Memorandum

Diabetes Care article reports real-world trends in diabetes prescriptions 2005-2016; SUs remain leading second-line therapy (though down from 60% to 46%); GLP-1 + SGLT-2 comprise ~14% of second-line prescriptions – November 16, 2017

Executive Highlights

  • Real-world data from a recent Diabetes Care article reveals the following trends in type 2 diabetes prescription dynamics over the past decade: (i) an increase in the use of metformin as first-line therapy and (ii) a very gradual diversification of second-therapy prescriptions, away from sulfonylureas toward newer agents like DPP-4 inhibitors, GLP-1 agonists, and SGLT-2 inhibitors. That said, SUs are still the most commonly prescribed type 2 diabetes treatment after metformin, comprising 46% of second-line diabetes prescriptions in 2016 vs. 20% for DPP-4 inhibitors, 17% for insulin, 7% for GLP-1 agonists and SGLT-2 inhibitors, and 4% for TZDs.
  • Use of newer, more advanced diabetes agents remains disappointingly low: Only 4% and 7% of patients in the study had ever been prescribed an SGLT-2 inhibitor or GLP-1 agonist at any point in their diabetes care. This pales in comparison to overall rates of prescription for metformin (79%), SUs (35%), insulin (26%), DPP-4 inhibitors (18%), and TZDs (11%).
  • Dr. Irl Hirsch attributed these patterns to cost, in that sulfonylureas are generic, cheap, and accessible where GLP-1 agonists and SGLT-2 inhibitors are still out of reach for many: “Between an increasing uninsured population, unaffordable co-pays, high-deductible plans, and the donut hole, we can't practice the evidenced-based medicine the RCTs tell us we should.”

An article published recently in Diabetes Care reports major trends in the real-world use of different diabetes therapy classes in the US. Most notably, over the past decade (2005-2016), (i) metformin has become a more popular prescription for first-line therapy (growing from 60% to 77%), and (ii) there’s been a gradual diversification of second-line therapy prescriptions, away from sulfonylureas toward newer agents like DPP-4 inhibitors, GLP-1 agonists, and SGLT-2 inhibitors, though SUs remain the most commonly prescribed type 2 diabetes treatment after metformin.

The increase in metformin as first-line has come largely at the expense of TZDs, with prescription rates falling from 11% in 2005 to 1% in 2016. We imagine this was likely driven by the FDA black box warning for congestive heart failure on rosiglitazone and pioglitazone, as well as other safety concerns surrounding the TZD class such as bone fractures and bladder cancer. Metformin also seemed to replace some first-line sulfonylurea prescriptions between 2005-2016 (which fell from 20% to 8%) – we wish we could celebrate this, but very unfortunately, sulfonylureas continue to be a mainstay as second-line diabetes therapy (down from 60% to 46% by 2016, but still far more common as second-line treatment than DPP-4 agents at 20%, insulin at 17%, GLP-1 agonists or SGLT-2 inhibitors at 7% each, or TZDs at 4%). The SU class comes with an unfavorable side-effect profile of weight gain, hypoglycemia risk, beta cell burnout, and possible CV harm. Many thought leaders, including Dr. Jay Skyler, have expressed disdain for this drug class, and yet sulfonylureas remain in play due to their low cost (and the inescapable fact that cost determines care for the vast majority of people with diabetes). Moreover, though we are heartened to see the uptick in SGLT-2 and GLP-1 prescriptions (both <1% in 2005, when Byetta was just-approved and SGLT-2 inhibitors were nowhere near the market yet), we find their current second-line prescription share of 7% each to be quite low, given the very compelling clinical benefits they offer – more potent glucose-lowering, weight loss, low hypoglycemia risk, and possible renal and cardioprotection.

In light of this and the well-deserved clinical enthusiasm surrounding these newer agents, we were saddened to learn that only 4% and 7% of patients in the study had ever been prescribed an SGLT-2 inhibitor or a GLP-1 agonist, respectively, at any point in their diabetes care. This pales in comparison to overall rates of prescription for other major diabetes therapies such as metformin (79%), SUs (35%), insulin (26%), DPP-4 inhibitors (18%), and TZDs (11%). That said, in the past 12 months, two advanced diabetes drugs have received new CV indications from the FDA – Lilly/BI’s SGLT-2 Jardiance and Novo Nordisk’s GLP-1 Victoza – and J&J has filed an sNDA to make Invokana the third amidst these ranks. The ADA recommends Jardiance or Victoza for people with type 2 diabetes at high CV risk in its 2017 Standards of Care. All this is to say – perhaps there is hope for accelerated uptake of these advanced therapy classes in the years ahead (the present data only goes through 2016, before these label updates/new ADA guidelines), especially since we know how much real-world HCPs rely on treatment algorithms and product labels to inform their practice.

Figure 1: Proportional Share of First-Line and Second-Line Diabetes Therapies by Class (2005-2016)

Source: Diabetes Care

  • Despite the small proportion of patients on a GLP-1 agonist, these agents showed promising signs of improved treatment outcomes as second-line agents vs. SUs. Patients prescribed a GLP-1 agonist second-line experienced a marginally longer time until intensification with a third diabetes agent (4.2 years vs. 3.9 years with an SU) or with insulin, specifically (6.6 years vs. 6.3 years with an SU). We note that many factors beyond the effectiveness of a medication regimen determine whether therapy is intensified (patient/provider discussions, possible clinical inertia, the role of lifestyle modification, cost/accessibility of drugs, etc.). No data of this kind was available for SGLT-2 inhibitors, but we’d be very curious to see if these agents follow a similar pattern as GLP-1 agonists. Despite improved outcomes with newer diabetes therapies, these agents showed higher rates of discontinuation: 25% and 20% within one year for SGLT-2 inhibitors and GLP-1 agonists, respectively, compared to 17% for DPP-4 inhibitors, 10% for SUs, 8% for metformin, and 5% for insulin. Again, no data was provided on reasons for discontinuation, but we suspect this trend was more so due to high cost/low access rather than side-effects (i.e. GLP-1 agonists may cause some GI symptoms and SGLT-2 inhibitors introduce risk for genital mycotic infections, but SUs come with a more severe side-effect profile of hypoglycemia and weight gain). 
  • We discussed these findings with Dr. Irl Hirsch, who attributed the comparatively low prescription share of GLP-1 agonists and SGLT-2 inhibitors to the notoriously high cost of these newer agents. He surmised: “Between an increasing uninsured population, unaffordable co-pays, high-deductible plans, and the donut hole, we can't practice the evidenced-based medicine the RCTs tell us we should.” The fact that such a vanishingly small percentage of people with diabetes can access the very therapies that are most highly recommended for them underscores the pressing need to address drug affordability in the US and elsewhere. We continue to believe that change will only come from increased collaboration and transparency from industry, payers, PBMs, and policymakers. We can only hope that these striking findings, illustrating the stark divide between real-world prescribing patterns and evidenced-based medicine, can serve as an impetus to kick-start these conversations, though we acknowledge that this will be a very long road ahead.
  • These real-world results leave us deeply concerned that prescription of sulfonylureas persists at such high frequency. And, even more troubling in our view is the lack of clear guidelines on which SU to prescribe when this class is the only option that patients can afford. There is no doubt that SUs are not the ideal therapeutic option, but they are sadly the only accessible option for many patients in the US with high-deductible or no insurance (not to mention in many parts of the world where newer diabetes agents simply aren’t available).  Dr. Hirsch noted his preference for glimepiride due to “old but soft” observational data that this agent confers the least amount of hypoglycemia in the class. Dr. Pablo Aschner shared a similar opinion at the CODHy Latin America meeting earlier this year, defending the CV neutrality of newer SUs such as glimepiride and gliclazide (though he conceded that these, like older SUs, still carry the risk of hypoglycemia). Given that SUs comprise 8% of first-line diabetes prescriptions in the US and 46% of second-line prescriptions (according to the latest 2016 figures shared in this paper), there is a pressing need to differentiate between agents in the class to identify the “safest” of the bunch. We’d love to see these differences reflected in future iterations of professional treatment guidelines.
  • Interestingly, despite the narrowing bar for TZD prescriptions (as shown in the graphs above), recent commentary from thought leaders suggests that low-dose pioglitazone could actually be a very effective therapy for many people with diabetes in the real world. At EASD 2017, Drs. Steve Edelman and Robert Chilton delivered powerful remarks in defense of pioglitazone, pointing to more durable A1c effects over the long term. The duo acknowledged TZDs’ bad reputation, but argued that many of these safety concerns don’t apply to low-dose pioglitazone – in fact, the IRIS trial found significant risk reduction for stroke/MI with the agent, and the TOSCA.IT study (also presented at EASD) found no signal for heart failure, bone fractures, or bladder cancer. We’ve heard similar endorsements for low-dose pioglitazone from Drs. Jay Skyler and Robert Eckel at CMHC, and we note that this drug is also now generic and therefore low-cost.

 

-- by Abigail Dove, Payal Marathe, and Kelly Close