Type 2 diabetes mellitus (T2DM) is a chronic disease that results from a combination of insulin resistance and insulin deficiency caused by progressive beta-cell failure . Treatment of T2DM should aim to control glycemia to preserve quality of life and reduce the risk of the microvascular and macrovascular complications of diabetes . Metformin, the most commonly used anti-diabetes agent, is recommended as first-line therapy for patients with T2DM [2, 3]. However, as glycemic control deteriorates, patients with T2DM usually require more than one anti-diabetes agent to control glycemia [4–6]. Sulfonylureas (SUs) are one of the treatment options recommended for second-line therapy of T2DM [2, 3] and are commonly used in clinical practice [1, 7], usually in combination with metformin.
Sulfonylureas in the treatment of T2DM
Although initially effective in controlling hyperglycemia, SUs have low durability [1, 2]. In the UK Prospective Diabetes Study (UKPDS), following an initial decline in glycosylated hemoglobin (HbA1c) in patients randomized to receive chlorpropamide or glibenclamide compared with patients who received dietary advice alone, a progressive increase in HbA1c was observed over the next 15 years, similar to the increase that occurred in patients randomized to dietary advice alone . Secondary failure rates with SUs may exceed those of other anti-diabetes agents, possibly due to increased loss of beta-cell function [1, 8]. In the UKPDS, beta-cell function assessed using the homeostasis model assessment (HOMA-B) was found to be inversely proportional to failure rates with SUs . In a study in newly diagnosed patients with T2DM, patients treated with an SU for up to 6 years showed a lower C-peptide response to glucagon than patients treated with insulin, suggesting a more rapid deterioration in beta-cell function and endogenous insulin production [10–12]. In a study of patients diagnosed with T2DM for more than 3 years, the duration of SU treatment was the only factor found to be independently associated with decreases in fasting C-peptide levels .
In addition to low durability, SUs are commonly associated with weight gain and hypoglycemia [2, 14]. In patients with T2DM receiving oral anti-diabetes agents, both weight gain and hypoglycemia are independently associated with lower treatment satisfaction and lower health-related quality of life . Hypoglycemic episodes lead to fear of further episodes, which may lead to patients eating more to avoid their blood glucose becoming too low, resulting in an association between hypoglycemia, fear of hypoglycemia and weight gain . The magnitude of impact on quality of life has been observed to increase with the severity and frequency of hypoglycemic events experienced over a 6-month period  and the level of weight gain over 12 months . Hypoglycemia and weight gain can also affect adherence to treatment. In a cross-sectional survey of 407 patients with T2DM, a potential weight gain of 2.3 kg over 6 months with a fictional anti-diabetes agent was associated with a 10-15% decreased likelihood of adherence compared with an agent that caused no weight gain; more than 2 episodes of mild-to-moderate hypoglycemia per month was also associated with a reduced likelihood of adherence . This is important, given that adherence to medication for the treatment of T2DM is poor. In prospective studies in patients with T2DM, rates of adherence to oral anti-diabetes agents, defined as the proportion of doses taken as prescribed, have been reported to be as low as 38% . Furthermore, in patients with T2DM, non-adherence to prescribed medication has been independently associated with all-cause mortality . In the ACCORD study, which investigated the effect of intensive versus standard glycemic control on cardiovascular (CV) events in patients with T2DM at high CV risk, symptomatic severe hypoglycemia was associated with increased all-cause mortality . The mechanisms by which hypoglycemia could precipitate a major vascular event include autonomic activation, primarily of the sympatho-adrenal system, provoking hemodynamic changes, such as increased heart rate and systolic blood pressure, increased myocardial contractility, stroke volume and cardiac output, to maintain glucose supply to the brain .
Microvascular complications such as albuminuria and decreasing estimated glomerular filtration rate (eGFR) are independently and continuously associated with an increased risk of CV events (CV death, non-fatal myocardial infarction, stroke) and renal events in patients with T2DM . There is limited evidence that SUs reduce the microvascular complications of T2DM [6, 22], but the evidence is not conclusive. In the UKPDS, after a median follow-up of 10 years, an absolute reduction of 2.8% in the incidence of microvascular endpoints was observed in patients who were randomized to intensive glucose control with an SU or insulin compared with patients randomized to dietary advice alone . However, most of this benefit was attributed to a reduction in retinopathy requiring photocoagulation (absolute reduction of 3.1%). Differences were detected in progression of retinopathy, microalbuminuria, and albuminuria in favor of SU/insulin, but there were no differences in blindness, visual acuity, or renal failure . This may have been because these patients were at an early stage of diabetes and so had a low risk of such complications. During the post-trial follow-up (median duration: 8.5 years), the reduced risk of microvascular events was maintained in patients who received SU/insulin compared to dietary advice alone, despite there being no difference between the groups in HbA1c after 1 year . In the ADVANCE study in patients with T2DM and a high risk of vascular disease, intensive glucose control provided by the SU gliclazide reduced the absolute incidence of microvascular complications compared with standard glucose control by 1.5% over a median follow-up of 5 years, a reduction that was primarily due to a reduction in new-onset microalbuminuria .
Most patients with T2DM die from CV-related causes . The effect of SUs on macrovascular disease is unclear. In the UKPDS, intensive glucose control with SU/insulin was not associated with a reduction in macrovascular complications compared with dietary advice after a median follow-up of 10 years . However, after a median follow-up of 16.8 years, significant reductions in myocardial infarction (by 15%), diabetes-related death (by 17%) and all-cause mortality (by 13%) were observed in patients randomized to receive SU/insulin compared with dietary advice alone during the intervention part of the study . In the ADVANCE study, intensive glucose control provided by gliclazide did not reduce the risk of major macrovascular events in high-risk patients over a median follow-up of 5 years compared with standard glucose control , while in a retrospective cohort study in 5795 patients with T2DM, all-cause mortality was higher in patients receiving monotherapy with higher daily doses of first-generation SUs (adjusted hazard ratio [HR] 2.1) or glibenclamide (adjusted HR 1.3) than in patients receiving lower doses, as was mortality due to acute ischemic events (adjusted HR 1.2 and 1.4, respectively) . A recent review of randomized controlled trials that evaluated the impact of SUs on CV outcomes found no increase in the incidence of CV events with SUs, but noted that the available data are limited and there have been no adequately powered trials addressing the CV safety of SUs .
Empagliflozin, an SGLT2 inhibitor
Empagliflozin is a potent and selective sodium glucose cotransporter 2 (SGLT2) inhibitor  in development for the treatment of T2DM. SGLT2, found in the proximal tubule of the nephron, is responsible for the reabsorption of ~90% of the glucose filtered through the kidneys . By blocking SGLT2, empagliflozin reduces renal glucose reabsorption, leading to excretion of glucose in the urine, thus reducing hyperglycemia in patients with T2DM . In Phase II and III studies, empagliflozin improved glycemic control in patients with T2DM when used as monotherapy [30, 31] or as add-on therapy [32–36]. In Phase III studies, treatment with empagliflozin was also associated with mean placebo-corrected reductions in body weight of 1.6 kg to 2.9 kg over 24 to 78 weeks [31, 33–36], likely due to the loss of calories (glucose) in the urine. Further empagliflozin was associated with mean placebo-corrected reductions in systolic blood pressure of 2.1 mmHg to 4.8 mmHg over 24 to 78 weeks [31, 33–36]; this may be due to a mild osmotic diuretic effect associated with urinary glucose excretion .
Empagliflozin is well tolerated in patients with T2DM [30–36, 38, 39]. As the mechanism of action of SGLT2 inhibitors is independent of the action of insulin [40, 41], empagliflozin is associated with a low risk of hypoglycemia [30–36, 38, 39]. It is tempting to speculate that the insulin-independent mechanism of action of empagliflozin may preserve beta-cell function and provide better durability of glycemic control than SUs. In an 8-week study in Zucker Diabetic Fatty (ZDF) rats, an animal model of T2DM, treatment with empagliflozin, but not with the SU glibenclamide, preserved beta-cell mass, increased insulin levels, and improved glycemic control . Furthermore, due to its body weight and blood pressure lowering properties, in addition to its effects on glycemic control, treatment with empagliflozin may have a beneficial effect on CV risk [43, 44]. A large dedicated CV outcome study is underway to determine the effect of empagliflozin on CV endpoints (NCT01131676).
Empagliflozin versus glimepiride: a 4-year phase III trial
The objective of this trial, the EMPA-REG H2H-SU™ trial, is to compare the effects of the SGLT2 inhibitor empagliflozin with the SU glimepiride given over the long term (4 years) as second-line therapy for T2DM in patients in whom metformin and diet/exercise has failed with respect to 1) glycemic control, 2) beta-cell function, 3) the CV risk factors: body weight, blood pressure and lipid levels, and 4) safety including prospectively adjudicated CV events and markers of renal function/damage. All patients have been recruited and randomized. Blinded baseline data are presented in this manuscript.