crucial to generate more transparency around PBMs. Countless diabetes thought leaders have criticized these “middlemen,” who take rebates and exacerbate pricing pressure without offering any clarity on their value to the healthcare system.

    • Looking forward, we suspect the ultra-rapid-acting insulin class will grow in number in the coming years. Lilly has an ultra-rapid acting candidate in phase 3 (wrapping up now) and Adocia's ultra-rapid BioChaperone Lispro is phase 3-ready; both would join Fiasp and MannKind’s inhalable Afrezza in this category. Speaking of Afrezza, its sales have been relatively low to-date, far below 1% of the rapid-acting class ($3.8 million in 2Q18). FDA added an ultra-rapid-acting claim to the Afrezza label last year, and we’re watching to see if this can boost uptake. We’re excited to see how newer, more efficacious agents in this class can benefit patients. To be sure, the current offerings for rapid-acting insulin aren’t good enough, or fast enough (it’s ironic that a class defined by competitive pricing pressure is actually an area in need of major innovation, since “pricing pressure” doesn’t exactly inspire major R&D investment). Faster-acting insulins also have potential applications in closed loop systems, which is another future direction for this therapy class.

    • Lastly, Sanofi has yet to break out sales for its first-to-market biosimilar mealtime insulin Admelog (biosimilar lispro), but we’ll be looking for this in future quarterly updates from the company. We estimate that Admelog brought in ~$18 million in 2Q18, following a 1Q18 US launch. This product was first approved and launched in Europe, under the brand name Insulin lispro Sanofi. We’re very curious to see how a biosimilar affects market dynamics on the rapid-acting insulin side – will this mirror patterns in the basal insulin market, where biosimilar Basaglar gained favor from payers over the original Lantus? Lilly management perceives little threat to its Humalog business (as CEO Mr. Dave Ricks explained at JPM 2018, price is already on the floor for rapid-acting insulins, so there’s little room for competition). On the other hand, we note that Medicaid utilization of Humalog is declining, while Sanofi’s first priority for Admelog distribution is the Medicaid channel.

    Human Insulin Market (Novo Nordisk, Sanofi, & Lilly) Fluctuates around ~$740 Million/Quarter; +4% YOY in 1Q18, -5% YOY in 2Q18

    • Human insulin sales from Novo Nordisk, Sanofi, and Lilly fell 5% YOY to $739 million in 2Q18 after rising 4% YOY to $746 million in 1Q18. Human insulin reflected 14% of the total insulin market in both quarters. Revenue from Novo Nordisk’s human insulin declined 7% YOY in 2Q18 to $366 million, while sales of Lilly’s Humulin dipped 3% YOY to $346 million and sales of Sanofi’s Insuman dropped 18% YOY to $27 million. In 1Q18, these reported sales were $391 million (-9% YOY), $326 million (+4% YOY), and $30 million (-11% YOY), respectively. As of 2Q18, Novo Nordisk’s human insulin leads the human insulin market with 50% share by value, followed by 47% share for Humulin and 4% for Insuman. As noted, as pricing pressures continue to hinder patient access to newer insulins, we have heard thought leaders such as Dr. Irl Hirsch predict that prescription volumes for human insulins will rise in 2018. Although sales figures do not reflect volume for human insulin, we do expect this class to post steady revenue for the foreseeable future.

    GLP-1 Agonists

    GLP-1 Agonists Rise ~25% YOY to $1.9 Billion, Matched by Volume Growth; Drive 29% and 34% of Industry Growth; Accumulating Evidence for CV Class Effect

    • The GLP-1 agonist class continued on its impressive growth trajectory in 1H18, rising 30% YOY to $1.9 billion in 1Q18 and then 20% YOY to $1.9 billion in 2Q18. Sequentially, the market fell 1% in Q1 and grew a modest 2% in Q2. While these results aren’t nearly as positive as the 19% sequential rise in 4Q17, the slowing down of sequential growth isn’t unexpected for a class that’s now been on the market for more than a decade. The financial performance of GLP-1s remains remarkable overall, in our view: the class is consistently growing by double-digits YOY from a high base, and in 2Q18, drove 55% of growth in the entire diabetes drug industry (it contributed a still very high 44% share of growth in 1Q18). That said, GLP-1s were even more powerful growth drivers last year, contributing a 74% share of growth to the diabetes drug industry in 1Q17 and a 67% share in 2Q17. For context, the GLP-1 class posted YOY growth rates of 38% in 4Q17, 25% in 3Q17, 29% in 2Q17, and 35% in 1Q17. This market is on the verge of topping $2 billion in pooled quarterly sales and could potentially hit $8 billion in total revenue for 2018 (up from $6.5 billion in the full-year 2017). Total prescription volume of GLP-1s shows no signs of slowing down, and it’s reassuring to see that volume increases are mirroring revenue increases (though there’s still so much more do be done in improving access to advanced GLP-1 therapies). What’s behind GLP-1 market growth rates? For one, manufacturers continue to innovate in this class, with oral administration next on the horizon (more on this below). PBMs/payers perceive more heterogeneity among GLP-1 products compared to insulin products (in part due to ongoing innovation), which results in less pricing pressure and lower rebates in this category. New CVOT data is highlighting the cardioprotective and renal protective and “beyond-A1c” benefits to GLP-1 agonists, and we have reason to believe that cardioprotection is a class effect of these agents. Lastly, the new market entry of Novo Nordisk’s Ozempic (semaglutide once-weekly) in 1Q18 spurred underlying class growth.

      • By product, Novo Nordisk’s Victoza (liraglutide) maintained market leader status in the first half of 2018, claiming 53% of the total GLP-1 market by value in 1Q18 ($989 million sales) and 47% in 2Q18 ($898 million). Lilly’s Trulicity (dulaglutide once-weekly) has experienced tremendous growth in recent quarters and is not far behind: Trulicity captured 36% value share in 1Q18 ($678 million sales) and 41% value share in 2Q18 ($780 million). It’s worth noting that Trulicity appears to be driving the majority of underlying class growth for GLP-1s, and that the gap between Victoza’s and Trulicity’s market share has been progressively closing (in 2Q17, these numbers were much farther apart at 56% and 30%, respectively). These two products are the established giants in the GLP-1 commercial landscape, but we expect to see disruption from Novo Nordisk’s second-gen Ozempic, which already holds 2% of the market by value as of 2Q18 (posting $31 million in its second quarter on pharmacy shelves and $11 million in the first quarter), especially as liraglutide’s patent expiry is coming up in 2022/2023. AZ’s once-weekly Bydureon (exenatide) held steady at 8% market share by value in 2Q18 ($155 million) and 7% in 1Q18 ($139 million), while twice-daily Byetta (exenatide) reflected 2% of pooled revenue in 2Q18 ($29 million) and in 1Q18 ($31 million). Finally, GSK’s Tanzeum (albiglutide once-weekly; recently discontinued – but more on this below) and Sanofi’s Adlyxin (lixisenatide) captured the remaining 1% of the GLP-1 market in both quarters.

    Pooled GLP-1 Agonist Sales (1Q06-2Q18)

    • The GLP-1 class continues to outpace the SGLT-2 inhibitor class, posting similar YOY growth rates (20% vs. 21% in 2Q18 for GLP-1s and SGLT-2s, respectively) from a much higher base ($1.9 billion vs. $1.0 billion in 2Q18). By no means is this a knock against SGLT-2 inhibitors, which themselves continue to show very meaningful growth despite key headwinds (e.g. amputation-related concerns for Invokana, concerns about UTIs and other adverse events early on). Rather, we note this comparison to highlight the truly impressive performance of the GLP-1 class, which is able to match the newer SGLT-2 class that we would’ve expected would’ve grown faster being earlier in their launch cycle.

      • We attribute this success in large part to ongoing innovation from GLP-1 manufacturers: Auto-injectors (e.g. the IDEO-designed Trulicity pen and Bydureon BCise) have made GLP-1 therapy more convenient and patient-friendly than ever before; the proliferation of once-weekly agents has lowered hurdles to GLP-1 use, namely by reducing injection burden; and the rollout of Ozempic is sure to stimulate additional class growth, especially because thought leaders agree that semaglutide is more potent than other GLP-1 molecules (in glucose-lowering and weight loss). In fact, there’s so much innovation happening in the GLP-1 category that many products in the class are quite early in their launch cycle, and different products are differentiated based on their device, dosing frequency, potency, etc. Going forward, the expected approval and launch of oral semaglutide could also serve as an important landmark in bringing even more patients into the fold though we are not clear what it will do to the market since pricing may be closer to an oral than a GLP-1. Novo Nordisk is in the midst of reporting phase 3 data on oral sema from the PIONEER program; the company is targeting 2019 for an FDA submission.

    • Novo Nordisk’s Ozempic (once-weekly semaglutide) is primed to shake up the GLP-1 market – we believe its ongoing launch will be crucial in stimulating at least volume growth for the class as a whole. Excitement around Ozempic’s launch was palpable in 1H18, with it being mentioned 50 (!) times in the span of Novo Nordisk’s hour-long 2Q18 earnings call. The product generated $31 million revenue in 2Q18 ($11 million in 1Q18) and it’s the question of the quarter for 3Q18 results, coming up soon. To be sure, Ozempic could very well reshuffle the current status quo within the class: SUSTAIN 7 showed Ozempic’s superiority to Lilly’s Trulicity (its greatest competitor in terms of once-weekly GLP-1 agonists) in terms of glucose-lowering and weight loss. Novo Nordisk has already indicated that it has prioritized marketing of Ozempic over Victoza, and we imagine commercial resources will continue to be shifted over as liraglutide approaches loss-of-exclusivity in 2022/2023. Importantly, we think Ozempic’s greatest influence on the GLP-1 market will be expanding the class rather than stealing share though we do think it will steal some share. There is so much room for growth, considering how few patients with type 2 diabetes are currently taking a GLP-1 agonist. We hope to see enhanced patient access/affordability to parallel this expansion of patient choice. Moreover, we want to emphasize how Ozempic’s market entry offers a unique opportunity for personalized diabetes treatment. We remain curious to see how HCPs will incorporate Ozempic into their practices and are eager to see payer coverage for Ozempic increase in the coming years.

    • As CVOT data on GLP-1s has accumulated, there’s more talk than ever before on GLP-1 agonists as cardioprotective medicine, offering key benefits beyond A1c reduction. HARMONY-Outcomes for Tanzeum (albiglutide) was the latest to make headlines. Indeed, at EASD earlier this month, we learned that Tanzeum reduced risk for three-point MACE (non-fatal MI, non-fatal stroke, CV death) by 22% vs. placebo (HR=0.78, 95% CI: 0.68-0.90) in a population with type 2 diabetes and established CVD. Notably, albiglutide is known for its lower-than-average potency on glucose and weight, which makes this evidence for a CV class effect that much stronger (see our EASD coverage for a much deeper dive on this class effect debate). GSK officially discontinued Tanzeum earlier this year, after “withdrawing support” from the business in 2Q17, and we’re eager to see another company pick it up in the wake of these positive CVOT results though we are not at all sure it will happen given that it is said to be expensive to manufacture and since the factory is now being used for something else. Though it’s never been a major commercial player in the GLP-1 market, we would like to see greater GLP-1 access for patients, especially those at high risk for CV morbidity/mortality.

      • Coming up is the REWIND CVOT for Lilly’s Trulicity, expected to read out in 4Q18. We’re excited to hear topline results. REWIND boasts a unique study design with a longer follow-up period and very large (69%) primary prevention cohort – that is to say, it’s well-positioned to offer insight on key lingering questions in the field, starting with “do GLP-1s have a role to play in primary CV prevention?”. Positive REWIND results and a CV indication for Trulicity could dramatically alter the landscape by broadening the patient population that could officially reap CV benefits from the class – that said, it is a different population being studied and we cautious against over-comparison of the CVOTs.

      • We are glad to see FDA taking a more progressive view in streamlining future CVOTs. In major news during Novo Nordisk’s 2Q18 earnings call, the company announced that FDA will allow it to conduct one superiority CVOT for both (i) oral semaglutide and (ii) injectable Ozempic. If the PIONEER 6 CVOT (data expected by year-end) demonstrates CV superiority of oral semaglutide vs. placebo, then FDA will accept aggregated data from SUSTAIN 6 and PIONEER 6 as evidence supporting a CV indication for both oral semaglutide and Ozempic. As a result, the originally planned post-market SOUL CVOT for Ozempic has been cancelled. We see this as a major win for Novo Nordisk and are extremely pleased to see FDA display this type of systems innovation to streamline the (costly) CVOT process. Furthermore, we wonder how FDA will approach the notion of a cardioprotective class effect for GLP-1s, should more consensus emerge on that front. Will the agency further relax the CVOT requirement if there’s agreement and scientific evidence suggesting a class-wide benefit? Will the CVOT requirement be dropped at the upcoming October 24-25 meeting?  

    • Oral semaglutide looms large, as all 10 trials of the PIONEER program are expected to report by end of 2018. If results from these trials are positive, Novo Nordisk plans to submit an NDA in 2019. Results from more than half of these studies (PIONEER 1, PIONEER 2, PIONEER 3, PIONEER 4, PIONEER 7 and PIONEER 10) have already read out and have been characterized as “positive across the board” by Novo Nordisk management. Expectations are sky-high for this first oral GLP-1: some analysts  believe that oral semaglutide could, once established on the market, achieve $5 billion (!) in annual revenue. For context, the current market leading Victoza brought in ~$3.6 billion in sales in 2017. We believe oral semaglutide will bring more patients to the GLP-1 class who are currently reluctant about injectable therapy.

    Basal Insulin/GLP-1 Agonist Fixed-Ratio Combinations

    Basal/GLP-1 Combos Bring in $147 Million Total in 1H18, Exceeding Total Sales for 2017; Driven by Xultophy in 1Q18, Soliqua 2Q18; Xultophy’s EU Label Gains DECLARE + LEADER Data

    • The two basal insulin/GLP-1 agonist fixed ratio combinations on the market – Novo Nordisk’s Xultophy (insulin degludec/liraglutide) and Sanofi’s Soliqua (insulin glargine/lixisenatide) – posted a combined $147 million in sales for 1H18, nearly tripling YOY from $53 million in 1H17 (the class’ first two quarters on the market). $147 million is also more than the class sold for the entirety of 2017 ($142 million), and we believe that both therapies have heaps of headroom for continued growth (more on that below). These numbers reflect 252% and 140% YOY growth in the first two quarters of 2018 from low bases; sequential growth was less impressive at 24% and 19% for 1Q18 and 2Q18 respectively. As such, we continue to be somewhat underwhelmed by the performance from this highly-anticipated class, particularly given how early Xultophy and Soliqua are in their launch cycles and how much they were praised by the KOLs. Certainly, sequential growth since both products’ launched has been relatively unimpressive at 75% in 2Q17, 12% in 3Q17, 45% in 4Q17, 24% in 1Q18, and 19% in 2Q18. That said, we were encouraged by a strong 2Q18 for Soliqua, which drove 69% of sequential class growth after two relatively flat quarters in 4Q17 and 1Q18, albeit from a low base; Soliqua sold $11 million in 1Q18 (+125% YOY, 0% sequentially) and $20 million in 2Q18 (+240% YOY, 89% sequentially). Xultophy drove 100% of class growth in 1Q18 with 228% YOY and 26% sequential growth to $56 million; in 2Q18, Xultophy rose 111% YOY and 13% sequentially to $60 million. We’re optimistic that these products are finding their footing in prescriber minds and treatment algorithms.

    Pooled Basal Insulin/GLP-1 Agonist Fixed-Ratio Combination Sales (1Q17-2Q18)

    • At the end of 2Q18 and on the tide of Soliqua’s surge, Xultophy accounted for 75% of this market by value, down from 83% in 2Q17 and 84% in 1Q18. We have also heard that Soliqua actually accounts for a majority of prescriptions in this market (as per Zealand CEO Ms. Britt Meelby Jensen), apparently reflecting significant differences in realized price between Sanofi and Novo Nordisk’s products.
    • In June, Novo Nordisk secured EMA approval to add LEADER (CV benefit with Victoza vs. placebo) and DEVOTE (hypoglycemia benefit with Tresiba vs. Lantus) data to the Xultophy label, following a positive CHMP opinion in 1Q18. We suspect this could meaningfully boost EU prescriptions/sales, particularly if Novo Nordisk can leverage the update into better access and reimbursement. An equivalent request has been submitted to FDA, but we’re less certain that the US agency will be receptive. Though we wouldn’t underestimate Novo Nordisk’s ability to engage in productive conversations with the agency, FDA is generally less flexible than EMA on this type of request. Moreover, in line with this label update, we know that Xultophy’s clinical profile is stronger than that of Soliqua (though no head-to-head RCT has been conducted) though some patients get around this by taking Soliqua slightly more often. Tresiba and Victoza, the two ingredients of Xultophy, are clinically superior to the Lantus and Adlyxin that comprise Soliqua. Victoza has a CV indication and DEVOTE demonstrated that Tresiba carries a 40% reduced risk of severe hypoglycemia vs. Lantus. Ultimately, however, there’s no doubt that Soliqua is a better option than basal-bolus regimens or taking GLP-1 and basal insulin separately. We’ve also heard some HCPs say that most of the gap between Soliqua and Xultophy can be made up by taking Soliqua at “hour 18,” though this is not an official recommendation and obviously places some additional burden on patients.

    • We passionately believe that faster uptake of these products is more than justified given their superior glucose-lowering efficacy, greater patient convenience/lower injection burden, and milder side-effect profiles. For example, a meta-analysis across the DUAL program found that ~80% of patients randomized to Xultophy achieved A1c <7% without weight gain or severe hypoglycemia compared to only ~37% of patients on standalone liraglutide and only ~15% on insulin glargine alone – this is truly exceptional! Moreover, fixed-ratio dosing of basal insulin and GLP-1 allows for lower doses of insulin especially and mitigates side-effects associated with each – that is, less hypoglycemia and weight gain than basal insulin alone (even weight loss with Xultophy) plus less GI discomfort than with GLP-1 alone. In addition to mitigating side-effects, the lower injection burden of fixed-ratio combination treatment (one injection with one pen per day) can also improve adherence, thereby increasing treatment satisfaction and leading to better patient outcomes. For example, PROs from DUAL VII favored Xultophy over basal-bolus in terms of willingness to continue treatment after 26 weeks, treatment burden, and diabetes management.

    • Despite the wealth of evidence supporting the safety and efficacy of basal/GLP-1 combinations, many barriers to uptake persist. As we see it, little of the thought leader enthusiasm for this class has translated to clinical uptake, and we’ve come to understand very real (if surmountable) logistical and psychological barriers to greater use. We would love to see Sanofi and Novo Nordisk invest collaboratively toward the goal of addressing prescriber hesitation around the rigidity of these combinations, limitations of indications in the US, and reimbursement, to name a few (we aren’t sure that would be legal – it sounds hard to plan). Sanofi has already begun engaging on this front, sponsoring peer-to-peer medical education to familiarize HCPs with the concept of a fixed-ratio combination, including its patient-friendly side-effect profile. In the words of Sanofi’s former EVP of Diabetes & Cardiovascular Mr. Stefan Oelrich, these programs have been undertaken to combat “clinical inertia.” In our opinion, the barriers against basal/GLP-1 combo uptake are even more complicated than this blanket term might indicate. We’ve identified four key factors keeping the class from growing more rapidly (and we certainly believe the time and effort required to overcome these is more than worth it):

      • HCPs are reluctant to prescribe fixed combination therapy, which is rigid in both its components and titration. Prof. Philip Home recently explained to us that some providers may want to combine Sanofi’s Lantus with Novo Nordisk’s Victoza, or even Lilly’s Trulicity, which are the most familiar and widely used molecules in each class. Indeed, we’ve heard from Novo Nordisk reps in exhibit halls that the company has, for the time being, de-prioritized Xultophy to first build familiarity with Tresiba and Victoza individually – not enough patients are being prescribed these efficacious therapies as it stands, much less a combination of the two (though, in our view, they fill different treatment needs).

      • Also according to Prof. Philip Home, many prescribers view GLP-1s as a “pre-insulin” therapy, due to relative ease of use and cleaner side-effect profile. As such, a fixed-ratio combination of a GLP-1 with insulin might seem illogical to some, though we also know many are now using GLP-1+ basal as a safer and superior alternative to basal-bolus regimens.

      • US labels (Soliqua; Xultophy) require that a patient already be on basal insulin or the specific GLP-1 component of the combination therapy. This is particularly prohibitive for Sanofi, since so few type 2s take Adlyxin (lixisenatide); in contrast, Victoza (liraglutide) is the market-leading GLP-1 agonist. However, ~two-thirds of fixed-ratio prescriptions do go to Soliqua (we’re curious how many come to combos from basal insulin vs. GLP-1). Notably, the EU indications (Suliqua; Xultophy) are much less restrictive – the combinations can be prescribed as second-line after metformin/oral therapy, without prerequisite basal or GLP-1 usage. 

      • Reimbursement. Historically and in our observation, Sanofi has invested quite a bit more than Novo Nordisk in promoting fixed-ratio combination therapy and in securing access for Soliqua, though our sense is that this is starting to change at Novo Nordisk. We know that the US comprises ~80% of Soliqua sales, but the breakdown for Xultophy isn’t clear in Novo Nordisk’s earnings materials. Anecdotally, we’ve heard that Xultophy is doing very well in European markets where it has gained access, particularly in France, leading us to suspect that better reimbursement in the US could aid uptake.

        • As it stands, Soliqua has the advantage of a meaningfully lower list price than Xultophy. At our local CVS, a five-pen box of Xultophy pens is priced at $1,170, compared to five Soliqua pens for $794 – of course, this only means so much before rebates/insurance and prices vary by pharmacy, but we were truly struck by this difference. A quick scan of ten pharmacies in different locations around the US confirms that the average list price difference between the two agents is ~$310. 

    SGLT-2 Inhibitors

    SGLT-2 Inhibitors rise ~31% YOY and Drive 31% of Industry Growth; Sequentially Flat at -2% and +4%; Amputation Concerns Hinder Class Despite Cardio- and Renal-Protection

    • The SGLT-2 inhibitor class surged 41% YOY in 1Q18 (from $715 million in 1Q17) and 21% YOY in 2Q18 (from $857 million in 2Q17), to just over $1 billion in sales for each quarter. Sequentially, however, SGLT-2s were flat in the first half, posting a disappointing 2% sequential decline in 1Q18 offset by 4% sequential growth in 2Q18. While these YOY numbers reflect encouraging growth for the class from 2H17, it’s clear that very little of that growth came in 1H18. Indeed, seriously slowed underlying class growth (to ~5% annually) was widely referenced by management across calls for SGLT-2 manufacturers in the first half, though different companies cited a variety of disparate reasons for this stagnation (see below). Interestingly though, we noted that these flatter sales for the half formed a trend similar to that seen in in 1H17; in 1Q17, SGLT-2s posted 18% YOY growth and a 15% sequential drop, followed by 20% YOY and sequential growth in 2Q18. While we don’t see the class returning to  the tremendous 44% YOY growth it saw 2016, we do think SGLT-2s could match the 24% growth they posted in 2017 overall. To be sure, growth remains meaningfully lower than we’d expect for a class with three positive CVOTs, one product with a CV indication on its label (Lilly/BI’s Jardiance), and a new product on the market (Merck/Pfizer’s Steglatro). Our sense is very much that safety concerns, particularly with regard to amputations, have hindered both patient and provider interest in the class. However, we’re optimistic that new outcomes trials supporting CV (DECLARE) and renal (CREDENCE) benefits can strengthen the tailwinds driving the SGLT-2 class; with time, we think more providers will come to accept and capitalize on these impressive and meaningful effect.

      • As it stands and by our calculations, SGLT-2 inhibitors comprised 8% of the entire diabetes market in both 1Q18 and 2Q18, also driving 19% of overall industry growth in each quarter. Notably, this roundup is the first to include four players in the SGLT-2 market: Lilly/BI’s Jardiance (empagliflozin), AZ’s Farxiga (dapagliflozin), J&J’s Invokana (canagliflozin), and Merck/Pfizer’s Steglatro (ertugliflozin).

    Pooled SGLT-2 Inhibitor Sales (1Q13-2Q18)

    • At the close of 2H18, Jardiance continued to lead the class with 43% value share at $445 million in sales, followed by Farxiga (33%, $340 million), Invokana (21%, $215 million) and newcomer Steglatro (4%, $37 million). This represents a slight shift from 1Q18, when Jardiance claimed 45% value share with $456 million in sales, Farxiga 30% ($299 million), Invokana 25% ($248 million), and Steglatro 1% ($8 million). Even more striking is how much this breakdown has shifted from one year ago, when Jardiance held 36%, Farxiga 29%, and Invokana 34%. Essentially, it seems as though Jardiance and to a lesser degree Farxiga are seriously benefitting from the decline of Invokana, which has been driven by the therapy’s association with lower limb amputations in CANVAS. Of course, the story has other critical nuances. Jardiance’s has almost certainly benefitted from gaining a CV indication in the US and EU, based on positive results from EMPA-REG OUTCOME; Lilly management also attributed Jardiance’s “flat-ish” half (+6% sequentially in 1Q18 but -3% sequentially in 2Q18) to changes in rebate estimates and discount normalization (which we presume to mean volume continues to grow). We’ll be interested to see if Invokana can be buoyed by its recently-approved CV indication in the EU (based on positive results from CANVAS); still, the indication remains under FDA review in the US, which is arguably more important to Invokana’s success. Growth would be a welcome change for Invokana, which has struggled commercially since FDA added black-box warning for amputations to the drug. J&J has cited increased discounts, higher rebates, and share loss in explaining Invokana’s US-driven 27% YOY decline in 2Q18, and we very much think these have been driven by J&J’s weakened negotiation position on the heels on amputation concern. Of note, another critical boon for Invokana and the SGLT-2 class should come from J&J’s renal outcomes trial for Invokana, CREDENCE, which was stopped a year early for meeting its primary endpoint ~1 year ahead of schedule (more below). In some ways, Farxiga has fared the best in 1H18, finishing the half with an all-time high revenue of $340 million, up 36% YOY and 14% sequentially. While AZ cited volume-driven growth OUS (Farxiga continues to lead the class with 41% volume share) and contract-driven growth in the US, it also seems that Farxiga’s gain in revenue came at the expense of lower participation in affordability and discount programs vs. 1H17 – a disappointing compromise to achieve growth, in our opinion. The newest player in the game – Merck/Pfizer’s Steglatro – generated an estimated $45 million in sales in 1H18, and we note that the product launched about one-third of the way through 1Q18 (late January for Steglatro + Steglujan, February for Segluromet).

      • We estimate total global sales for the Jardiance franchise, since only Lilly’s share (and not BI’s) is reported publicly. We’ve based our calculations on the assumption that Lilly’s portion represents ~33% of total sales, given that BI listed global net sales for the franchise at €165 million, or ~$183 million, in a 2015 diabetes update. In that same year, Lilly collected $60 million in franchise revenue (we recognize this may be slightly imprecise!).

    • By volume, Farxiga leads the class at 41%, followed by Jardiance at 36% and Invokana at 16%. This is according to AZ’s 1Q18 presentation (slide 17), which shows that Jardiance’s volume share has been climbing as Invokana’s falls, with Farxiga holding relatively steady; Farxiga also held 41% volume in 2Q18. We imagine Jardiance outpaces Farxiga in value because Farxiga continues to garner the majority of sales OUS, where realized price is lower; similarly, Invokana maintains a slightly higher value share through a higher proportion of US sales. Volume changes match the value trends noted above, in which Jardiance is “stealing share” from Invokana as Farxiga grows at pace with the overall market.

    • Currently, SGLT-2s face a range of strongly opposing forces influencing the class’ growth (cardio-renal protection, amputations, DKA risk), creating a storm of headwinds and tailwinds that has somewhat becalmed the class. On the one hand, the convenience of taking an oral tablet offering the benefits of glucose-lowering, weight loss, blood pressure reduction, and CV and (now) renal protection is hard to beat. That said, we do very much think that commercial performance is lagging far behind thought leader enthusiasm for SGLT-2s. Our sense is that the class overall has been hindered by ongoing safety concerns, particularly with regard to the amputation signal seen in CANVAS, and now a class-wide label warning for Fournier’s gangrene – an exceedingly rare but highly concerning complication. There were several notable developments on these fronts in 1H18:

      • Topline results from DECLARE (AZ’s CVOT for Farxiga) were announced in September, revealing that the SGLT-2 inhibitor was superior vs. placebo on one co-primary composite endpoint of hospitalization for heart failure and CV death. On the other co-primary endpoint of three-point MACE, Farxiga was non-inferior to placebo (according to AZ, results trended in favor of Farxiga); as a reminder, DECLARE enrolled a majority (59%) of participants without baseline CVD. AZ will reportedly file for an indication based on these results, but no specifics on the potential label change were given. As it stands, current SGLT-2 CV indications for J&J’s Invokana (approved in EU only) and Lilly/BI’s Jardiance (approved by both EMA and FDA) are exclusively for those with established CVD. However, because the majority of the population enrolled in DECLARE did not have established CVD (they did have CV risk factors), it’s very possible that regulatory agencies will be willing to include primary prevention patients in an indication; what the precise indication will be for (CV death and/or heart failure) remains to be seen. Either a HF indication or extension to primary prevention would be a huge development and big win for patients and the class, in addition to allowing for a much wider prescribing base for Farxiga compared to competitors. Of note, all-important safety data were not shared in AZ’s press release, apart from a broad statement that DECLARE confirmed the well-established safety profile of Farxiga. Amputations and DKA are the primary concern with SGLT-2s, and we tentatively expect that DECLARE is likely to demonstrate safety on these endpoints given the overall healthier enrolled population. Full results will be presented November 10 at AHA 2018.

        • With DECLARE reading out positively, the needle skews ever-further in favor of cardioprotection as a class effect of SGLT-2 inhibitors. The lack of superiority on three-point MACE will likely be an enormous point of discussion, but our early sense is that this result can – as in EXSCEL for AZ’s GLP-1 Bydureon – be attributed to an overall healthier enrolled population. In addition to DECLARE, the EMPA-REG OUTCOME (Jardiance) and CANVAS (Invokana) CVOTs support cardioprotection for SGLT-2s, while real-world evidence from AZ’s CVD-REAL and CVD-REAL 2, as well as J&J’s EASEL offer further support.

      • CREDENCE, J&J’s renal outcomes trial for SGLT-2 inhibitor Invokana, was stopped over a year early (in July) due to overwhelmingly positive results. It’s hard to overstate the implications of these results; diabetic kidney disease (DKD) is notoriously challenging to treat, and SGLT-2s could fill an enormous unmet need by slowing the progression of renal dysfunction. Moreover, this CREDENCE result bodes well for Lilly/BI’s EMPA-KIDNEY (Jardiance in CKD) and for AZ’s Dapa-CKD (Farxiga in CKD). J&J management stated their intention to file for a DKD indication in 2019 in a recent investor call, and we could not be more eager to see the full results (though no timeline has been given).

      • SGLT-2 safety concerns remained at the forefront of discussion in 1H18, as the FDA added a class-wide warning for necrotizing fasciitis of the perineum (Fournier’s gangrene) to the labels for all SGLT-2 containing compounds. Moreover, a provocative cohort study identified a class-wide amputation signal with SGLT-2s compared to GLP-1 agonists. The former, which received coverage from major news outlets (CBS, Drugwatch, Forbes), was based on 12 cases of the life-threatening bacterial infection in patients taking an SGLT-2 inhibitor between March 2013 (when the first class member was approved) and May 2018. While some may view a class-wide label change as a drastic measure for so few cases of a complication (the FDA warning noted that an estimated 1.7 million people in the US received a dispensed prescription for an SGLT-2 inhibitor in the US in 2017), we are very glad that this complication was identified and that this information is out to patients. Our sense is that progression to such a severe complication could be prevented with better patient education and access to care. On amputations, at ESC 2018, Karolinska’s Dr. Peter Ueda presented an observational cohort study (N=34,426) identifying ~double the risk of lower limb amputations (HR=2.23, 95% CI: 1.37-3.91) among patients starting on an SGLT-2 inhibitor vs. a GLP-1 agonist. What’s more is that the study was primarily comprised of patients on Jardiance and Farxiga, with only 1% on Invokana. Thought leader opinion on amputations as a class effect remains mixed; for example, Dr. Jay Skyler has said that he prefers to prescribe any SGLT-2 inhibitor other than Invokana due to the amputation signal seen in CANVAS, while others have postulated a possible class effect. While the question of class effects remains hotly debated, our understanding is that the vast majority of KOLs consider these safety concerns manageable with proper patient selection, education, and monitoring.

    DPP-4 Inhibitors

    DPP-4 Inhibitors Achieve ~5% YOY Uptick in 1H18 (+1% and +9% YOY), Remain Flat Sequentially; Manufacturers Emphasize US Pricing Pressure Holding Sales Steady at ~$2.5 Billion Quarterly

    • The DPP-4 inhibitor market saw a slight uptick in sales, rising 9% YOY in 1Q18 to $2.4 billion and 1% YOY in 2Q18 to $2.5 billion in revenue. Sequentially, though, revenue was essentially flat: Sales rose 6% in 2Q18 to balance a 6% sequential drop in 1Q18. Of course, these fluctuations around a steady class-wide pooled revenue of ~$10 billion annually have come to characterize the DPP-4 class over the past several years. Each quarter, the class hovers around the $2.5 billion mark, with an ever-so-slight trend toward positive growth trend. The DPP-4 class remains an established giant in the diabetes landscape and comprised 45% of total branded, non-insulin diabetes drug revenue in both 1Q18 and 2Q18 – this represents the highest share of the market of any drug class (including GLP-1s, SGLT-2s, glucagon, and GLP-1/basal insulin combos). We do note, however, that this 45% mark is the lowest for the class since 2010, and a far cry from the 72% it held in 2013. Of course, this reflects the rise of newer therapy classes rather than lower utilization of DPP-4s; our understanding is that DPP-4 sales by volume actually continue to rise. But as both GLP-1s and SGLT-2s have demonstrated benefits beyond glucose-lowering, such as cardioprotection, renal protection, and enhanced weight loss – while also providing more efficacious glucose-lowering – it’s not surprising that DPP-4 growth would be somewhat hindered and lose market share. Indeed, the class-wide theme has emerged that pricing pressure in the US has hindered revenue growth (pushing down realized price as volume increases), while OUS revenue has grown modestly. Either way, these agents certainly retain a prominent position in modern diabetes treatment algorithms. In particular, DPP-4s have an important role to play early on in the course of diabetes treatment and are considered by many to be the most tolerable diabetes therapy on the market. The clean side effect profile of the class makes DPP-4s very easy to prescribe and will firmly root them in treatment algorithms for the foreseeable future – and we expect the class will only become more commonly used once these agents go generic, particularly early in disease progression.

      • Merck’s Januvia franchise continues to be the familiar frontrunner among all DPP-4s, posting $1.4 billion in 1Q18 and $1.5 billion in revenue in 2Q18. By our calculations, Januvia commanded 60% of total DPP-4 class revenue in 1Q18 and a similarly impressive 61% in 2Q18. Lilly/BI’s Tradjenta followed with 16% of value share in both 1Q18 and 2Q18, with revenues of $392 million and $394 million, respectively. Novartis’ Galvus accounted for 13% of whole class sales in both 1Q18 and 2Q18 ($318 and $332 million), even though it is not marketed in the US, while AZ’s Onglyza held steady at 5% in both quarters ($129 and $126 million). Takeda’s Nesina controlled 5% in both quarters as well ($115 and $127 million).

    Pooled DPP-4 Inhibitor Sales (1Q06-2Q18)

    Diabetes Technology

    CGM

    Global CGM Sales of ~$1.2 billion in 1H18, an Impressive ~60% YOY Rise = Faster Growth from a Higher Base; 1st Ever Half-Year Crossing $1 Billion! Freestyle Libre (Est) Drives ~60% of CGM Category Growth

    CGM Sales by Geography, Estimated (1Q12-2Q18)

    Dexcom, Abbott (estimated), Medtronic (estimated), Senseonics

    By our estimates, the worldwide CGM market grew to ~$1.2 billion in the first half of 2018, up an impressive ~60% YOY from 1H17. This represents faster growth from a much higher base of sales – 1H17 saw sales of ~$759 million rise ~50% YOY. We estimate global CGM sales were ~$561 million in 1Q18 (up ~55% YOY) and a record-smashing ~$652 million in 2Q18 (up ~64% YOY) – the latter was a milestone for the category, crossing $600 million in quarterly sales for the first time in our CGM model. Notably, 4Q17 was the first time crossing $500 million in quarterly sales, meaning 1H18 has been quite the half year for the category. The acceleration is a testament to what could be an inflection point in the field: CGM may finally be moving from an early adopter technology to a broader user base, especially with four excellent sensors on the market, two factory calibrated sensors, two AID systems in the US, balanced US and OUS momentum (especially OUS sales for FreeStyle Libre), and rising clinician enthusiasm. Both 1Q18 and 2Q18 saw international CGM sales exceeding US sales – the first time we’ve ever seen this (more geographic details below). Assuming seasonality, CGM is on pace to handily cross $2.5 billion in sales for all of 2018!

    • By company, we estimate Abbott has provided the majority of the CGM category’s 1H18 growth: ~61% of 1Q18 growth and 58% of 2Q18 growth. Abbott has seen remarkable expansion in its user base, which is now over 800,000 globally and makes it the #1 CGM worldwide by user base. Our model estimates Abbott has provided the largest share of CGM category growth for eight straight quarters now. Dexcom has also had a strong start to 2018, providing ~21% of 1Q18 category growth and 28% of 2Q18 growth. Finally, Medtronic drove ~17% (1Q18) and ~12% (2Q18) of the category’s growth, while Senseonics has provided ~1%. Our model includes Dexcom’s reported sales, our best estimates for Medtronic and Abbott’s FreeStyle Libre (clarified below), and Senseonics’ reported sales. This is the third roundup where we’ve included sales estimates for Abbott’s FreeStyle Libre in the CGM category, which we pegged at roughly ~$212 million in 1Q18 and ~$263 million in 2Q18 – assumptions are below, and we caution these might over- or underestimate FreeStyle Libre’s actual global sales. (Our CGM model shown above includes historical sales estimates back to Libre’s launch in Europe in fall 2014. The BGM section further below reports our estimate for Abbott’s traditional fingerstick business, subtracting out our estimate for Libre.) Meanwhile, Medtronic told us in mid-2017 that its CGM sales are now ~25% of its Diabetes sales, up from ~20% two years ago. We include a more detailed geographic and company-by-company analysis below.

    • International CGM growth continues to be a story in 1H18, following strong momentum in 2017. We estimate a record-high ~53% of CGM sales came from outside the US in 1H18 (~$642 million), marking two consecutive quarters where international sales were greater than US sales. Our model has never before seen this in the CGM category, as shown in the graph above (green line overtaking red line). That said, both US and international sales are seeing sustained category growth, as the slopes of the lines are almost identical.

      • We estimate international CGM sales were ~$305 million in 1Q18 (+81% YOY) and ~$337 million in 2Q18 (+71% YOY), mostly driven by FreeStyle Libre. CGM sales outside the US drove an impressive ~68% of 1Q18 CGM category growth, but then softened to only ~55% of the category’s 2Q18 growth – the latter was the lowest seen in the past two years, reflecting very strong quarters in the US for Dexcom and Medtronic and expanding uptake for FreeStyle Libre in the US. By company, we estimate Abbott provided ~75% of 1H18 CGM category growth outside the US, on par with its 2017 performance. Dexcom supplied the second-largest share of 1H18 OUS CGM growth (~13%), followed by ~10% from Medtronic and ~2% from Senseonics. Clearly, FreeStyle Libre continues to dominate the OUS market and drive most of the category’s growth; indeed, FreeStyle Libre’s OUS CGM sales (estimated) of $219 million in 2Q18 were more than quadruple Dexcom’s OUS business ($53 million) and more than triple Medtronic’s estimated OUS CGM business (~$62 million).

    • US CGM sales were an estimated ~$256 million in 1Q18 (+33% YOY) and ~$315 million in 2Q18 (+56% YOY). The US provided 45% of the CGM category’s growth in 2Q18, its highest point in eight quarters.

      • US growth was led by Dexcom, but showed more balance across the companies vs. 2017. We estimate Dexcom provided 43% of 2Q18 CGM category growth, a notable downtick from the 53%-73% share of growth Dexcom had in the US in in 2017 and the 80%-97% (!) share of growth it had in the US in 2016. As expected, this now reflects a far more competitive US standalone CGM market, which in 1H18 included FreeStyle Libre real-time and Medtronic’s Guardian Connect – products not available in the US this time last year. Indeed, we estimate Abbott provided a strong 39% share of US CGM category growth in 2Q18, while Medtronic contributed 18%. (Both of these are highly estimated, so we could be under- or over-reporting.) We note that Senseonics’ Eversense will start to show up in 2H18 sales in the US, as the first Eversense insertions occurred in July.

      • We expect CGM sales to accelerate significantly in the US in 2H18 for several reasons: (i) Dexcom had minimal contribution from G6 in its already-strong 2Q18 (see below); (ii) Abbott just secured FDA approval for 14-day wear FreeStyle Libre real-time with a 1-hour warmup; (iii) Medtronic’s Guardian Connect only launched in mid-June; and (iv) Senseonics Eversense had the first insertions in late July. Plus, seasonality always means 2H18 is stronger than 1H18 in the US. With all that mind, we expect a larger share of category growth from the US in 2H18, even as CGM adoption continues outside the US.

    CGM Sales By Company, Estimated (1Q12-2Q18)

    Global Sales: Abbott (estimated), Dexcom, Medtronic (estimated), Senseonics

    • We estimate worldwide sales of Abbott’s FreeStyle Libre were roughly ~$212 million in 1Q18 (+116% YOY) and ~$263 million in 2Q18 (+90% YOY) – both came on extremely difficult comparisons to estimated global sales more than doubling in 1Q17 and 2Q17. We estimate FreeStyle Libre’s OUS sales were ~$219 million (+90% YOY) with ~675,000 users in 2Q18, while US sales were ~$45 million in 2Q18 (+100%) with an estimated ~125,000 users. These sales figures reflect our best estimates based on a number of assumptions – many of which could be incorrect. We detail our thinking and assumptions below, beginning with a graph and data table we’ve used to inform our model. It’s been excellent to see the success of this well-priced, easy to prescribe, easy to order, factory calibrated CGM. We can’t wait to see how FreeStyle Libre fares in 2H18, now that it has 14-day FDA approval with a one-hour warmup in the US, and now that Dexcom’s G6 is launching globally with competitive factory calibration. Plus, we look forward to updates on the FreeStyle LibreLink apps in the US and the next-gen FreeStyle Libre with continuous communication. At FFL 2018, Bigfoot CEO Jeffrey Brewer said the latter is coming “sooner than people realize. On Abbott’s 2Q18 call, CEO Miles White confirmed previous guidance for $90-$100 million in 2018 US FreeStyle Libre revenue. Assuming our estimates are not far off, Abbott might actually exceed $1 billion in Libre revenue globally in 2018, along with >1 million patients – that would be something! Read our Abbott 2Q18 report.

    FreeStyle Libre User Base and Estimated Quarterly Sensor Revenue

    • In the graph below, pale colors and text represent projections based on CEO Miles White’s 2Q18 statement that the user base will reach one million patients by the end of 2018. Estimates on the right side assume equal distribution of new users across the remaining quarters of the year and between US and OUS. Given recent patient growth of +150,000-+200,000 per quarter, Abbott seems very well positioned to exceed one million users by the end of 2018 – perhaps even in Q3.

     

     

    1Q17

    2Q17

    3Q17

    4Q17

    1Q18

    2Q18

    FreeStyle Libre User Base

    Global

    “about 300,000”

    ~350,0003

    400,000+

    ~450,000

    650,000+

    800,000+

     US

    0

    0

    0

    ~3,0001

    50,000+

    ~125,0004

     OUS

    ~300,000

    ~350,000

    400,000+

    ~447,000

    600,000+

    ~675,0004

    Estimated sensor-only sales, assuming ~90% utilization

    Global

    ~$90 million

    ~$115 million

    ~$130 million

    ~$146 million

    ~$212 million

    ~$263 million

     US2

    $0

    $0

    $0

    ~$1 million

    ~$18 million

    ~$45 million

     OUS3

    ~$90 million

    ~$115 million

    ~$130 million

    ~$145 million

    ~$194 million

    ~$219 million

    1. Close Concerns’ estimate; 2. Assuming ~10% of US patients obtain FreeStyle Libre through Medicare (~$240/user/month sensor pricing) and 90% obtain it commercially (~$120/user/month sensor pricing); 3. Assuming ~$120/user/month sensor pricing, not including readers; 4. Assuming increase in user base was evenly split between US and OUS given ~50/50 division in share of total revenue growth. We note these sales figures exclude reader sales.

    • FreeStyle Libre Revenue Assumptions:

      • User base (~800,000 globally): We’ve used the publicly announced global FreeStyle Libre user base figures when possible. For this roundup, this is 650,000+ global users (1Q18) and 800,000+ users (2Q18). By geography, we’ve estimated 2Q18 included ~125,000 US FreeStyle Libre users (up 150% sequentially from “>50,000” in 1Q18) and ~675,000 OUS FreeStyle Libre users (up 13% sequentially from 600,000 in 1Q18). It’s unclear how Abbott defines a “user,” which would obviously have utilization and revenue implications.

      • Pricing: Our 2Q18 model assumes ~10% of US Libre users are obtaining the system through Medicare at the higher pricing of ~$8/user/day (Libre received Medicare reimbursement in January), and Abbott realizes ~$4/user/day in revenue for the remaining US and international users. Combined with the user base figures noted above, this puts Libre 2Q18 global revenue with 90% utilization at ~$263 million, a 90% YOY rise. The $4/user/day pricing estimate for commercial/cash-pay users (i.e., non-Medicare) sensors may be overshooting Abbott’s realized revenue, since we’ve previously reported sensor net price to be <$3.60 per day. That said, our revenue calculations do not take reader sales into account, so the inflated sensor revenue estimate may be somewhat compensatory. The pricing obviously also does not account for discounts, different geographies/currencies, payer relationships, etc. Still, we note that our FreeStyle Libre sales estimate roughly backs out when comparing to Dexcom’s revenue/patient, and our US Libre model is tracking well to CEO Miles White’s guidance for $90-$100 million in US Libre revenue for all of 2018.

      • Utilization: We’re assuming that 100% Libre sensor utilization is unlikely – i.e., every user outside the US buying two sensors per month, and every US user buying three sensors per month. Our model estimates 90% utilization, which reflects the vast majority of users (but not all) wearing Libre 100% of the time.

      • What we still don’t know or haven’t estimated: How does Abbott define a “user” – are all of these people currently wearing and ordering FreeStyle Libre, or is this base those who meet broader conditions – e.g., those who have ordered from Abbott at least once, but may not be wearing the system now. If the latter, revenue would obviously be less. How widely used is FreeStyle Libre Pro, and does the worldwide “user” number include professional sensor wear? (We assume not, but aren’t sure.) We’re also unsure of how pricing differs (if at all) in cases where FreeStyle Libre is reimbursed – presumably Abbott receives less than ~$120/month in these cases. Last, we haven’t accounted for sales from readers (~$60 each), though assume they are small – especially with FreeStyle LibreLink apps available outside the US. In the 2Q18 call, the geographic split of Libre users was not provided. We assumed that the increase in Libre user base was evenly divided between US and OUS markets, given that the share of growth was roughly 50/50 (US: 53%; OUS: 47%).

    • Dexcom’s sales totaled $184 million in 1Q18 (+30% YOY) and a blowout record-high $243 million in 2Q18 (+42% YOY). Dexcom has now had three record sales quarters out of the last four, with 2Q18 clobbering the previous $221 million record in 4Q17. Worldwide sales grew a remarkable 32% sequentially in 2Q18, the largest Q2 sequential gain Dexcom has seen since 2010. Notably and surprisingly, the acceleration came from G5, as G6’s full US launch was only ~three weeks long in Q2 (it started in early June) – this bodes extremely well for 2H18. Both geographies saw record-highs: (i) US sales of $190 million grew 35% YOY and 30% sequentially in 2Q18, Dexcom’s strongest US growth in two years and driving 68% of 2Q18 growth; and (ii) international revenue of $53 million grew a robust 78% YOY as reported (+71% operationally) and 36% sequentially, hitting a record 22% of Dexcom’s total sales and driving 32% of growth. In a massive guidance raise, 2018 revenue is now expected at “approximately $925 million,” reflecting 29% YOY growth from 2017 and up a remarkable ~$70 million from 1Q18’s guidance range for $850-$860 million (+19%) in full-year sales. Dexcom has its hands full with the G6 global launch and has not shared expectations for any major new pipeline launches in 2018; the biggest one beyond G6’s rollout is Tandem launching Basal-IQ in August. Read our Dexcom 2Q18 report.

    • We estimate Medtronic’s worldwide CGM sales totaled ~$161 million in 1Q18 (up ~26% YOY) and ~$143 million in 2Q18 (up ~27% YOY). Estimated US CGM sales were ~$93 million in 1Q18 (+22% YOY) and ~$81 million in 2Q18 (+33% YOY), slightly larger than international sales of ~$69 million in 1Q18 (+32% YOY) and ~$62 million in 2Q18 (+20% YOY). Medtronic does not specifically break out CGM vs. pump sales, but the company did confirm with us last year that roughly ~25% of its Diabetes sales come from CGM, up from ~20% as of late 2015. The mix presumably varies from quarter to quarter and by geography; we’ve applied a straight 25/75 split to the US and international sales to derive the CGM/pump estimates in our model. If we had to guess, we may be underestimating Medtronic CGM sales, given the faster growth in this area. Indeed, Medtronic reported in 2Q18 that it saw “nearly 50%” YOY sales growth in CGM – much higher than our model has – but it was unclear if that referred to professional/ standalone CGM only (i.e., iPro 2, Guardian Connect) vs. all CGM sales including those that pair with pumps (e.g., 640G/Enlite 2, 670G/Guardian Sensor 3). In any case, Medtronic has a number of tailwinds that should pave the way for strong CGM sales in 2H18. First, the MiniMed 670G/Guardian Sensor 3 saw several milestones in Q2: (i) >97,000 trained, active users, a 38% gain from 1Q18’s >70,000 (strong momentum); (ii) a CE Mark announced in June (launch “this fall” in 10 EU countries – likely the first AID system to the EU market); (iii) FDA approval for 7-13 year-old patients (likely some pent-up demand here); and (iv) announcement of the $25,000 670G Outcomes Guarantee for Payers (“strong payer interest”). In addition, the Guardian Connect mobile CGM and paired Sugar.IQ app with IBM Watson launched in June in the US, marking Medtronic’s first foray into the competitive standalone US CGM market. CEO Omar Ishrak said in 2Q18 that the standalone CGM has “launched extremely well.” No metrics were provided, but we have to imagine the company will capture at least some US market share; how much will be fascinating to see, as it will likely rest on pricing, payer relationships, and great marketing of the Sugar.IQ app and Guardian’s predictive alerts. (Relative to FreeStyle Libre and G6, Guardian Connect has several disadvantages: it needs two fingersticks per day, seven-day wear, it is not on Android, and it doesn’t have Medicare reimbursement.) Read our Medtronic 2Q18 report.

    • Senseonics reported revenue of $2.9 million in 1Q18 and $3.6 million in 2Q18, marking six straight quarters of sequential gains after $300,000 in initial 4Q16 sales (the first quarter when revenue was reported). 2Q18 sales quadrupled from 2Q17’s $0.8 million. Strong EU performance (38% installed base growth) and early commercial progress in the US prompted a slight 2018 sales guidance raise by $1 million on both ends – now $19-$21 million (tripling YOY). July 31st marked the first US Eversense insertions, and we estimate that the global installed base is just under 4,000 users (up 38% in 2Q18). BCBS NJ is the first US payer to cover Eversense (UHC’s recent Medical Policy deems it “unproven” and “not medically necessary”), and >80% of US users to date have come from Dexcom (though as of the August call it was only one week into the launch and arguably not that representative of what’s to come). Internationally, 75% of 2Q18 revenue came from Italy, Germany, and Sweden, 70% of the installed base is on the 180-day XL sensor, and the Roche distribution agreement expires at the end of the year (we’d guess it will be renewed, since this is Roche’s best CGM play at this stage). On the pipeline, Senseonics planned to submit a single PMA supplement to the FDA for non-adjunctive dosing and one-calibration per day – as of the August 8 call, the submission was expected “within a month.” IDE submission for a 180-day Eversense XL US clinical trial is expected this month, with the study anticipated to start “in the coming months” (i.e., fall) and go well into 2019. Excitingly, Dr. Goodnow also mentioned that a trial with a 365-day Eversense is expected to begin in 1H19, referencing “tremendous progress” on in-vivo stabilization over one year – that would be remarkable, though is likely a late 2020 approval at the very earliest. Obviously, it is still early days for Senseonics, though the company has executed nicely and there is certainly a big market to tackle! Read our Senseonics 2Q18 report.

    BGM

    Roche, J&J, and Abbott Pooled BGM Sales Total ~$2.0 Billion in 1H18, Down ~1% YOY; Has the OUS Market Stabilized?

    BGM Sales by Geography, Estimated (1Q12-2Q18)

    Roche, Abbott, J&J

    Pooled global BGM-only revenue for Roche, J&J, and Abbott totaled an estimated $985 million in 1Q18 (+2% YOY) and an estimated $1.0 billion in 2Q18 (-4% YOY). Both came on easy comparisons to declines of 3% and 8% in 1Q17 and 2Q17, respectively. 1H18 estimated BGM sales for the three companies totaled ~$2.0 billion, falling ~1% YOY from 1H17. BGM revenue has continued to drop steadily since its peak in 2011 – we estimate that pooled sales for these three companies have decreased ~36% between 2Q12 and 2Q18. J&J has been hit the hardest, declining ~43% between 2Q12 and 2Q18, followed by Abbott (-37%) and Roche (-29%). Still, there are signs of potential recovery, particularly in the international market, which saw three consecutive quarters of positive YOY growth (3Q17-1Q18) albeit followed by a 1% YOY decline in 2Q18. The US BGM market continues to struggle, with seven consecutive quarters of YOY declines – punctuated by an 11% YOY decline in 2Q18. Overall, we estimate Roche accounts for the lion’s share of BGM revenue for these three companies, contributing ~45% and ~46% of total sales in 1Q18 and 2Q18, respectively.

    • These BGM figures are highly estimated for Abbott, J&J, and Roche and do not include privately-held Ascensia (formerly Bayer and now part of Panasonic/KKR) or private companies (e.g., Livongo). In addition to BGM, Abbott has FreeStyle Libre CGM, which we estimate is now running over $200 million per quarter and larger than its legacy BGM business (see CGM section above). Roche has Accu-Chek pumps and revenue from Senseonics distribution and app revenue from mySugr, among other sources. Finally, our model estimates a tiny bit of residual Animas revenue (~$2 million) from pump supplies distribution outside the US (given the Animas exit in October and Medtronic taking over in the US, we’ll model this at zero going forward). See the CGM and pumps sections for specifics on our assumptions. BGM growth and sales might look better or worse than what we’ve reported here, depending on the accuracy of our assumptions within these other business segments.

    • International BGM sales totaled an estimated ~$743 million in 1Q18 and ~$776 million in 2Q18, rising ~6% YOY and falling ~1% YOY, respectively. Both came on easy YOY comparisons to 4% and 9% declines in 1Q17/2Q17. Until 2Q18’s 1% YOY decline, OUS revenue did see an encouraging three consecutive quarters of positive YOY growth. The graph above suggests the business is pretty stable – not declining significantly, but also not rebounding significantly. Pooled 1Q18 growth was driven by strong performances from estimated Roche BGM (+9% YOY) and Abbott BGM (+ 11% YOY). By our estimates, J&J BGM declined ~3% YOY. In 2Q18, Roche was the only company to see positive YOY growth in BGM (+3% YOY), with J&J faring the worst (-10% YOY) and Abbott in the middle (-1% YOY). 

    • The US BGM market continues to struggle, with estimated pooled sales of $242 million in 1Q18 (-9% YOY) and $270 million in 2Q18 (-11% YOY). The 1Q18 decline came on an easy comparison to flat YOY growth in 1Q17, and the 2Q18 decline came on an easy comparison to a 5% YOY decrease in 2Q17. US BGM sales have fallen or remained flat YOY in all but two of the past 26 quarters, highlighting the damage imparted by CMS’s Competitive Bidding program, and likely the rising pressure in the BGM segment as more high-frequency strip users move to CGM. Roche was the only company to see positive YOY growth in the US, reporting estimated ~18% and ~6% YOY increases in 1Q18 and 2Q18, respectively. We assume the Accu-Chek Guide and Simple Pay Savings Program launched in the US last May is helping (very strong cash pricing for accurate strips), alongside the mySugr unlimited strips bundle with coaching (see AADE). Abbott and J&J saw US BGM declines in both 1Q18 and 2Q18, by our estimates – Abbott was down ~9% YOY and ~17% YOY, while J&J dropped ~19% YOY and ~15% YOY.

    BGM Sales by Company, Estimated (1Q12-2Q18)

    Global Sales: Roche, Abbott, J&J

    • Roche had the strongest performance in 1H18, with estimated BGM sales of ~$924 million rising ~7% YOY. Sales were an estimated $439 million in 1Q18 (+11% YOY) and ~$485 million in 2Q18 (+4% YOY). See our Roche 2Q18 report.

    • J&J saw the largest declines in 1H18, with estimated BGM sales of ~$691 million down ~10% YOY. Sales were an estimated $337 million in 1Q18 (-9% YOY) and ~$354 million in 2Q18 (-12% YOY). See our J&J 2Q18 report. As a reminder, J&J has accepted Platinum Equity’s offer to acquire the LifeScan business for ~$2.1 billion; the transaction is expected to close by the end of 2018, at which point LifeScan will be private under Platinum and we won’t have visibility on this business.

    • Abbott reported ~$416 million in 1H18 BGM revenue by our estimates, down 2% YOY. Sales were an estimated $209 million in 1Q18 (+3% YOY) and ~$207 million in 2Q18 (-6% YOY). Abbott does not break out FreeStyle Libre vs. BGM revenue, but our model now estimates FreeStyle Libre is the larger business (>800,000 users globally); see the CGM section above. Read our Abbott 2Q18 report.

    • The traditional BGM market’s future is uncertain, with several factors pushing companies to rethink their strategy and approach. CMS’s Competitive Bidding Program is partially responsible for the market’s decimation, although rising CGM adoption amongst high-frequency users and the broader use of glycemic-dependent agents in type 2 diabetes (with much lower risk of hypoglycemia) may also be playing a role. With four standout CGM options now on the market, two of which are factory calibrated and indicated for non-adjunctive insulin dosing (Abbott’s FreeStyle Libre and Dexcom G6), BGM companies will have to refocus and pivot – especially pushing earlier in the type 2 diabetes treatment paradigm where CGM has a long way to go. (Of course, professional CGM is also gaining ground in this group – yet more competition.) Business models may need to shift to being focused around devices – e.g., coaching, connectivity, education, insulin titration, and population management. More outcomes-based approaches will also be important, whereby BGM is provided as one part of a broader outcomes-driven program that combines or curates several tools. Smaller players like mySugr, Livongo, One Drop, and Dario have done great work to rethink the model, though a big question is how the larger legacy businesses will adapt and how quickly. (Roche has obviously taken a lead here by acquiring mySugr.) In our view, forming partnerships with payers and health systems will be key – using traditional BGM data and connectivity to fuel tools and generate insights around insulin titration, coaching, population health – rather than selling strips. Traditional BGM is not disappearing anytime soon, but traditional models will have to be creative in a declining market as real-time and professional CGM move towards standard of care.

    • Given the tough market drivers at play, Roche, Abbott, and J&J have been pushed to rethink their business strategies. Roche has taken a clear stance by investing in its digital ecosystem, acquiring mySugr and partnering to distribute Senseonics CGM. Plus, Roche has the soon-to-be-launched patch pump, Solo (slated in 2Q18 for a controlled EU launch “for the remainder of the year), alongside its ambitions for automated insulin delivery with its tubed Accu-Chek Insight pump. Roche is also reportedly developing a professional version of mySugr, which was first introduced at ATTD. The platform could emerge as a competitor to Glooko – what might Roche add to the increasingly crowded connected care landscape? Meanwhile, Abbott has firmly shifted attention to FreeStyle Libre, with no mention of its BGM business for quite some time; no surprise there. And last, J&J has decided to leave diabetes devices entirely; it accepted a ~$2.1 billion offer from Platinum Equity to acquire LifeScan in June (the transaction is expected complete by the end of 2018) and CeQur acquired Calibra Finesse (OneTouch Via) from J&J in July. The future of LifeScan will be fascinating to watch under the control of Platinum Equity – how will it compete? Where will it innovate? Despite struggles, the LifeScan BGM business still accrued net revenue of $1.5 billion in 2017 – that sized business is not going to zero any time soon. What might Platinum Equity, which lacks significant healthcare experience, do with the business? Will Platinum invest in a digital ecosystem to compete with Roche? Will it acquire players like One Drop, Livongo, Glooko, or WellDoc? Will it double down on the popular OneTouch Reveal app at the center of the diabetes business? Will Platinum acquire an earlier-stage CGM – see our competitive landscape for some candidates? What will happen to the WellDoc/LifeScan partnership? With the right investment, we think LifeScan can succeed; however, we’ll lose the ability to track sales come 2019.

    Insulin Pumps and Delivery Devices

    1H18 Sales of ~$1.3 billion up 20% YOY; Industry Record High in 1Q18 (~$679 Million); Strong 1H18 for Medtronic (~72% of industry growth)

    Insulin Pump Sales by Geography, Estimated (1Q12 – 2Q18)

    Medtronic (estimated), Insulet, Tandem, Animas (estimated), Roche (estimated), Valeritas, and Cellnovo

    We estimate the insulin pump and delivery devices market rose ~20% YOY in both 1Q18 (~$679 million) and in 2Q18 (~$630 million), marking three consecutive quarters of positive YOY growth for the industry. This includes sales from Medtronic (estimated), Insulet, Tandem, Animas (estimated), Roche (estimated), Valeritas, and Cellnovo. Both YOY increases came on easy comparison to declines of 3% and 7% in 1Q17 and 2Q17, respectively. 1Q18 was actually an industry revenue record (~$679 million), handily surpassing the previous 4Q17 record of $651 million. 1H18 pump sales totaled ~$1.3 billion, rising 20% YOY. With the obvious exception of Animas, which has exited the pump market (October 2017), we estimate every company saw growth in both 1Q18 and 2Q18. Medtronic has carried the field in 2018, with an ~70% market share by sales and supplying ~73% of 1H18 pump industry growth – its highest contribution in nearly five years! Insulet and Tandem have also seen very strong revenue growth in 1H18, though with Medtronic’s gains, both companies have made smaller contributions to global pump industry growth (~14% and 9%, respectively) than we saw in 2H17; see a company-by-company recap below. Overall, it’s promising to observe this rebound from a very soft pump industry performance in 2017 – last year saw the first estimated yearly decline in pump industry revenue in our model going back to 2005, with most of the weakness coming in 1H17 (-6%) followed by a more moderate decline in 2H17 (-1%). We still see the pump market as somewhat fragile, as evidenced by Animas and Roche ceasing sales in the US in 2017. Still, with Medtronic, Insulet, and Tandem seeing excellent starts to 2018, two AID systems now in the US (670G, Basal-IQ), lots of international runway, and much more innovation to come (e.g., type 2 focused devices, mobile apps, next-gen AID systems with less burden, open protocol systems?), we expect pumps to see continued strength in 2H18 and beyond. In 2019-2021, we are also likely to see several potential new entrants in the US, including BD’s type 2 patch pump (by September 2019), Bigfoot (~2020), Lilly (~2019-2020), and Beta Bionics (~2020), and potentially Ypsomed (FDA review ongoing), Sooil DANA, Cellnovo, and others. On the type 2-focused delivery devices front, Valeritas had a record 2Q18, and CeQur acquired J&J’s Calibra Finesse (OneTouch Via) bolus patch device in July (launch expected in mid-2019).  

    • Estimated US pump revenue totaled ~$382 million in 1Q18 (+19% YOY) and ~$362 million in 2Q18 (+28% YOY). Both increases came on easy comparisons to declines of 6% and 13% in 1Q17 and 2Q17, respectively. (The easy comparisons were especially true for Medtronic, who still drives most of the pump industry’s sales.) 2Q18 does mark three consecutive quarters of positive US growth, and notably, the US provided 74% of the industry’s growth in 2Q18 – its strongest showing since 1Q15! Sales were driven by an especially strong half-year from Medtronic, who reported US growth of ~22% YOY and ~33% YOY in 1Q18 and 2Q18, respectively. Though Medtronic’s YOY comparisons were easy in 1H18, these strong gains supplied a very strong 69% of the US pump industry’s growth in 2Q18. Medtronic’s US pump growth, driven by strong MiniMed 670G uptake, seems poised to continue – per the company’s 2Q18 call, there are >97,000 active 670G users (+38% from 1Q18) and it now has pediatric approval in 7-13 year olds. Of course, Medtronic will have strong competition with Tandem’s Basal-IQ/Dexcom G6 in 2H18 – that will be a fascinating battle to watch! Meanwhile, we estimate Tandem and Insulet each provided ~14% of the US pump industry’s growth in 2Q18. Tandem saw 1H18 US sales of $61 million grow an impressive ~58% YOY, while Insulet reported 1H18 US sales of $148 million, up a slightly more modest 19% YOY on the higher base of sales. Insulet did see record-high US Omnipod sales of $78 million in 2Q18, though the new Dash PDM won’t make a material impact on revenue until 2019. We’re likely to see Tandem’s impressive performance continue, as Basal-IQ began shipping in August and we imagine many are excited to software upgrade to the Control-IQ hybrid closed loop currently in its pivotal trial (launch in Summer 2019). Read more company-by-company analysis below.

    • Estimated international pump sales hit a record ~$298 million in 1Q18 (+21% YOY) and totaled ~$268 million in 2Q18 (+12% YOY). Both gains came on fairly easy comparisons to 2% and 1% YOY growth in 1Q17 and 2Q17, respectively. 1H18 marks a return to double-digit international pump growth after four consecutive quarters of single-digit YOY increases. Even more so than in the US, Medtronic carried the OUS pump industry – its estimated sales were ~$206 million in 1Q18 (+32% YOY) and ~$186 million in 2Q18 (+20% YOY). Our model estimates Medtronic drove an impressive ~81% of OUS pump industry growth in 2Q18, even outpacing its 69% OUS pump market share (by sales). We should see continued Medtronic OUS pump growth during the latter half of the year, now that the 670G has a CE Mark and is expected to launch in 10 EU countries this fall. We estimate Roche contributed the second-largest share of OUS pump growth (~14%) driven by estimated international pump growth of ~12% YOY in 2Q18 (~$52 million; this is highly estimated). Insulet’s OUS business was marked by two very different quarters in 1H18: 1Q18 saw impressive 53% YOY growth ($38 million), while 2Q18 saw weaker 7% growth ($29 million) on inventory issues as it assumed direct EU distribution from Ypsomed.

    Insulin Pump Sales By Company, Estimated (1Q12 – 2Q18)

    Medtronic-only Global Sales (Pumps Only):

    Global sales – excluding Medtronic – for Animas, Roche, Insulet, Tandem, Cellnovo:

    • We estimate Medtronic’s 1H18 insulin pump sales totaled ~$913 million, rising 27% YOY. Sales were an estimated ~$484 million in 1Q18 (+26% YOY) and ~$429 million in 2Q18 (+27% YOY). (As noted in the CGM section, this is based on an estimate that ~75% of Medtronic’s sales are from pumps.) As of Medtronic’s 2Q18 call in August, the MiniMed 670G now has >97,000 active US users. Medtronic finally obtained CE Mark approval for 7+ years, and in the US, 670G approval for 7-13 years came in June. Per CEO Omar Ishrak’s comments during the 2Q18 call, Medtronic is “now just beginning to commercialize the 670G in international markets.” The June timing expected an EU launch in 10 countries this fall, which should help continue the momentum seen in 1H18. See our Medtronic 2Q18 report.

    • Insulet’s 1H18 sales totaled $215 million, rising 22% YOY. Sales were $109 million in 1Q18 (+28% YOY) and $107 million in 2Q18 (+16% YOY). In August, management reduced full-year 2018 sales guidance to $542-$562 million (+18%-21% YOY), down ~$18 million at the midpoint from previous guidance of $565-$580 million (+22%-25%). This deficit reflects international inventory headwinds related to Insulet taking over direct distribution from Ypsomed (see our 2Q18 report for the full story). Per the 2Q18 call, Insulet expects the global Omnipod user base to hit ~167,000-177,000 by the end of 2018, reflecting growth of ~15%-26% - roughly in line with revenue. We’ll be watching Insulet’s US performance in 2H18 closely, as Tandem’s Basal-IQ and Medtronic’s MiniMed 670G will be tough competition. (Insulet would argue these systems aren’t tailored for MDIs, which may be a fair point and we’ll excited to watch the dynamics...) The next-gen Omnipod Dash system commenced its limited launch in July following FDA clearance in June, though it’s a fairly small launch and full rollout isn’t expected until early 2019. See our Insulet 2Q18 report.

    • Tandem has had an impressive return to growth, reporting 1H18 revenue of $61 million, up 58% YOY. Sales were $27 million in 1Q18 (+56% YOY) and $34 million in 2Q18 (+60% YOY). Although the YOY comparison to a 7% decline in sales in 2Q17 was easy, 2Q18 revenue represented the second-highest sales in Tandem’s history, only shy of the blowout record $40 million in 4Q17. An impressive 5,447 pumps were shipped in 2Q18, a 59% YOY gain and a very strong 23% sequential rise. This brings the estimated “active installed base” to over 66,000 users (pumps shipped in the last four years), and total pumps since inception to over 78,000. “About half” of Tandem’s organic new customers (i.e., non-renewal) came from MDI, similar to previous quarters and a sign it is still expanding the market. The strong Q2 momentum came from early excitement for the software upgrade to Basal-IQ/G6 (launched in Q3 in mid-August), conversions from Animas and Medtronic patients, and 800 pump renewals from existing customers (more than doubling YOY). We expect to see continued strong growth from the company in 2H18, as Basal-IQ is the only Dexcom-G6 integrated pump in the US and brings a number of advantages over 670G (the PLGS interface/operation is much simpler, no alarms (if desired), no modes, touchscreen, factory-calibrated G6 CGM, remote software updates, Bluetooth, CGM remote monitoring via G6 app, etc.). Tandem also expects to commence international sales in 2H18 – while those won’t be transformative, they will certainly be additive. See our Tandem 2Q18 report. We’ll hear more about its AID and international strategy on Tuesday at the company’s first-ever Analyst Day.

    • We estimate Roche saw ~8% YOY pump growth in 1H18 on estimated insulin delivery sales of ~$104 million. Sales were an estimated ~$51 million in 1Q18 (+7% YOY) and ~$52 million in 2Q18 (+9% YOY) – all from outside the US. Roche has never split out pump sales specifically, so we’re relying on measuring a bar graph from Roche’s 2Q18 slide deck, which shows its “Other” category (mostly pumps as far as we know) as contributing ~10% of total 1H18 revenue. It’s possible that we have overestimated Roche’s 1H18 pump sales here and/or underestimated pump revenue in 1H17; if either is true, then its pump growth would obviously be lower. As of July, Roche’s Solo patch pump was slated for a pilot launch “in the coming weeks” in Austria, Poland, Switzerland, and the UK, followed by further launches in 2019 and filing for FDA approval “in the future.” We have yet to hear any updates on the pilot launch and will be eager to see if and when it expands to full commercial scale. Read our deep dive on Solo for feature details and screenshots, and see our Roche 2Q18 report.

    • Valeritas reported $13 million in 1H18 revenue, up 34% YOY. Sales were $6.1 million in 1Q18 (+32% YOY) and a record-high $6.5 million in 2Q18 (+35% YOY). Although Valeritas has yet to report international revenue, the 24-hour disposable basal-bolus insulin patch device V-Go launched in Australia and New Zealand in August, with a subsequent launch in Italy expected later this year. On the company’s 2Q18 call, CEO Mr. John Timberlake stipulated that “meaningful revenue” from the New Zealand launch (and likely Australia, which followed it) is not expected in the near term. We’d expect these international launches might eventually boost sales in 2019, as will Valeritas’ connected V-Go SIM (Simple Insulin Management) accessory – US availability with Glooko integration is expected “by the end of 1H19.” Read our Valeritas 2Q18 report.

    • Cellnovo reported $1.0 million in 1H18 revenue, rising 83% YOY. Sales were $0.5 million in 1Q18 (+123% YOY) and $0.5 million in 2Q18 (+57% YOY). Cellnovo finally completed the transition to large-scale manufacturing of insulin cartridges with Flex in June, enabling a 12-fold increase in annual cartridge production. It remains to be seen whether demand will rise to meet the expanded supply, or whether other headwinds, including competition from Medtronic, Insulet, and Roche will push back. Earlier this month, Cellnovo announced the EU launch of its Gen 3 system, which moves the handheld controller from a custom medical device to a locked down, consumer-grade color touchscreen Android phone. The company has historically had a hard time scaling, but perhaps these latest developments will get things in line for expansion. Read our Cellnovo 2Q18 report.

    • While the MiniMed 670G is still the only hybrid closed loop system on the market, the automated insulin delivery landscape is going to shift quite a bit over the next two years. In the US, Tandem launched Basal-IQ (PLGS) with Dexcom G6 in August, including a free software update for current t:slim X2 users; a US launch of Control-IQ, Tandem’s hybrid closed loop system with automatic boluses, is expected in “summer 2019.” Bigfoot’s US pivotal trial is expected to start in 2Q19 with an anticipated 2020 launch, assuming everything stays on plan, while Beta Bionics’ home-use bridging studies are currently underway at MGH with Senseonics’ Eversense and at Stanford with Dexcom’s G5. We assume Lilly’s plan to launch its hybrid closed loop in ~late 2019-late 2020 is still on track. Internationally, the Medtronic 670G received its CE Mark and is initiating OUS commercialization now/soon, per 2Q18 remarks. As of ADA in June, Diabeloop expected a launch in late 2018 in France, the Netherlands, and Sweden. Meanwhile the EU pivotal trial for the Roche/Senseonics/Type Zero AID system is slated to begin testing patients in Europe in 3Q18, although we’re not sure how Dexcom’s acquisition of TypeZero impacts this trial.

    • While the type 2 pump market is still extremely underpenetrated, there are several promising patch insulin delivery devices designed for type 2 on the horizon. CeQur’s acquisition of Calibra Finesse (One Touch Via) from J&J, the highly anticipated, bolus-only, disposable insulin patch device is slated for a US launch in mid-2019. We’re very excited about this device and thrilled for it to be out of J&J, where the pharmaceutical and consumer giant was not investing in it. CeQur’s own basal-bolus PAQ device is expected to launch in the US in ~2020-2021. Meanwhile, BD’s type 2 patch pump and Insulet/Lilly’s U500 Omnipod (both with connectivity) are expected to launch by “September 2019” (BD) and in “2019” (Insulet) in the US – we expect both to be very strong and well-received devices and we will continue to watch these timelines. Sensile Medical, Verily, and Sanofi also announced a partnership this summer to develop and commercialize an “all-in-one” pre-filled, connected type 2 patch pump; no development timeline has been shared. We also wonder if Roche will move the Solo patch pump into type 2, given the emphasis on ease of use and sleek form factor. Can costs get low enough? Medtronic’s type 2 focus has pivoted to CGM, a smart move since the potential market is far wider.

    Obesity

    Obesity Drug Sales Climb ~29% YOY for $359 Million in 1H18, Driven by Surges in Saxenda Revenue; CAMELLIA Supports CV Safety for Belviq; Nalpropion Acquires Contrave

    • Sales of branded obesity drugs totaled $359 million in total sales for 1H18, up 34% YOY (from $268 in 1H17) and 23% since 2H17 (from $291 million). By quarter, revenue climbed 43% YOY and 13% sequentially to $174 million in 1Q18, followed by 27% YOY and 6% sequential growth to $185 million in 2Q18. Included in these calculations are Novo Nordisk’s Saxenda (liraglutide 3.0 mg), Nalpropion’s (formerly Orexigen’s) Contrave (naltrexone/bupropion), Vivus’ Qsymia (phentermine/topiramate), and Arena/Eisai’s Belviq (lorcaserin). Importantly, this analysis assumes $27 million in revenue for Orexigen’s Contrave in both quarters, which is the last-reported revenue (4Q17) before Orexigen’s March 2018 bankruptcy filing and subsequent acquisition by Nalproprion Pharmaceuticals (bid April 2018, finalized August 2018; more below). By value, Saxenda claimed 73% ($127 million) and 75% ($138 million) of the market in 1Q18 and 2Q18, respectively, continuing a trend of steadily-rising value-share for the therapy since its launch in 1Q16. Still assuming $27 million in Contrave revenue (16% and 15% market value in 1Q18 and 2Q18, respectively), Qsymia followed with 6% value share in both 1Q18 ($10 million) and 2Q18 ($11 million). Belviq rounded things out with 6% ($10 million) and 5% ($9 million) value share in 1Q18 and 2Q18, respectively. Accordingly, Saxenda drove 83% and 84% of YOY class growth in 1Q18 and 2Q18, though all four therapies saw YOY growth in both quarters, save Qsymia in 1Q18. Overall, we continue to see enormous potential for growth in this market: Novo Nordisk has estimated the proportion of patients receiving obesity treatment at 2% of the 600 million people living with obesity. Significant barriers remain, chief among them stigma, poor reimbursement, under-diagnosis, and under-appreciation of obesity as a treatable medical disease. Moreover, as better treatments, such as semaglutide for obesity, reach the market, we think that physicians and patients alike will become more optimistic about their ability to treat obesity – and we’re optimistic that this would help drive better uptake.

      • If Contrave is excluded, pooled sales of major obesity drugs (Saxenda + Qsymia + Belviq) climbed 21% YOY and 20% sequentially to $147 million in 1Q18 and increased a more modest 9% YOY and 7% sequentially to $158 million in 2Q18. Within this analysis, Saxenda claimed 86% of the market by value in 1Q18, while Belviq and Qsymia split the remaining 14% evenly. In 2Q18, Saxenda sales made up 87% of the market by value, followed by Qsymia at 7% and Belviq at 6%.

      • Excluding Saxenda, total obesity sales drop to $47 million for both 1Q18 and 2Q18; revenue growth similarly slows to 5% YOY and 9% sequentially for 1Q18, and 17% YOY and 0% sequentially for 2Q18, if Contrave sales are held at $27 million. Of course, these numbers reflect Saxenda’s dominance in the market with respect to value. That said, the product actually claims a small minority of prescriptions. Novo Nordisk has revealed that, when accounting for other branded drugs and generic obesity prescriptions, Saxenda holds only ~1%-2% of total obesity prescriptions (but an impressive ~50% of value share). Branded obesity medications claim ~15% of the value (slide 81 of Novo Nordisk’s 1Q18 slide deck). We do suspect that many are using Victoza (and now Ozempic) off-label for weight loss, given that reimbursement as a type 2 therapy is much better. And while Saxenda’s commercial success is impressive, we’re eager to see this drug get into far more patient hands.

    Total Obesity Market Sales (1Q13-2Q18)

    • Nalpropion finalized its acquisition of the recently-bankrupt Orexigen in August 2018 ($74 million), and we’re waiting for news on when and how the company will bring Contrave back to market. As detailed above, we don’t know if or how much Contrave has sold since Orexigen declared bankruptcy in March, though we find it hard to imagine sales have increased. In our opinion, Nalpropion scored a good deal with this acquisition, as Contrave brought in more than the sale price in 2017 alone ($88 million with 77% YOY and 42% sequential growth in 4Q17). That said, we expect the company will have to meaningfully invest in Contrave to revitalize the franchise, and Contrave has already been excluded from Express Scripts’ 2019 formulary, a new exclusion since 2018. We do see meaningful growth potential for this therapy in particular, given it was actually on a clear upward trajectory before Orexigen’s bankruptcy – in contrast to Belviq and Qsymia.

    • Cardiovascular trials for obesity therapies – completed and pending – came into the light during 1H18. The CAMELLIA CVOT suggested resounding CV neutrality for Belviq, Vivus presented a retrospective analysis suggesting cardiovascular safety for Qsymia, and Novo Nordisk moved toward initiation of the ambitious SELECT CVOT for semaglutide in obesity.

      • Full CAMELLIA (n-12,000) results were presented at ESC 2018 (following topline data released by Eisai in July 2018), indicating non-inferiority vs. placebo on 3-point MACE (HR=0.99, 95% CI: 0.85-1.14) over a median follow-up of 3.3 years. Results were similar on the secondary composite endpoint of CV death, MI, stroke hospitalization for heart failure or unstable angina, or coronary revascularization (HR=0.97, 95% CI: 0.87-1.07). Belviq was associated with very modest improvements in CV risk factors, including systolic and diastolic blood pressure (mean=1 mmHg), reduced heart rate (mean=1 BPM), lowered cholesterol (mean=1 mg/dl), and reduced triglycerides (12 mg/dl) – disappointing effect sizes given the topline release highlighted significant improvements in CV risk factors. Notably, incident diabetes was the only (!) individual outcome on which Belviq conferred benefit (HR=0.81, 95% CI: 0.66-0.99). This includes weight loss, which was significant after one year of treatment (~6.2 lbs greater with Belviq than placebo, p<0.001) but dwindled to only ~4.2 lbs at the end of the ~three year trial.

      • Vivus presented a retrospective analysis of MarketScan Commercial Claims and Medicare Supplemental Data (n=846,042) at the 34th International Conference on Pharmacoepidemiology and Therapeutic Risk Management, suggesting CV safety for adults taking Qsymia. The crude MACE incidence rate (IR) was 0.39 (95% CI: 0.01-2.15) for those currently taking Qsymia, compared to 0.92 (95% CI: 0.19-2.70) for those currently taking both phentermine and topiramate independently. Additionally, the IR was 0.91 (95% CI: 0.57-1.38) for those currently taking only phentermine and 3.37 (95% CI: 2.94-3.85) for those currently taking only topiramate, all versus unexposed time in former users (>60 days without exposure to any study medication). To be sure, these results don’t carry anywhere near the same weight as a long-term, randomized, controlled CVOT, though Vivus is certainly hoping that data such as these will help sway FDA away from requiring a full CVOT of the company (indeed, FDA does seem to be reconsidering CVOT requirements overall). In its announcement of these results, Vivus referenced its ongoing dialogue with FDA to reduce or eliminate the CVOT requirement for Qsymia while adding a label modification that would allow for the “safe and effective short-term utilization of Qsymia,” which the company referenced in both 1Q18 and 2Q18.

    • Novo Nordisk’s SELECT trial (n=~17,500) for semaglutide 2.4 mg in obesity is set to start later this year and powered for superiority. With this trial, Novo Nordisk hopes to shift the obesity treatment paradigm and demonstrate that meaningful weight loss can confer a benefit on CV outcomes and a reduction in mortality. As Dr. Donna Ryan has argued, SELECT could give the obesity field the “legitimacy” it needs, connecting weight loss to improvements in hard outcomes – an association that has somewhat remarkably not been demonstrated to date. We’ll most likely have to wait until at least 2023 for results, however, so we can’t expect SELECT to drive better reimbursement and uptake of obesity pharmacotherapy for some time.

     

    -- by Adam Brown, Ann Carracher, Martin Kurian, Brian Levine, Peter Rentzepis, Maeve Serino, Payal Marathe, and Kelly Close