Comparative Effectiveness of Dapagliflozin Versus DPP-4 Inhibitors

Background & rationale Sodium-glucose co-transporter-2 inhibitors (SGLT2i) have become available for the treatment of type 2 diabetes (T2D) in the U.S. in April 2013 (canagliflozin), January 2014 (dapagliflozin), and August 2014 (empagliflozin). In Italy, first-in-class dapagliflozin received marketing authorization approval in March 2015. SGLT2i lower glucose levels by increasing urinary glucose excretion. In turn, glycosuria results in a significant reduction of body weight and is accompanied by osmotic diuresis and reduction in blood pressure. In the pre-marketing phase III clinical development program, dapagliflozin was evaluated versus placebo as monotherapy or as add-on to metformin, sulphonylurea, DPP-4i, or insulin, and versus active comparators (monotherapy versus metformin, or add-on to metformin versus sulphonylurea or DPP-4i). In meta-analyses, dapagliflozin 10 mg reduced HbA1c by 0.5-0.7% compared to placebo at 24 weeks, was non-inferior to glipizide and saxagliptin, and overall showed sustained glucose lowering effects over periods of 48-102 weeks. In most dapagliflozin phase III RCTs, body weight was reduced by 2-3 kg and a consistent reduction of about 3-4 mm Hg in systolic blood pressure was observed compared with other glucose lowering medications. In a phase III RCTs of initial 52 week duration, dapagliflozin compared to glipizide provided a sustained systolic blood pressure reduction up to the 4 year extension. Dapagliflozin phase III RCTs also consistently report an increase in HDL cholesterol by about 0.1 mmol/l although with a mild raise in LDL cholesterol. However, a study dedicated to explore the effects of dapagliflozin on HDL function failed to identify any significant effect.

Thus, owing to glycaemic and extra-glycaemic effects, SGLT2i are ideal second-line agents for the treatment of patients with T2D, who are often overweight/obese and affected by arterial hypertension. Indeed, in the EMPA-REG Outcome trial and in the CANVAS program, a significant reduction in the risk of major adverse cardiovascular events (MACE) and hospitalization for heart failure were reported among patients randomized to empagliflozin or canagliflozin, respectively, versus those randomized to placebo. Empagliflozin also reduced cardiovascular death and death from any cause, whereas the effects of canagliflozin on mortality were not statistically significant and canagliflozin conferred a significant 2-fold increased risk of amputations. While the cardiovascular outcome trial for dapagliflozin is still ongoing, a retrospective real world study conducted in the U.S. and Europe confirmed that SGLT2i as a class can improve cardiovascular outcomes, including MACE, hospitalization for heart failure and death. In addition, both the EMPA-REG Outcome trial and the CANVAS program showed that empagliflozin and canagliflozin, respectively, exerted significant renal protective effects. Based on these data, there is great promise and mounting evidence that SGLT2i help in the comprehensive management of T2D, with the ultimate goal of delaying complications.

RCTs in the field of diabetes have limitations inherent to the fact that design, setting, and patient characteristics may be poorly transferrable to clinical practice. Thus, evidence from studies using real world data are increasingly valued. RCTs enrol limited numbers of highly selected, relatively young participants, who are very well motivated and compliant, educated on the disease and instructed in drug use, mostly free from co-morbid conditions, and regularly followed-up. On the other side, the real world contains all patients who may receive a given drug, irrespectively of age, education, compliance, concomitant medications and co-morbidities, and who are followed according to local practice, often with resource limitations.

A number of real world studies on dapagliflozin have been initiated in several countries. To understand how dapagliflozin may improve hard outcomes in clinical practice, it is first critical to describe the clinical characteristics of patients at the time they initiate the drug. For instance, a study conducted among general practitioners and diabetologists in Germany aimed at evaluating the clinical characteristics of patients using dapagliflozin: as compared to patients on other glucose lowering medications: dapagliflozin users were younger, more often males, had similar diabetes duration, but higher baseline HbA1c. Second, real world studies can evaluate efficacy on glycaemic and extra-glycaemic endpoints. Among 1169 T2D patients started on dapagliflozin in Germany, 77% of whom was followed by primary care physicians, HbA1c declined by 0.8% at 3 and 6 months, body weight decreased by 2.2-2.5 kg and systolic blood pressure by 2.2-2.3 mm Hg. In this type of studies, it is theoretically possible to search for baseline clinical characteristics that modulate dapagliflozin effectiveness. Interestingly, except for baseline HbA1c, being followed by a diabetologist was associated with larger glycaemic improvement than being followed by primary care physicians in Germany. Regarding background medications, an observational study conducted in Spain reported that the glycaemic improvement at 6 and 12 months was lower when dapagliflozin was initiated in patients on a GLP-1RA containing regimen than in patients on non-GLP-RA containing regimen. A retrospective chart review study on Indian patients with uncontrolled T2D on insulin therapy (baseline HbA1c 10.3%) found that initiation of dapagliflozin was followed by a reduction in HbA1c by an average of 2.1% and of body weight by 2.1 kg, with a significant insulin-sparing effects of up to 20% the initial dosw.

Finally, in countries where it is possible to link registries of prescriptions, claims, hospitalizations and causes of death, valuable real world studies can deliver data on hard endpoints in patients exposed to different medications. A retrospective study conducted in Sweden and comparing 2047 patients initiated on dapagliflozin versus 4094 matched patients initiated on insulin during 2013 to 2014 reported a 56% reduction of all-cause mortality and a 49% reduction in fatal and nonfatal cardiovascular events among dapagliflozin users. In addition, The Health Improvement Network (THIN) database, which is based on a retrospective collection of general practice data in the UK, shows that the risk of death among 4444 patients exposed to dapagliflozin was about 50% significantly lower than the risk of death among 17,680 non-exposed patients. Finally, results of the multinational CVD-Real study showed a 39% lower risk of hospitalization for heart failure and a 51% lower risk of death among about 150k patients initiated on SGLT2i (more than 90% of whom were on dapagliflozin in European countries) than among 150k matched patients initiated on another glucose lowering medication.

So far, real world studies on dapagliflozin have widely confirmed the glycaemic and extraglycaemic effects observed in RCTs and provided initial evidence in support of protection against cardiovascular disease and mortality. Unfortunately, most large retrospective real-world studies on hard outcomes have been performed by linking various types of administrative databases but typically lack several details on HbA1c, body weight and lipid profile at baseline and how they change over time.

The DARWIN-T2D study was a nationwide real word study designed to describe the baseline characteristics of patients who started dapagliflozin at Italian diabetes outpatient clinics. Building on the DARWIN-T2D study, we herein propose a retrospective longitudinal comparative assessment of the effects of dapagliflozin versus DPP-4i on combined efficacy endpoints.

Research hypothesis Owing to their glycemic and extraglycemic effects, SGLT2i are becoming ideal second-line agents for the treatment of T2D. However, SGLT2i are not devoid of side effects. Because of glycosuria, SGLT2i increase the risk of genito-urinary tract infections (GUTI) and may favour dehydration or volume depletion, especially in patients taking diuretics [22]. In addition, SGLT2i can precipitate diabetic ketoacidosis (DKA), especially when used off-label in type 1 diabetes or in T2D patients with poor beta cell function. Furthermore, based on final results of the CANVAS program, a boxed warning was added to the canagliflozin label regarding an increase in the risk of amputations. For these reasons, although the cardiovascular benefits of SGLT2i are clearly delineating, their widespread use as second-line agents may be contended by other oral glucose lowering medications which are perceived to be provided with a more neutral safety profile, namely DPP-4 inhibitors (DPP-4i). DPP-4i as a class lower HbA1c by 0.5-0.7% and exert minor or no effects on body weight, blood pressure, and lipid profile. In addition, three large RCTs showed no benefit of sitagliptin, saxagliptin, and alogliptin on cardiovascular outcomes, with an isolated signal that saxagliptin might increase the risk of hospitalization for heart failure.

Importantly, a sub-analysis of the CVD-Real retrospective study has shown that dapagliflozin compared to DPP4i is associated with lower risk of cardiovascular events and all-cause mortality.

Therefore, the present study proposal aims at providing real world data on the comparative effectiveness of SGLT2i versus DPP-4i on a composite endpoint of HbA1c, body weight and blood pressure reduction. The study has the potential to demonstrate multiple benefits of SGLT2i in the routine clinical practice, as compared to DPP-4i, which are perceived to be safer but are mostly devoid of extraglycemic effects. We hypothesize that dapagliflozin is superior to DPP-4i in the attainment of a composite endpoint of HbA1c, body weight and blood pressure reduction.

Benefit/risk and ethical assessment There are no patient's risk associated with this retrospective protocol. In Italy, retrospective studies conducted on database-extracted anonymized records are not subjected to ethical approval. The protocol just needs to be notified to the competent Ethical committee (see Determina AIFA 20/03/2008). The benefit is clear in terms of generating additional data, eventually challenging results of clinical trials, and providing new hypotheses.

Study Objectives

Primary objective The primary objective of the study is to compare the efficacy of the SGLT2i dapagliflozin versus DPP-4i on a combined endpoint composed by simultaneous changes in HbA1c, body weight, and systolic blood pressure.

Secondary exploratory objectives To compare variations in the overall complication burden during therapy with dapagliflozin versus DPP-4i, defined as new occurrence or worsening of any microangiopathy and new occurrence or worsening of any macroangiopathy.